PROJECT HERMES TU1-2000A

CONTRACT NO. DA-30-115-0RD-23 REPORT NO. R52A0510

FINAL REPORT, PROJECT HERMES V-2 MISSILE PROGRAM

A Report of
Guided Missiles Department
Aeronautic and Ordnance Systems Divisions
Defense Products Group
Schenectady, New York

GENERALS ELECTRIC
GENERALWELECTRIC
SCHENECTADY, N. Y.

TECHNICAL INFORMATION SERIES

TITLE PAGE

Author Subject Classification No.

R52A0510
WHITE, L. D. MISSILES, GUIDED
Date
SEPTEMBER, 1952

Title

FINAL REPORT, PROJECT HERMES V-2 MISSILE PROGRAM

Abstract

THIS IS THE FINAL REPORT ON V-2 OPERATIONS CONDUCTED AT THE WHITE SANDS PROVING
GROUND AS PART OF PROJECT HERMES. PRIMARY ATTENTION IS GIVEN TO THE PERFORMANCE
OF THE MISSILE AND THE COMPONENTS. OPERATIONAL AND TEST PROCEDURES AS USED AT
WSPG ARE DISCUSSED. THE PROGRAM WAS CONCLUDED IN JUNE 1951 AFTER 67 ROCKETS USING
V-2 COMPONENTS HAD BEEN CONSTRUCTED, TESTED, AND LAUNCHED.

G. E. Class. Reproducible Copy Filed at No. Pages

4
REPORTS & DOCUMENTS OFFICE 18/,
Gov. Class. GUIDED MISSILES DEPARTMENT

Conclusions

THE PROGRAM PROVIDED: (1) BALLISTICS DATA, (2) VEHICLES FOR UPPER ATMOSPHERE
RESEARCH PROJECTS, (3) VEHICLES FOR EXPERIMENTS DIRECTLY CONCERNED WITH THE DESIGN
OF FUTURE MISSILES, (4) VEHICLES FOR OPERATIONAL TESTS OF FUTURE MISSILE COMPONENTS,
AND (5) EXPERIENCE IN THE HANDLING AND FIRING OF LARGE MISSILES.

Information Prepared for PROJECT HERMES. U. S. ARMY ORDNANCE

Tests Made by SERVICE ENGINEERING DIVISION
-- '
> J
Countersigned DR. R. W. PORTER Div. GUIDED MISSILES DEPARTMENT

Divisions A. & O. S. Locatio n SCHENECTADY, NEW YORK

FOREWORD

This report is offered as a final account of V-2 (A-4)

operations performed by the Service Engineering Divisions
of the General Electric Company at the White Sands Prov-
ing Ground under the Hermes Project. It is intended to
provide information and experience, concerning the V-2
program, which is not known to be recorded elsewhere.
Information which is readily available in other publications
has, in general, been omitted from this report. Appropriate
references are included.
TABLE OF CONTENTS

Page

OBJECTIVES 1

GENERAL SUMMARY 3

PERFORMANCE SUMMARY (1) 5

1.1 GENERAL STATISTICS 5

1.2 RESULTS OF PROPULSION UNIT CALIBRATION 8

1.3 FAILURES 9

1.3.1 Definition 9
1.3.2 Investigation 9
1.3.3 General Distribution 9
1.3.4 Time Failure
of 9
1.3.5 Allocation of Fault 10
1.3.6 Effect of Low Temperature Due to Lox 11

MISSILE COMPONENTS (2) 13

TEST, REPAIR AND ASSEMBLY (3) 19

3.1 GENERAL 19
3.2 PROPULSION UNIT CALIBRATION 20
3.3 STRUCTURE 21

GROUND EQUIPMENT (4) 27

4.1 GENERAL 27
4.2 MECHANICAL EQUIPMENTS 27

4.2.1 Meilerwagen 27
4.2.2 Launching Platform 28
4.2.3 Liquid Oxygen Trailer 28
4.2.4 Hydrogen Peroxide Trailer 28
4.2.5 Alcohol Pump Trailer 29
4.2.6 Compressor Trailer 29

4.3 ELECTRICAL EQUIPMENTS 29

4.3.1 Firing Desk 29

4.3.2 Cast-off Cables 29

EQUIPMENT AND PROCEDURE MODIFICATIONS (5) 31

5.1 CALIBRATION OF PROPULSION SYSTEM AS A UNIT 31

5.2 OMISSION OF ALCOHOL TANK PRESSURIZATION ON RE-ENTRY 32
5.3 SECONDARY SOURCE OF AIR PRESSURE FOR VENT VALVE 32
5.4 LOADING OF OXYGEN AFTER HYDROGEN PEROXIDE 32
5.5 SUBSTITUTION OF ELECTRIC MOTOR FOR GAS ENGINE DRIVES 32
5.6 SUBSTITUTION OF AUDIBLE LEVEL INDICATOR FOR FLOAT SWITCHES 32
5.7 ADDITIONAL VENTING OF LOX VALVE CONTROL CHAMBER 33
5.8 HYDROGEN PEROXIDE LOADING BY GRAVITY 33
5.9 CHANGE OF MISSILE POWER RELAY 33
5.10 COMMAND- CURRENT CIRCUIT CHANGE 33
5.11 CHANGE IN GROUND CUT-OFF RELAY CIRCUIT. 34
5.12 COMPUTER CHANGES 34
5.13 CHANGE IN STEERING TEST 34
5.14 ADDITION OF ROLL ERECTION MONITOR FOR PITCH GYRO 34
5.15 ADDITION OF ROLL ERECTION MONITOR FOR ROLL- YAW GYRO 34
5.16 RELOCATION OF VANE-BALANCE POTENTIOMETERS 34
5.17 ADDITION OF ISOLATION SWITCH IN PITCH CONTROL 35
5.18 NEW SERVO MOTORS 35
TABLE OF CONTENTS (CONT'D)

Page
MISSILE BREAK-UP INSTALLATIONS (6) 37

V-2 OPERATIONS OUTSIDE WSPG (7) 41

7.1 AT SEA 41
7.2 AT LONG RANGE PROVING GROUND, COCOA, FLORIDA 41

PERSONNEL (8) 43

8.1 GENERAL 43
8.2 SCHENECTADY WORKS PERSONNEL 44
8.3 GERMAN PERSONNEL 44
8.4 MILITARY PERSONNEL 44

SAFETY (9) 45

9.1 HAZARDS INVOLVED 45

9.1.1 Liquid Oxygen 45

9.1.2 Alcohol 45
9.1.3 Hydrogen Peroxide 45
9.1.4 Explosives 45
9.1.5 High-pressure Gas 45
9.1.6 Experimental Equipment . 46
9.2 SAFETY MEASURES 46
9.2.1 Liquid Oxygen 46
9.2.2 Alcohol 46
9.2.3 Hydrogen Peroxide 46
9.2.4 Explosives 47
9.2.5 High-pressure Gas 48
9.2.6 Gantry Crane 48
9.2.7 Launching Area 50
9.2.8 Component Testing 50

9.3 SAFETY RECORD 50

BIBLIOGRAPHY 51

APPENDICES 53

APPENDDC A, PROPULSION SYSTEM COMPONENTS 55

A.l OVER-ALL SYSTEM 55

A.2 COMPONENTS 56

A.2.1 Jet Vanes 56

A.2. 2 Burners 58
A.2.3 Gas Bottles 60
A.2.4 Heat Exchanger 60
A.2.5 Lines and Fittings 61
A.2.6 Turbopump 62
A.2.7 Alcohol Main Valve 68
A.2.8 Alcohol Preliminary Valve 70
A. 2.9 Oxygen Fill Valve 71
A.2.10 Switch Battery 71
A.2.11 Burner Drain Valve 72
A.2.12 Ram Charger Valve 72
A.2.13 Alcohol Tank Drain Valve 72
TABLE OF CONTENTS (CONT'D)

Page

APPENDIX A, PROPULSION SYSTEM COMPONENTS (CONT'D)

A. 2 COMPONENTS (CONT'D)
A.2.14 A-3 Check Valve 72
A.2.15 Oxygen Vent Valve 72
A.2.16 Oxygen Main Valve 73
A.2.17 Steam Plant 75
A.2.18 Main Tanks 78

A.3 TEST INSTRUCTIONS 78

A.3.1 Carbon Vanes 78

A.3. 2 Burners 79
A.3.3 Air Bottles 81
A.3.4 Heat Exchanger 81
A.3. 5 Important Pipes and Fittings 81
A.3.6 Alcohol Main Valve 83
A.3. 7 Alcohol Preliminary Valve 83
A.3.8 Oxygen -filling Valve 85
A.3. 9 Switch Battery 85
A.3.10 Alcohol Drain Valve 85
A.3. 11 Ram Charger Valve 85
A.3. 12 Alcohol Drain Valve 85
A.3.13 A-3 Check Valve 86
A.3. 14 Oxygen Vent Valve 86
A.3.15 Oxygen Vent Valve 86
A.3. 16 Steam Plant 87
A.3.17 Reducer Safety Valve 90
A. 3. 18 Hand-operated Valve on Reducer Assembly 90
A.3.19 High-pressure Hand Valve 90
A.3. 20 Peroxide and Permanganate Bleed Valves 91
A.3. 21 Peroxide and Permanganate Drain Valves 91
A.3. 22 Peroxide and Permanganate Check Valves 91
A.3.23 25-ton Valve 92
A.3. 24 PE-4 Control Valve 92
A.3.25 Eight-ton Valve 92
A.3. 26 High-pressure PE-10 Valve 92
A.3. 27 Z Contact 92
A.3. 28 Throttle in Line to Preliminary Alcohol Valve 93
A.3. 29 Check Valve on Cross Piece 93
A.3. 30 Alcohol Tank 93
A.3.31 Oxygen Tank 94

APPENDIX B, PROPULSION UNIT CALIBRATION 95

B.l GENERAL 95
B.2 METHODS OF CALIBRATION 95
B.2.1 Hydraulic Testing of Component Parts 95
B.2.2 Water-flow Test of Complete Propulsion Unit 95
B.2.3 Cold Propellant Test of Complete Propulsion Unit 95
B.2. 4 Static Firing 96
B.3 CALIBRATION AT WSPG 96
B.4 DETAILED INFORMATION ON THE FESTAL CALIBRATION PROCEDURE 98
B.4.1 Steam Plant Test 98
B.4. 2 Combustion Pressure Simulating Orifices 98
TABLE OF CONTENTS (CONT'D)

Page
APPENDIX B, PROPULSION UNIT CALIBRATION (CONT'D)
B.4 DETAILED INFORMATION ON THE FINAL CALIBRATION PROCEDURE (CONT'D)
B.4.3 Instrumentation 99
B.4. 4 Preparation for Calibration Test 103
B.4. 5 Test Data Reduction 106

B.5 CALIBRATION AND FLIGHT DATA 114

B.5.1 Average Values from Calibration Data 114

B.5. 2 Static Firing Data 114
B.5. 3 Flight Data 114

B.6 SPECIAL V-2 PROPULSION UNIT TESTS 115

B.6.1 Introduction il5

B.6. 2 Test Procedure 115
B.6. 3 Results 120
B.6.4 Conclusions 120

APPENDDC C, STEERING SYSTEM COMPONENTS 121

C.l GYROS 121

C.l.l Anschutz Type 122

C.l. 2 LGW Type 122
C.1.3 LGW Type (American) 123
C. 1.4 Test Procedures and Specifications 124
C.1.5 Special Vibration Test i26

C.2 MIX COMPUTER (AUTOPILOT SERVO AMPLIFIER) 126

C.2.1 American Version of German Computer 129

C.3 SERVO 136

C4 TIME SWITCH i40
C.5 INVERTERS AND REGULATORS 141

APPENDK D, MISSILE FAILURES i43

D.l SUMMARY 143

D.2 INDIVIDUAL REPORTS 143

MISSILE 2 144
MISSILE 8 144
MISSILE 10 144
MISSILE 11 145
MISSILE 14 146
MISSILE 16 146
MISSILE 18 149
MISSILE 20 150
MISSILE 24 151
MISSILE 26 152
MISSILE 27 152
MISSILE 29 153
MISSILE 30 154
MISSILE 32 156
MISSILE 37 157
MISSILE 38 158
MISSILE 39 161
MISSILE 40 162

vi
TABLE OF CONTENTS (CONT'D)

Page
APPENDIX D, MISSILE FAILURES (CONT'D)
D.2 INDIVIDUAL REPORTS (CONT'D)
MISSILE 42 164
MISSILE 45 165
MISSILE 46 167
MISSILE 50 168
MISSILE 52 172
MISSILE 54 172
MISSILE 55 175
MISSILE 57 176
MISSILE BUMPER 2 177
MISSILE BUMPER 4 178
MISSILE BUMPER 6 179
MISSILE BUMPERS 7 AND 8 180
MISSILE SPECIAL 181

REFERENCES 183

DISTRIBUTION lg4
LIST OF ILLUSTRATIONS

Figure Page
1 V-2 Mounted for First Static Test at WSPG 3
2 First V-2 Static Test in Progress at WSPG 3
3 V-2 Rocket Shortly After Launching at WSPG 4
4 Propulsion Unit Calibration in Progress 8
5 V-2 Being Loaded with Oxygen 11
6 V-2 Burners in Test Preparation Area 14
7 Missile Assembly Building at WSPG 19
8 Missile Assembly Building at WSPG 19
9 V-2 Steam Plant Being Tested Prior to Assembly in Rocket 20
10 Burners Being Prepared for Test 21
11 Propulsion Unit Calibration Stand 21
12 Control Cables in V-2 Midsection 22
13 V-2 Tailsections at WSPG 22
14 Tailsection Fin-alignment Work Sheet 23
15 Bumper Missile Launching Sequence 25
16 Meilerwagen 27
17 German Hydrogen Peroxide Trailer 28
18 V-2 Firing Desk Built at WSPG 30
19 Rear View of V-2 Firing Desk Relay Box 30
20 V-2 Being Erected for Launching 30
21 Propulsion Unit Assembly Nearly Completed at WSPG 31
22 Impact Crater, Missile 34 38
23 Midsection and Tailsection of Missile 26 After Impact 39
24 Tailsection of Missile 26 After Impact 39
25 Midsection of Missile 26 After Impact 40
26 Spectograph Being Removed from V-2 After Impact 40
27 Gantry Crane 49
28 Jet Vane Descriptive Terms 57
29 Cutaway View of Burner and View of Burner-piping Installation 59
30 Technical Data on V-2 Turbopump Assembly 62
31 Cutaway View, V-2 Turbopump Assembly 63
32 V-2 Turbopump Overspeed Trip 67
33 Turbopump Test Report Form 68
34 Cutaway View, Main Oxygen Valve 74
35 V-2 Steam Generating Plant 75
36 Apparatus Arrangement for Heat Exchanger Test 82
37 Heat Exchanger Test Report Sheet 82
38 Alcohol Preliminary Valve Cutaway View and Test Report Sheet 84
39 Lox Valve Test Report Sheet 88
40 V-2 Steam Plant Test Report Sheet 91
41 Alcohol Tank Test Report Form 93
42 Oxygen Tank Test Report Form 94
43 Propulsion Unit Calibration Stand 97
44 Stainless Steel H2O2 Tank 97
45 Propulsion Unit Calibration Control Desk 100
46 Schematic Diagram, Control Desk Electrical System 100
47 Float and Recorder Used in Measuring Flow from Calibration Stand Tanks 101
48 Interior View of Calibration-stand Instrument House 102
49 Propulsion Unit Installed in Calibration Stand 103
50 Schematic Diagram of V-2 Propulsion Unit Calibration Arrangement 104
51 Combustion-pressure Simulating Orifices (A) 105
52 Combustion-pressure Simulating Orifices (B) 105
53 Oxygen Pump Performance Curve 108
54 Alcohol Pump Performance Curve 109
55 Pressure Regulator Correction Curves for Mixing Ratio Changes 113
56 Pressure Regulator Correction Curve for Total Flow Error 112

viii
LIST OF ILLUSTRATIONS (CONT'D)

Figure Page

57 German- and American- made Gyros Used at WSPG 121

58 Gyro Test Table and Panel 125
59 Gyroscope Tilt Stand 125
60 Computer Output Windings 131
61 Schematic Diagram Output Transformer 132
62 Inductance in Windings 1 and 2 versus Resistance in Windings 5-7 and 8-10 133
63 Volt-ampere Curves for Control Windings 1-2 with Other Windings Open 135
64 Volt-ampere Curve with 40 Volts, 500 Cycles on Windings 3-4; Normal Load on
Windings 5-7 and 8-10 134
65 Servo Test Sheet (WSPG) 138
66 Servo Test Sheet (WALDORF) 138
67 Servo Being Tested with Ldad Test Panel 139
68 Time Switch and Test Panel 140
69 Inverter Regulator and Test Panel 141

PICTURE CREDITS
The illustrations noted below are published courtesy of U. S. Army Ordnance, White Sands Proving
Ground: Figures 4, 6, 7, 8, 9, 11, 22 through 26, 29 left, 31, 32, 34, 35, 38, 43, 44, 45, 47, 48, 49, 51, 52, 57,
58, 59, 68 and 69.

LIST OF TABLES

Table Page
I V-2 Launchings 6
V-2 Propulsion Unit Calibration Data facing 110
HI V-2" Flight Data 116
IV Data from Special Tests on V-2 Propulsion Unit 118
Normal Portion of Special V-2 Propulsion Unit Tests "
119
VI Special Portion of Special V-2 Propulsion Unit Tests 119
VII Laboratory and Rocket Tests, German and American Computers 130
VIII Laboratory and Rocket Tests, American-made V-2 Computers 131
K Test Results, German and American Transformers 132
OBJECTIVES

1. To obtain experience in the handling and firing of large missiles:

The V-2 missiles offered opportunities not only for the training of military personnel
but also for obtaining experience which would be of value in the design of ground equip-
ments for future missiles.
2. To provide vehicles for experiments directly concerned with the design of future missiles:

3. To provide vehicles for operational tests of components for future missiles:

4. To obtain ballistics data:

The V-2 missiles offered opportunities not only for obtaining ballistics data for high-
altitude trajectories but also for developing and proving various means for tracking
and measuring these trajectories.

5. To provide vehicles for upper atmosphere research projects:

At the time this program was initiated the V-2 was the only rocket in existence capable
of carrying heavy payloads to a high altitude. This high-load capacity offered agencies
the advantage of being able to conduct many related high-altitude experiments simul-
taneously. Since the V-2 provided many cubic feet of space for experimental apparatus,
it was not necessary for the experimental agencies to expend the time and effort neces-

sary to reduce the size of their apparatus to meet the dimensions of smaller missiles.
GENERAL SUMMARY

A large quantity of equipment and components for the German V-2 (A-4) missile was captured in the
European Theatre of Operations in 1945. Many trainloads of this material were shipped to Las Cruces, New
Mexico, for use at the White Sands Proving Ground (WSPG).

The General Electric Company was assigned the task (as an addition to the existing Hermes Project) of
firing anumber of V-2 rockets which were to be constructed from the captured components. The work was
defined more specifically as follows:

"In general, this work will consist of the firing of a number of German rockets Also
included is the necessary work in connection with the actual firing such as transporting, handling,
unpacking, classifying (identifying), reconditioning and testing of components of German rockets
as well as assembling and testing subassemblies ahd complete rockets, manufacture of new parts,
modification of existing parts, conducting special tests, constructing temporary test equipment
not available at the Proving Ground, procuring and handling of propellants and supervision of the
launching of rockets."

The captured materiel was unloaded at Las Cruces in August, 1945; with the assistance of military
personnel and German specialists, the first rocket was static fired on March 15, 1946 (Fig. 1 and 2). The
first rocket was launched on April 16, 1946 (Fig. 3).

By June 30, 1951, the General Electric Company had supervised the construction, test and launching of
67 rockets using V-2 components. General Electric Company participation in the V-2 program was termi-
nated by agreement on June 30 ; 1951.

Fig. 1 V-2 Mounted for First Fig. 2 First V-2 Static Test
Static Test at WSPG in Progress at WSPG
Fig. 3 A V-2 Rocket Shortly After Launching at WSPG
PERFORMANCE SUMMARY

1.1 GENERAL STATISTICS

During the program a total of 67 V-2 missiles were launched (Table I). In three of these missiles, the
steering system was modified to meet the requirements of the experimental agency. All three showed
steering trouble during flight. Since these modifications placed the missiles themselves in the experimental
category, they are not included in the performance data that follow.

Fifty percent of the remaining 64 missiles performed normally. In arriving at this percent, any missile
which showed any malfunction was counted as a failure, regardless of the adequacy of the trajectory or of the
success of the experiments carried. Actually a number of missiles, thus classified as failures because of
some known malfunction, were useful from an experimental standpoint. For example, missile 30 failed to
steer properly but reached an altitude of 99 miles. Experimental results were reported as excellent. The
average altitude of ten of the missiles classified as failures was 80 miles.

To meet the needs of the experimental agencies, 71 percent of all missiles launched were above design
weight. The empty weight of the standard V-2 was 8800 pounds which included 2200 pounds of payload (war-
head). The average empty weight of all missiles launched was 9218 pounds. This represented an increase
of 19 percent in terms of payload. As the program advanced the experimental agencies progressively took
advantage of the "work-horse" ability of the V-2 as shown below (these figures do not include the Bumper
missiles).

YEAR POUNDS OF PERCENT

ADDED WEIGHT ADDED
(ABOVE 8800 LB) PAYLOAD
1946 150 6.8
1947 400 18.2
1948 527 23.8
1949 1036 47.0

To accommodate experimental needs, there were major contour modifications on 24 missiles. Thus,
36 percent of all missiles launched departed from the design contour in some major respect. There were
minor modifications around the nose tip on 11 other missiles. The major modifications became more
prevalent as the program advanced.

YEAR PERCENTAGE OF MISSILES

WITH MAJOR CONTOUR
MODIFICATIONS
1946
1947 7
1948 41
1949 75
1950 80
1951 100
|

«^
II
Sg °
s S 15 *

i §* «o
M 1<

4)

It ;2«
81 |
o « o fas P
s e S3 p ft ||S
'CM
c
o

fl I
bo
•S 13 I I
ai

If SiH .2
I 4> bp a,
« 55
8 I 8 £
<u

55 a: 55

So O OOOOO <5 w to 5o ss ,5 5 5
UIU « V
cc !c5 big * c
3ix3rt-O
Of u £ Q. I-
^ 36 J 3 3 3 13 | a
3 3 J CQCQiCOK i
« « 3 3 5 3 3 3 .§ 3 3
qcc£££ mHwmK
•£ -a -s *S
£ H t
!

ffl

fl
2 v
bo * oo oo
OOOO OOOOI

5 = 'OOOO
U^U U^ldU
ot-mo ocMinoo oc-moo oooin
U U W> 5 £ £ £ W
nOtOH ^cOcDi-Hi-h* WCOOCQt- M N H H

ooo rrooioo o^rooo

O CD CO O ~* I

o C- irt o CO o
m O ! ! ! ! 3 OlOCOH
<
H
>
.H CM
bp

c E °. °. °. ss o ia ©m m ©m ocoomo
<£>

£ P
ta o oo
CO in

Or-OiO IfiKOh,
Ed 3:

01 s s s s
a,
ss
a. a. a,
s

HHNO -h O C- K5

t- c- t- *T v «. oo •» oo rr-<r
<»• •«•

A to 6o>n oo ct> o in in -. omo o

ooioStMoo eNicnicn n p-
«
i

C- r- CO

u OOU -I J se
Jjj o _j j j o j _ j j oj
a«e.3H '
as K S
On" Ed" Ed OSK Pi W W S BS W J OS W S Eii Oh' Oh"

S z« 5 § 2" < Z* O <"

o z' z" < d 6 < z d d co z" d < < <i

2| CO CO O -Hi

31
I a o3
S 8 8 c g

S S g| S58' 8 £ So'8
. 5 3 SS **« 2 sb I SIS 3™
I 1 * u HI
553
* ££ *
22
*i*
I
> I J- I gj |||
! ft IS] 111 1 II »|ljf
£ 13.sfS.sf 32:2: S a ^S | 1
ft. H S3 >B ft. ft. 11,
-a
H WWOJE-otH ji -a
?g

S O w o «n © O <» O O w56 tS 3
_
_ _
II
<-> nhHKH ibPcbcqS co H u< u. m w m P m co ££mwm wwos
abo "8= to e- co -< in w
eg t- 00 » o ho*b •"
j ] ] ] ]

5 <3 » 10 »' 1 in cs in eo to co' co' 1

j

1 c-* 1 co 00 co to'
j

1
[

1 c-' 1
[

1
|

1
j

1
|

8
I
fa,*S
t,
^oo
° p
co co oooom oomoo ooooo Jo ] o ooo
V
[

u "3 c- r- c- *r i> t> i> t-^ ca i>i>CNit^t^ c-^i>t>c-"i> i i t- i it- t-^ c-^ r-"

1 of
OCINH
inoifi ooooo oowmci o c- o o m co [
ro J^h co co
d M N ^'
[
*
CO t* CD (N -S"* *•* co o" o" «" o oi rH* 1 1 _* 1 o* o" o o"
w
bo 001000 oomom rr ot *r o in m o co in h m e«a -<r

©© w nm
| J

§ eo ca o? m" o* r-* ~ esi o' o* —* t-" o* t-' d J

i co i o' o' o ©"

•u
3
*j
^« co
o ^ co n
en o
©
ojtocot-io own
t- c- a> co o co - 10 co
co
co <* c- ct>
en c— c- en
m iicoi
i i in b — < «o^
iu jj

*i o o cm m o ooinoo
OlflNNCO CD CO N * O
© o o o c-
(Difi IDlrt O
co
CO (O
om om
o CO tN
i i ©
© m O worn
co o
CN
5 NCNCOCO «N t-OCft^TOO V CO © <«• *** t- OS N f CO I

I
I

I (O h CO CO CO
'J'
tsi
>
<« eg o 10

c^, m co $£,
°- ^ ifl
S ifl ifl
S in in co co o t-io VOO en t- © in o
V d ^' do
[
J

"eg in co" co ^^rt 1060 © in oi cd m* t- -4* in co in in i ci

^§2
i
q^^J-
3 H co co" © ^
CQ co in co co
11*
:>.*-
wD-af
£.
co cn <n tN
oicoNino m co^mnn
in co co ifl co
h o h o m co
r- co m o t-
co ** w o
o>c*hhO in in co t- co
c- cd co o en
t-
© tN
en co
h
© O ^" O ~H CO CM in CO C- 0> m CD CN N *• CM © mcDCDCOO)
Ifi Cft 1-1 t-
o> o co en o en en © o en co co o co ai o '•J*
c<u co co co co co co co a) a> oo en en if :!
O co

sssss ssl^s ^|sss

< < < < B.<^B,
sss|| ss|ss sss
p,<p B.<!<:„o <<„»<& 0.<Q, 0,
s a
t-moo^* inoocNO^* ^
D, <
I
-* vint-wifl a conno aimoN<r cococo
^ 9. r. 9.
inr-coaot-
«?. ?» <?. oon ~
9.
coos^H^eo h 9. r. r.
n in con
•?. «^
r. 9. 9.
taoiHH ^^n "H
^ r. ?. r.~
6" 9. ifl
oococm
01 to
to S> o S-
oo
«i^-^^i^
coo) O) 01 ck en en oioi
e»
ai o
^ *ifl ooom^ o D > §
«J
co co co
co^co^^" ^j"v^«oi^r
o>
V* i
*-<^h
inininiin nmin
i?
u icnio*-i
in — « co
^cBcot-^' — ,h «
icN~4Cq cn « cn m —
t i coai^cot-
^ ^r en ^ h co i^rco
shn l i|S
O —
i

cm -. cm cm co
,

co
_
i

cocnen
I i
i

^ „ h^*
^ CM
mm co^^mto
i I I I

oo>hhn
i

—i CM
i

i I

c-c-co^^h
I i
•

I III
cococo I d-ss
g Sq.So So
o j jj q o oJ oj ujj ujd
S kh'oHM u J k p) « S J bi u S S k u S « u u S oj k S k S
•«7 z ci oj ci ci ti ca z iJ 6 < co 6 ci < <z m < z d d < co z' < z" «ti § s
«il o n« m co oot-
^^
og «3nB»
eo'eoj.,' * N in co
*^>*'*P h o cot-
•fl-mSv'cr
' ' Nononn DDiniom
n*«in«i J.J.-h.-i'V f
in
in
in in
n 1a
o CQ,
«Zs
CO (B CO CQ CQ CD
I
, v
&
1.2 RESULTS OF PROPULSION- UNIT CALIBRATION
1947 the calibration of propulsion units (Fig. 4) was started at WSPG. The data given below offer
In late
some improvement in performance which resulted. It should be noted that the data does not in-
idea of the
clude any correction for differences in take-off weight, trajectory or missile configuration. The lack of such
correction should not introduce any serious errors in the averages.

Comparison has been limited to missiles which: (1) burned to propellant exhaustion and (2) gave no
evidence of malfunction. Records were available on 19 missiles which satisfy these conditions. The follow-
ing data are based on eight missiles which were not calibrated at WSPG and on 11 missiles which were
calibrated at WSPG.
ALTITUDE EMPTY WEIGHT
Miles Pounds
Calibration by German test records 89.1 9064
Calibration by WSPG test 88.5 9764

It is noteworthy that the units calibrated at WSPG were able to carry an average of 700 pounds additional

load while the average altitude remained essentially constant.

The WSPG calibration also produced much more consistent results. In terms of altitude the deviation
from average was +18 and -11 percent (79 to 104 miles). Using German test records, the deviation from
average was +30 and -29 percent (63 to 116 miles). For many experiments this improvement was of con-
siderable value.

Fig. 4 Propulsion Unit Calibration in Progress

8
1.3 FAILURES
1.3.1 Definition

arriving at general performance figures the missiles are divided into two classes, "Normal" and
In
"Failures."A missile was classified as a, failure if there was any known malfunction or if the missile
performance indicated the probability of a malfunction. This classification is based on performance as a
missile and not as an experimental vehicle. As previously noted, a number of missiles, here classed as
failures, produced very useful flights from an experimental viewpoint.

1.3.2 Investigation

Based on the above definition, 32 missiles were classed as failures. In each case, a thorough in-
vestigation was conducted in an attempt to identify the cause of failure. In general, the results of these
investigations were disappointing. In most cases it was possible to determine the general nature of the
fault but the exact origin was seldom located with certainty. A detailed report of each failure is included
in Appendix D, section 2, this report.

There were three primary sources of information concerning faults in flight: telemetry, recovery,
and optics. Both recovery and optics contributed very valuable information on occasion, but these
occasions were somewhat rare. Telemetry provided a large percentage of the total information.

Normally, six telemetry channels were assigned to missile performance. Although four of these
channels were usually sub-commutated to monitor twelve functions, only a small fraction of all the pos-
sible sources of trouble could be covered. On no occasion were the channel assignments and timing
ideal for giving a complete picture of the malfunction.

1.3.3 General Distribution of Malfunctions

In view of the limited information available, it was seldom possible to identify the exact component
which failed. However, the data which follow should present a fair picture of the types of failure en-
countered.

Over the entire program the missile failures were divided almost equally between steering and pro-
pulsion (steering 14, propulsion 15, and miscellaneous 3). It should be noted, however, that this division
was not typical of all phases of the program. After the twenty-seventh launching, most of the German-
made components of the steering system were replaced by components of domestic manufacture. The
main exception was the servo, which was completely overhauled and equipped with a US-made motor. In
addition, all German cables were replaced with new cables made of stranded wire. These changes re-
sulted in a pronounced improvement in steering performance. Prior to these changes, 46 percent of the
missiles showed steering trouble. After the changes only five percent showed this trouble.

1.3.4 Time of Failure

A
review of time-of-failure data offers some indication of the factors which caused the failures. For
example, if a missile was in trouble from the instant of lift, there is a strong presumption that the initiat-
ing conditions were present, or at least set up, prior to lift. The factors which would allow such condi-
tions would include stray potentials, inadequate test and personnel errors. On the other hand, if a missile
performed normally for a number of seconds of flight, there is a strong presumption that any subsequent
failure was produced by in-flight conditions. It seems probable that vibration accounted for a high per-
centage of in-flight failures.

In most cases the time of failure has been established with reasonable accuracy. It was not always
possible, however, to detect a steering fault in the first few seconds of flight. The following data are
based on the assumption that any steering fault detected within the first five seconds was present at lift.
NUMBER PERCENTAGE
OF TOTAL
FAULTS AT LIFT
Steering 6 19
Propulsion 3 9
Miscellaneous 3 9
TOTAL 12 37

FAULTS IN FLIGHT
Steering 8 25
Propulsion 12 38
TOTAL 20 63

As shown in the table below, the in-flight failures were rather uniformly distributed throughout the
powered flight. There is a slight suggestion of concentration in the speed-of -sound zone (20 to 30
seconds) but these few samples are not sufficient to establish this point.

TIME OF FAULT TYPE OF FAULT

Seconds
8.0 Propulsion
10.7 Propulsion
13.0 Propulsion
13.9 Steering
15.5 Propulsion
22.0 Steering
24.3 Steering
25.7 Propulsion
27.0 Propulsion
28.2 Propulsion
28.5 Propulsion
33.0 Propulsion
36.6 Propulsion
38.0 Steering
43.4 Propulsion
45.2 Propulsion
47.5 Propulsion
48.4 Steering
56.4 Steering
57.5 Propulsion

1.3.5 Allocation of Faults

As stated previously, the precise origin of failure often remained in doubt. Therefore, any allocation
of faults is subject to question. The following list is an estimate based on available information.

SOURCE OF FAULT AT LIFT IN FLIGHT

Components 4 7
Wiring 7
Piping 1 5
Stray potentials 2
Personnel errors 3
Unidentified 1 2

10
1.3.6 Effect of Low Temperature Due to Liquid Oxygen
It was general practice
minimize the time during which lox was present in the missile, but many
to
factors introduced unavoidable delays. The time from start of lox loading (Fig. 5) to take-off varied
from 80 to 536 minutes. There has been considerable interest in the relation of this time to the probability
of a successful flight.

Reliable time figures are available for 40 missiles. These data indicate that there was no appreciable
change in percentage of successful flights for the period from 80 to 120 minutes. Beyond the two hour
wait-time, there was a marked drop (65 to 39 percent) in the percentage of successful flights.

The value of these figures is limited since: (1) forty samples are not enough to establish any definite
data and (2) all failures were counted, although it is highly improbable that they were all due to low-
temperature effects. In view of the limitations, no definite probability curve is justified. It would appear,
however, that a wait-time time in excess of two hours should be avoided whenever possible.

t MEILERWAGEN

Oo EXHAUST

Fig. 5 V-2 Being Loaded with Oxygen

11
MISSILE COMPONENTS

The following is a brief summary of each major V-2 component covering quantities received, condition,
performance and other general comments.
Detailed descriptions of the components may be found in reports by Project Hermes(l) and the British
Special Projectiles Operation Group(2). information on the use of these components at WSPG is noted in tne
Appendix of this report.

WARHEAD
Quantity : Approximately 50
Condition : Good, no repairs required
Performance : Good
Comments Very few of the German warheads were used. For the housing of experimental equip-
ment the German units were unnecessarily heavy and offered very poor access. Many
of the experimental warheads were made by the Naval Gun Factory and supplied to the
program by the Naval Research Laboratory. There were also a number of special
warheads constructed for special applications.

CONTROL CHAMBER (section containing missile control equipment)

Quantity Approximately 115
Condition Varied from good to very poor, considerable repair required
Performance No known cases of failure

MIDSECTION
Quantity Approximately 127 sets
Condition Good to poor, moderate repairs required
Performance No known failures

THRUST FRAME
Quantity Approximately 100
Condition Generally good, few repairs required
Performance No known failures

TAIL SECTION
Quantity Approximately 90 in usable condition
Condition Good to poor, considerable repair required on many
Performance No known failures
Comments Toward the end of the program it was necessary to have eight units built in the U.S.A.

PROPELLANT TANKS
Quantity Approximately 180 for each reactant
Condition Generally good, some welding repairs required. Leaks were usually found at support
attachments requiring additional rivets.
Performance No known failures
Comments Workmanship, particularly welding, was excellent, practically no deterioration noted
after approximately seven years.

13
TURBINE AND PUMPS
Quantity Approximately 200
Condition Generally excellent, only minor repairs were made.
Performance Good, one known failure where bearing seized.
Comments Oxygen seals (three, steel, 120-degree segments with garter springs) required con-
siderable attention; cleaning and lapping were required. Turbine blades were
exceptionally sensitive to steam temperature; much cleaning was required after
calibration runs.

BURNER
Quantity Approximately 215
Condition Varied from good to unusable. Included were obsolete types, unfinished and untested
burners and rejects.
Performance Good, no known case in which burner failed. Most burners were recovered and in-
spected carefully. Evidence of one small defect which did not ruin flight was found.
There was possibility, unsupported by evidence, that small cracks in head might have
contributed to one or more tail explosions.
Comments Although burners were stored in the open for years, there was remarkably little
evidence of deterioration. Considerable effort was required, however, to remove
scale and accumulated dirt from interior of burners. Calibration tests (Fig. 6)
were adequate to establish proper mixing ratio and flow rate. No known method,
other than static firing, was considered adequate for checking the effects of pressure,
temperature and vibration.

aM mmV t

tM IP

11 f 91 ^rlFil If

J Mm Wit "1 P^^^k^t-^p^i —v 9* 1 —*/ / ^|V a

'^H • j J ^H

Fig. 6 V-2 Burners in Test-Preparation Area

14
HYDROGEN PEROXIDE TANK
Quantity Approximately 200
Condition Excellent, with quantities available, no repairs were required.
Performance Excellent, no known failures.
Comments These tanks were made of steel which was not compatable with hydrogen peroxide. To
prevent decomposition, the tanks were lined with a special protective coating. The
Germans suspected that age would cause this coating to crack and become useless.
One coated tank was used for 13 calibration runs. After these runs it was beginning
to show signs of "distress" but was not considered to have reached the danger point.

PERMANGANATE TANK
Quantity Approximately 200
Condition Excellent, with quantities available, no repairs were made.
Performance No trouble experienced

HEAT EXCHANGER
Quantity Approximately 100
Condition Good to poor, some required considerable repair.
Performance No trouble experienced

AIR BOTTLES
Quantity Approximately 115 sets
Condition Bottleswere generally good mechanically but rusted inside. Manifolds frequently were
poor condition; some manifolds were constructed.
in
Performance No trouble experienced
Comments Rust was cleaned from inside the bottles by rotating while partially filled with abrasive
fragments. Bottles, that were pressure-tested to destruction, showed a remarkably
consistent failure both as to pressure and type of rupture. The fail-point was approxi-
mately twice the working pressure.

STEAM GENERATOR
Quantity Approximately 200
Condition Excellent, very little repair required.
Performance No known trouble

VALVES (GENERAL)
In general, there were enough valves of each type available (lox vent and alcohol drain
excepted) to allow valves with serious defects to be discarded without attempting
major repairs. As described in Appendix A, many of the valves required minor re-
pairs. In addition, there was some concern about the effects of age on the rubber
seals used in some valves. It seemed certain that deterioration of the rubber would
eventually render those valve useless. With this in mind, a program was arranged
to develop, test and install new seals in a limited number of valves.

OXYGEN VALVE
Quantity Approximately 200
Condition Varied from good to unusable. After the best valves had been used, considerable effort
was required to provide acceptable valves.
Performance Generally satisfactory, but there were a few instances in which these valves were sus-
pected of having contributed to propulsion failures.
Comments It was definitely established that there was excessive leakage past the rubber seal of
this valve. A new seal was developed. In addition, a metal-to-metal seal also showed
excess leakage. Various corrective methods were tried but none were particularly
successful. Additional details are noted in Appendix A.

15
ALCOHOL MAIN VALVE
Quantity Approximately 200
Condition Varied from fair to poor, many rejected.
Performance No known trouble during flight.
Comments The most common defects were porous cases and rough stroke; for details see
Appendix A.

ALCOHOL PRELIMINARY VALVE

Quantity Approximately 190
Condition Generally good, although many required minor repairs.
Performance No known failures in flight. On one or two occasions the valve was suspected of closing
prematurely, but this was more likely to be a failure of the controls than of the valve
itself.

OXYGEN VENT VALVE

Quantity Approximately 80, it was necessary to manufacture additional units in US.
Condition Generally fair, adjustments and minor repairs were required.
Performance No known failures in flight.
Comments Details are rated in Appendix A.

GYROSCOPES
Quantity : Approximately 50, (two required per missile)
Condition : Varied from poor to unusable; many of the usable gyros required extensive re-
conditioning.
Performance : Generally good; although steering troubles were frequent during the time these gyros
were in use, the trouble was seldom attributed to the gyros.
Comments : The German gyro, in good condition, gave excellent performance. From a design
viewpoint it had two weaknesses of secondary importance. The torque-motor spiders
were of a material which distorted and produced binding when hot and the program
motor was slightly weak. For details, see Appendix C.l.
It was necessary to procure 140 additional gyros of domestic manufacture. These were
essentially copies of the German gyro with minor improvements and changes to US
standards. In performance they were fully equal to those of German origin.

MIX-COMPUTER (AUTO-PILOT SERVO AMPLIFIER)

Quantity : Approximately 70
Condition : Varied from fair to unusable; useful units required test, adjustment and replacement of
defective components.
Performance Poor, there were five flights in which the computer was suspected of causing steering
trouble.
Comments The computer failures were not attributed to poor design but rather to deterioration of
components. The tubes in particular had a high "mortality rate" under vibration
tests.
It was necessary to procure 80 computersdomestic manufacture. These were es-
of
sentially a copy of the German design US-made components. The only design
using
change of any importance reduced the maximum voltage within the computer. The
performance of the US computers in flight was excellent. For details see Appendix
C.2.
TIME SWITCHES
Quantity Approximately 350 (two types)
Condition Good, corroded solder joints were the principal defects.
Performance Excellent; no known trouble in flight.

16
SERVOS
Quantity Approximately 500
Condition Varied from fair to unusable. After the best servos had been used, it was necessary to
install new motors and completely overhaul the remaining units.
Performance Apparently good, while there were seven flights in which the observed steering trouble
could have been caused by a servo failure, there was, in each case, some other type
of failure which seemed more probable.
Comments For further comment on servos, see Appendix C.3.

INVERTERS
Quantity Approximately 600 (two or more required per missile).
Condition Good; with quantity available, little repair was required.
Performance Good, no known failures in flight. It was occasionally necessary to replace an inverter
at the launching site, but much of this trouble was attributed to dust and sand.

REGULATORS (FOR INVERTERS)

Quantity Approximately 600 (one required per inverter).
Condition Good, very little repair necessary.
Performance Excellent
Comments An excellent design, held inverter frequency very close over a wide range of input
voltage and a wider range of load. The unit had no moving parts and occupied little
space.

MAIN DISTRIBUTORS (MISSILE RELAY AND JUNCTION BOX)

Quantity Approximately 70
Condition Varied from good to unusable. Some units were unwired. All units required modifi-
cation.
Performance Generally good, no known failures, although several malfunctions could have
of the flight
originated in the distributors.
Comments The majority of the German relays were of good design and workmanship and showed
little response to vibration. The wiring showed good workmanship but was subject to
breakage because of the use of small, single-strand wire. Electrical connectors were
fair.
It was necessary to manufacture approximately 30 distributors. These were essentially
copies of the German design, although it was necessary to substitute some relays of
US manufacture. The most important change was the use of stranded, instead of
single-strand, wire.

ELECTRICAL CABLES
Quantity An adequate supply for 100 missiles
Condition Fair, connectors and workmanship were good but small, single-strand wire was used.
Performance Questionable, it is probable that some of the early missile failures were caused by wire
breakage.
Comments The use of the single-strand wire appeared so questionable that all main German cables
were scrapped in 1947. Stranded wire was used in new cables.

JET VANES
Quantity An adequate supply for 100 missiles (three types received)
Condition Varied from good to unusable; many vanes were rejected in test.
Performance Good, there were only two flights in which there was any suspicion of vane failure.
Comments Good performance could be expected from properly tested and protected vanes but the
percentage of rejected vanes was high. Detailed information on the jet vanes is pre-
sented in Appendix A.

17
TEST, REPAIR AND ASSEMBLY

3.1 GENERAL
There appears to have been a fairly widespread impression that many missiles were
received virtually
complete and ready for flight. This was completely erroneous. No missiles were
received in anything re-
sembling a flyable condition. If complete missiles had been received, the first step
would have been to dis-
assemble them so that the individual components and subassemblies might be tested
properly (German ex-
perience proved that there was a large increase in in-flight failures when assembled
missiles were stored
for an extended period). All missiles launched under this program were
assembled (Fig. 7) at WSPG from
basic components.

All basic components were individually tested and inspected for performance
and condition prior to
assembly (detailed test instructions will be found in the Appendix). Repairs and
adjustments were made as
required, after which, tests were repeated. All basic components met
established test specifications before
being assembled in larger subassemblies or in the missile itself.

Larger subassemblies received complete tests before being assembled in the

missile. One of the most
elaborate of these tests was the calibration (Fig.
4, p. 8 ) of the propulsion unit. This test is summarized
under a separate heading and is presented in detail in Appendix B.

The completely assembled missile was given two over-all tests before leaving the
Missile Assembly
Building (Fig. 8). At the launching site, one over -all test was completed
prior to launching day. The same
test was repeated on launching day immediately before the loading
of propellants. It was an established
rule that no connections could be broken after final test.

Fig. 7 Missile Assembly Building at WSPG Fig. 8 Missile Assembly Building at WSPG
with Three V-2's Nearing Completion
V-2 in Foreground is Being Tested
19
3.2 P ROPULSION UNIT CALIBRATION
Propellant ratio and flow rate are influenced by three major components: the steam plant (Fig. 9), the
turbine-pump unit and the burner (Fig. 10). In Germany these components were tested separately. From
the test records it was possible to select the proper orifices with acceptable accuracy.

Test records were available at WSPG for a limited number of components but the results obtained by this
method were far from consistent. Two missiles, fired within a 30-day period, and having empty weights with-
in one-percent of the nominal design value, gave altitudes of 63 and 116 miles. This was probably the result
of changes in the components, particularly the burner, after the tests in Germany.

It was apparent that local tests would be required if consistent results were to be obtained. With the as-
sistance of German specialists, a calibration stand (Fig. 11) was designed and constructed. This equipment
was arranged to calibrate the entire propulsion system as a unit, contrasted to the German method of sepa-
rate tests on individual components.

As experience was gained, a number of modifications and improvements were made (Appendix B). The
final arrangement was considered to give very acceptable results. Flight performance and static burning
tests indicated that there was little to be gained by further refinements.

Fig. 9 V-2 Steam Plant Being Tested Prior to Assembly in Rocket

20
Fig. 10 Burners Being Prepared for Test Fig. 11 Propulsion Unit Calibration Stand

3.3 STRUCTURE
The primary structural components, exclusive of the warhead, are listed below. A description of each of
these components will be found in the "Backfire," report(^) Vol. 2, on the pages indicated. The weights given
are for the bare structural elements before the installation of any other equipment.

COMPONENT "BACKFIRE" WEIGHT

REFERENCES Pounds
Control chamber page 9, par 10 365
page 94, Fig. 56
Midsection page 10, par 11 985
(Two, half shells) page 93, Fig. 46 and 47
Thrust frame page 90, Fig. 39 155
Tail unit page 14, par 13 1110
page 14, Fig. 4 and 5

was not necessary to make major repairs to critical members of the various structural
In general, it
components. The wooden parts of many of the control chambers were in poor condition but the steel frame-
work, and particularly the important longitudinal members, were in relatively good condition. It was neces-
sary to construct new plywood crosses for about 75 percent of the control chambers. Since these crosses
provided an appreciable percentage of the strength of the chamber, care was taken to see that the new crosses
were fully as strong as the original German units.
21
The midsection shells (Fig. 12) were generally in fair condition. In many cases the skin was torn, and
patching was frequently required, but the ribs and longitudinal members seldom showed serious damage.
Since there was a surplus of about 25 percent from which to select, major repairs were not required.

In the case of thrust frames, there were barely enough to complete the program. Fortunately, the ones
available were generally in good condition. The alignment of this frame was a matter of some importance and
the construction of new frames or the extensive repairs of old ones would have called for the manufacture of
a fairly elaborate fixture.

A few extra tail units (Fig. 13) were received but several, out of the total, were damaged beyond repair.
Some obviously had been damaged in Germany prior to being crated, certain units were damaged by handling
and a few were damaged by wind and weather.

A high percentage of the damage consisted of injury to the fins (mostly the buckling of structural mem-
bers and misalignment of the fins). This type of damage could not be repaired properly without a large and
fairly accurate fixture to position the fins with respect to the body of the tail (any appreciable misalignment
would result in excessive demands on the steering system). In addition, special welding apparatus would have
been required, as well as facilities for the construction of the fins themselves. Consideration of the various
factors lead to the conclusion that it would not be feasible to set up the necessary facilities at WSPG. Conse-
quently, an order was placed with the Douglas Aircraft Company, Inc. for the construction of eight new tail
units. These consisted of the basic structure only. All the related accessories and equipments were provided
and installed by the Hermes Project.

All tail units, old and new, were tested at WSPG for fin alignment to avoid the use of any tail which might
overload the steering system. Since the fins of the old tail units had been positioned by fixture in Germany, it
was assumed that they were still located, with sufficient accuracy, at 90 degrees to each other. Consequently,
no elaborate check was made in this respect. The test at WSPG was primarily concerned with the detection
of any misalignment which might have occurred after the original assembly.

This test was made by means of a Wild theodolite which was capable of measuring angles to one second.
The theodolite was set up on a line through a pair of opposite fins, such as 1 and 3, at a distance of 20 feet
from the outer bottom edge of the fin to be examined. Selected reference points on the tail unit were used to
establish the long axis of the tail. The tail was then shifted until this axis was vertical and the "zero-point"
of the fin fell in a plane through the tail axis and the theodolite. Zero-point of the fin was defined as the point
at which the top of the fin merges with the body of the tail. With these conditions established, the deviation of
points along the edge of the fin can be measured in terms of angle. Readings can be converted to linear terms
if desired. This process was repeated for each fin.

Fig. 12 Control Cables Being Laid Fig. 13 V-2 Tailsections at WSPG

in V-2 Midsection
22
A survey work-sheet for the alignment test is shown in Figure 14. Point A on this sheet is the point
previously defined as the "zero-point" of the fin. Point B was located at the center of the knee of the fin.
Point C was the extreme outer and bottom (aft) point of the fin. Intermediate readings were taken when the
need was indicated. From the work sheet it will be noted that the maximum allowable deviation for any given
fin was ±6.75 mm at Point B. The total allowable deviation at this point for all four fins (with due considera-
tion to direction) was 9.75 mm. At Point C the maximum allowable deviation for any given fin was ±15.0 mm.
Total allowable deviation at Point C for all four fins was 20.0 mm.

C. Surveying of Tail Fins

Point A Point B Point C

Fin I

Difference between point A - B A-C

Distance of Transit ft. = mm
Difference in mm =

Distance of Transit in mm X .

Angle
,

57.3 x 3600 .

in sec

Fin III
Point A Point B Point C

Difference between point A - B A - C

Distance of Transit ft. mm

Difference, in mm =

Distance of Transit in mm X .
Angle
,

57.3 x 3600 .

in sec

Finll
Point A Point B Point C

Difference between point A - B A - C

Distance of Transit • ft. = mm

Difference in mm =
1

Distance of Transit in mm X . ,
1

57.3 x 3600 ^S 16
in sec

FinlV Point A Point B Point C !

Difference between point A - B A - C •

Distance of Transit It- mm

Differ e nee in mm = !

Distance of Transit in mm

Allowed difference in mm for each fin

Allowed Total Diff. for all four fins

Measured Total difference in mm

Tested by:

Fig. 14. Tail Unit Fin-alignment Work Sheet

23
Although about 50 of the standard German warheads were received, very few were used in this program.
There were two objections to the use of German warheads for the housing of experimental equipment: (1) they
were heavier (550 pounds) than required for the purpose and (2) they offered very poor access to equipment
mounted inside. Many of the experimental warheads were made by the Naval Gun Factory and supplied to the
program by the Naval Research Laboratory. In addition, there were a number of special warheads constructed
for special applications.

A series of modified V-2 Missiles, known as Blossom were launched with the Air Force, Cambridge Re-
search Laboratories, as the cognizant agency. In addition to the warhead of special construction, these mis-
siles were further distinguished by an increase in length of one caliber (approximately 65 inches). Standard
German midsection shells were delivered to the Air Force and the structural modifications were carried out
under the direction of the Franklin Institute Laboratories for Research and Development. The development,
design and construction concerned with this modification are thoroughly covered by a two-volume report,
F-2106, issued by the Franklin Institute Laboratories for Research and Development

Seven missiles were included in this particular Blossom series. All were heavily loaded. The lightest
had an empty weight of 9781 pounds (normal empty weight of the standard V-2 was 8800 pounds). The heaviest
had an empty weight of 10,683 pounds and the average was 10,232 pounds. Of the seven missiles, four resulted
in successful flights. The velocities and altitudes attained were about normal for the weights involved. There
was no evidence that the change in structure had any appreciable effect on the trajectory. It is of interest to
note, however, that all three failures were of the same type: tail explosions early in the flight (missile No. 32
at 10.7 seconds, No. 57 at 15.5 seconds and No. 52 at 8.0 seconds).

The forward portion of the V-2 structure was modified for a series of two-stage missiles known as Bumper
(Fig. 15). Over -all responsibility for these missiles was given to the General Electric Company and included
in the Hermes Project. The Jet Propulsion Laboratory of the California Institute of Technology was assigned
responsibility for the theoretical investigations required, the design of the second stage and basic design of the
separation system. The Douglas Aircraft Company was assigned responsibility for fabircation of the second
stage and detail design and fabrication of the special V-2 parts required. A general report on the Bumper
vehicle is given in General Electric Company - Project Hermes report R50A0501. Structural aspects are
covered in Douglas Aircraft Company Reports 12266 and SM-13178. The latter contains references to related
reports by the Jet Propulsion Laboratory.

A total of eight Bumper Missiles were launched. As far as V-2 performance was concerned, three were
successful, three failed and two were partially successful. In the case of the latter two, the program was some-
what influenced by a "sneak" circuit, but this was in no way connected with the structural changes. The three
failures conceivably could have been caused by excessive vibration due to the changes, but there is little evi-
dence to support this possibility. Two of the faults were of a type previously experienced on missiles of more
conventional construction. The third failure, a tail explosion at 28.5 seconds is of more interest. The only
other known tail explosions occurred in the Blossom series of missiles, which, like Bumper, involved major
structural changes.

Two missiles were prepared for launching at sea under Operation Sandy. The only structural modifica-
tion consisted of a moderate reinforcement of the fins. Since the fins carry the full weight of the loaded mis-
sile, it was considered desirable to add strength at this point to provide for any added load which might be
introduced by the ship's motion. Two triangular plates of 1/8-inch steel were attached to the bottom of each
fin, one plate on each side of the fin. The outboard edge of the plate was attached to the main vertical mem-
ber of the fin by self-tapping screws. The upper edge was attached to the bottom horizontal member of the
fin; the remaining edge was attached to the diagonal member extending from the bottom of the fin to the base
of the fin near the edge of the burner. A strap of 1/8-inch steel, about two inches in width, was welded to the
plate and extended along the main vertical member up to the bend in the fin. Since the ship had virtually no
roll or pitch after the missile was loaded, the effectiveness of the modification was not tested.

24
Fig. 15 Bumper Missile Launching Sequence

25
The two missiles for Operation Sandy were shipped from WSPG to Norfolk, Virginia by rail. Special ar-
rangements, regarding the careful handling of the shipment, were made in advance and a detail of military
personnel rode with the missiles. It appeared, however, that the shipment received rough treatment. When
the missiles arrived at Norfolk, it was found that both tail sections had been damaged. In each case, the body
of the tail had buckled a few inches aft of the forward end of the tail on the side which was down during ship-
ment. It is believed that the damage was primarily due to the type of support used at the base of the burner
and aggravated by the rough handling.

Bumper missiles 7 and 8 were shipped to Florida by military truck. The shipping cradle was exactly the
same type as used on Operation Sandy, with one exception: the rigid tail support located at the base of the
burner was replaced by a partially-inflated truck tire. The Army vehicles were driven with care and both
missiles arrived in Florida in excellent condition. This trip provided a clear demonstration that large mis-
siles of this type can be transported for long distances without damage, structural or otherwise.
German test records on new V-2 missiles indicated that the structure could withstand lateral accelera-
tions in the order of 3g. The structural components used at WSPG
were far from new and had been subjected
to weather and much handling. On this basis, some structural failures might have been expected. However,
on the favorable side was the fact that a much smaller program was normally used at WSPG. The typical
German program was 47 degrees, while that at WSPG was usually 10.3 degrees or less. In any event there is
only one known case, in this program, of a missile having been lost because of a structural failure. In that
one case, the failure is believed to have resulted from excessive heat due to a very unusual trajectory.

Although the usual program was 10.3 degrees or less, there were three flights for which the program was
70 degrees or more. In each case the standard German midsection, thrust frame and tail unit were used. No
attempt was made to provide extra strength at any point. All three missiles made this relatively hard turn
(70 degrees or more) successfully and without any evidence of trouble, structural or otherwise.
GROUND EQUIPMENT

4.1 GENERAL
The captured equipment included an adequate but limited amount of essential mechanical ground equip-
ment. Among the major items were Meilerwagens, launching platforms, lox trailers, hydrogen peroxide
trailers, alcohol pump trailers, and compressor trailers. In general these items were in poor condition and
required considerable reconditioning before using.

The situation with respect to electrical ground equipment was poor. Virtually every item required was
built at WSPG with the help of German specialists.In a way, this was fortunate since it presented an oppor-
tunity to learn the equipment in detail.

4.2 MECHANICAL EQUIPMENTS

4.2.1 Meilerwagen' 4 )
The primary functions of this device (Fig. 16) were to erect the missile on the launching platform and
to provide working platforms after erection. The erecting force was provided by a hydraulic system, in-
cluding telescoping pistons.

In general, this device was well designed for field use, but it had two objectionable features. First, the
hydraulic system required excessive maintenance, partially due to the exposed pistons. Second, it did not
always position the missile properly on the launching stand. In an attempt to insure proper positioning,
the Meilerwagen was coupled to the platform prior to erection. If the coupling was tight enough to position
the missile accurately, it was difficult to mesh with the Meilerwagen. If it was loose enough for easy mesh-
ing, the position of the missile might be unsatisfactory. Thus, either extreme, or any compromise in be-
tween, was likely to result in excessive erection time.

Fig. 16 Left to Right, Meilerwagen used for Transporting and for Erecting the V-
Figure 20 shows this Unit in the Vertical Position.

27
4.2.2 Launching PlatfoiW 5 )

This device was essentially an adjustable table mounted over a flame deflector. It provided facilities
for leveling and rotating the erected missile. The platform was simple and gave no trouble, however, two
features were added. Detachable plates were provided to prevent the missile from "walking" off the table
by action of wind gusts. In addition, provisions were made for discharging carbon dioxide into the burner
by remote control, in case of emergency.

4.2.3 Liqu id Oxygen Trailer 6)

This trailer Was used by the Germans to transport lox from railroad tank cars to the launching site.
Since lox was delivered direct to the missile(Fig. 5, p. 11 )by the supplier, the trailers were not required
for their original purpose at WSPG. Instead, they were used for the mixing and transport of alcohol.
These trailers had one outstanding weakness, the outlet at the bottom of the tank was so constructed that
it could not withstand severe road shock.

4.2.4 Hydrogen Peroxide Trailer^')

This trailer (Fig. 17) had provisions for warming the peroxide and for pumping it to a measuring tank
on the Meilerwagen. The peroxide was surrounded by a water jacket; water was heated by a gasoline
heater. Pumping was by a hand pump which was a constant source of trouble.

Since the heater was not required at WSPG and since the pump was poor, the use of the trailer was dis-
continued. Peroxide was ordered in exact quantities so that two drums made one charge. These drums
were hoisted to a platform about ten feet above the missile peroxide tank and the peroxide loaded by gravity
feed.

tsmm

Fig. 17 German Hydrogen- Peroxide Trailer Used to Load V-2 Missile at WSPG
28
4.2.5 Alcohol Pump Trailer ( 8 )

This trailer (used to load alcohol into the missile) contained a gas-engine driven pump, a filter and
metering equipment. After a short time the gas engine was replaced by an electric motor. The meter was
checked periodically and found to be accurate and reliable.
9)
4.2.6 Compressor Trailer
The Germans used this trailer to compress gas to about 3500 psi for use in the missile. It contained
its own silica gel dryers which were reactivated by heat from the exhaust of the compressor engine. It
also contained a rack of storage bottles.

When V-2 operations were started at WSPG, there were no other 3500 psi compressors available. With
the help of German specialists, these compressors were reconditioned and placed in service. They were
fairly simple to operate and performed moderately well considering their condition. Since practically no
spare parts were available, it was something of a problem to keep them in operating condition. Further,
there was not too much confidence in the reliability of their safety devices. When domestic compressors
became available, use of the German compressors was discontinued.
From later experience it was learned that the German dryers were not fully effective. Some regulator
troubles at the launching site were attributed to inadequately -dried air. A dew point indicator showed that
further drying was desirable. A series of tests indicated that the dew point could be brought down to about
-100°F by silica gel, but to do so required considerable more baking of the gel than was previously thought
necessary.

4.3 ELECTRICAL EQUIPMENTS

4.3.1 Firing Desk

Among the principal electrical items were

the firing desk (Fig. 18) and its associated relay box (Fig.
19). These were built at WSPG using German
materials for the most part. The desk and relay box were
built separately and connected by plug-in cables. This was suggested by the German arrangement in which
such division was desirable for portability.

Connections between the desk and the relay box were a source of considerable trouble. The contacts
in the German "Flak" plugs were good but there was insufficient space in the plugs for fanning out from
the cable to the contacts. In addition, the wires in the German cables were insulated with a thin cotton
cover which very readily became loose exposing the wire. The result was many shorts and grounds.
Later, a new desk was constructed with the relay panel built into the desk. This eliminated the intercon-
necting cables and a major source of trouble. The new equipment contained a much larger percentage of
US-made devices.

The German was designed with the intent of placing minimum requirements on the
firing equipment
operator. Many security were included to interrupt the firing sequence automatically in the event
circuits
of improper conditions. The desk was provided with a number of indicating lights to show the source of
such a hold. These provisions added to the complexity of the ground equipment. Approximately 120 wires
were required for control and monitoring of the missile; some 60 relays were used in the ground control
equipment.

4.3.2 Cast-off Cables

To keep missile weight at a minimum, the ground control wires were connected to the top of the mis-
sile by external cables (Fig. 20) which were cast off just prior to lift. The cables terminated at the lower
end in Flak plugs of the type previously described. At the missile end a "Stotz" plug was used. A special
cable, having exceptional flexibility, aided in the process of casting the plugs free.

The Stotz plug had adequate space for fanning the cables to the contacts, and the contacts themselves
were excellent. An appreciable amount of trouble was experienced with these plugs, but it should be noted
that they were subjected to very severe duty. The release mechanism was simple and effective. The only
unfavorable feature was the fact that its gear sector showed excessive wear.
Fig. 18 V-2 Firing Desk Built at WSPG

Fig. 19 Rear View of Firing-Desk Relay Box

Fig. 20 V-2 Being Erected for Launching

30
EQUIPMENT AND PROCEDURE MODIFICATIONS

Since the primary purpose of the V-2 program was to provide an experimental vehicle, not a tactical mis-
sile, no attempt was made to carry on development leading to changes in the standard German components.

The capabilities of the standard system and of the standard components had been established, with fair
accuracy, through some thousands of launchings in Germany. It was felt that the factors of time, economy and
reliability were opposed to any major departures from these standards.

When the program was extended to cover one hundred rockets, it became necessary to procure certain
components of domestic manufacture. These components included the gyros, computer, main distributer,
oxygen vent valve, tail structure and most of the piping (Fig. 21). In general, these components retained all
the basic features of the originals. Only minor changes were made to utilize domestic materials, methods
and standards. The one exception to this was the computer, and even in that case the change was not of a
fundamental nature. However, during the program experience brought out certain beneficial changes which
could be made without sacrifice of reliability.

5.1 CALIBRATION OF PROPULSION SYSTEM AS A UNIT

The procedure in Germany
consisted of testing the steam plant, the turbo-pumps and the motor as sepa-
rate units. From were selected by calculation. At WSPG the turbo-pump, the
the data thus obtained, orifices
motor and the associated piping were calibration-tested as a unit after being assembled (Fig. 21) for flight
(Appendix B.2).

Fig. 21 Propulsion Unit Assembly Nearly Completed at WSPG

31
5.2 OMISSION OF ALCOHOL TANK PRESSURIZATION ON RE-ENTRY
was pressurized after burning ended to prevent collapse on re-entry into
In tactical use, the alcohol tank
the lower atmosphere. This required three air bottles, plus associated valves and piping. Since air break-up
was desired on most WSPG flights, this equipment was omitted, thus providing additional space for experimen-
tal equipment and reducing the dead weight by about 65 pounds.

5.3 SECONDARY SOURCE OF AIR PRESSURE FOR VENT VALVE

The oxygen tank vent valve was held open by air pressure from a ground supply. If the ground supply
should fail with oxygen in the tank, the relief valve was the only safety measure remaining. For safety of
personnel and protection of equipment, provisions were made to allow a quick switch to a separate, secondary
source of pressure.

5.4 LOADING OF OXYGEN AFTER HYDROGEN PEROXIDE

In tactical operations, oxygen loading was started before hydrogen peroxide loading. This was a reason-
able procedure since oxygen loading required only about 12 minutes and provisions were available for topping,
if required, up to the time of lift.

At WSPG, oxygen loading required 35 minutes and topping-to-lift facilities were not available. Although
these conditions were somewhat less favorable than those for German launchings, the German sequence of
loading was used successfully for many launchings at WSPG.

During operations in Florida, condensation of moisture proved to be a source of some difficulty. Under
such conditions, there was a definite advantage in loading oxygen as late as possible. The practice of loading
oxygen last was used initially in Florida and later continued at WSPG. There were two secondary benefits.
First, there was less oxygen evaporation; second, there was less probability of having to remove the oxygen
because of delays. The latter was of some importance, since the removal of oxygen would cause condensation
of moisture within the oxygen system.

5.5 SUBSTITUTION OF ELECTRIC MOTOR FOR GAS ENGINE DRIVES

The alcohol pump and the hydraulic pump of the Meilerwagen were originally driven by gas engines. These
engines were replaced by an electric motor to: (1) reduce the fire and explosion hazard, (2) reduce maintenance
and (3) increase reliability.

5.6 SUBSTITUTION OF AUDIBLE LEVEL INDICATOR FOR FLOAT SWITCHES

Twofloat switches were used originally to give warnings as the level of alcohol reached the overflow
point.These switches were not entirely reliable because of the possibility of a sticky float, a loss of signal
power or an open circuit. On one occasion, the launching of a rocket was delayed when the failure of a float
switch resulted in an overflow of alcohol.

To avoid such occurrences, a warning device was developed which had the advantages of no moving parts,
no power supply and no electrical connections. In addition, it gave a continuous indication that the alcohol
level was rising.

The device consisted of a two-tone whistle which was accurately positioned in the tank at the time the
volume of the tank was measured. Two openings in the whistle intake caused a characteristic tone to be pro-
duced by the air which was being expelled by the rising alcohol level. When the alcohol level closed the lower
opening, the tone changed abruptly, giving a preliminary warning. When the alcohol level closed the upper
opening, the whistle ceased to function. This was the signal to stop pumping.

32
5.7 ADDITIONAL VENTING OF LOX VALVE CONTROL CHAMBER
During the latter part of 1948 it became evident that the composition seals on certain valves were show-
ing the effects of age.

One seal on the lox valve was particularly troublesome. This seal was intended to prevent leakage be-
tween the control chamber and the lox passages of the lox valve. Leakage of liquid oxygen into the control
chamber would result in pressure tending to force the lox valve closed. During flight, the control chamber
was vented through several feet of small diameter piping but this vent was too small to prevent a pressure
build-up in case of heavy lox leakage.

The development of a new seal was started but an interim measure was considered desirable. The bottom
plug in the control chamber of the lox valve was tapped to take one of the German 25-ton hydrogen peroxide
valves. This valve was pneumatically operated through a pilot valve whose coil was connected in parallel with
the coil of the pilot valve for the main alcohol valve. The outlet of the 25-ton valve was piped outside the tail,
through 18 inches of 25 mm
tubing. Thus, the vent area was increased by about 600 percent while the length
was decreased about 50 percent. It is believed that this modification was effective in preventing malfunctions
due to the leakage described above.

5.8 HYDROGEN PEROXIDE LOADING BY GRAVITY

In tactical operations inGermany, H2O2 was transferred from a bulk container to a loading dolly. From
this dolly it was transferred by a hand pump to a measuring tank mounted on the Meilerwagen and then loaded
into the rocket tank by gravity feed.

After the above procedure had been used at WSPG for a number of V-2 missiles it became apparent that
much better procedure could be devised. This consisted of hoisting the original
(with the facilities available) a
shipping drums to an upper level of the Gantry crane; from there, the H2O2 was gravity-fed into the rocket
tank. This method reduced the hazard, maintenance, loading time and cost.

5.9 CHANGE OF MISSILE POWER RELAY

In the original German design, the relay B2y( 10 ), which connected the main power battery to the rocket
bus, was de-energized in flight. Vibration tests showed that this relay, when de-energized, was particularly
susceptible to vibration. There was a decided improvement in this respect when the relay was energized.
Since even the momentary opening of "back" contacts on this relay could ruin the flight, and since there was
some cause to suspect that this had happened on one flight, the circuit was changed so that B2y was energized
in flight.

As added insurance that B2y would remain energized in flight, its seal-up circuit was paralleled by a
switch which was closed at lift. This switch was of a type that was virtually vibration-proof.

5.10 COMMAND-CURRENT CIRCUIT CHANGE

command currents to the servos were routed through milliammeters in the re-
In the original design, the
mote steering desk until the Stotz plugs were dropped. To do this, the direct connections from the computer
to the servo coils were opened when relay R2x was energized. Since the coil of R2x was energized from the
ground, the direct connections from computer to servos were supposedly re-established by "back" contacts
of R2x when the Stotz plugs dropped. If any of these failed to make adequate contact, due to dirt, vibration,
etc., the rocket would be launched without steering control.

To avoidthe possibility of trouble from the fault described above, a change was made in this circuit.
Resistors of 6.8 ohms were connected across the contacts of R2x, and the milliammeters in the steering desk
were replaced by millivoltmeters. The low value of resistance was selected to insure that there would be
negligible effect if it should be left in series with the servo coil. With this arrangement, no trouble would be
expected if R2x should fail to close any, or all, or its back contacts.

33
5.11 CHANGE IN GROUND CUT-OFF RELAY CIRCUIT
After experience had indicated the need, the Germans added an auxiliary relay, A90z, to permit cut-off
from the ground after the normal control cables had been disconnected from the rocket. This relay was not
included in the main distributor but was mounted separately. As originally installed, the closing of a "make"
contact of A90z would energize the cut-off relay, A9z, from the rocket bus. If vibration caused the-contact
of A90z to close, premature cut-off would result. Although there was never any direct evidence proving this
circuit to be at fault, some cut-off trouble was experienced. As a result, the circuit was modified so that
A90z received power to energize A9z from a ground source, only. Thus cut-off would not occur in flight if the
contacts of A90z should close.

5.12 COMPUTER CHANGES

The US-built computers contained no changes in basic circuits. However, minor changes were made to
lower the maximum voltage within the computer thus reducing the possibility of arc-over- at high altitude.

5.13 CHANGE IN STEERING TEST

procedure the response of the steering system was tested by tilting the missile through the
In the original
use of jacks on the launching platform. This method was slow and did not give strong signals. At WSPG the
gyro plate was unbolted from the rocket and tilted by hand. This allowed rapid motion of the gyros which
produced strong signals that were easily recognized. It should be noted that the cables to the gyros were made
long enough to allow this operation without disconnecting any of the final flight wiring.

5.14 ADDITION OF ROLL ERECTION MONITOR FOR PITCH GYRO

design there was no provision for monitoring the roll erection of the pitch gyro. The prob-
In the original
ability of trouble this source was believed to be too small to justify any elaborate means of checking. It
from
was found, however, that this function could be monitored in a very simple way. Low-energy coils of two re-
lays, located at the launching desk, were connected in parallel with the erecting coils of the gyro. The contacts
of these relays were used to control signal lights. Thus, normal drift-and-erection could be observed at the
control desk.

5.15 ADDITION OF ROLL ERECTION MONITOR FOR ROLL-YAW GYRO

was learned by experience that a short circuit (or an open circuit) could exist at any point between the
It

rollgyro pick-off and the Computer input without being detected at the steering desk. Under this condition a
rocket could be launched without roll control.

A very simple method of correcting this situation was used at WSPG. The pick-off voltages had always
been brought back to the control desk to energize relay Rl2p. By connecting a zero-center voltmeter across
the coil of R12p, it was possible to watch the normal drift-and-erection sequence. The absence of such action
would serve as a warning.

5 - 16 RELOCATION OF VANE-BALANCE POTENTIOMETERS

In the original design, the vane-balance potentiometers were spaced around the base of the
missile each
6par e hatCh eCUred by Seven SCrews A" excessive
amount of time was required for final vane
halTnl* o
ba ance Since ^ f
the operation
'

was performed after lox loading, it was desirable to reduce this

mum. To accomplish this, all four potentiometers were moved time to a mini-
to one location. The pot shafts were brought
out flush for ease of operation; the adjustment
openings were closed with clip buttons.These changes reduced
the required time by several minutes. B

34
5.17 ADDITION OF ISOLATION SWITCH IN PITCH CONTROL
The standard steering system included a circuit for synchronizing the pitch vanes. No means were pro-
vided for making the circuit inoperative during vane balance. The result was that drift in one vane would
cause similar motion of the opposite vane. It was not possible to determine readily which vane originated
the motion. There was, therefore, an equal chance that the drift would be stopped by introducing a balancing
drift in the opposite vane. Under normal conditions this would have no adverse effect on the flight. If, how-
ever, the synchronizing circuit should become inoperative during the powered flight, there was a strong prob-
ability that the rocket would roll.

To avoid simple isolation switch was installed so that the synchronizing circuit could
this possibility, a
be opened during vane balance. Thus, each pitch vane was adjusted independently for zero drift. With this
method there was a strong probability that roll would not result if the synchronizing circuit became inopera-
tive. It should be noted that the isolation switch was of the spring-return type so that the synchronizing cir-
cuit could not be left open by error.

5.18 NEW SERVO MOTORS

The powerof the German servo motor was low for the application. When full load was applied, the speed
would drop toapproximately 50 percent of the no load speed. With a new servo mechanism in excellent condi-
tion, the motor was barely adequate. Most of the servos received at WSPG had begun to show the effects of
usage. Cylinders, pistons and gear pumps showed wear and the leakage, particularly with hot oil, was appre-
ciable. This, coupled with poor speed regulation, produced a very marginal condition.

An aircraft-type motor of domestic manufacture was modified to replace the original German motor. The
current requirement of the new motor was somewhat higher, but the speed regulation was virtually flat from
no load to full load. In addition to the new motor, the cylinders, pistons and gear pumps were re-worked to re-
duce leakage. These modified servo units had more than adequate power.

35
MISSILE BREAK-UP INSTALLATIONS

Very early in the program, explosives were installed in the missile for the purpose of separating the war-
head from the missile body. Both range safety and the recovery of experimental apparatus, required warhead
separation. For range safety it was considered desirable to be able to alter the trajectory by destroying the
missile contour. Separation was essential to the recovery of experimental equipment. No recovery could be
expected from a missile that was not broken prior to impact. Figure 22 shows the crater and recovered parts
from missile 34 which landed without separation.
Various methods of separation were tried. In one of the earlier attempts, strands of primacord were se-
cured to the tail skin about one foot aft of the junction with the midsection. Recovery showed that the skin had
been cut as desired but the tail had not separated from the rest of the missile. It appeared that cables, piping
and other structural elements had held the tail in place.

The most effective means of break-up was found to be the destruction of the four longitudinal members of
the controlchamber. Initially, two pounds of TNT were used at each member. This proved to be more than
required. One pound of TNT per member was found to be adequate and this amount was used for most of the
missiles.

On a number of missiles the TNT was placed at the forward end of the control chamber and satisfactory
separation usually resulted. Later the TNT was moved to the aft end of the control chamber and even better
results were obtained. Explosives at the aft end of the control chamber not only destroyed the longitudinal
members but also blew out the forward end of the alcohol tank.
It was found that best results were obtained when the explosives were detonated before the missile re-
entered the denser atmosphere. There was photographic evidence which showed that the warhead could re-
main in position to impact when the explosives were detonated after the missile was nose-down in dense atmos-
phere. Unless experimental requirements dictated otherwise, it became standard practice to effect warhead
separation at 40 miles, or higher, on the downward leg of the trajectory.

With proper separation, the impact of the after end of the missile was surprisingly gentle. Figures 23,
24 and 25 show the midsection and tail of a missile which landed from an altitude of 76 miles after a success-
ful separation.Recovery of experimental equipment from such impacts was generally excellent. Figure 26
shows recovery of a Naval Research Laboratory spectrograph which was mounted in the fin of a missile that
had reached an altitude of 97 miles. A spectrograph of this type survived two impacts, with only moderate
repairs needed. However, the spectrograph was damaged beyond repair on its third impact.

Since the explosives were a possible source of danger to personnel, safety precautions were given care-
ful attention. The details of these procedures are given under "Safety," Section 9.

37
Fig. 22 Crater After Impact of Missile 34.
Recovered Parts Can Be Seen at Lower Right.
Fig. 23 Midsection and Tail of Missile 26 After Impact

Fig. 24 Tailsection of Missile 26 After Impact

39
Fig. 26 Spectograph Being Removed From V-2 After Impact

Fig. 25 Midsection of Missile 26 After Impact

40
V-2 OPERATIONS OUTSIDE WSPG

7.1 OPERATION AT SEA

In 1947 two missiles were assembled at WSPG for launching at sea under Operation Sandy. These
missiles were constructed essentially the same as those launched at WSPG, with only minor modifications
to meet unusual requirements. The tail structure was reinforced to a moderate extent and certain modifi-
cations were made in the steering system.

The missiles were transported to the east coast by rail. Arrangements were made to provide careful
handling in transit and a special detail rode with the missiles. Even after these precautions, the missiles
apparently suffered considerable damage enroute.

On board ship other agencies provided special equipment for holding the missile and for transferring
oxygen. One missile was launched on September 6, 1947. A discussion of the launching appears in Chief of
Naval Operations report OpNav P57-110, "Report of Operation Sandy," September 6, 1947.

7.2 OPERATION AT LONG RANGE PROVING GROUND, COCOA, FLORIDA

Two missiles were assembled at WSPG for this operation. Both were of standard Bumper (Fig. 15, p 25)
construction except for program modification.

The missiles were transported to Cocoa, Florida, by standard Army tractors and flatbeds. The cradles
were essentially the same as those used for Operation Sandy. The main modification of the supporting cradle
consisted of a partially inflated truck tire located near the base of the burner. This provided a non-rigid
support for the tail. Both missiles arrived at their destination in excellent condition.

In general, the conventional V-2 ground equipment was used. The one major change was in the type of
working platform used to service the upper levels of the missiles. The platforms were made up of standard
commercial iron pipe scaffolding of the type commonly used by painters. These assemblies were mounted
on casters. The scaffolds, extending to about 55 feet above the concrete pad had sufficient strength and
rigidity for the purpose. It is felt that this type of service platform should find increasing application in
missile work, since it: (1) can be disassembled into small, light sections and transported in regular military
vehicles, (2) can be moved into position (or removed) rapidly when assembled, (3) can be used with a variety
of missiles since the basic structure can be assembled to suit the needs of the job and (4) is relatively
inexpensive.

The first launching attempt was unsuccessful. Collected moisture caused a fuse to blow when the main
stage signal was given. This resulted in a complete shut down of the propulsion system. The missile was not
damaged, but it was necessary to return it to the hangar to be checked and dried properly.
Two steps were taken to reduce the probability of further condensation troubles: (1) silicone grease was
applied at vulnerable points and (2) the loading sequence was reversed to load lox after loading hydrogen
peroxide. These measures proved adequate in two subsequent launchings.

The first missile was launched July 24, 1950, and the second July 29, 1950. Results of these launchings
are discussed in Patrick Air Force Base - Long Range Proving Ground Division, Technical Report No. 1,
"Bumper Missiles 7 and 8," September 29, 1950.

41
PERSONNEL

8.1 GENERAL
In the early days of the program a tentative goal of 25 missiles was established. It
was thought that this
work could be accomplished by approximately ten General Electric Company engineers working with German
specialists and military personnel.

As the program advanced there were two upward revisionsin the total number of rockets to be fired.

In addition, as found that the original schedule of one rocket per week did not allow sufficient time for the
it

reduction and analysis of data. It became evident that the program would continue for a much longer time
than originally planned. For these and other reasons it was decided that G-E would arrange for a group of
sufficient size to assemble, test and launch missiles, on a one-every-two-weeks schedule, without
"outside"
assistance. In keeping with this decision, German specialists were gradually replaced by G-E personnel and
by the spring of 1947 all German specialists had been replaced.

By the end of 1947 the number of G-E personnel had reached thirty-four. From that time to the end of

the program there was no significant change in the number of personnel. The following was the typical
division of assignment:

Engineers & Engineering Assistants 17

Shop Personnel, Mechanical 11
Shop Personnel, Electrical 3
Clerical & Administrative _3
Total 34

should be noted that G-E activities at WSPG were not confined entirely to the assembly, test and
It

launching of the missiles themselves. Almost all of the missiles carried a large amount of experimental
apparatus. The General Electric group devoted an appreciable percentage of its time to the modification
of the missiles to accommodate this apparatus. Considerable time was also spent in the installation and
wiring of this equipment.

Experimental work associated with the missiles required the extensive use of telemetry equipment.
From September of 1947 to the spring of 1949, approximately 60 man-months were devoted to telemetry work.

Additional activities included assisting the Navy in the assembly and test of three special V-2 missiles.
These missiles were not complete in all respects but did involve (due to special requirements) the ex-
penditure of an appreciable amount of time by G-E personnel.

Two operations were conducted away from WSPG. In one operation, a V-2 was launched at sea from an
aircraft carrier (Operation Sandy). On another occasion two missiles were launched from the Long Range
Proving Ground in Florida. Each of these operations involved an expenditure of man-hours many times
greater than that normally required for the launching of a comparable missile at WSPG. Six G-E engineers
from WSPG accompanied the V-2 for Operation Sandy. Eighteen engineers assisted directly in launchings
at the Long Range Proving Ground.

Although it was general policy to schedule missiles at two-week intervals, there were never more than
17 missiles fired in any one year. This reduction from a nominal 26 missiles per year did not result from
a lack of personnel. The most important factors contributing to the reduction were: (1) experimental re-
quirements and (2) operations away from WSPG. It is believed that a force of 34 would have been adequate
to sustain a schedule of 26 V-2 missiles per year.

43
In a continuing operation of this type the turnover of personnel has an important bearing on
the effec-

tiveness of the group. This program was very fortunate in having an exceptionally low turnover.

8.2 SCHENECTADY WORKS PERSONNEL

The preceding comments refer to personnel regularly assigned to WSPG. It should be noted that this by
no means covers all G-E people contributing to the V-2 program. The group at WSPG received a large
amount of support from various divisions of the Company. The greater part of this support was provided
by the Aeronautic and Ordnance Systems Divisions and the Service Engineering Divisions in Schenectady.
Within these units there were men with full-time assignments to the V-2 program. In addition, a much
larger number devoted a part of their time to the program. The engineering assistance received from
these sources was of prime importance. These divisions also manufactured or procured the bulk of the
components, equipments and supplies which were required. On many occasions assistance on special
problems was received from other divisions, particularly from the General Engineering and the Research
Laboratories.

8.3 GERMAN PERSONNEL

The first German specialist arrived at WSPG in October 1945 and by March 194S German personnel
numbered 39. Thereafter this number was steadily reduced; by the spring of 1947 all Germans had been
replaced.

This group was undoubtedly selected with care since it included representatives of almost all phases of
German V-2 activities. Among these were scientists, engineers, technicians and manufacturing personnel.

From the manufacture, test and launching of several thousand V-2 missiles, they had acquired a wealth
of background experience as well as detailed "know how." This experience was passed on to General
Electric personnel as rapidly as circumstances permitted. The co-operation was excellent. The exchange
of information was limited at the start by language difficulties but this was overcome rapidly through the
efforts of the Germans.

There is no way to assign figures to the value of the assistance provided by the Germans at the start of
theprogram. It is certain, however, that the information received from this source was responsible for an
advance of many months in the program.

8.4 MILITARY PERSONNEL

Military personnel provided assistance in many phases of the operation at WSPG. In the early stages of
the program the use of enlisted men was confined, in general, to activities requiring limited skill and ex-
perience. This was necessary because of the extremely high turnover of personnel which resulted from rapid
demobilization. As the situation became more stable, it was possible to utilize military personnel on more
responsible work.

It was a matter of general policy to use as many enlisted men as possible to provide the maximum op-

portunity for training. In assembly and test, men were assigned to work directly with G-E personnel. After
a time it was possible to delegate, with a minimum of supervision, considerable work on ground equipments.
This work included maintenance of mechanical ground equipment, transport and erection of missiles, mixing
and loading of alcohol, storage and transport of hydrogen peroxide and assistance in the loading of peroxide.

In the course of the program, valuable assistance was received from the enlisted personnel. They, in
turn, received useful training and experience in the handling of large missiles.

44
SAFETY

9.1 HAZARDS INVOLVED

The possibility of injury to personnel is usually present during the test and launching of a rocket. By
design: (1) a maximum amount of energy is stored in a minimum space, (2) propellants are selected on the
basis of a violent reaction if mixed and/or ignited and (3) material safety factors are kept low.

Normally, the V-2 was loaded with: (1) 5 1/2 tons of liquid oxygen, (2) 4 1/2 tons of 75 percent ethyl
alcohol, (3) 370 pounds of 80 percent hydrogen peroxide, (4) four pounds of TNT and (5) 1.7 cu ft of air at
3200 psi.
9.1.1 Liquid Oxygen

Liquid oxygen, at -183°C, would produce severe injury if allowed to remain in contact with any part
of the body for an appreciable time. Ordinarily, this is not a serious hazard since lox is not usually
present in open containers. Care should be taken to see that clothing, such as loose boots, does not pro-
vide pockets where lox could collect. Care should also be taken to avoid allowing a body to remain in
contact with metal which is near lox temperature.

A more probable hazard is the possibility of clothing becoming saturated with oxygen gas. Such
saturated clothing will ignite readily and will burn so rapidly that severe injury would certainly result.

The presence of 5 1/2 tons of oxygen in proximity to 4 1/2 tons of alcohol is a general fire hazard.
More important, if alcohol and oxygen become mixed as liquids, a severe explosion may result.
9.1.2 Alcohol

A large quantity of alcohol is always a potential source of fire or explosion. Both of these hazards
are increased by the presence of oxygen. Otherwise, the handling of alcohol is comparable to the handling
of gasoline and introduces no additional or unique hazard to personnel.

9.1.3 Hydrogen Peroxide

of 80 percent purity requires very careful handling, if danger to personnel is to
Hydrogen peroxide
be avoided. Many substances will ignite immediately, if brought into contact with 80 percent H202- Per-
haps the greatest hazard to personnel lines in burns from clothing set afire by hydrogen peroxide. Gross
contamination by any one of many substances will cause a rapid decomposition which can reach explosive
violence if confined.

9.1.4 Explosives

Four pounds of TNT were normally carried in the forward end of the V-2 for range safety and re-
covery purposes. Detonation was initiated through a radio cut-off receiver which applied potential to a
detonator. The detonator set off the TNT through two or more strands of prima cord.

Accidental detonation of theTNT would undoubtedly cause severe injury to those working near the
forward end of the missile. The damage would not be confined to this area, however, because the TNT
destroys the forward end of the alcohol tank. Both fire and explosion could be expected to follow
immediately.

9.1.5 High-pressure Gas

Almost two cubic
feet of air was stored in the rocket at about 3200 psi. At this pressure, a break in
the piping or the fracture of a storage bottle would endanger personnel.

45
9.1.6 Experimental Equipment
On some missiles, the experimental apparatus introduced additional hazards. This category included
acid, grenades, ejection equipment and spin rockets.

On the Bumper missiles, inherent hazards of the WAC were added to those of the V-2. Any potential
danger was increased by the fact that much of the servicing of the WAC took place some 50 feet above
the ground.

9.2 SAFETY MEASURES

The potential dangers involved in the servicing and launching of the V-2 were recognized to some extent.
A determined effort was made to minimize any potential danger. It was obviously impossible to appraise any
particular hazard and draw a precise and accurate line between safety and danger. It was, therefore, a mat-
ter of policy to "lean-over backward" in the direction of safety in the hope of staying outside the questionable
zone. During the course of the program, safety measures were adopted or developed as rapidly as the need
appeared. Many of these measures were common practice industry. Others were developed to meet special
conditions.

9.2.1 Liquid Oxygen

For some missile hazards, there are mechanical aids which may be used to advantage. For liquid
oxygen, however, vigilance is the primary protection. This starts with assembly and continues until the
missile is launched. There must be assurance that: (1) the tanks and piping are mechanically sound and
have no leaks (2) the system is clean and contains no combustible substance or any substance which is
shock-sensitive in the presence of lox (many anti-seize compounds are dangerous in this respect), (3) the
system contains no mechanical parts which will become excessively brittle at lox temperature and (4)
there is adequate freedom of motion to allow for contraction. At the launching site, a careful watch must
be maintained so see that: (1) there is minimum exposure of personnel to liquid or gaseous oxygen, (2)
combustible materials are present only as absolutely required, (3) sources of ignition are kept to a mini-
mum (no smoking, soldering, drilling and use vapor-proof drop lights in or near missile) and (4) the
oxygen tank vent remains open at all times after lox loading starts.
9.2.2 Alcohol

Protection against injury to personnel by an alcohol fire or explosion consisted of reducing to a

minimum the possibility of: (1) the alcohol coming in contact with an oxidizer, (2) the accumulation of
alcohol vapor and (3) ignition. In the normal operating procedure, neither lox nor H2O2 was present
during alcohol loading. A drip pan was used at the alcohol tank inlet to insure that no alcohol could fall
within the missile. Firemen stood by to dilute and wash away any outside spillage. After loading, a very
careful check (using a vapor -proof light) was made to insure that there were no leaks within the missile.
An effort was made to keep possible sources of ignition to a minimum.
9.2.3 Hydrogen Peroxide
Hydrogen peroxide was stored in an air conditioned (by evaporation cooler) pill box in an isolated
location. Thermocouples were provided to determine the temperature of each drum before personnel
entered the storage room. Standard procedure called for daily inspection.

In the normal-German loading procedure, H2O2 was first placed in a pumping unit from which it was
hand pumped to a measuring container. From there, it was gravity-fed to the missile tank. To avoid un-
necessary handling(with the attendant possibility of contamination and spillage) a new procedure was
adopted. The original H2O2 shipping drums were hoisted to an upper level of the crane and placed in an
aluminum pan. From there, the H2O2 was gravity-fed directly into the missile tank. Thus, external
sources of contamination were reduced to one short section of hose.

46
The missile tank, itself, was a possible source of trouble. The tank, made of steel, was lined with a
protective coating. One tank began to cause some decomposition of H2O2 after being used 13 times at the
calibration stand. Inspection showed that very little of the protective lining remained. A series of tests
indicated that even with this used tank, the decomposition rate was not rapid enough to constitute a serious
hazard. Similar tests on an unused tank showed practically no decomposition. Although these tests indi-
cated that considerable reliance could be placed on unused German tanks, three safety measures were
practiced: (1) the protective coating of the tank was carefully examined, prior to assembly, (2) during
and after loading, the tank was checked frequently by hand to detect any appreciable rise in temperature
and (3) a quick-action valve was connected to the outlet of the tank, a hose from this valve led to a
tank containing enough water to provide about 15 to 1 dilution. Thus, the tank could be drained quickly
and safely at the first sign of trouble. During H2O2 handling and loading, full suits of protective clothing,
including helmets, were worn. These suits had one undesirable feature, being impermeable by design,
they allowed very little air to enter. On a hot day it was necessary to remove the hood at about ten-minute
intervals.

9.2.4 Explosives

system, the most probable source of trouble was the electrically fired detonator used
In the explosive
to detonate the through prima cord. A relatively small amount of energy was required to fire this
TNT
detonator. Several steps were taken in an attempt to prevent the detonator from being fired through in-
duced potential, electrostatic potential, leakage or direct contact. The circuit was isolated as far as pos-
sible. No terminal blocks were permitted in the circuit and connectors were kept to a minimum. A sepa-
rate, ungrounded battery of 67.5 volts was used. A time-switch contact held the battery circuit open until
20 seconds after take-off.

Another protective feature was a short-circuit across the detonator. This was completed through a
pull-away plug so that it remained effective up to the instant of lift. Heavy wire was used to keep the re-
sistance low and to withstand high current in the event of direct contact. It was intended that the current-
carrying capacity of the short would be greater than that of any other wires connecting to the detonator.
In addition, a dropping resistor was connected directly ahead of the detonator and short.

A was used at the end of the program. A small motor, controlled from
different type of short-circuit
the blockhouse, was used open and close a heavy-duty switch. This switch was connected across the
to
detonator by very heavy wires about six inches long. The position of the switch was monitored. It is be-
lieved that this switch arrangement was better in that it provided a minimum of resistance and a mini-
mum of exposure to direct contact or pickup.
Various safety procedures were followed during and after installation of the detonator. The installa-
tion was scheduled as late as possible to minimize the time of exposure. A radio-silence and no-switching
period was established before the installation was started. The wires to be connected to the detonator
were tested to ground with a megger and were tested with a voltmeter to see that there was no voltage be-
tween wires or from either wire to ground (before and after this test, the voltmeter was proved operative
by connection to a dry-cell). The open-circuit resistance between wires was measured and the value of
the dropping resistor checked. Resistance of the short-circuit was measured. The short was removed
and replaced to insure that the value measured was actually that of the protective short. Thereafter, the
access port to the short was locked to prevent tampering. All of these checks were witnessed by a second
person.

Immediately after these checks, the detonators were connected in the circuit. While being connected,
they were housed in a strong, steel safety box in which suitable vents were provided. Thus, no injury
would result if they should fire at the instant of contact, which was one of the more probable occasions.
After the connections had been completed, the detonators were removed from the safety box and taped to
the prima cord.

47
Beginning in December 1949, the detonators were housed in a destructor block which offered added
protection. This device was developed by the Naval Ordnance Test Station. The design was such, that
when the destructor block was in the disarmed position, the firing of the detonators would not detonate the
prima cord and TNT. It was armed by withdrawal of a safety wire when the cast-off plugs fell. The de-
structor block was regarded as an additional safety measure and not as a replacement for those measures
previously used.

At times it was necessary to return to the missile for further work after all normal launching
preparations had been completed. On such occasions, the first step was to remove the detonators from
the prima cord and to disconnect them from the blow-off circuit.

9.2.5 High-pressure Gas

general, the high-pressure lines within the missile were of small diameter and adequate wall
In
thickness to offer a reasonable safety factor.

Since the strength of the storage bottles was less obvious, three bottles were tested to destruction.
There was remarkable consistency both in the failure pressure and in the type of rupture. The failure
pressure was nearly twice the working pressure and failure was by a split in the tank rather than by
fragmentation.

Prior to assembly, all storage bottles were tested hydraulically at 4500 psi. Although these tests
gave moderate assurance of safety, one further safety measure was followed. The pressure in the missile
was not brought above 2700 psi, until the missile was cleared for launching. On occasion, it became
necessary to return to the missile after the bottles were charged to 3200 psi. This was undesirable, but
could not be avoided.
Experience leads to the conclusion that flexible hose offered greater safety than copper tubing for
temporary high-pressure test connections. Copper tubing tends to work-harden after a few bends and
appears to be more susceptible to damage. Standard procedure called for periodic tests of all flexible
hose. To avoid excessive "whipping" in the event a fitting (at the end of a high-pressure hose) was lost,
an anchoring device was utilized. A clamp was attached to the hose a few inches from the end fitting and
a short chain used to secure this clamp to the closest suitable anchorage.

9.2.6 Gantry Crane

Many preparations for launching were made at 40 feet, or more, above the ground. A gantry
of the
crane (Fig. 27) was provided by WSPG for this work. Safety- wise, the crane offered many advantages over
the German field equipment, but a few problems still required attention. Since it was necessary for men
to work at the bottom of the missile while others worked at the top, one of the most common dangers was
falling objects. Personnel were urged to use care in handling tools and equipment overhead. Side boards
were added to the gantry platforms to keep objects from rolling off or being pushed off. A safety net was
suspended below the upper platform and "hard hats" were worn by those working on the ground.
In the course of the program, it became evident that the regular ladders on the gantry might not pro-
vide a safe escape route in the event of certain troubles, such as a burning V-2 (never occurred at WSPG).
An escape cable, with sliders, was secured to an upper platform and was anchored to a post located about
200 feet from the gantry. This provided a means of escape which was fast and which continually increased
the distance between the missile and the man.

It was apparent that if a man were severely injured on the upper levels of the crane, it would be ex-

tremely difficult to get him to the ground by way of the ladders without aggravating the injury. To meet
this problem, special stretchers were obtained. These were designed so that a man could be strapped in
securely and lowered by one of the gantry hoists.

On occasion, it was necessary to move the gantry, under power, until a fixed platform came within a
few inches of the erected missile. It was conceivable that a mechanical or electrical failure in the control
equipment might cause the crane to continue in motion and push the missile from its stand. To provide for
this possibility, a master disconnect switch was located within easy reach of the crane operator. During
close operations, the operator was instructed to keep one hand on this switch.

48
Fig. 27 A Gantry Crane Facilitated Work on The V-2 at WSPG
49
9.2.7 Launching Area
A warning siren was installed to clear the launching area in case of danger. Siren switches were
located at suitable points in the area.

Standard Proving Ground procedure called for an ambulance and one or more fire trucks to remain
within the areawhen launching preparations were in progress.
9.2.8 Component Testing
Testing of the propellant tanks required caution.

Although the tanks were of considerable volume (162 cubic feet each) they were of light construction,
made of an aluminum magnesium alloy. Wall thickness of the oxygen tank was 0.08 inch and the alcohol
tank 0.048 inch. Of the two, the oxygen tank was probably more dangerous since the alcohol tank had a
tendency to pull rivets and relieve itself.

The potential danger of the oxygen tank was demonstrated in a destruction test. The tank top blew out
at 52.6 psiand the tank traveled 8 feet before striking a guard rail. A two- inch iron pipe upright (set in
concrete) was broken and a 14 1/2 foot section of double two-inch pipe railing bent outward. Pieces of
tank were thrown 150 feet.

Since the oxygen tank was pressurized to 21 psi at the launching site, it was necessary to test it to a
higher value prior to assembly. Hydraulic tests at 32.7 psi were made in a cleared area, outside the Mill
Building. These were followed by an air test at 21 psi within the building. During the latter tests, the
following precautions were taken: (1) all personnel, other than testers, were cleared from the area, (2) a
relief valve was installed on the tank, (3) two independent gages were used to indicate tank pressure, (4)
gages were observed through a slit in a heavy steel barrier and (5) one man remained at the supply valve
until the tank had been charged and the supply hose disconnected.

9.3 SAFETY RECORD

This report covers a period of 69 months. During this period, the average number of G-E personnel
stationed at WSPG was 29. Total lost time for this group due to accidents was 9.5 man-days.

50
BIBLIOGRAPHY

The following technical reports on the V-2 program were prepared by Project Hermes for external
distribution:

DF-71369 The Missile A-4, Series B

R 45770 German A-4 Electric Hydraulic-servo. Broome, J. W.
R 45779 Report of A-4 Trajectory Calculations Preliminary to A-4 Missile Tests at White Sands' N. M.
Moore, J. R., Reehl, G., Maher, R. T., Crawford, J. E., Anderson, A. M.

R 55127 Theoretical Vertical Braking Trajectories and Temperatures for Booster WAC Corporal.
Meylach, B.

R 55128 Summary Report on Tests of Models of the German Missile "A-4" and "Wasserfall" for
Mach No. 1.28 in the Aberdeen Bomb Tunnel and Comparison with the German Data.
Street, R. E.

R 55137 3rd Report, Aberdeen Wind Tunnel Tests of Missile A-4 and Wasserfall Effect of Roll Angle
and of Removing Tail-fins and/or Wings at M = 1.72. Klima, O.

R 55138 2nd Report on Aberdeen Wind Tunnel Tests of Missiles A-4 and Wasserfall Drag Lift and
Center of Pressure Results for the Complete Missiles at - 1.72. M
Sinclair, E. M.

R 55145 Dynamic Stability of A-4 WAC Corporal Combination Missile. Reehl, G. H.

R 55149 A Theoretical Comparison of the Standard German and a General Electric Steering Control
for the A-4 Missile. Barbour, P. K.

R 55204 Report on A-4 Rocket No. 19. Cunningham, H. A.

R 55208 Test and Analysis on the A-4 Rocket Control System. Broome, J. W.
R 55227 Trajectory Calculations for the A-4 Bumper WAC Missile. Crawford, J. E.

R 55256 Report on A-4 Missile No. 27 Including Skin Temperature Measurements to Mach No. 5.
Haviland, R. P.

R 55258 Supersonic Convective Heat Transfer Correlations from Skin-temperature Measurements

During Flights of V-2 Rockets No. 27 and No. 19. Fischer, W. W.

R 55272 Report on A-4 (V-2) Missile No. 36. Botkin, C. C.

R 55273 Report on Special Test A-4 (V-2) Launched 20 November 194/. Haviland, R. P.

R 55289 Investigation of Unsymmetrical Forces Due to Jet-entrained Air Produced During Launching
of the V-2. Gregg, A. B.

R 49A0500 Report on A-4 (V-2) Missile No. 39. Botkin, C. C.

R 40A0527 Report on A-4 (V-2) Missile Number 44 and 46. Botkin, C C.

R 50A0501 _P_rogress Report on Bumper Vehicle a Two-stage Rocket-powered Test Vehicle.

Haviland, R. P.

R 50A0523 Skin Temperature Measurements During Bumper No. 5 Flight. Fischer, W. W.

R 51A0513 Aerodynamic Load Distribution Over The Tail Fins of the V-2 Missile Provided as Basis for
Structure Design. Voutsas, A. M.

51
APPENDICES

To avoid repetition material already available in other reports, the appendices are written on the
of
assumption that the reader is well informed on the V-2 system or has access to: (1) General Electric
Company - Project Hermes report DF-71369, "The Missile A-4 Series B" February 1, 1945 (called A-4
Manual for brevity in appendix references) and (2) British Special Projectiles Operation Group, "Report
on Operation Backfire," November 7, 1945 (called Backfire for brevity in appendix references).

53
APPENDIX A
PROPULSION SYSTEM COMPONENTS

A.l OVER-ALL SYSTEM

Ithad been mentioned by some of the German personnel that after the V-2 went into production it was very
difficult tohave any changes made. There were two reasons given for this: first, changes necessarily slowed-
down production and, second, (and this reason was referred to time and time again), since they already had a
missile that would fly, they were hesitant about changing to something that had not been flight-proven. Thus,
many changes and improvements that were developed and tested were never used on a missile. Two examples
of this are the removal of the check valve in the steam generator and the incorporation of the heat exchanger
into the turbine. Both of these changes were worked out and tested, but were never used.

However, a great number of changes were made after production had started, as indicated by the many dif-
ferent types of components received at WSPG and by the differences in components received at WSPG compared
to the descriptions and pictures in the German A-4 manual. This includes valves, steam plant, steam plant pip-
ing, turbine, and propulsion piping. Possibly some of these differences were due to the fact that more than one
factory made the same component; however, many differences could not be explained in this manner.

The propulsion system had been fairly reliable and was (in most respects) well designed. However, both
main valves and the lox vent valve were considered to be designed wrong; it is felt by WSPG launching-personnel
that these valves should have been designed to be fail-safe pneumatically for personnel (closed with pneumatic
pressure in the case of the lox vent valve, and vice versa in the case of the main valves).

The lox tank pressurizing valve and pressurizing control switches were not flown but were left on the
stand; thiswas a step in the right direction, as was the design of the lox topping arrangement (following re-
mote topping and draining while adding only one check valve as a flyable component). The control for the open-
ing and closing of the valve was left on the ground.

In only one application (the large seat of the lox main valve) did the Germans use anything but a metal-to-
metal seat for the sealing of liquid oxygen. Considering the amount of trouble experienced at WSPG with leak-
age past these metal-to-metal seats, it is possible that such an arrangement was used simply because the
Germans had no suitable sealing material for use with liquid oxygen. A WSPG test, though inconclusive, indi-
cated that the rubber -like seat used on the lox main valve was not sealing well with liquid oxygen. It is pos-
sible that an alignment problem was the reason this seat was not also of the metal-to-metal type.

It was obvious the V-2 was not designed with an eye toward easy repair in the field. Access hatches were
placed only where necessary for fueling and de-fueling and for last minute adjustments. This was probably
part of an over -all policy decision of allowing a minimum of time on the launcher, whether launched or not.
Conversely, access for hangar repair was very good, since the removal of the tail took only about an hour and
allowed maximum accessibility of all parts. The design of the steam plant as a sub-assembly allowed the
complete unit to be changed in a very short time and made it possible to test it as a separate unit. This ar-
rangement also made possible a speedier initial assembly of the missile, which of course reduced the over -all
assembly area needed. Furthermore, the propulsion section itself was a sub-assembly, so there was no neces-
sity of bringing it into the main assembly area with the midsection, control section, and tail (also sub-assem-
blies) until it was tested and ready.

55
A.2 COMPONENTS
A.2.1 Jet Vanes

Approximately 20 percent of the vanes were discarded because of faults discovered in the examination
of theX-rays. These faults included cracks, voids, surface holes, inclusions and porosity. A number of
vanes had stripped holding threads and could not be used.

The carbon vanes may be divided into two general types: the mill-marked type, and the smooth or non-
mill-marked type. According to a German specialist these two types were manufactured to different speci-
fications. The smooth vanes were softer and showed a greater tendency to pull away from the backing plate.
Conversely, the milled vanes were more brittle and more likely to break off in the jet. However, the Ger-
mans considered both types suitable for firing.

To
verify this information, the studs holding the vane to the backing plate were torqued down until the
carbon threads failed on a number of unflyable vanes. It was found that the threads in the smooth vanes
pulled out at a lower torque than did the threads on the milled vanes. It was also found that vanes with the
prefix or suffix "712" in the serial number withstood the greatest torque of all vanes. The results were:

Minimum torque necessary to strip threads: Zero (probably stripped prior to test).

Maximum torque necessary to strip threads: 570 inch-pounds

Average torque necessary to strip threads: 270 inch-pounds

Average torque necessary to strip threads on smooth vanes: 200 inch-pounds.

Average torque necessary to strip threads on mill-marked vanes: 290 inch-pounds.

Average torque necessary to strip threads on mill-marked vanes with a prefix or suffix "712" in the
serial number: 340 inch-pounds.

Average torque necessary to strip threads on mill-marked vanes that do not have a prefix or suffix
"712" in the serial number: 230 inch-pounds.
Some carbon vanes had red or white stripes approximately 1/2 to 1 inch wide painted on each side. This
marking was a German code. White paint indicated that the vane was satisfactory; red paint indicated that
the vane was unsatisfactory and should not be used.

Carbon vanes were classified A, B, and C. Class A and B vanes were suitable for firings; class C vanes
could be used only for static tests. Although originally classified in Germany, vanes were later X-rayed and
classified in the United States. The following are instructions for inspection and classification of vanes:

a. Any unusual appearance should be noted and, if possible, correlated with visual inspection of the vanes.
The inspector should be especially careful to look for the following faults:
1. Porous material gives a filamentary appearance to the negative, as if heavy lint were on it

2. Cracks show up as thin dark lines on negative

3. Voids, Chipped Plates, Surface Holes appear as dark spots on negative

4. Inclusions appear as light spots on negative

b. Faults should be characterized as "weak," "distinct" or "prominent," and in some cases as "round,"
"elongated," or "irregular." The sizes of round areas should be described as follows:

Smaller than 1/32-inch diameter... "very small"

1/32 to 1/8-inch diameter... "small"

1/8 to 3/1 6 -inch diameter... "medium"

3/16 to 1/4-inch diameter... "large"

Larger than 1/4-inch diameter... "Very large"

56
In thecase of elongated or slightly irregular faults, the sizes should be described as for a round fault of
equivalent area.

c. The location of a fault is also important, particularly if it is in a critical area, such as the leading

edge of the main part of the vane or of the toe, particularly near the outboard corners. The descriptive
terms in Figure 28 are suggested.
d. Although not intended to eliminate the need for exercising judgment in each individual case, the follow-
ing rules are offered as a guide:

1. Class A (Excellent): This class of vane contains no "large" voids or inclusions, nor more than
three "medium" or "small" voids or inclusions unless these are weak and not located in a critical area.
The vanes may contain a large number of "very small" speckled inclusions.
2. Class B (Second Choice): This class of vane has more or larger defects than A vanes, but none are
located in critical areas, the vane otherwise being satisfactory.

3. Class C (Questionable): Any cracks or noticeable porosity, as evidenced by filamentary appearance

places vanes in this classification. "Very large" voids or inclusions, "medium" or "large" voids or in-
clusions in critical areas, "small" prominent voids in critical areas, or any unusual appearance which can-
not be explained by visual inspection will also place vanes in this category.

The Germans were not greatly worried if the carbon side strips were defective, i.e., if they had one or
two inches chippedoff one end. It was the policy at WSPG, however, to change these strips if they were
broken.

The clearance between the vane and the backing plate had to be at least the thickness of ordinary writ-
ing paper, or the vane could not be used.

The carbon vane holding studs were torqued to 110 inch-pounds both before and after the load test. They
were tightened beforehand so that all studs would take an equal load and were tightened afterward to make sure
there had been no failure of the threads.

Not earlier than one month before the firing date, the vanes were given a static load test by applying 2300
pounds on one side of the vane, then turning the vane over and applying the same load on the other side. A
special machine was used for this test. The 2300-pound force was obtained by using a 224-pound weight and
a lever arm. The force was applied longitudinally over 8 inches of the vane, 6-1/2 inches from the outer
edge of the backing plate. Of course, measurements or computations were not necessary during the test; the
vane was simply bolted into place on the machine, and the weight gently lowered. No vanes were broken at
WSPG during this test, but the test probably detected weaknesses in some holding threads, since a few stripped
threads were detected after the test had been run.

LEADING EDGE

INBOARD OR INNER EDGE

MAIN PART

TOE

TRAILING EDGE

OUTBOARD OR OUTER EDGE, THICK PART

Fig. 28 Jet Vane Descriptive Terms

57
The Germans considered that a vane was still useable for a firing even if the threads for one holding stud
were stripped, provided the defective part was located on the lower half of the vane. The policy at WSPG, how-
ever, was not to use such a vane if other "A" or "B" vanes were available.

To minimize the danger of breakage, vanes were not installed on the missile until the day of the launching.

It was the policy at WSPG

discard vanes which had been exposed to flame by a misfire. The possibility
to
of the vanes being damaged by the flame was small, so the procedure was merely precautionary. As time
elapsed after the firing, however, it became more important to change the vanes, since the vanes were hygro-
scopic and could have absorbed moisture after the protective lacquer coat had burned off. This moisture might
have caused the vane to break open as it turned to steam in the jet. It was felt that the vanes should definitely
be changed if they had been sprayed with water at any time or if they had been sprayed with CO2 while hot.

When vanes were put on the missile, a notation was made of the serial numbers and the corresponding fin
numbers. Thus, if a vane fell off and was recovered, it might have been possible to determine which fin it had
been on.
Just prior to firing, the vanes were covered with a 3/32-inch thick cardboard envelope which was lightly
wired in place. This was a precautionary measure to protect the vane from the possible impact of the igniter
as it was expelled from the burner. These covers were expelled from the burner prior to lift.

It is believed that on one occasion a missile failure was caused by a piece of a vane, or a complete vane,
breaking off. This incident occurred before vane covers were used and before X-rays and load tests were
made.
Although vanes were considered fragile, and were always handled as such, on some occasions vanes were
found intact at impact. In these cases the warhead had been blown off and the tail section had landed on one
side. Thus, the vanes that survived had not come into contact with the ground.

A.2.2 Burners

References: A-4 Manual pp. , 17, 88, 101, 105a, 105b and 223; Backfire, Vol. II, pp. 19, 20, 21, 25, 30,
89, 113, 114 and 115.

An attempt was made at WSPG to boost the cooling jacket test pressure from 255 to 325 psig. This re-
sulted in the rupturing of the inside-dome wall in three of the four burners tested at the higher pressure.
The following information was obtained from conversation with German personnel:
a. It was possible to tell, by visual inspection, which burners had been static fired. Static firings left
darkened streaks on the alcohol nozzles, and, because of the temperature differences, left discolored areas
around the ZK nozzles and the cooling nozzles.
b. Test runs were made with the upper row of film cooling-holes blanked off and it was found that this row
was not necessary.
c. Practically all burner failures detected because of static firings occurred on the inside of the burner.

Of these, some inner walls bulged inward, but most failed on one of the inside wall welds.
d. At first, burners were manufactured with only two expansion folds in the outer wall, and with no expan-
sion loops in the cooling supply lines. It was later found that with this design, unequal expansion of parts was
still causing failures; in subsequent missiles another expansion fold was added and expansion loops were put
in the cooling lines.

e. All burners used by the Germans for flight purposes were static fired. Because of the expense in-
volved in static firings, a series of tests was made to determine if powered -flight conditions could be dupli-
cated. In these tests, burners were subjected to extreme vibration while pressurized. The burners were
then inspected, and if no cracks or other faults had developed, they were static fired. Since some burners
that had successfully passed the vibration-pressure test developed cracks when static fired, it was decided
that such a test was not adequate. If these tests had proved successful, it was planned to affix a plate to the
throat of the burner, and pressurize the upper combustion chamber at the same time the outer shell was
tested. This would protect the inner wall of the dome from excessive pressures while giving the outer jac-
ket a high pressure test. The throat instead of the outlet would have been used for blocking because it was
believed that the lower part of the burner would not withstand the necessary internal pressure.
58
f. During the latter part of the program, burners were static fired for only 10 seconds. This procedure
was initiated after it was found that practically all burner failures occurred during the first few seconds of
test.

g. On was no covering on the bottom skirt of the burner. It was

the first burners manufactured, there
was reaching a temperature sufficient to ignite alcohol that often leaked from
later found that this skirt
loose couplings. A sheet metal cover with glass wool insulation was then added, which greatly reduced
the number of tail explosions.

Applying a hot-air blower to remove moisture from a burner would often damage the ZK nozzle plastic
gaskets sufficiently to cause a gas leak when the burner was later pressurized. When burners with ZK
nozzles installed were hydrostatically pressurized to the 255-psi test-value, Wood's metal in the nozzles
would often blow out. This necessitated the use of blank ZK nozzles for such tests. In almost every case
in which a missile reached preliminary stage and did not subsequently take off, it was found that the Wood's
metal had melted out of some of the ZK nozzles and alcohol was leaking into the combustion chamber.
When this happened (and it was still possible to fire the missile) the holes were plugged with sharpened
pieces of wood or matchsticks.

One burner, after having been flight-tested was found to have a small hole burned completely through
the inner jacket. This was caused by a plugged hole directly above it in the cooling ring. At least one
other burner showed evidence of eroding, but the burner metioned above was the only one noted in which a
hole had been burned through the jacket.

Many of the burners at WSPG had mounting brackets knocked out of line, cooling lines mashed and
expansion joint fittings and cooling line fittings (Fig. 29) knocked off. Almost every burner used had some
defect requiring a welding -repair. One burner used at WSPG had a flattened main valve seating ring. This
was probably caused by mishandling some time after manufacture. In addition to these defects, some bur-
ners had to have loops welded into the cooling lines. Also, some sheet metal skirt covers were rusted
through, and had to be repaired or replaced.

A
Fig. 29 Cutaway of V-2 Burner ^ left) and Burner-piping Installation (right)
59
All missiles fired at WSPG had paper cups installed over the oxygen nozzles. These cups were used
"In order to prevent a premature mixing of the oxygen and alcohol which would cause local explosion on
ignition, or freezing " (**' German personnel believed these cups were used to keep oxygen fumes out
of the upper alcohol chamber of the burner. Any oxygen that leaked past the lox main valve would be di-
rected downward away from the alcohol nozzles.

A. 2. 3 G as Bottles

References: A-4 Manual , p. 17 and 80; Backfire, Vol. II, p. 22 and 23.

To obtain an indication of the weight and volume range of the air bottles a small number of bottles were
weighed and volume measurements taken. The weight varied from 18 to 21 pounds, and the volume from
7.04 litres to 7.45 litres.

Five sets of bottles on hand at WSPG had been wound with wire; they weighed between 20 and 21 pounds
each. According to German personnel, these were experimental bottles never used on the V-2. Three were
hydrostatically tested to rupture. The wires began to break at 7000, 7100 and 7750 psi; final rupture was at
4200, 4500 and 7000 psi. In all cases, the rupture was a longitudinal break in the side of the bottle.

During the firing program it was discovered that some of the air bottles were rusty inside. After that,
before any bottles were used in a missile, the manifolds were removed and each bottle was checked for
cleanliness. If a bottle was rusty, the rust was removed. This was accomplished, after some experimenting,
by removing the bottle adapter, filling the bottle about 30 percent full with Mecha Finish #4, filling the voids
between the Mecha Finish particles with water, and rotating the bottle for two hours. The rotating mecha-
nism was designed so that the bottle could be tilted at intervals during the cleaning process and thus clean
the ends of the bottle. After the two hour cleaning period, the Mecha Finish was removed and the bottle was
water flushed and dried. Mecha Finish #4 was an abrasive rock-like material distributed by the Mecha
Finish Corporation of Sturgis, Michigan. The pieces had to be broken apart before they were small enough
to be put into the bottles. Prior to the time that Mecha Finish was used to clean the bottles, scrap nuts
and screws were used. Although this did not clean the bottles as well as did the Finish it did remove
most of the rust.

After the bottle cleaning program was started, it was found that most of the manifold nipples were being
deformed during retightening, many to the extent that there was concern that they might possibly pull through
the nut. New nipples were made to replace the German units. The German nipples were cut off at the
original weld and the new nipples were slipped over the manifold tubing and lap welded. The curvature of
the seats remained the same as before, but the nipple shoulder was made slightly wider. The distance be-
tween the shoulder and the seating circle was made shorter, so that a maximum of threads could be utilized.
On a few manifolds, the nipples were not being deformed. These manifolds were easily detected before-
hand, since they were painted a color different from the others.

Bottles, with adaptors installed, were hydrostatically pressurized to 4500 psi after they had been cleaned.
Manifolds were also hydrostatically pressurized to 4500 psi. During most of the program, the assembled
bottles and manifolds were pressurized to 3500 psi. During the latter part of the program, however, only
3000 psi test air was available. It was considered safe to use bottles tested to this pressure, however,
since they had previously been tested hydrostatically to a much higher pressure. On no occasion did an
air bottle fail in test, except in cases where the purpose of the test was to find ultimate strength. There
were occasions of leaks in manifold welds and at connections, but no leakage was ever found through the
bottle proper.

A.2.4 H eat Exchanger

References: A-4 Manual ,
p. 17, 88 and 139; Backfire, Vol. II, p. 25 and 32.

Oxygen flow through the heat exchanger was about 0.3 kg per sec, according to German personnel and
the A-4 Manual.

60
A test was run at WSPG on each unit to determine if the orifices in the orifice block were of the cor-
rect size. This was determined by applying a constant air pressure (7.1 psi) to the liquid side of the unit
and allowing the air to exhaust to atmosphere while measuring the flow by means of an orifice and mano-
meter. A flow deviation of about 10 percent above and below the optimum flow was allowed. In many
instances, this flow range could not be met and the orifice block was replaced with a block made at WSPG.

All cases were hydrostatically pressurized to 28 psi. No leaks were found in any case,

The tubing coils were pressurized to 71 psi. A number of coils were found with pinhole leaks.

According to one of the Germans, a heat exchanger was designed to fit inside the turbine. This was
never used on a missile.

It was reported that the Germans had considerable difficulty with pressure fluctuation on their first

heat exchanger models. It was finally determined that oxygen gas was forming an insulating coating on the
inside of the tubes, and a slight jar would disturb the coating, causing a sudden increase in gasification, and
therefore, a pressure surge. This problem was solved by constructing and mounting the heat exchanger
coils in such a manner that they would vibrate easily. Thus the missile vibration was enough to keep the
gas film broken up, resulting in a fairly constant gas output.

A.2.5 Lines and Fittings

References: A-4 Manual , p. 52 and 114; Backfire, Vol. II , p. 18 and 19.

Beginning with missile 56, all lox and alcohol pressure lines and fittings were hydrostatically tested to
500 psi. On missiles prior to this time, the lines and fittings were checked for leakage, but not for strength.
It was found that the American-made alcohol three-fold couplings would not withstand the 500-psi hydro-

static test pressure. Two were tested and each ruptured at a crotch weld at 410 psi. These two, and all
other American-made alcohol three-fold couplings were returned for re-welding. It was found that the
German-made alcohol three-fold couplings on hand would withstand the 500 psi hydrostatic test-pressure
without failure. Only German couplings were used throughout the rest of the program. About 15 percent
of the American-made alcohol feed lines leaked air when pressurized to 100 psi. About 50 percent leaked
air when pressurized to 100 psi after a 500 psi hydrostatic test. This indicated that 500 psi was opening
the welds, although in no case was there any noticeable sign of cracks in the welds, except in cases
where such cracks were evident before any tests were run. It is possible that the holes in the welds were
plugged with flux or oxide and a pressure higher than 100 psi was necessary to clear them. In some lines,
there was leakage in cracks alongside the weld, but these cracks were evident before any tests were run.
In no instance was there leakage at any place except in or alongside a weld. In most instances, the leakage
was at a pronounced crater in the weld, or at a spot where a weld had been ground down. One representa-
tive line was returned to the vendor, the others were repaired at WSPG.

In contrast to the alcohol feed lines, the American-made oxygen feed lines had smooth welds, and less
than five percent leaked when subjected to the same tests given the alcohol feed lines.

The live steam line and the lines from the alcohol two-fold fitting to the alcohol three-fold fitting were
too rigid to allow bending for alignment with mating parts. Since these lines were stocked disassembled
flanges not welded on), the procedure for mounting was to bolt the flanges in place, cut the line to size,
(i.e.,
and tack weld the line to the flanges. The assembly was then removed and the weld completed in a more
convenient location.

The stripping of threads on the lox three-way couplings and the lox lines was at first quite serious.
Removal of these lines to facilitate the removal of the lox main valve resulted in damage to 40 to 50 percent
of lines and couplings on the first two or three calibrated units. Later it was found that the lines could be
sprung back sufficiently to allow removal of the lox main valve; this procedure was followed on the remain-
ing units.

The live steam line was hydrostatically tested to 600 psi. The line, including all protuberances, was
insulated by lagging with asbestos and coating the asbestos with water glass. Flanges were insulated by
covering with glass wool or glass rope held in place by sheet metal covers. Many air lines for the missile
had to be made up at WSPG. These were made of copper tubing with steel silver -soldered fittings.

61
A.2.6 Turbo pump
References: A-4 Manu al, p. 59, 69, 71, 73 and 222, Backfire, Vol. II, p. 18, 27, 29 and 111.

a. Cleaning

of the firing program, pumps were cleaned by pouring water into the alcohol
During the first part
pump and carbon tetrachloride into the oxygen pump and allowing the liquid to stay in the pumps for about
30 minutes, agitated frequently by rotating the unit. The pumps were then dried with a hot-air blower. As
the program progressed and units were calibrated, the pumps were cleaned by wiping when disassembled,
then, after final assembly, carbon tetrachloride was poured into the lox pump and allowed to set for about
30 minutes to thoroughly degrease all parts. The pump was then dried with a hot-air blower. This proce-
dure was followed until a shaft in a turbine that had been assembled for about six weeks was found to be
frozen. Disassembly showed that the steel parts in the lox pump were very badly rusted. From that time
until the end of the program, all cleaning was done by wiping while pumps were disassembled, except that
the lox pump was filled with carbon tetrachloride and soaked for about 30 minutes before final disassembly
to degrease parts that would not be accessible for hand cleaning after disassembly.

b. Disassembly Policy
If turbopumps were disassembled during the first part of the firing program it was only to the
at all
extent necessary to decrease leakages. Later in the program when calibration of propulsion units was
initiated, all turbopumps were disassembled after calibration to dry and to clean residue off the turbine
parts. Shortly after that time, it became standard routine to disassemble all turbines before calibration,
and use only those turbines which appeared to be easily repairable in the event a part should become
defective during calibration. Such disassembly was considered advisable on other grounds also, since an
increasing amount of foreign matter was being found in turbopumps.

c. Corrosion

No
rust or other corrosion was ever noticed in a turbopump until after it had been calibrated. The
firstfew turbines calibrated were very badly rusted when disassembled. The procedure at that time was
to dry the complete propulsion unit with a hot-air blower before the turbine was disassembled. Often a
week or more had elapsed between the calibration run and the disassembly of the turbopump. Later in the
program, the turbopump was removed from the unit immediately after the run, and was, in most instances,
completely disassembled. Each steel and iron piece was dried the same day the calibration run was made.
Even with this procedure, there was some rusting. When the turbopump was allowed to set even overnight
without being disassembled, considerable rusting occurred.

Alcohol pump: Stationary blades:

Outside diameter of impeller 12.74 inches Number -4 sets of 28 112
Inside diameter of case 13.4 inches Height at inlet 0.68 inch
Number of vanes -seven Height at outlet 0.83 inch
Thickness of vanes at outlet 0.22 inch Pitch 0.32 Inch
Outlet angle of vanes 32 degrees Outlet area (per segment) 3.6 square inches
Width at outlet 0.6 Inch Outlet angle 28 degrees
Inside diameter of inlet 5.92 inches Moving blades:
Oxygen pump: First row -
Inside diameter of impeller 10.53 inches Height at inlet 0.63 Inch
Inside diameter of case 10.6 inches Height at outlet 0.63 inch
Number ofvanes seven Pitch 0.32 Inch
Thickness of vanes at outlet 0.15 inch Blade inlet angle 24 degrees
Outlet angle of vanes 18 degrees Blade outlet angle 24-26 degrees
Width of outlet 0.66 inch Second row -
Inside diameter of entry 5.57 inches Height at inlet 0.98 inch
Turbine: Height at outlet 1.10 inches
Mean diameter of blades and nozzles . . .18.5 inches Pitch 0.32 inch
Number of nozzles 16 Blade inlet angle 30 degrees
Throat area per nozzle 0.059 inch Blade outlet angle 26-28 degrees
Mean height of nozzles 0.04 inch
Mean pitch of nozzles 1.6 inches
Outlet area per nozzle 0.21 square inch
Nozzle angle 20.5 degrees

Fig. 30 Technical Data on V-2 Turbopump Assembly

62
One lox pump had rusted so badly that the shaft could not be rotated. This pump was supposedly ready
for launching and had been assembled to its turbopump for about six weeks when the condition was discovered.
It was never determined what caused the rusting. A lox pump that had never been calibrated was given an
identical cleaning treatment and to have rusted after two months storage. Some possible
was found not
damage to the protective coating during calibration might have caused the rusting in the lox pump men-
tioned first, but the most logical explanation is that a considerable quantity of water had mixed with the
carbon tetrachloride used for the final degr easing. From paragraph 1422 of the A-4 Manual "The steel
,

parts (of the turbopump) are protected against the action of moisture by a special polishing and hardening
process using corrosion protective oil. The lox pump is specially degreased during the last test run at the
supply firm." Since water was never admitted to any lox pump until it was on the calibration stand, it is
not known what effect large quantities of water (but no calibration) would have on the special-process parts.

No rusting of parts occurred in the alcohol pump since all steel parts were thinly coated with grease
in Germany. The seals on the alcohol side of the turbine rusted appreciably on only the first few turbo-
pumps calibrated; on later turbines the seals were coated with a thin film of grease during the initial dis-
assembly. The seals on the lox side of the turbine were not coated with grease because of their close
proximity to the lox pump.

Turbine and lox pump seals were cleaned by using crocus cloth. Fine emery cloth was used on the
other steel parts in the lox pump.

Fig. 31 Cutaway View, V-2 Turbopump Assembly

63
d. F urther Effects of Calibration

The first two or three turbdpumps that were calibrated showed considerable erosion of the turbine
buckets. Subsequent reduction of the steam temperature completely eliminated this erosion.

All turbopumps had permanganate residue left in the turbine after calibration. Most of this was re-
moved with wiping cloths and a wire brush, although no special effort was made to make the surfaces com-
pletely clean.

On many turbopumps, the turbine and lox seals held pressure better after calibration than before
calibration. This was probably caused by the wearing-in of the seals, or, in some cases, by the fine rust
that was formed on the seals.

e. Component Difficulties

Since many turbopumps were available at WSPG, the units which could not be repaired easily were not
used; the parts were stocked as spares. Included in this category were turbopumps that: (1) did not have
an alcohol return line fitting, (2) had lox seals different from the usual seals, (3) had defective parts, and
(4) had leakages difficult to correct.

The iron seals on the turbine and lox pump gave much more trouble than did all other components.
During the first part of the program, leaky seals were repaired. The sides and ends of the seal segments
were, where necessary, ground-down with a fine stone, and then lapped. The inside radius was ground
down by using lapping compound between the segment and the corresponding cylindrical part from a dis-
carded turbine. Such repair work was very time-consuming, therefore, as the program progressed and
spare parts were accumulated, seals were replaced instead of repaired. During the latter part of the
program, virtually no seals were repaired. A large stock of spare seals was necessary, since as many
as a dozen seals might be tried before one was found that would fit a particular shaft.

Another cause of leakage past the turbine seals was distortion of the seal plates. In most turbines,
these plates were fairly thin and some would be badly distorted. The plates were replaced when better
spares were available. If better units were not available, the distorted plates were "trued" on a lapping
plate or by using a bearing scraper. Defects were found by using a surface plate and "Prussian blue."
Even if "trued-up," the plates would often distort when tightened on the turbine. A few of the turbines
had extra-thick seal plates. These plates were quite superior to the thin ones, since they could be put
into place very tightly without distortion.

In many turbines, the area on which the seal plate seated had been raised at the tapped holes (prob-
ably caused by improper tapping). The raised metal was removed with a bearing scraper to reduce leak-
age between the turbine and the seal plate.

Rubber alcohol seals often leaked after calibration, ft was found that if these seals were replaced
with unused German seals before calibration, the new seals would withstand calibration without leaking.
This indicated that the seals had lost their resiliency after the prolonged storage in a flexed position.

The studs on the oxygen cover plate were made of steel in most turbopumps, but on a few, the material
was aluminum. It was difficult to obtain a good seal between the cover and the pump casing on turbopumps
with the aluminum studs, since the studs could not be torqued down tightly without breaking.

A number of turbines leaked air between the nozzles and manifold and/or between the nozzles and
turbine cover. In some cases, gaskets were missing from these places. Two such turbines were repaired,
but in the process, so many threads were stripped in the turbine cover that it was decided not to use similar
turbines for the rest of the program.

The lox-pump cheek pieces were sometimes temporarily or permanently distorted when the holding
screws were torqued too tightly. This resulted in leakage between the cheek piece and the pump.

64
f. T urbopump disassembly procedure: A prick punch was used in all cases where alignment marks
were needed, except that seal covers were aligned by scribing. Care was taken to insure that punch or
scribe marks were not made in or adjacent to any area that contacted any other part of the turbopump.
Also, personnel were careful not to mar any bearing or seating surface. In general, the procedure noted
below was followed.

The steam ring was removed after first removing the insulating covers from the connections. Gaskets
between the ring and the turbine were not removed.
The two cheek pieces between the lox pump and the turbine were marked at adjacent spots on their outer
periphery. These marks were made lightly and as far as possible from the juncture of the two cheek pieces,
to keep from raising an area on the seating surface. The lox pump was then lifted from the assembly after
first removing the four holding nuts. Then, before either shaft was turned, both sides of the flexible coup-
ling were marked at points adjacent to the marks on the cheek piece. This "round-about" method of align-
ing the flexible coupling pieces was necessary, since the pieces were not accessible before the lox pump
was removed.
Four spacers between the turbine and the lox pump were marked to provide for correct positioning
when replaced.
The flexible coupling holding nut on the turbine side was removed, and the shaft and flexible coup-
lingwere marked. The flexible coupling was removed, and a scribe mark made to align the seal cover.
To avoid later damage, the seal cover and seal were removed at this time.
The turbine cover holding nuts and the two alignment pins were removed, and the turbine cover lifted
off. During removal, the cover would often strike against the shaft. Therefore, it was decided to remove
the turbine seal before this operation.

The turbine wheel was marked, and the reversing bucket segments were loosened from their align-
ment pins by prying upward against the ends of the segments (not between the segments and the case). A
bearing puller was used to remove the turbine wheel.

The four turbine -casing holding nuts were removed and the turbine case lifted free from the alcohol
pump. Care was taken not to strike the case against the shaft, since this would damage the turbine seal.
The four spacers were marked to insure correct positioning when replaced. The turbine seal bushing was
removed from the shaft and the turbine seal was removed from the turbine case after first marking the
cover alignment with a scribe.

The overspeed case and holding nut were removed, and the overspeed disk was removed from the
shaft. When it was necessary to pry the overspeed disk off, it was done carefully, since bending of the
disc edges could result in a changing of the trip setting. The grease seal adjacent to the overspeed was
removed by prying on the sides (not bottom) of the seal.
The alcohol leak line and the alcohol pump cover were removed. The cover could not be tilted while
being removed, since this would jam the bearings. Care was taken to avoid stretching the sealing ring
gasket. The bearing was easily removed from the cover with finger pressure if care was taken not to tilt
the bearing.

Next the bearing plate on the turbine side of the alcohol pump was removed. The alcohol-impeller
holding nut was removed, and the shaft and impeller were marked. A nut was screwed on the end of the
shaft, and the shaft was removed by tapping lightly on the nut with a plastic mallet.

Alcohol seals were removed by "driving-out" with a 3/8-inch round drift pin. Care was taken not
small dent in the metal seating surface could cause leakage. Some turbines
to let the drift slip, since a
had a snap ring and seal holder on the turbine side of the pump. These had to be removed before the
smaller seal could be driven out.

The nut on the lox pump flexible coupling piece was removed; the coupling and shaft were marked
and coupling removed.

65
The cheek piece on the lox pump was removed, care being taken not to break the gasket or allow the
seals to strike the shaft. The adjacent lox pump bearing and thrust piece were then removed. Seal covers,
seals, and spacers were removed from the cheek piece after first marking the cover alignment with a
scribe mark. Patience was required in removing these units, since even the slightest tilting of the spacers
would cause jamming.
The lox cover holding nuts were removed and the cover forced off by screwing two studs into the
holes provided in the cover. Gasket material was scraped from the lox cover and lox pump. The impeller
and shaft were removed as a subassembly and no further disassembly made.

g. Turbopump assembly procedure: Most washers on the turbopump were curved lock washers although
they appeared at first glance to be plain flat washers. Therefore, if any were lost or broken, they were re-

placed with like washers taken from another turbopump.

In general, the following assembly procedures were used after the parts were cleaned and, where
necessary, degreased.

The lox impeller and shaft were installed in the lox pump casing, and the lox pump bearing and thrust
piece were placed. Seals and spacers were replaced in the cheek piece, making sure that the segment
joints were staggered. The seal plate was lined up with the scribe mark and screwed into place; the cheek
piece gasket was replaced. The lox pump bearing was rotated so that its pin would line up with the slot
in the seal plate, and the cheek piece was installed and tightened. Overtightening was avoided to prevent
buckling of the cheek piece.

Lox pump seals were centered, and the flexible coupling was lined up and replaced. If the coupling
did not slip easily into place, it was removed and the seals again centered (any attempt to force the coup-
ling would result in seal damage). When the coupling was in place, the center Jiut was replaced and locked
by driving some of the outer skirt of the nut into the two slots provided for that purpose.

Gasket areas on the lox pump cover and lox pump casing were covered with a water-glass and feldspar
powder mixture, the gasket put in place, and the lox cover replaced and tightened. No pressure test was
made for at least 24 hours to insure that the gasket-seal mixture was dry.
Alcohol pump seals were driven into place with a hardwood block. The seal holder and snap ring were
installed (on turbines so equipped) before the large seal was installed.

Both alcohol pump bearings were cleaned and repacked with light grease. The alcohol pump shaft was
put in place and the bearing on the turbine side of the pump was installed. The bearing plate was put on and
tightened evenly.

The alcohol pump impeller was installed on the alcohol pump shaft and the nut was installed and tight-
ened. To lock the nut in place, the locking washer was bent upward.
The alcohol sealing ring gasket was put in place on the pump cover. If the ring had stretched, it was
shortened and the free ends cut at an angle and joined by rubber cement.

A special conical assembly tool was placed over the overspeed end of the alcohol shaft to protect the
alcohol pump seals; the alcohol pump cover was lined up and pushed partly into place. Before the cover
was put completely on, a visual check was made to insure that the inside seal spring had not slipped off
the seal. The cover was then pushed into place and bolted down, the overspeed assembly tool was removed,
and the bearing, grease seal, overspeed disk and overspeed nut were put in place. The overspeed case
was installed and the four holding nuts tightened down evenly to insure proper alignment of the case; the
leak line was then replaced.
The turbine seal bushing was put on the shaft with the groove toward the alcohol pump. Four spacers
installed on the alcohol pump and lined up with the punch marks. The seal was installed in the
turbine
were
case and the seal cover aligned and tightened. The turbine case was then installed on the alcohol pump,
making certain that the seal did not strike the shaft and that the seal was centered so that it would slip
easily over the seal bushing. Any attempt to force the case into position would result in damage to the
seal. After it was certain that the spacers were fitted correctly into the alcohol pump and the
turbine
case, sealing washers were slipped over the four alcohol pump studs and the nuts were tightened. These
nuts were locked in place by driving a part of the nut skirt into a hole provided for that purpose.

66
The turbine wheel and reversing segments were aligned and lowered into place, and the segments
tightened. Two alignment pins on the outer circumference of the turbine were placed in position, the
turbine gasket replaced and the turbine cover positioned and tightened. The seal was put in place and
the seal cover aligned and tightened. The flexible coupling piece was aligned, slipped into place and
tightened with the center holding nut. If the coupling could not slip easily into place, it was i emoved and
the seal re-centered (any attempt to force the coupling into place would result in seal damage). The nut
was locked by driving some of the skirt of the nut into the spaces provided for that purpose. Both parts
of the flexible coupling were rotated to the exact position they occupied when the turbine was disassembled
(as determined by the punch marks on the flexible coupling pieces and the cheek pieces). Four spacers
were installed on the turbine studs and aligned with the punch marks. The lox pump was then put in place,
tightened and the spacers examined to make certain they were seating correctly on the turbine and the lox
pump.
The steam ring was replaced and tightened, and the insulating covers installed.

The turbine was then rotated to determine if it was turning freely. If it was binding, the center cap
on the lox cover was removed and the lox shaft tapped (using a plastic mallet and drift pin) until the lox
pump shaft was heard to seat against the turbine shaft. The center cap was then replaced and tightened.
h. Testing specifications: The lox pump was pressurized to 28.4 psi with the air source cut off; the
pump failed test if the pressure fell below 9.9 psi in 120 seconds. Original specif ications allowed this drop
in 60 seconds, but this was changed later in the program. With 28.4 psi pressure in the pump, leakage was
not allowed between the turbine and cover nor at any fittings.

The alcohol pump was pressurized to 25.5 psi; no leakage was allowed. All joints and fittings and
the leak line were checked with soap solution.
The turbine was pressurized to 22.7 psi. With the air source cut off, the turbine failed test if the
pressure fell below 7.1 psi in 60 seconds. With 22.7 psi pressure in the turbine, no leakage was allowed
at any place except the seals.

i. Over speed Assembly: A test determined the speed at which the overspeed trip ring began to move
out as compared to the speed at which the overspeed trip ring had moved completely out. It was found
that once the trip ring had begun to move out, it would probably move completely out at that same speed
and would definitely be completely out at 200 additional rpm's.

Fig. 32 V-2 Turbopump Overspeed Trip

67
Some overspeed cocking leverscould be placed, by accident or otherwise, in such a position that elec-
would not be made. However, a jar or vibration would cause the contact to be made. All
trical contact
assemblies tested, would function properly if the cocking lever was pulled completely out to the stop.
Each complete overspeed assembly was installed on a test rig and the speed at which the trip was
actuated was maintained between 4800 and 5200 rpm. During the latter part of the program, the tripping
range allowed was changed to between 5000 and 5200 rpm. The trip point could be adjusted by disassem-
bling the overspeed disk and changing the number of shims under the spring.

A WSPG test report sheet for the turbopump is shown in Figure 33.

A. 2. 7 Alcohol Main Valve

References: A-4 Manual , p. 116, 121 and 133; Backfire, Vol. II, p. 43, 45, 101 and 104.

Many porous cases were found during main valve. This may have been a manu-
tests of the alcohol
facturing defect, but could have been caused by aging or corrosion. Disassembly of such valves showed
numerous pits on the inner surface of the case.

TEST REPORT
-2-

TURBQ PUMP

4. Cleaning

Ho. of Pump:
(a) Oxygen Pump:
Built Into Rocket No.:
(1) Cleaned with air

1. Visual Examination (2) Cleaned with solvent

(a) Were the Intakes and outlets of turbine, oxygen pump and alcohol pur ap closed? (3) Dried

(b) Alcohol Pump:

(b) Was there any visible damage? If so, where?
(1) Cleaned with air

2. Disassembly
(2) Cleaned with solvent
(a) Was the turbo pump disassembled? If so, how far?
(3) Dried and oiled

(c) Turbine:
(b) Remarks
(1) Cleaned with air

3. Pressure Test
Remarks
(a) Turbine:
5. Overspeed Cut-off
(1) Test pressure (2) Pressure decrease to

(a) Release tested

In seconds. (3) Leaks? Where?
(b) Electrical switch tested

(b) Oxygen Pump:

G. Could the turbo pump be easily rotated by hand?

(1) Test pressure (2) How often was test repeated?

Pressure decreased to atm. in seconds.

(3) 7. Final Operation

(c) Alcohol Pump: All lines and fittings closed dust-tight

(1) Test pressure (21 Leaks at the leak line ?

(3) Pressure decrease to atm. in seconds.

(4) Leaks? Where? Tested by: Date:

Fig. Turbopump Test Report Form

68
Although many and varied defects were noted, there was no instance in which the valve failed to operate.
There were, however, many instances of very rough strokes. Disassembly of these valves showed a scoring
on the center cylinder where the stem-nut moved. In some instances, this scoring was quite severe (pieces
of aluminum had been rubbed off the cylinder). This was probably caused by manufacturing errors, either
in tolerances or in alignment. To make certain that sticky "U" cups were not causing rough strokes, two
valves with rough strokes were disassembled and a graphite base lubricant ("Aquadag") was applied to the
surfaces that would be in contact with the "U" cups. When reassembled, the valves continued to operate
with rough strokes. A rough stroke in this valve was serious for two reasons. First, it indicated that a
rubbing between two metal surfaces existed which might result in a complete seizure at a later time. Second,
it might cause a missile failure at takeoff (explained in the oxygen main valve section of this report). A
minor design change to increase the clearance between the center cylinder and the stem nut, to provide no
contact between the two units, would probably have eliminated the rough stroke. This would have necessitated
inserting a thrust washer on top of the spring and increasing the inside diameter of the stem.

The vertical air inlet fitting elongated badly with moderate torque. Although no failures occurred, there
was considerable necking; replacement parts of steel were utilized. If it was thought necessary to use alum-
inum for the new fittings, smaller diameter elongated holes placed vertically so that the metal cross-sec-
tional area through the holes' was greater, would have been utilized. Also, a heavier -walled fitting could
probably have been adopted.

At one time, control air leakage was consistently appearing at the large nut on top of the valve, indi-
cating that the gasket between the cylinder and case was not sealing correctly. It was noted that the
installation and tightening of the return line rotated the cylinder to the extent allowed by the Woodruff
key, this loosened the nut by that amount. Tightening the large valve nut after the return line was tightened
corrected the difficulty.

There was one instance main seat of the valve. This was caused by a flat spot on
of leakage past the
the knife-edge of the burner. A
second valve was installed and no leakage was detected past the seat.
Although no hardness tests were made, it was assumed that the rubber composition was softer on the
second valve than it was on the first.

At one time, these valves included a switch which operated when the valve went to the preliminary
stage position. This switch was used as an interlock to prevent main stage operation if the valve had not
operated. It is not known why this switch was later omitted, but it seems that an operator could be ex-
pected to wait for a preliminary flame before giving main stage, in which case the switch would not be
necessary.

Figure 49 on Page 133 of the A-4 Manual notes that this is a "pressure switch for control of the pre-
stage motion." Actually, this was not a pressure.switch in domestic terminology, but a mechanical switch
operated by a push rod which rode on the valve stem.

It was possible to adjust the preliminary stroke by turning the valve disk and locking nut on the stem.
The Germans suggested the removal of the alcohol return line, and the capping of the top of the valve.
They considered that the original reasons for this design, that is, the reduction of: (1) water-hammer and
(2) the possibility of burner explosion, were not sufficient to warrant continued use of the system. Tests
at WSPG did not indicate the severity of the supposed water-hammer if the valve were used with the by-
pass blocked. Some tests, however, showed that the valve would lack 1/8-inch of closing completely under
control air pressure with the by-pass line in position. With the by-pass line blocked off, the condition
would be changed considerably, but effective pressure areas indicated that with 500 psi control pressure,
220 psi alcohol pressure would begin to open the seat if there were no burner pressure. This showed that
there would be no continuous complete closure of the valve, although the restriction to flow caused by the
valve might be enough to cause a water-hammer effect or cause excessive pump-outlet pressure. Possi-
bilities of a burner explosion would be increased by the blocking of the return line. All air in the burner
would be entrapped during alcohol filling. During the rest of the firing preparation, this air would have
time to mix with vapors given off by the alcohol, with the resultant possibility of an explosion in the com-
bustion chamber when the alcohol valve was opened. If this happened, there would have been a good possi-
bility of the flame propagating back through the nozzles and causing an explosion in the jacket.
A. 2. 8 Alcohol Preliminary Valve

References: A-4 Manual , p. 56 and 123; Backfire, Vol. II, p. 34, 41, 42, 104 and 105.
The primary function of this valve was never entirely clear, although there are a number of possible
purposes.

a. The Germans pressurized the alcohol tank after burnout to prevent the collapse of the tank upon its
re-entry into the atmosphere. There would have been some loss of air through the Z.K. nozzles if there
had been no preliminary valve.

b. If control air was last prior to launching, the preliminary valve would have prevented the dumping
of tank alcohol, although the alcohol in the piping would have been dumped.

c. In the event of a misfire in which the Z.K. nozzles had been melted, the preliminary valve would
prevent prolonged burning caused by alcohol being supplied through the Z.K. nozzles.

d. Possibly this valve was originally designed into the system to confine the alcohol to the tank until
just before firing thus preventing the freezing of alcohol in the tube that runs through the lox tank.

The was extremely dangerous to personnel not thoroughly acquainted with its
test of this valve
operation. one instance (not at WSPG) a test man was severely injured when his hand was
In at least
caught between the valve body and the valve cone.

Travel of the indicating switch was adjusted by adding shims beneath the push rod. Too many
shims, however, would cause damage to the switch. The switch was held in place by a cover nut with a
rectangular rubber grommet between the cover and the switch. This arrangement was generally unsatis-
factory for, over a period of time, the grommet would flow around the switch causing it to move away
from the push rod far enough, in some cases, to become inoperative. In other instances, the switch
would be floating within the grommet and would operate only part of the time, or would operate only after
a short time lag. In all instances in which switch trouble occurred after the tanks were put into the mid-
section, a pressure switch was installed in the valve pressurizing control line to perform, as closely as
possible, the same function as did the valve switch. Later in the program, failures were practically
eliminated by using good grommets from valves that had never been completely assembled and assembling
the valves only a short time before they were to be used. It was also determined that a small torque on
the switch cover nut would effectively seal against alcohol leakage, and would lessen the tendency for the
grommet to flow. Approximately 50 percent of the switches were cracked or broken. Possibly many of
these were damaged by excessive tightening of the switch cover. There were other cases of inoperative
switches -caused by corrosion and contact misalignment. Experience at WSPG indicated that heavier
duty switches were needed for this application.

The valves were assembled in Germany with solid conductor switch leads. These were changed to
stranded wire after the first 20 to 25 firings at WSPG. The aluminum tubes were replaced with copper
tubes for ease in installing the new wire.

In the initial assembly of the valve (during manufacture) a small hole was drilled through the top
locking nut and into the cylinder to take the spring locking clip. In one instance, the hole was drilled
completely through the cylinder. This defect was not noticed until alcohol was observed leaking from the
midsection.

In many instances, the 10 mm

threads in the cylinder stripped out while tightening the adapters.
Stronger 12 mm threads could have been used without difficulty.

The gasket between the "U" cup holder and the cylinder was fibre in most instances. The fibre
units caused considerable trouble and many had to be replaced. However, no trouble was encountered in
the few valves that had aluminum instead of fibre gaskets.

During one valve bench-test, a threaded nut on the bottom of the valve stem pulled away from the
valve cone. To determine the average strength of these connections, three valves were tested by secur-
ing the covers to the bodies and applying control pressure. In all three cases, 2000 psi control pressure
was insufficient to break the stems loose from the cones. It was never determined why the one stem nut
was so weak. The valve stems on about 30 percent of the valves were pitted and/or corroded.
70
The plug in the control pressure passage sometimes leaked control air. This had to be checked when
the case and cone were removed.
In almost every valve, the "U" cups had taken a permanent set and would not seal control air.Spare
German "U" cups were available at WSPG. These were invariably good and would seal, except when the
cylinder or push rod was defective.

A throttle valve in the control line of this valve provided a means for the slow closing of the valve.
The only reason advanced for this slow closing was that it would protect the rubber seat from being slammed
against the tank. With 510 psi control pressure, an average of 7.0 seconds elapsed between the de-energiza-
tion of the control valve and the complete closure of the valve. The valve stem did not begin to move to the
closed position until 1.8 seconds before the complete closure of the valve. This was determined by a bench
test; actual flight conditions would be somewhat different.

A.2.9 Oxygen Fill Valve

References: A-4 Manual ,

p. 57 and 125; B ackfire, Vol. II, p. 53, 55 and 103.

Twenty to thirtypercent of the valves on hand at WSPG had cracks between the bolt bosses and the
cases. This was caused by excessive tightening of the mounting bolts, coupled with the design weakness of
the casting.

Considerable trouble was experienced with leakage past the seats of these valves. Although re-turn-
ing of seats and lapping with a lapping compound was tried, it was found that lapping the two surfaces to-
gether with plain lubricating oil was most effective. Even when following this procedure, some valves
would leak lox after filling. This could sometimes be remedied (at the time) by tapping the seat very lightly
with a wooden hammer handle, or equivalent. If this did not stop the leakage, the cap was put on and tight-
ened, leakage was then confined to the valve. This was satisfactory, except that it was very difficult to re-
move the cap for lox topping. Some caps stuck so tightly that it was believed the piping might be broken in
the process of removing the cap. On later missiles, a 5/16-inch OD section of copper tubing was run from
a fitting installed in the cap to the lox vent line. This line carried away the leakage, makin it no longer
necessary to tighten the cap excessively.
There were a few instances of porous cases.
A.2.10 Switch Battery (Pilot Valves)

References: A-4 Manual , p. 141 and 149; Backfire, Vol. II, p. 51, 52, 54, 56, 110 and 111.

Paragraphs 58 and 59 in Backfire Volume , II mention that the cast-body-type valve is obsolete. This,
however, was the type used at WSPG.

The switch battery was one of the three most troublesome valves. The following faults, in order of
frequency, were found: (1) leakage past external aluminum washers, (2) leakage past outside rubber ring
washers, (3) leakage at servo bleed holes, (4) leakage at main bleed holes, (5) slow operation, (6) no oper-
ation, (7) leakage by electrical pins, and (8) low insulation resistance. At. least one instance of each
of the above faults could be found by testing a half dozen switch batteries; faults (1) and (2) could be
found many times. Some of the leaks could be stopped by tightening, however, only about one-third to one-
half of the switch batteries tested could pass tests. No attempt was made to overhaul the defective valves
since our supply was adequate. There were no instances of leakage through the case.

During the first part of the program, a number of switch batteries had to be changed at the launching
site due to slow operation. This was reduced during the latter part of the program by passing (on the basis
of bench tests) only those valves that were very rapid in their operation. A slight increase in operating
time would not necessitate the changing of the valve.

Often leakage would occur after a missile had been loaded with lox, but these leaks were almost al-
ways stopped by tightening.

Quick-disconnect terminals were soldered to the electrical pins on later missiles, since there was
concern that the original plugs might not be making good contact.
In one instance, it was believed that moisture collecting around the electrical pins caused a misfire
when the resistance across the pins became so low, a fuse was blown in the firing desk.

71
A.2.11 Burner Drain Valve

References: A-4 Manual ,

p. 141; Backfire, Vol. II, p. 55.

Most burner drain valves used at WSPG were American-made. About 20 percent of these were
of the
defective in that the hand wheel could be turned completely off. When this happened, reassembly was a
major job. No valves showed any leakage.
A.2.12 Ram Charger Valve
References: A-4 Manual ,
p. 56 and Backfire, Vol. II, p. 37, 103 and 104.

The tube from this valve to atmosphere was, in most cases, run through the warhead. On some mis-
siles at WSPG the tube was run through a door in the control compartment.

Approximately 15 percent of valves tested leaked control air. In most of these defective valves, the
leak was past the metal-to-metal top seal. Five valves manufactured in the United States were tested; three
leaked control air.

A.2.13 Alcohol Tank Drain Valve

References: A-4 Manual ,

p. 57; Backfire, Vol. II, p. 43, 44, 106 and 107.

Trouble was not experienced with this valve except for leakage (in a few instances) past the aluminum
washer on the spring-holding eye bolt.
A.2.14 A-3 Check Valve

References: A-4 Manual ,

p. 114 and 128; Backfire, Vol. II, p. 25 and 33.

This valve prevented gas from being lost through the burner before and after burning, as well as pre-
venting liquid oxygen from filling the heat-exchanger coils before steam was available. Considerable
trouble was experienced with leakage past the gaskets. A number of valves were found with rusty springs
and/or rusty seats.

A.2.15 Oxygen Vent Valve

References: A-4 Manual ,

p. 57 and 129; Backfire, Vol. II, p. 43, 47, 102 and 103.

Practically all valves tested, leaked at the steel control chamber cap. Except in isolated instances,
this leakage could be stopped by additional tightening of the cap.

In most instances the safety feature of the valve was out of adjustment. If the spring compression was
not adjusted carefully, the spring would be subjected to a torque which would cause the setting to change
after the valve had been operated a few times. A special wrench was used to keep the spring from turning
when the adjusting nut was being set.
An estimated five percent of the valves tested had one or more holes or pores through the case.

A few valves had leakage past the plastic piston, but in practically every instance this was due to nicks
or pits in the seat of the case.

Many on the leather facings of the large piston that looked and felt some-
of the valves had a coating
what This coating seemed to give a better seal against leakage, but it was removed as a
like soft soap.
safety precaution. Many of the valves leaked more than the allowable amount (2.54 cfm) past the leather
seats. This was due to the pitted and scarred condition of the leathers.

The Germans considered that the relief feature of this valve could not be depended upon for personnel
safety. According to one source, the Germans once actually shot holes in the lox tank to relieve pressure
after control airwas lost. During a test in Germany, the automatic cutoff did not function during lox tank
pressurizing and the lox tank blew up because the vent valve could not pass enough -gas to keep the pres-
sure low. A test was made at WSPG to determine the amount of pressure rise that could be expected with
a higher input volume. This was of course a bench test and no tank was used. With 2500 psi applied
through 1/4-inch OD thin-wall copper tubing, a lox vent valve that had passed all regular bench-tests held
a chamber pressure of 54 psi. This would be sufficient to blow up a lox tank. During this test the valve

72
box had a restriction in the lox tank pressurizing line which reduced the flow. However, it is possible that
the valve box in the incident referred to by the Germans did not have this restriction. The specification for
a regular bench test was that flow resulting from a pressure of 355 psi supplied through a 0.8 orifice, mm
should give a valve chamber pressure of from 28.4 to 31.2 psi. It is not known how this volume was ar-
rived at; possibly it is the calculated rate of lox vapor given off by lox boiling.

The normally-open valve in the valve box that controlled the vent valve was designed for quick bleed-
ing; the vent valve would slam shut and therefore seat better.

Tests were run to determine the effect of heating on the plastic control pressure pistons. It was found
that temperatures as low as 65°C might render the valve permanently inoperative.
The following information was obtained from German manufacturing directions: (1) spring and lock
rings were zinc coated, (2) small pores were allowed in the case, (3) guide ribs and seats were machined
in one setup, (4) rings were made of leather from the underside of the animal and were degreased with
trichloroethylene, (5) rings were cemented into the piston by applying the glue, installing the leather and
applying pressure by putting the piston in the corresponding casing and then loading the piston, (6) outside
diameter of the leather rings could not exceed the outside diameter of the piston and (7) oil or grease were
not used during the manufacture of any part of the valve.

American-made vent valves were manufactured to German specifications and were used on many of the
later missiles. Many of the valves leaked control pressure past the steel cap. In the later shipments, this
was corrected by using a copper washer between the cap and case. There was no instance of a porous case
on any of the valves. Leakage past the leather seats was very small. The plastic piston would withstand
about 10°C more than the piston in the German valve.

On a number was picking up metal where it rubbed against the case. It was
of valves, the large piston
later found that the clearance between pistons and cases was between zero and 0.006 inch, as compared to
0.009 inch or more on the German valves. All American-made valves were therefore disassembled and the
pistons turned down to give a clearance of from 0.011 inch to 0.012 inch before use. This eliminated the
trouble completely. In examining telemetry records, it was noticed that lox tank pressure would often
reach a peak then drop down 5 to 10 pounds before starting to rise again. Since this condition was not evi-
dent after the valve clearance was increased, it was considered possible that the valves were binding slightly
if the clearance was small. These results could not be duplicated on the ground, even by cooling the valve
with liquid oxygen boil -off.

A.2.16 Oxygen Main Valve

References: A-4 Manual , p. 121 and 131 and Backfire, Vol. II , p. 47 and 48.
Figure 34 is a cutaway view of the main oxygen valve.

The electrical connection leaked control air in most valves and had to be tightened. In some valves, the
plug had to be removed and the gasket replaced. A few valves leaked air past the steel gasket located below
the metal-to-metal seat.

During the latter part of the program, valves were used that had leakage past the metal-to-metal seal
(twice the amount allowed on the first valves fired). Since this leakage seemed to get worse with time, it
was suspected that corrosion was pitting the surfaces. Although several methods were used in an attempt
to decrease this leakage, none were successful. The methods included: (1) grinding and lapping the sur-
faces with various compounds and by various methods, (2) making new seating disks and (3) copper -plating
the seating surface on the piston.

The item that caused the most trouble of all propulsion components was the double lip seal in this
valve. If this seal were too loose, lox would leak past it and vaporize in the control chamber. If the
vaporization was so rapid that it could not be relieved through the switch battery, the resultant pressure
would start to close the valve. As the pressure in the control chamber was relieved through the switch
battery, the valve would start to move to the full open position which would admit more lox to the control
chamber and start the cycle again.

73
If the seal were too tight, the stroke would be rough, which could cause a missile to topple at takeoff.

According to the Germans this happened at times in Germany. The theory behind this is as follows: a
smooth operating alcohol main valve combined with a sticky oxygen valve may result in a condition in
which the oxygen valve sticks at the position where the thrust just overcomes the weight of the missile;
the main alcohol valve would continue to open, and as more alcohol entered the burner, the thrust was
decreased and the missile would settle back.
In the latter part of 1948, it was found that the double lip seal was loosening when the valve was cooled
with liquid oxygen. Tests at Fort Bliss and WSPG showed that control air leakage increased greatly; lox
vapors from the switch battery were observed. It was found that tightening of the seal nut when the valve
was cold would decrease the leakage. However, if the seal nut was tightened too much, a rough stroke
would sometimes result. Additional cold tests at Fort Bliss indicated that it was the temperature dif-
ference between the seal and the cylinder that was causing the leakage. Cold tests were run with a calrod
unit wrapped around the bottom of the valve; this reduced the leakage considerably. The calrod power
was supplied by a variac which was adjusted to insure that the calrod would not become uncomfortably hot
to the touch. Immediately after the double lip seal trouble was discovered, missiles were supplied with
a 25-ton valve on the bottom of the main valve to aid in venting the control chamber. The vent lines on
the 25-ton valve and on the switch batterywere run to the side of the missile. Also, the calrod was utilized
and control air leakage was monitored until access ports were closed to make sure the seal was not loosen^
ing to a dangerous degree.

Fig. 34 Cutaway View, Main Oxygen Valve

74
Meanwhile, the General Electric Research Laboratory at Schenectady was developing new type seals,
since it was believed that the trouble could be attributed to the old seals having lost much of their elas-

ticity. Several development seals were tested at Fort Bliss and sent to WSPG. These seals were slightly
smaller in diameter, and considerable stickiness was experienced until the seals and cylinders were coated
with "Aquadag," a collodial graphite compound. It was found that the amount of torque used on the seal
nut did not effect the leakage appreciably. The leakage with the new seals was quite low, even when the
calrod was not used. These new seals were used on the last few missiles; on the last four missiles, the
calrod was not turned on.

A. 2. 17 Steam Plant
References: A-4 Manual p. 17, 78 through 91, 121, 127, 128, 135, 142 through 153 and 222; Backfire,
,

Vol. n, p. 34 through 40, 53, 57, 59 through 72, 84, 104 through 111 and 116 through 119.

Fig. 35 V-2 Steam-generating Plant

75
a. Reducers
At least two reducers stuck at the launching site after lox was filled. It is believed this was caused
by moisture freezing between the shaft and the brass adapter. In both cases, the reducer became operative
when the pressure was dropped somewhere below 300 psi. On later missiles, the heater was turned on at
the start of lox filling. The heater remained on until takeoff, unless the firing was delayed, in which case
the heater was turned off if the reducer got uncomfortably hot to the touch.

Ahalf dozen or more reducers were changed at the launching site. The reason for this was always that
the differential pressure (difference in pressure between small bleeder valve open and small bleeder valve
closed) was too great. A differential pressure greater than 28 psi indicated that the reducer was getting
dangerously sticky or the seat was being pitted.

During the first part of the program, the reducers were set immediately before leaving the missile.
On later missiles, the reducers were set before liquid oxygen was filled after first having the heater on
for about 10 minutes. The inlet pressure was kept as near 2500 psi as practicable, although some reducers
were set with inlet pressures as low as 2100 psi. Reduced pressure was monitored at the missile until the
access hatches were replaced and on later missiles was monitored in the blockhouse until launching. It
was found from bench tests that the static pressure and the dynamic (small bleeder valve open) pressure
were erratic when the reducer was cooled, but the full flow pressure remained the same. Therefore, the
reducer was set before lox filling.

A
very few of the reducers fluctuated badly during the flow test. This indicated stickiness between
the shaft and the brass adapter in the open position. About a third of the reducers had too great a differen-
tial pressure during the initial steam plant test and were replaced. Some reducers stuck partly open or
closed during the initial flow test.

All of the reducers disassembled had a thin coating of grease on the inside. When this coating was
removed, the differential pressure increased. Also, the difference between dynamic pressure and full flow
pressure increased.
To minimize a permanent set of the spring, the adjusting screw was removed when pressure was not
required. This was merely precautionary, since experience on the calibration stand indicated that practi-
cally no permanent set occurred over a period of two to three months. Many of the adjusting screws did
not fit well into the spring case and some seized.

The spring cover was vented on all reducers. Some reducers had a hole drilled in the end and others
had a small check valve on the end of the adjusting screw. When it was necessary to change damaged adjust-
ing screws, care was taken not to put a blank adjusting screw into a case that had no relief hole. There
were instances of leakage past the diaphragm as evidenced by leakage out of the relief hole or relief check
valve.

When
a reducer stuck closed, it would cause an over pressure when it became free which would actuate
the relief valve. The relief valves had to be adjusted before they were operated, since they had been assem-
bled so long that the rubber seat had flowed around the knife edge, resulting in a much higher initial relief
pressure.
b. Gages
The N4R contact on the 0-60 kg per cm gage was not consistent in its operation. After the first 20
or 25 missiles, a line was tapped into this gage line and an American built pressure switch was installed.

The glass covers on both gages were replaced with plastic covers which had 1/8-inch diameter
holes drilled through. This was done because the bourdon tubes would sometimes fail and the pressure
would shatter the glass front.

c. High Pressure Hand Valve

This valve would often stick open. The check-valve part of the valve gave no trouble. Pressure dif-
ferential across the valve varied from zero to eight atmospheres.

76
d. Peroxide and Permanganate Bleed Valves
There was no known instance of any malfunction of these valves.

e. Peroxide and Permanganate Drain Valves

These valves often leaked pressure at the metal-to-metal juncture of the two halves of the case. The
external drain coupling would sometimes leave the seat tilted to one side, preventing the valve from sealing
against liquid.

f. Steam Generator Check Valve

According to one of the German engineers, the check valve was installed to avoid an explosion in the
steam generator (which might flash back into the peroxide tank) if the permanganate line should become
clogged and then cleared. During the latter part of World War II, the Germans made tests to determine if the
valve could be eliminated, however, no conclusive results were obtained.

g. Peroxide and Permanganate Check Valves

There were two different spring-types in these valves. One valve unit had a steel spring. In the other,
the spring was machined out of the same piece of aluminum or brass as was the seat, and was still attached
to the seat. The steel spring valves worked well, but many of the brass and aluminum springs had taken a
permanent set in the compressed position and were worthless.
h. Twenty-Five Ton Valve
Some of these valves leaked control air and were replaced. A few were rejected because of rough
strokes. Tests run on three valves showed that the valve would begin to open at about 25 psi if there were
no tank pressure, and that once open would remain open against 600 psi tank pressure even if the control
pressure dropped to 50 psi.
i. P. E. 4 Valve

Some of these valves would not pick up at the required voltage (18 volts). Pneumatically, the valves
were very good, and only an occasional defect was noted. One valve was energized with 30 volts for 7
hours without burning out; the outside of the coil reached a temperature of 120°C. Terminal covers were
not securely fastened and would often be blown off. In one instance (in the test room) the cover shorted
across the terminals when the holding strap kept the cover from being blown clear. After this, all such
covers on the missile were attached down with friction tape as an extra precaution.

j. Eight Ton Valve

About 30 percentof these valves leaked past the seat and were rejected. One of the valve coils was
energized with 30 volts until it burned out. It began smoking in 40 minutes and burned out about 2 hours
later; the case reached a temperature of 255°C. At normal operating pressure, the valve would open with
about 13 volts. Flow through this valve was regulated by an orifice in the peroxide inlet line.

k. Main Stage (P. E. 1. 0) Valve

This valve was one of the three most troublesome valves. Approximately a half of the units were de-
fective and had to be repaired or discarded. The plastic pilot scored and pitted easily, causing leakage at
the bleed holes, often rendering the valve inoperative.

Many of the "U" Cups had taken a permanent set and were replaced with spare cups. The rubber
seats were often pitted and some came out completely. With normal operating pressure, this valve would
open with about 10 volts. Two of the valve coils were energized with 30 volts until they burned out. Both
began smoking after about 45 minutes, and burned out immediately after. Outside case temperatures were
172 and 120°C.

77
A.2.18 Main Tanks (Alcohol and Oxygen)

References: A-4 Manual , p. 51 and 221; Backfire, Vol. II, p. 24, 26, 93 and 111.

On some tanks, the brackets were welded on, not riveted. These were not used at WSPG until the brackets
were riveted on. Leaks at rivets were repaired by re-riveting.

The side plates used to steady the tanks in the midsection were slotted to allow for tank expansion and
contraction.

Most lox tanks had holes or weak spots that had to be welded.

An attempt was made pressurize an alcohol tank to rupture. However, the skin pulled away from the
to
stiffening rings, leaving holes in the skin which leaked so much air that the pressure could not be held.
Maximum pressure reached was 41.5 psi.
An oxygen tank was pressurized to rupture; it ruptured across the top end at 52.6 psi with explosive
violence.

A. 3 TEST INSTRUCTIONS
The following are test instructions for the propulsion components. Original instructions gave test pres-
sures in kilograms per square centimeter. Valve travel is still listed in Millimeters since almost all valves
used were original German units. Leakage in "bubbles per second" refers to the leakage from a 1/4-inch OD
thin-wall copper tubing immersed in about three inches of water (it is admittedly a very crude measurement).

A. 3.1 Carbon Vanes

a. Examine x-rays and compare faults and classification with inspection sheet. The vane must be
Class A or Class B.

b. Examine vane for broken tips and other obvious defects. Discard any vane with red stripe painted on
the surface.

c. Examine side strips and replace broken units.

d. Using a wire brush, clean rust from the backing plate.

e. Check the clearance between the vane and the backing plate. This clearance must be equal to or
greater than the thickness of ordinary writing paper.

f. Torque all holding studs to 110 inch-pounds; do not go above this value. Start with the center studs
and work outward.

g. Bolt vane in tester and gently lower the weight so that pressure is on the vane. Leave weight on for
one minute, then raise weight, turn vane over, and apply pressure on the other side for one minute.

h. Remove vane from fester and again torque all holding studs to the 110 inch-pounds, starting with
center studs and working outward; do not torque vane studs above 110 inch-pounds.

i. Complete and sign two vane test sheets; one for filing and the other to be placed with vanes.

j. Install holding ears, nuts and locking strips on vanes.

k. Seal vane box to discourage tampering.

78
A.3.2 Burners

a. Visual Examination

1. Use only burners that have been static fired.

2. The burner must have three expansion folds (in some units the middle fold is missing).

3. The cooling lines must have an expansion loop.

4. The burner must have a thrust-frame connection.

Inspect the entire body of the burner for cracks in the welding. Burners should be rejected
5. if

cracks are found at places difficult to work on, such as a burner head inside or outside, combustion
chamber and nozzle inside.
b. Cleaning Process

If there is much dirt on the outside or

1. in the combustion chamber, remove it with a wire brush
and water, or by any similar' method.

2. Cleaning with compressed air

all large particles of dirt (dust, earth, metal shavings) blow compressed air in the
To remove
opposite direction of normal operation into the cooling-jacket, cooling lines and nozzles. To obtain
thorough cleaning:

(a) Close the oxygen intakes with blind flanges.

(b) Close the opening for the alcohol valve with a blind flange.

(c) Remove all pressure reducing throttles from the cooling lines.

(d) Remove screw plug at the alcohol intake fold.

(e) Connect burner pressure gage to the blind flange of the alcohol valve opening.
Place burner on top of the test plate and fasten it with turn-buckles. Compressed air is fed
(f)

through the test plate. Pressure of compressed air in burner is not to exceed 14 psi. Before and during
the cleaning, hammer the cooling jacket and the cooling lines with a wooden or plastic hammer to remove
loose particles of dirt. Repeat procedure three to four times (four to five seconds of blowing each time)
until the escaping air appears clean.

3. Flushing with water

Water should dissolve and flush dirt which is firmly encrusted in the cooling-jacket. To flush:

(a). Attach the flushing device to the opening of the valve seat.

(b) Open the six alcohol intakes.

(c) Open the flushing-screws

at every expansion fold. Flush the water through at a pressure of
about 150 psi for one or two minutes (two to three times).

c. Pressure Test
1. Subject cooling jacket and lines to 265 psi water pressure

For reasons of safety, make these tests with a water pressure pump. Cooling jacket and cooling
lines should be tested separately.

(a) Cooling jacket

(1) Place the testing device in the opening for the alcohol valve.

(2) Close the six alcohol intakes with blind flanges.

(3) Fasten the pressure line to the gage-fitting on the alcohol intake fold.

79
(4) Exchange ZK nozzles for blind screws.

(5) Fasten manual drain valve at alcohol intake fold.

(6) Pressurize to 265 psi.

(7) When pressure is removed, check burner for evidence of leakage,

(b) Cooling lines

(1) Exchange the hollow screws for blind screws.

(2) Exchange throttles for blanking screws in the orifice rings.

(3) Fasten the pressure supply ring and the pressure line at the orifice nipples on the top ring.

(4) Pressurize to 265 psi.

2. Repeat the same test with 225 psi air pressure. After emptying the burner of all water, subject
the cooling-jacket and cooling lines to 225 psi air pressure to detect any small points of leakage. Test all
critical points, such as welding seams and connections, with soap solution. Testing procedure is the same
as 1. except that air is used instead of water.

3.Burners should be dried. After finishing the above tests, drain the water thoroughly. To remove
the water from the expansion loops of the cooling lines, lead compressed air through the cooling line intakes.
Remove moisture by blowing hot air into the venturi. Thorough drying takes from 6 to 8 hours. The fol-
lowing are the necessary preparations:

(a) Replace the blind flange (in the opening for the alcohol valve) with the testing device.

(b) Exchange the blind screws for hollow screws in the intakes of the cooling lines.

(c) Install the original throttle orifices in the rings (5 mm orifices in the two bottom rings, and
4.5 mm orifices in the two top rings). Install all orifice cover plugs except the units in the top ring.
(d) Remove the drain plugs at the expansion joints.

4. Pressure-test the venturi-protector with 14-psi compressed air. Check all welding seams with
soap and water. If a leak is not indicated, close the pressurizing valve. If the pressure as measured at
the testing gage decreases, a leak does exist and is between the venturi-protector and the lower cooling
chamber.
5. Pressure-test the entire burner with 14-psi compressed air.

(a) Close orifice cover plugs of top ring with good gaskets.

(b) Put a drain valve or a blind plug on the drain opening at the alcohol intake fold.

Place burner on test plate and connect it to the compressed air supply.
(c) Check the safety
valve of the system and make sure that it opens and releases pressure at 17 psi.
(d) Maintain test pressure of 14 psi accurately. Specifically, check burner with soap and water
for the following:

(1) Leaks at welding seams at the outer head of the burner

(2) Leaks at welding seams of the injection cooling chambers

(3) Leaks at plugs which were open during the drying process

Rising pressure at the outlet opening of the venturi protector. This is a check of the weld-
(4)
ing seam between the space of the venturi protector and the bottom cooling chamber. After leaks are re-
paired, perform a test for constancy of pressure. A pressure of 14 psi must remain constant for 15 min-
utes with the pressurizing valve closed. With the pressurizing valve closed, the pressure (14 psi) should
remain constant for 15 minutes.
d. Final Work
After the testing and cleaning is finished, close all openings carefully with blind flanges or plugs and
good gaskets, if they were not already closed perfectly during the test. After, protect the burner from all
dirt, especially that which might get in through the venturi. The blind flanges at the oxygen inlets stay in.

A.3.3 Air Bottles

a. Pressurize manifold hydrostatically to 4500 psi. Release pressure and dry.

Pressurize bottles (with adapters installed) to 4500 psi hydrostatic.

b. Keep pressure on for one
minute, release, drain and dry.

c. Assemble bottles to manifold and rack, connect pressure hose and immerse assembly in water tank
in steam-plant test room. Pressurize with air to 3000 psi. Observe for leaks through test room window.
DO NOT ENTER TEST ROOM WHILE BOTTLES ARE AT THIS PRESSURE. If it is necessary to enter
the room determine location of leak, bleed pressure to 1500 psi. If manifold requires welding to repair
to
leakage, another 4500-psi hydrostatic test will be required as well as another 3000-psi assembly air test.

d. Cap off manifold inlet for storage.

A.3.4 Heat Exchanger

a. Visual Examination

Examine unit and clean if necessary. Watch especially for the following:

Heat exchanger must show no traces of grease or oil. If such substances are detected, the unit
1.

must be thoroughly cleaned with a grease-dissolving liquid to avoid danger of explosions.

2. The pipe coils and especially the nozzles must not be clogged.

b. Pressure Test
1. Subject casing to 28-psi water pressure

Subject pipe coils to 71 -psi air pressure.

2. For this test it is advisable to submerge the unit under
water to detect small leaks better.
3. After testing, dry the heat exchanger carefully with warm air to prevent oxidation.

c. Performance Test
Test arrangement is shown in Fig. 36. Carefully open valve until the precision pressure gage shows
7.1 psi. Differential pressure at the U-tube (filled with water) must remain constant at 260 ± 25 mm.

d. Final Operations

Close all intakes and outlets to prevent entrance of dust. A test reports sheet for the heat exchanger
is shown in Fig. 37.

A.3.5 Important Pipes and Fittings

a. Test procedure for alcohol two-fold fittings, alcohol three-fold fittings, alcohol lines from two-fold
to three-fold fittings, alcohol feed lines, alcohol return lines (both parts), oxygen lines and oxygen three-
fold fittings is noted below.

1. Pressurize hydrostatically to 500 psi.

2. Drain and remove cover plates and let dry at least 12 hours.
3. Pressurize with air to 100 psi; check for leakage both by immersing in water and with soap
solution.

4. Any leakage necessitates rewelding and complete retest.

81
.212"

U TUBE MANOMETER
PRECISION GAGE
t

j s
= 1 ,1

g _J56I"_ 8;

AIR INLET 2 1
2 <
HEAT EXCHANGER <
Ul
IH
DC
rr

Q. o
o 1

ORIFICE

Fig. 36 Apparatus Arrangement for Heat Exchanger Test

TEST REPORT

HEAT EXCHANGER

No. of Heat Exchanger:

Used in Rocket No.

1) Visual Examination

a) General Condition
b) What repairs were effected?
c) Was the heat exchanger cte. ned?

2) Pressure Test
a) Casing:

Pressure (2 atm. water pressure)

Leakage
b) Pipe Coils:

Pressure (5 atm. air pressure)

Leakage
c) Drying

3) Performance Test
a) Pressure in front of nozzle (0.5 atm.)

b) Differential Pressure

4) Final Operations

All lines and fittings closed dust-tight

Fig. 37 Heat Exchanger Test Report Sheet

82
b. Test procedure for steam line:

1. Pressurize hydrostatically to 600 psi, watch for leaks.

2. Any leakage necessitates rewelding and complete retest.

c. Test procedure for alcohol intake pipe with expansion joint:

1. Blank off assembly and put holding clamp on expansion joint.

2. Pressurize with gas to 29 psi and check for leakage with soap solution.

d. Test procedure for expansion joints:

1. Blank off expansion joint and affix tie rods.

2. Pressurize with gas to 29 psi; check for leakage by immersing in water on side and rotating so
that all soldered joints can be scrutinized. No leakage is allowed.

A.3.6 Alcohol Main Valve

a. Plug openings on side of auxiliary case and on by-pass line. Apply 170 psi to the bottom of the aux-
iliary case. The main stage must have a stroke of from 28 to 30 mm. Repeat this test five times.
b. Apply 170 psi to the bottom of the auxiliary case. Check the by-pass line for leakage. No more than
three bubbles per second are allowed. Check the control pressure connection pipe for leakage. No more
than three bubbles per second are allowed.

c. Apply 570-psi control pressure. The prestage must move immediately and the minimum stroke must
be 5.5 mm. Repeat this test five times.
Apply 570-psi control pressure. Apply soap solution around case threads and both plugs. Leakage
is not allowed.

d. Apply 570-psi control pressure. Plug off air 7 inlet at bottom of auxiliary case. Check the by-pass
line for leakage. No more than two bubbles per second are allowed.
e. Apply 570-psi control pressure. Check the auxiliary case for leakage at the side. No more than
two bubbles per second are allowed.

Apply 570-psi control pressure and apply 21 psi at bottom of auxiliary case. Check for leakage at
side of auxiliary, case. No leakage in addition to the leakage in e. above is allowed.

A. 3. 7 Alcohol Preliminary Valve

When operating this valve, use test stand. Never place hands between the valve body and the cone
when applying control air. Keep hands clear and exercise extreme caution. Improper use can result in
serious injury.

a. Remove switch, rewire with stranded wire, and replace, using good grommet. Do not overtighten.

b. Exchange "U" cups for units that have not been used.
c. Apply 570-psi control air. Apply 15-psi pressure at the vent line. Check for leakage with soap
solution at vent fitting and at all fittings on top of valve. Remove the 15-psi pressure before the 570-psi
control air is removed or the center "U" cup will be damaged.

d. Apply 70-psi control air. The valve cone must open smoothly. In the open position, the switch
must make contact. Bleed the control air. The valve cone must close smoothly. Repeat test five times.
e. Apply 570-psi control air. Check the leakage at the breather pipe connection. No leakage is
allowed. Check for leakage at the push rod connection. Leakage is not allowed. Repeat test five times.
A test report sheet for this valve is shown in Fig. 38.

83
PRELIMINARY ALCOHOL VALVE BP L50

Test 1 Test 2 Test 3

By
Date

1. SMOOTHNESS OF STRIKE, OPENING

2. SIGNAL LAMP

3. SMOOTHNESS OF STROKE, CLOSING

4. LEAKAGE AT BREATHER PIPE

5. LEAKAGE AT PUSH ROD

a. Packing gland (inner seal)

b. Gasket (outer seal)

REMARKS:

Fig. 38a Alcohol Preliminary Valve Test Report Sheet

Fig. 38b Alcohol Preliminary Valve

84
A.3.8 Oxygen-filling Valve

a. Apply 42 psi to the tank side of the valve. Leakage greater than four bubbles per second at the inlet
side is not allowed. Soap the valve case. No leakage is allowed through pores in the case.

b. It must be possible to open the valve piston 20 mm without feeling any friction.

c. Check for cracks between the bolt bosses and the case. Cracks are not allowed.

A.3.9 Switch Battery

a. Install switch battery on test fixture. Tighten the two electrical connections, the control pressure
connection, the two pressure outlet connections and the four screw connections on bottom of switch
battery.

b. Apply 570-psi control pressure. With soap solutions, test all screw connections including: (1) outer rim of
four screw connections at bottom of switch battery, (2) electrical connections, (3) screw connections on pressure
outlets, (4)screw connections on control pressure inlet, (5)screw connection on plug opposite control pressure
inlet and (6) area connecting the two halves of switch battery. No leakage is allowed at any of these locations

c. Apply 5 70-psi control pressure. Turn electric switch "ON." Maximum voltage is 18 volts. The
switch battery must operate within 0.5 second (with automatic timer used, light will go on if time is too
great). Repeat this test five times. Do not leave switch on more than one minute and do not use more
than 18 volts.

d. Apply 570-psi control pressure. Throw switch "ON." Check for leakage at the main bleed hole
and at theservo bleed hole. No more than two bubbles per second at either hole is allowed.

e. Apply 570-psi control pressure. Throw switch "OFF." Check for leakage at the servo bleed hole.
No more than four bubbles per second are allowed. Check for leakage at the main bleed hole. Only two
bubbles per second are allowed.

f. Disconnect electrical plugs. Apply 40 atmospheres control pressure. Test the electrical connec-
tion for leakage. No more than two bubbles per second are allowed.
g. The resistance of insulation may not be less than five megohms.
A.3.10 Alcohol Drain Valve

a. Apply 355 psi to the inlet side. No more than one bubble per second is allowed at the discharge side.

b. Apply 28 psi to the inlet side. Block the discharge side and open the valve. No leakage is allowed
at the pin.

c. Open the valve all the way and then close it. The valve must seat if more than five revolutions
are made.

d. The minimum distance between the case and the handwheel must be 1 mm in the closed position.

A. 3. 11 Ram Charger Valve

a. Apply 570-psi control pressure. The piston must seat quickly and smoothly. The minimum stroke
must be 14 mm. Repeat this test five times.

b. Maintain 570 psi and check the valve for leakage. No more than two bubbles per second measured
at the intake side are allowed.

c. Maintain 570-psi control pressure. Apply 20 psi above the piston. Check the leakage at the intake
side. No more than two bubbles per second in addition to the bubbles found in b. above are allowed.
A.3.12 Alcohol Drain Valve (Tank)
a. Apply 28 psi to the auxiliary case. No leakage is allowed anywhere on the drain side.
b. Smooth movement of the valve should be checked by hand. The minimum stroke must be 4.5 mm.

85
A.3.13 A-3 Check Valve

a. The valve piston must have a minimum stroke of 4 mm.

b. Apply 35 psi at the discharge side. No more than 0.05 cfm leakage at the inlet side is allowed.

c. Block the discharge side and apply 20 atmospheres pressure at the inlet side. The threads of the
case must be tight.

A. 3. 14 Oxygen Vent Valve

a. Apply 100 psi through the control pressure connection. The valve must open quickly and smoothly.
Repeat five times.
b. Insert a metric scale from the outlet side of the valve so that the scale rests on the center stud in-
side the valve. The scale will then be inside the valve spring. Using a straightedge for ease in reading,
take the initial reading. Apply 100 psi at the control pressure connection. Take a second reading. The
difference in the initial and second readings (valve travel) must be at least 10 mm. Repeat five times.

c. Place the test flange with a rubber gasket on the inlet side. Apply 355 psi through the 0.8 mm
orifice on the inlet side. The pressure in the chamber must remain constant between 28.4 and 31.2 psi
measured with a precision pressure gage.

d. Leaving the test flange on the inlet side, install an additional test flange on the outlet side. Use a
rubber gasket. Apply 575 psi through the control pressure connection. Install the bubble indicator. No
more than four bubbles per second are allowed. Make sure the openings on the test flange at the inlet
side are plugged.

e. Remove bubble indicator and install flowmeter. Apply 15 psi at the inlet side. Increase the pres-
sure until the chamber pressure is 17 psi (actual inlet pressure may have to be increased to as much as
100 psi to get 17 psi in the chamber). Read the flowmeter. Maximum allowable flow leakage past seat
is 2.54 cfm.

A. 3. 15 Oxygen Main Valve

a. Hot Tests

1. Replace the double lip seal with an American-made seal, lubricate seal and cylinder with
Aquadag and reassemble.
(a) Apply 570-psi control pressure. Check the leakage at the fitting mounted on one of the six
flanges. No more than four bubbles per second are allowed.
2. Apply 570 psi at control pressure line and check electrical plug for leakage. In most cases this
plug will have to be tightened or the gasket replaced. Be careful not to strip the aluminium threads.

3. (a) Bleed the control pressure line. Blank off the side fitting on the flange which was open in
1. and apply 170-psi pressure through the main flange (top). The minimum stroke of the main stage must
be 31 mm. No binding is allowed. Valve should open quite smoothly, repeat test five times.
(b) Maintain 170 psi and check the control pressure connection-pipe for leakage. No more than
four bubbles per second are allowed.

(c) Check case and bottom gasket(between housing and center sections of valve) for leakage.
No leakage is allowed.

4. (a) Bleed the 170-psi pressure and apply 570 psi at the control pressure line. Apply 21 psi
through the main flange. Check the leakage at the side fitting. Maximum leakage allowed is seven
bubbles per second in addition to the leakage obtained in 1. above.
b. Liquid Oxygen Test

Place special potentiometer on bottom of valve to measure the operation of the main stage stroke.
1.

Fasten valve to the lox container. Supply air to the control pressure line by means of a set of three air
bottles (21 litres). This set of bottles is pressurized with dry air to about 2000 psi. A regulator and
switch battery is mounted on the bottles. A recorder is connected to the potentiometer. Dry air is sup-
plied through a regulator and solenoid valve to pressurize the tank. A solenoid-operated bleed valve is
mounted at the heat-exchanger connection (on the top of the tank). All these valves are operated elec-
trically.

2. Procedure
(a) Open hand valve on set of three bottles and supply 5 70-psi control pressure to main oxygen
valve.

(b) Operated solenoid valve to close off connection on heat exchanger.

Start recorder and operate switch battery, which will open preliminary part of the valve.
(c)
Operate solenoid valve to close top vent and start pressurization. Main stage valve should open at 100 psi
maximum. Recorder will show the smoothness of opening. No sudden jerks should be present at normal
temperatures. Also, while valve is open, check the presence of severe leaks. After valve is pressurized
to 100 psi, stop pressurizing and stop recorder. Bleed air from tank by operating bleed valves and de-
energize switch battery.
(d) Connect flowmeter to heat exchanger after removing bleed valve. Record the pressure in
the three air bottles. Put in lox at top of tank and fill completely. Take readings for one hour, recording
flowmeter and pressure drop in air bottles every three minutes.
(e) Flowmeter reading should not exceed 2.0 cfm. Next, remove flowmeter and connect bleed
valve again. Pressurize tank also. This time the valve must start opening at 100 psi and be completely
open at 150 psi. The stroke on the recorder will be quite rough. Check for any long period of stickiness.
(f) Remove valve and dry thoroughly. Install electrical contact and set preliminary stroke to
operate between 2.5 and 3.5 mm. Make hot tests again. Install calrod unit. Megger electrical connection
and calrod unit. Assemble adapter and 25-ton valve on bottom of lox valve. With 570 psi control air,
check gaskets and adapter on valve.

A test reports sheet for this valve is shown in Fig. 39.

A.3.16 Steam Plant

A. 3. 16.1 Preparation

a. Inspect interior of peroxide tank for dirt, rust arid corrosion. Tank must be clean and must have
black coating.

b. Clean exterior of steam plant with a wet cloth.

c. Check all electrical connections and insulate leads to the solenoid valves and regulator heater.

d. Replace the PE-10, Z contact and eight-ton valve with parts tested previously.
Replace the glass on both the 250 and 60 atmosphere gages with plexiglass and drill a small hole
e.
for venting in case the bourdon tube should break.

f. Make sure all fittings are reasonably tight.

g. Install a special cross in the control line that runs from the regulator-gage line to the PE-4
lines. Plug off that side of the cross in which the 0.017-inch diamter hole is located and one other side.

h. Calibrate the recording gages.

Connect gages to steam plant. Connect high-pressure line to intake of steam plant. Plug off
i.

steam generator with special orifice. Connect valve to allow air to escape through this orifice when
testing steam plant.

87
j. Make sure a clean filter is in the air supply line.

k. Remove safety valve NO-10, test according to instructions, and reinstall before making any
steam plant tests.

1. Megger and check resistance of each valve.

OXYGEN MAIN VALVE 441 3 C

TEST SHEET

VALVE NUMBER

1. A. Stroke of Pre-Stage

B. Signal Lamp

C. Leakage of Control Pressure

2. Leakage Past Seat

a. Leakage Past Inner Seat

3. A. Length of Stroke

Smoothness of Stroke

B. Leakage Into Control Pressure Chamber

4. Insulation Resistance

5. Leakage Past Electrical Connection

6. Leakage at Case Threads

7. Leakage Between Valve & Case

8. Case Leakage

REMARKS:

DATE: TESTED BY:

Fig. 39 Test Report Sheet for Liquid Oxygen Valve

A.3.16.2 Leak Tests

a. Fill both air batteries to approximately 1500 psi. Close high-pressure hand valve on steam
plant.

b. Close switch 2 on test box. This action supplies air from one set of air bottles.

c. Open high-pressure hand valve.

d. Set pressure regulator to 550 psi.

e. Close steam plant bleeders, Dlh.' 1 ^)

f. Close bottom of steam generator.

g. Open main stage valve Dh. Steam plant will start to pressurize.

h. When pressure levels off, shut off Dh valve and open up eight- and 25-ton valves. Check record-
ing gages for anypressure drop which would indicate leakage of air. CAUTION: Never enter the steam
plant room when a pressure of this magnitude exists in the tanks.

i. Bleed all air in the tanks (through the steam plant bleeders) by opening Dlh.

j. Reset regulator to 485 psi.

k. Pressurize tanks again to 300 psi.

With soap and water, check all fittings, valves, lines, bleedholes and welds for leakage. Make
1.

sure eight- and 25-ton valves are open. CAUTION: Always keep away from steam plant bleeders; loss
of power or severe leakage of air will cause bleeders to open.

m. Bleed air from steam plant and repair leaks.

n. Repeat steps 1. through m. until all leaks have been eliminated.

A.3.16.3 Valve Checks

Close high-pressure hand valve, bleed air through the small hand valve on the regulators.
a.
Determine if the pressure on low-pressure gage drops to zero and bleeding stops. This is a test for
leakage past the high-pressure hand valve.

b. Open high-pressure hand valve again and pressurize steam plant to 300 psi. With 18 volts,
check for quick and smooth operation of solenoid valves.

c. De-energize Dh. Break the connection between Dh and the Z and T check valves. If there is
any leakage at this break, either the PE-10 or the two check valves are faulty.

d. Leakage from the 25-ton valve past the discharge side can be checked by closing the 8-ton
valve and capping off the line on top of the permanganate tank. Check for leakage after removing cap
on line just below the 25-ton valve. No more than two bubbles per second are allowed.

e. Replace and tighten all fittings. Check for leaks with soap and water. Close high-pressure
hand valves.

A.3.16.4 Operational Tests

a. Start heater on reducer, and operate at 28 volts for five minutes.

b. Fill air bottles between 2850 and 3125 psi.

c. Close off compressor supply and open valve 2.

d. Open high-pressure hand valve.

e. Set regulator to 469 psi on recorder (with small needle valve open). With needle valve closed,
pressure should not exceed 490 psi.
f. Ink recorders.

g. Close bleeders on steam plant.

h. Close bottom of steam generator.

i. Open Dh valve (pressurization has started).

j. Open eight- and 25-ton valves

k. Throw switch to air bottles 1 and cut off 2.

1. Throw switch for 110 volts to recorder.

m. Open bottom of steam generator; start recorder.

n. Allow air bottles to drop to 700 psi.
Close 25-ton valve. Steam pressure should not drop below 285 psi.
o.
Close eight-ton valve.
Steam pressure should not drop below 125 psi.
p. Close Dh valve.

q. Stop recorders.

r. Open steam plant bleeders (test has ended).

s. Remove ink from recorders.
t. Take pressure off regulator.

u. Megger all valves.

v. Remove test lines.

w. Put protective fittings on all open ports and safety-wire all fittings.

A test report sheet for the steam plant is shown in Fig. 40.

A.3.17 Reducer Safety Valve

a. Apply 490-psi pressure. The seat must be absolutely tight.

b. Increase the pressure slowly. Between 500 and 555 psi a blowing must be heard. As the pressure
is increased the blowing must also increase.

c. Reduce the pressure to 285 psi. Between 555 and 285 psi the valve must close and be absolutely
tight.

A. 3. 18 Hand-operated Valve on Reducer Assembly

a. Block the discharge side and apply 570 psi to the inlet side. Open and shut the valve three times.
These operations must be done easily by hand. One revolution of the handwheel is a minimum
necessary to open or to shut the valve.
b. Open the valve; no leakage at the packing gland is allowed.

c. Close the valve; leakage may not exceed one bubble per second at the discharge side.

A. 3. 19 High-pressure Hand Valve

a. The minimum measured stroke of the fill piston must be two mm.
Block the discharge side and apply 2850 psi to the inlet side. Close and open the valve several
b.
times by turning the handwheel. The wheel must move easily and smoothly a minimum of three revolu-
tions.

c. Maintain 2850 psi and check for leakage at the case threads and at the packing. No leakage is
allowed.

90
d. Close the valve, open the discharge side and apply 2850 psi to the inlet side. Check the discharge
side and the fill connection for leakage. No more than one bubble per second at either place is allowed.
A. 3. 20 Peroxide and Permanganate Bleed Valves

a. Apply 570 psi to the control pressure side. The piston must move quickly and smoothly. Minimum
stroke of the piston is 6mm. Bleed the control pressure side. The piston must return quickly and
smoothly. Repeat this test five times.

b. Apply 570 psi to the control pressure side. Check for leakage at the discharge side. No more
than two bubbles per second are allowed.

c. Maintain 570 psi at the control pressure and inlet sides. Test the leakage at the discharge side.
No more than two bubbles per second are allowed in addition to the leakage in b. above.

A. 3. 21 Peroxide and Permanganate Drain Valves

a. Apply 57 psi to the working-pressure side. No more than one bubble per second is allowed at the
valve seat. No leakage is allowed at the case threads.

A. 3. 22 Peroxide and Permanganate Check Valves

a. The check valve must have a minimum stroke of 2mm. The force necessary to open it must not be
more than 1200 grams.
b. Apply 570 psi at the discharge side. Test the leakage at the inlet side. No more than two bubbles
per second are allowed. No leakage is allowed at the case threads.

TESfREPORT
STEAM PLANT

No. of Steam Plant:

Built into Rocket No.:

1) Visual Examination

a) Result of interior inspection of tank

b) What parts were missing?

c) What parts were replaced?
d) General condition of the plant

2) Test for Leakage

a) Test Pressure (I

c) Necessary repairs

3) Performance & Operation Test

a) Setting of pressure reduction valve

1) Pressure with ventilal closed (static)

2) Fluctuation of pre-set

b) High pressure:

1) Start of test (2) End of

c) Lou pressure (2)

d) Peroxide tank pressure

e) Steam pressure

f) Steam pressure of the 6- stage

4) Final Operation

a) Pressure reduction valve released

b) All gauge fittings closed

c) All other fittings protected

Tested bv:

Fig. 40 Test Report Sheet, V-2 Steam Plant

91
A. 3. 23 25-ton Valve

a. Apply 70-psi control pressure, bleed and repressurize 10 times quickly. Listen for rattle-free
moving of the piston. Minimum stroke is 5.5 mm.
b. Apply 570-psi control pressure. Check for leakage at the bleedhole (two mm diameter). No more
than three bubbles per second are allowed.

c. Block the discharge side and the pipe-connection side of the valve. Apply 570 psi to the inlet and
control pressure sides. No more than one bubble per second at the two mm bleedhole is allowed.

d. Bleed the control pressure. Apply 570-psi at the inlet side. Check for leakage at the discharge
side. No more than one bubble per second is allowed.

A.3.24 PE-4 Control Valve

a. Apply 570-psi to the inlet side. Blank off outlet and soap the bleedholes. No leakage is allowed.
Remove blanking plug.

b. Maintain 570-psi control pressure. Connect outlet to 1 litre pressure flask. Throw the electric
switch to the "ON" position (maximum voltage, 18 volts). The valve must operate without delay. Throw
the switch to the ''OFF" position. Now the valve must bleed quickly and smoothly.

c. Apply 570 psi to the inlet. Throw the electric switch to the "ON" position. No leakage at the bleed-
holes is allowed.

d. The resistance of the insulation must be more than five megohms.

A.3.25 Eight -ton Valve

a. Apply 570 psi at the inlet side. No more than one bubble per second at the discharge side is
allowed.

b. Maintain 570-psi pressure. Throw the electric switch to the "ON" position (maximum voltage,
18 volts). The valve must operate without delay.
c. Block the discharge side. Maintain 570-psi pressure. Throw the electric switch to the "ON"
position. No leakage is allowed at the case threads. No leakage is allowed between the coil and the case.
d. The resistance of the insulation must be more than five megohms.
A. 3. 26 High-pressure PE-10 Valve
a. Apply 570 psi to the inlet side. No leakage is allowed at the valve seat or at the bleedholes.

b. Maintain 570 psi at the inlet side. Throw the electric switch to the "ON" position (maximum voltage
is 18 volts). The valve must operate immediately. A delay of more than two seconds is not allowed (check
this by the sound). Throw the switch to the "OFF" position. The valve must shut immediately. Repeat
this test five times.

c. Block outlet side. Apply 570 psi to the inlet side. Throw the electric switch to the "ON" position.
No leakage at the bleedholes is allowed.

d. The resistance of the insulation must be more than five megohms.

A. 3. 27 "Z" Contact

a. Apply 570 psi to the inlet side. No leakage is allowed at the case threads.
b. Connect a signal lamp in series with the pressure-switch contacts. Apply pressure to the inlet side
and observe the gage. At a pressure of 18.5 to 24 psi the lamp must light. Repeat this test three times.

92
A. 3. 28 Throttle in Line to Preliminary Alcohol Valve

a. Connect: (1) a measured volume of 0.9 to 1.1 litres and (2) a gage to the discharge side of the
throttle. Apply 570 psi to the inlet side rapidly. The gage must show 570 psi within two seconds. Bleed
quickly; the measured volume must be empty in 10 to 15 seconds. Repeat this test five times.
b. Close the discharge side of the throttle. Apply 570 psi to the inlet side. No leakage is allowed at
the case threads.

A.3.29 Check Valve on Cross Piece

a. Connect: (1) a measured volume of 0.9 to 1.1 litres and (2) a gage to the discharge side of the
check valve. Apply 570 psi to the inlet side rapidly. After two seconds, the pressure in the measured
volume must be 525 psi. Repeat this test three times.
b. Apply 570-psi pressure to the discharge side. No more than two bubbles per second is allowed at
the inlet side; no leakage is allowed at the case threads.

A.3.30 Alcohol Tank

a. Test sequence (test report sheet for this tank is shown in Fig. 41).

1. Visual examination.

2. If tank is in good condition, rivet according to Drawing No. P-SK-336.

3. Air-test at 1.5 psi and repair leaks.

4. Hydrostatic-test at 21 psi and hold for 15 minutes.

5. Air-test at 12.8 psi and hold for 15 minutes. Check for leaks with soap solution.
6. Repair leaks and repeat d. and e. until leaks do not occur.

TEST REPORT

ALCOHOL TANK

No. of Alcohol Tank:

Used In Rocket No.:

1) What repairs were effected?

2) Inside of tank was cleane d thoroughly

Was found o be In perfect shape

S) Preliminary Valve Test:

V»lve worked perfectly

inside control and vent lines In tank were checked

4) Tested Fuel Level Guage wasbuiltin

S) Last Pressure Test:

Test Pressure

Leakage

6) Alter final test all openings (especially measuring fitting, fueling filling and outlet fitting
were closed

Tested by: Dale:

Fig. 41 Alcohol Tank Test Report Form

93
A.3.31 Oxygen Tank

place whenever air pressure is being applied to the oxygen tank. It shall be
The vent valve must be in
the sole duty of one manreach of the air-supply hand valve and to watch the pressure gage.
to stay within
This safety precaution will be observed from the moment the air hose is connected to the tank until the air
is completely bled from the tank.

Flexible hose will be used in all pressurization.

Two gages (each on separate pressure tops) will be connected to the tank when air is being applied.
These gages V-2 personnel immediately prior to pressurization.
will be tested by

Barricades will be used when the fluctuating-air test is being made at the calibration stand and when
the tank pressure is being brought up to 21 psi in the hangar.

The area will be cleared and roped-off and adjacent offices will be vacated, when 21 psi tank pressure
is used during test in the hangar.

When the air test is being made at the calibration stand, the area will be cleared and roped-off.

a. Test Sequence

1. Visual examination.

2. If in good condition, rivet according to Drawing No. P-SK-335.

3. Air-test at 1.5 psi and repair leaks.

Do not start the hydraulic test (step 4. below) until it has been determined that the midsection will be
ready when the tests (through step 7. below) are finished.

4. Hydraulic-test at 32.7 psi and hold pressure for 20 minutes.

5. Apply pressure for one minute over range of 4 to 21 psi. Repeat five times.
6. Repeat hydraulic test at 32.7 psi. Hold for 20 minutes.

7. Bring tank into hangar and air test at 21 psi, and hold for 20 minutes. Check for leaks with soap
solution. If any leaks develop, repair and start test again at step 4.

8. Put tank in midsection.

A test report sheet for the lox tank is shown in Fig. 42.

TEST REPORT

OXYGEN TANK

Fig. 42 Oxygen Tank Test Report Form

94
APPENDIX B
PROPULSION UNIT CALIBRATION

B.l GENERAL
obtain optimum missile performance, it was necessary to insure that the propellants were supplied
To
to the combustion chamber in the proper quantities and proportions. Variations in pump characteristics and
in pressure drop through the motor jacket were large enough to make individual calibrations of each pro-
pulsion unit necessary.

In the German procedure, each major component (the turbine-pump assembly, the motor and the steam
plant) of the propulsion unit was tested separately. From the data thus obtained, calculations determined the
correct size of flow-regulating orifices to be used in the main propellant lines. These orifices regulated the
relative flow in the two propellant lines and thereby determined the mixing ratio. The total-flow rate was
adjusted by varying the pressure of the gas supplied to the steam plant. From the steam-plant test data and
the size of the orifices selected, a calculation determined the gas pressure required to produce the desired
flow and thrust.

At the start of the V-2 program at WSPG, test papers were available on enough turbo-pumps and motors
to complete thirty missiles. From these data, calculations were made and orifices were selected for the
first 30 missiles. Results by this method of selection were not consistent. It seems probable that the in-
consistency was due primarily to changes in the pressure drop in the alcohol jacket of the motor. Many
months had elapsed since the German tests and some change in the surfaces of the inner walls might be
expected.

Complete test data were not available for the steam plants. Therefore, an air test was necessary to
determine the characteristics of the steam plants. From these data, the setting of the gas-pressure
regulator was determined.

To arrive at a mixing ratio of acceptable accuracy (for those missiles beyond the thirtieth) it was
necessary to establish calibration facilities and procedures at WSPG.
B.2 METHODS OF CALIBRATION
Among the various methods of calibration, the following were considered for use at WSPG.
B.2.1 Hydraulic Testing of Component Parts

method, each of the major components was water-flow tested under specified conditions. The
In this
test results would then be used, as in the German procedure, to calculate orifice sizes. This method
offers certain advantages when mass production is involved, but these advantages could not be fully
realized on a relatively small number of missiles at WSPG.

B.2. 2 Water-flow Test of Complete Propulsion Unit

The results obtained by this method are more accurate than those obtained by section B.2.1 above, due
to the fact that the sum of the measuring errors for individual components is reduced substantially. This
method also requires less special test equipment.

B.2. 3 Cold Propellant Test of Complete Propulsion Unit

This method is the same as B.2. 2 above except that actual propellants would be used in the place of
water. A slight increase in accuracy should result, since some conversion factors would be eliminated
from the calculations. The cost would be considerably greater and a certain amount of fire hazard
would be involved.

95
B.2.4 Static Firing

An actual burning test would undoubtedly produce the most accurate results. All other methods require
experimental corrections which must be obtained by static firings. Unfortunately, this method involves
the greatest cost, the greatest hazard and the most preparation time.

B.3 CALEBRATION AT WSPG

The calibration procedure selected for use at WSPG was a combination of methods outlined in B.2.2 and
B.2.3. The procedure consisted of operating the steam plant with 78 percent hydrogen peroxide and with
32.5 percent sodium permanganate, as in normal flight. Combustion pressure was simulated by means of
18 orifices in the oxygen system and two orifices in the alcohol system. Alcohol (75 percent) was used in
the alcohol system but water was substituted for oxygen in the oxygen system. The oxygen nozzles were
mounted in special extended housings, to which 1.5-inch fire hose was attached to carry the water to the
18 combustion-simulating orifices. These orifices were located around the periphery of a double funnel
which was used to keep the test fluids separate. The combustion-simulating orifices for the alcohol system
were located at the flanged connection of the three-way fittings in the alcohol feed lines. The calibration
installation is shown in Fig. 43 and in Fig. 4, p. 8.

The first calibration using this method was made on May 23, 1947. This procedure was followed for
approximately one year with no basic changes. In general, the results were satisfactory, but it was obvious
that improvement was possible in several respects. Preparation time was excessive, primarily due to the
necessity of keeping two fluids separate. Accuracy was influenced by the fact that the pressure drop was
recorded for only one of the 18 oxygen orifices. The use of alcohol involved expense as well as a fire
hazard. It was necessary to disassemble parts of the steam plant to clean it properly after the calibra-
tion run.

In July, 1948, a new method of calibration was used f or ihe first time. Water was used as a substitute
for both alcohol and oxygen. Combustion pressure was simulated by two orifices; one located in the dis-
charge line of each pump. A separate steam plant was installed as a permanent part of the calibration
stand.

Major objections to the original calibration method were eliminated by these changes. The use of a
single fluid reduced the setup and preparation time by about 75 percent. Water in the place of alcohol
eliminated the expense and fire hazard involved. The use of two orifices in place of 20 increased the ac-
curacy of calibration. A separate steam plant reduced the effort required to condition the steam plant for
flight and also allowed a 50 percent saving in the cost of hydrogen peroxide.

When the original procedure was adopted, it was felt that the ideal method required the calibration to
be made with all propulsion unit components installed in final flight condition. In theory this was probably
correct, but experience proved that it was not feasible. The steam plant had to be cleaned, requiring dis-
assembly. Thus, one of the advantages of the complete- unit test was lost when the piping was disturbed.
The advantage of using the flight steam plant for determining total flow rate was of secondary importance
since (within limits) flow rate was of much less importance than mixing ratio. Furthermore, a large number
of tests had demonstrated that the existing air test on the steam plant gave results well within the required
accuracy and well within the consistency limits of the gas-pressure regulators. Therefore, there was no
real benefit to be realized through the use of the flight steam plant. On the other hand, there were definite
benefits in the use of a separate steam plant. It had been found that a calibration run of 45 seconds allowed
ample time to obtain all the desired data, provided a small hydrogen peroxide tank could be used (the
regular tank, when only half filled, introduced a starting transient of excessive duration). With a stainless
steel tank of 17-gallon capacity (Fig. 44) runs were made with one-half the normal quantity of peroxide. In
addition, many man-hours were saved by the reduction in cleaning and retest time.

The new calibration method proved very satisfactory and no basic changes were made during the rest
of the program. Details of this final system are described in section B.4.
Fig. 44 Stainless Steel I^C^Tank
Mounted on Calibration Stand
97
B.4 DETAILED INFORMATION ON THE FINAL CALIBRATION PROCEDURE
B.4.1 Steam Plant Test

All steam plants used at WSPG had been adjusted previously in Germany. The purpose of the WSPG
air test was to see that all valves operated satisfactorily and that pressure drops throughout the system
were within specified limits. Steam plants with pressure drops outside these limits were set aside and
used as a source of spare parts.

The steam plant during actual firing except that compressed

air test simulated the operation of the air
was used instead of hydrogen peroxide and sodium permanganate. Steam pressure was simulated by a
4.25 mm orifice mounted in the discharge of the steam generator.

To pressurize the complete steam plant, the downstream side of this orifice could be blocked by a
solenoid valve. The pressure regulator was set at 470 psi with the bleed valve open. The difference
between this pressure and the pressure with no flow should not exceed 30 psi. With air flowing and with
both the 25- and 8-ton valves open, the low air pressure should be between 430 and 465 psi, tank pressure
between 410 and 445 psi and steam pressure from 375 to 425 psi.
With the 25-ton valve closed, the steam pressure should drop to between 2 75 and 325 psi. With both
the 8- and 25-ton valves closed, the steam pressure supplied by flow through the permanganate tank
should be from 115 to 165 psi.

B.4. 2 Combustion Pressure Simulating Orifices

The combustion chamber pressure for a normal V-2 flight or static test is 202 psi. In calibrating a
propulsion unit it is advantageous to operate the steam plant and turbo-pump under normal load. Since
there is no positive burner pressure, the normal pressure is simulated by means of orifices inserted in
the discharge line of the alcohol and oxygen pumps. The calibration test must be made under normal flow
conditions since the pressure drop across an orifice varies as the square of the flow, whereas the com-
bustion chamber pressure (for a hot run) varies directly with the flow. In addition, these orifices contain
other correction factors for the alcohol and oxygen systems as noted below (all values in psi).

OXYGEN
Combustion pressure +202
Pressure at oxygen nozzle +14
Experimental correction -7
Tank not pressurized -20
Eight feet added head at calibration +4
AP with oxygen 193
AP with water 193/1.14 169.3

Injection pressure is 7 psi higher for calibration than

for firing.

ALCOHOL
Combustion pressure +202
Experimental correction +7
Eight feet added head at calibration +3
AP with alcohol 212
AP with water 212/0.86 246.5

The cooling jacket pressure drop at burning is higher

than at calibration even though the cooling nozzles and
ZK nozzles are blocked during calibration.
Alcohol flow through the cooling lines and ZK nozzles comprise about 10 percent of the total alcohol
flow. This flow is blocked during calibration because of the difficulties involved in attempting to simulate
the pressure drop which occurs across the cooling nozzles during actual combustion. The experimental
corrections were obtained from the results of numerous calibrations and static tests at Peenemunde,
Germany^"*), and have subsequently been confirmed at WSPG.
The orifices used to simulate combustion pressure are calculated by the equation:

W = 0.668 AK yo^p as D = —906W (1)

[J^AP
where A = area, inches^
D = diameter, inches
K = flow coefficient (coefficient of discharge and approach factor)
3
f - density of medium flowing, pounds per cubic foot
AP = differential pressure, psi
W = flow rate, pounds per second
Orifice diameter may be calculated on the basis of either water or actual propellant provided the
corresponding flow rate, density and AP are used:
OXYGEN
OXYGEN WATER UNITS
FLOWING FLOWING
w 152.5 133.8 pounds per second
AP 193 169.3 psi
P 71.1 62.4 pounds per cubic foot
Sp. Gr. 1.14 1.0

ALCOHOL
ALCOHOL WATER UNITS
FLOWING FLOWING
W 123.5 143.6 pounds per second
AP 212 246.5 psi
53.7 62.4 pounds per cubic foot
Sp. Gr. 0.86 1.0

Instrumentation

Two types of equipment are included under this heading. This section will cover the equipment
needed and the equipment required to record test data.
to control the test

The control desk (Fig. 45) was located in the control house about 50 feet from the calibration stand
(Fig. 11, p.21).

The sequence of operations noted below was controlled from this desk. The turbine overspeed device
(TOS in Fig. 46) was equipped with a normally closed contact. With the rotary switch in position P-l and
TOS reset, the start button completed the circuit to energize the "Line Energizer" relay (LE) and supply
power for the test. Turning the switch to P-2 operated R-l to close the bleeders of the steam plant
through action of valve Dlh. Moving the switch to P-3 operated R-2 to open the alcohol preliminary
valve (Slh). The opening of Slh caused its contact, Sir, to complete a circuit to open the oxygen valve
03h to its preliminary stage position. Water then flowed through the system under gravity head. Going
to P-4 operated R-3, provided the oxygen valve had opened and thereby closed its contact 03r. In turn,
R-3 opened the steam plant pressurizing valve (Dh) which caused the peroxide and permanganate tanks
to be pressurized. When pressurization was completed, pressure switch D2r closed its contact and ap-
plied power to open the 8- and 25-ton valves, D8h and D25h. The opening of these valves allowed peroxide
to enter the steam generator where it was mixed with permanganate to generate steam and start the tur-
bine. The test could be terminated at any time by de-energizing LE through operation of the STOP
button or by action of TOS in the event of overspeed. Operation of R-3 also closed contacts to apply
115 volts (ac) to the recorders, thus starting them simultaneously with the start of the turbine.
-y^-P-
I

I
|
R -' w
(7)
/yi£ 5V
II

60r
1
R-3

_
"5V
M.S.
To RECORDERS
j
|R 3

uSir
R-2 Slj)<'

03h
'
'
©
v^P-4 I iR-2
¥AaR"3 03r

IR-3 Oh.

/
/
,rD2r

D25h

Fig. 45 and 46 (Top to bottom) Propulsion Unit Calibration Control Desk.

Schematic Diagram Control -Desk Electrical Systems
100
One of the most important measurements was the rate of flow from each tank. This was accomplished
by means of a float which followed the surface of the liquid: a chart (Fig. 47) on which a mark was made
every second was attached to the float. From previous calibrations, the quantity of liquid versus float
position was known. The rate of flow could therefore be determined. A secondary flow-measuring system
was also employed. Six float switches were placed at known intervals in the tank. Each switch caused a
pip to be recorded as the liquid level passed that switch. Since both the amount of water between switches
and the recorder chart speed were known, the flow rate could be determined.

Turbine speed was measured by two methods. The most accurate determination was made by means
of a contact on a 100:1 gear reduction unit attached to the turbine shaft. This contact caused a pip to be
recorded every 100 revolutions. By counting the pips and knowing the recorder-chart speed, an accurate
average of turbine speed could be obtained. The second method was the use of a tachometer-generator
attached to the turbine shaft. Although less accurate, the second method was valuable in that it gave a
continuous record, thus indicating any speed variations.

Steam temperature and steam exhaust temperature were measured by means of sheathed thermo-
These values were recorded on a photoelectric recorder.
couples inserted in the lines.

Fig. 47 Float and Recorder Used to Measure Flow from Calibration Stand Tanks

101
The pressures noted below were measured by means of Brown circular-chart recording pressure
(Fig. 48), located in a small room on the first level of the calibration stand (Fig. 11, p. 21).

Oxygen Pump Inlet Pressure Alcohol Injection Pressure

Alcohol Pump Inlet Pressure Alcohol Cooling Jacket Pressure
Oxygen Pump Discharge Pressure High-pressure Air Supply
Alcohol Pump Discharge Pressure Low-pressure Air Supply
Oxygen Pressure below Combustion-simulating Orifice (pressure regulator setting)
Alcohol Pressure below Combustion-simulating Orifice Hydrogen Peroxide Tank Pressure
Oxygen Injection Pressure Steam Pressure
Steam Exhaust Pressure

Not these measurements were required in the calculation of orifice size and pressure regulator
all of
setting. However, all were useful in determining that all components of the propulsion system were op-
erating properly. They were also useful in the analysis of troubles discovered during calibration runs.
As an example, severe erosion of turbine blades was encountered during early runs. It was found that
the blades were rather sensitive to temperature. In those early runs the use of 80.5 percent hydrogen
peroxide resulted in an average steam temperature of about 440°C. When the concentration of the
hydrogen peroxide was reduced to 78 percent, the temperature of the steam dropped to an average of
about 370°C. By holding the temperature below 400°C, blade erosion troubles were eliminated.

The data obtained during calibration runs were also compared with similar data obtained during flight
by means of telemetry. This comparison proved useful in accounting for missile performance.

Fig. 48 Interior View of Calibration-stand Instrument House

102
B.4.4 Preparations for Calibration Test

The procedure outlined below was followed in preparing the propulsion unit for calibration.

The turbopump assembly was dismantled and all bearings and seals were greased to protect them
from corrosion during the test (the assembly was completely cleaned and all grease removed after
calibration). The propulsion unit was completely assembled (as for flight) with the following exceptions:

a. Special main oxygen and alcohol valves were installed. This was necessary because (during cali-
bration) the simulated combustion-pressure drop occurs ahead of these valves, resulting in insufficient
pressure to operate the normal flight valves. The special oxygen valve had a lighter spring and the
alcohol valve was bolted open. These modifications did not affect the validity of the test results.

b. A steam plant was not mounted in the unit (a separate plant was used for calibration).

c. The four alcohol cooling lines were blocked off with blank plugs at the head of the motor.

d. One special oxygen feed line, containing the oxygen injection-pressure tap, was installed.

e. Heat exchanger and the exhaust steam lines were omitted.

f. Blank nozzles were substituted for the regular open "ZK" nozzles.

The propulsion unit was installed in the calibration stand as shown by Fig. 49 and 50. The turbo-
pump assembly was raised five feet above its normal position and two special lines, containing the
combustion-pressure orifices, were installed (Fig. 51 and 52).
A 125-mm
(open line) flow-regulating orifice was inserted at the discharge flange of the alcohol pump.
A special 78.1 mm,
stainless-steel orifice was inserted at the discharge flange of the oxygen pump. From
previous tests it had been found that these particular sizes were close to the average of those originally
selected and were, therefore, most likely to be near the correct size.

Fig. 49 Propulsion Unit Installed in Calibration Stand

103
X
Xr
[T-STEAM
T15ENERAT0R
I
STEAM

PRESSURE TAPS
1

2
PUMP INLET
2 ALC PUMP INLET
3 2
PUMP
4 ALC. PUMP
5
2
AFTER ORIFICE
6 ALC AFTER ORIFICE
7 2 INJECTION
8 ALC
9 ALC COOLING JACKET
10 STEAM
11 EXHAUST STEAM

TEMPERATURES
12 STEAM
13 EXHAUST STEAM
ORIFICES
14
2
COMBUSTION
15 ALC. COMBUSTION

2 FLOW REGULATING
16
17 ALC FLOW REGULATING

Fig. 50 Schematic Diagram of V-2 Propulsion Unit Calibration Arrangement

104
Fig. 51 Combustion-pressure Simulating Orifices Fig. 52 Combustion-pressure Simulating Orifices
and Turbopump, View A and Turbopump, View B

105
Steam, steam exhaust and measuring lines were attached and the necessary electrical connections
made. A preliminary test was run to check operation of the valves and measuring devices and to test the
steam plant for leaks. The tanks were filled with filtered water, the recorders were inked and the steam
plant was loaded with permanganate and peroxide. The system was then ready for the actual test.

Following the test, the excess permanganate was drained from the tank. After cooling sufficiently,
the turbine was flushed by filling the peroxide tank with water and forcing the water through the turbine
under air pressure. The turbine was then disassembled immediately and cleaned. The motor unit was
dried immediately by means of a hot-air blower.

B.4.5 Test Data Reduction

It should be noted that mixing ratio or mixture ratio as used in this report means the ratio of alcohol

flow rate to oxygen flow rate (Wf/W ). Most current propulsion unit calculations refer to the mixture
ratio as the ratio of oxygen flow rate to alcohol flow rate (W /Wf). However, the authors believe the
Wf/W ratio is more appropriate to this report since all V-2 propulsion unit calculations in Germany
were based on the Wf/W term.
To apply the calibration test data to the performance curves, it was necessary to correct the test
valves to standard alcohol and oxygen flow rates.

The V-2 propulsion unit was designed for optimum performance under the following conditions:

... alcohol 123.5 lb per sec

= n 01
. . .

Mixing
° ratio, : go = ,, 0.81
oxygen 152.5 lb per sec

Total flow, 276 lb per sec

B.4.5.1 Gravity Corrections (due to difference in head between gage and point of measurement)

Measurement Correction in Psi

Oxygen Alcohol

Pump Inlet Pressure -1.4 -0.5

Pump Discharge Pressure +1.2 +1.2
Pressure after Burner Orifice +2.2 +2.2
Injection Pressure +3.8 +4.0
Cooling Jacket Pressure +5.1

B.4.5. 2 Conversion of Corrected Pressures (with water flow to standard alcohol and oxygen flow rates)

Alcohol System Pc =^ 2
Uj^ Lj (2)

where P<j> = Corrected test pressure, psi

Ww = Test flow rate of water, lb per sec
0.86 = Specific gravity of 75 percent alcohol at 60°F
123.5 - Standard alcohol flow rate, lb per sec

152.5) 2 /i
Oxygen System Pc ^—J j^j
|P n
J

where Pt Corrected test pressure, psi

=
Ww Test flow rate of water, lb per sec
=
1.14 = Specific gravity of liquid oxygen
152.5 - Standard oxygen flow rate, lb per sec

B.4.5.3 Correction of Pump Pressure to Firing Conditions (for use with performance charts)
Oxygen Pump Pressure PF = Pc + 20 - 4 = Pc + 16 (4)

106
where Pp = Pump pressure for firing, psi
P£ = Corrected pump pressure from calibration, psi
20 = Correction (tank not pressurized at calibration)
4 = Correction (added head at calibration)
Alcohol Pump Pressure Pf = Pc - 3 (5)

where Pf = Pump pressure for firing, psi

Pq - Corrected pump pressure from calibration, psi
-3 = Correction (added head at calibration, psi)

B.4.5.4 Turbine Speed Correction

Since an orifice of 125 mm

(open line) was used in the alcohol system, the turbine speed, corrected
to standard flow, may be calculated as outlined below (provided the test results do not indicate the need for
a change in the alcohol orifice).
_ S T x 123.5
Sc
~W TA x0.86 (6)

where Sq = corrected turbine speed, rpm

S T = test turbine speed, rpm
123.5 = standard flow rate of alcohol, lb per sec
W
TA = test * low rate °f water in alcohol system, lb per sec
0.86 = specific gravity of alcohol

If an orifice is required in the alcohol system, the turbine speed is corrected by using the ratio
of the total flows in accordance with the following equation:
St x 276.0
SC "
W TA x 0-86 + W TO x 1.14 W
where Sq = corrected turbine speed, rpm
St = test turbine speed, rpm
276.0 = standard total flow rate, lb per sec
Wta = test f l°w rate °f water in alcohol system, lb per sec

Wto = test flow rate of water in oxygen system, lb per sec

0.86 = specific gravity of alcohol
1.14 = specific gravity of liquid oxygen

The latter method, using the ratio of total flows, is not completely accurate in that it does not take
into consideration the difference in pump characteristics. However, the err"or is not great, being in the
order of two percent.

B.4.5.5 Calculation of Orifice Size and Pressure Setting

The correct synchronizing orifice size is determined by applying the corrected test data to the
pump performance curves (Fig. 53 and 54). These curves are based on experimental data and are used to
simplify the calculation of orifice size. The curves are a translation of the German Work Sheet 033, ex-
cept that the average pump characteristic slope has been included in the pump pressure co-ordinate. A
detailed description of the curves is presented in Archive 57/13.

The calibration test is made with an orifice of 78.1 mm

in the oxygen system and an orifice of
125 mm (open line) in the alcohol system. If the results of the test show the mixing ratio or the total flow
to differ from the desired values by more than five percent, a second test is recommended. The limits
are: 262 to 290 pounds per second for total flow and 0.77 to 0.85 for mixing ratio. These limits are set
because the pump performance curves are correct only at the proper flow rates.

107
62-
en
ou
63- /^ £
55- i^ ^5
64-
-a.
^ ^i>^
i § £ x,X

65-
50-Q-- .,*••
^
%?
^ ^
<J
^ x> ,# >^!

66-
40 1^ i iiIi l ^
o^ # § i P ^ g ^ ^ ^
<5f

# 1 i1 ^ p ^ p
67-
40 f \<£
^^ ^*^
68-
X ^i Ii p p ^^^ ^
^ P ^^
07
69-
05 1 i 1 i1 p
UJ 70-
4^ 1 ^ ^
ou 1 1^ .06

P8
_

-
W-
.05
25
*

-
™ 04
cSJ

80 : 15 03

:
I0 02
z

5 ,01 — —
:
I00

180 200 220 240 260 280 300 320

PUMP PRESSURE, PS I

Fig. 53 Oxygen Pump Performance Curve

108
210 230 250 270 290 310 330 350
PUMPPRESSURE,PS.I.

Fig. 54 Alcohol Pump Performance Curve

109
B. 4. 5. 5.1 Mixing Ratio Correction (by change of orifice size)

The orifice pressure drop required to give the proper ratio is determined by applying the cor-
rected oxygen to the oxygen pump performance curve (Fig. 53). Proceed up the pump
pump pressure
pressure line to the mixing ratio error line, then read directly across to the AP required (this AP is
negative for ratios greater than 0.81 and positive for those less than 0.81). This is the AP which must
be added to give the correct ratio. A AP of 17.7 is already present (78.1 mm
orifice), hence the AP must
be added algebraically to the 17.7 psi to give the net AP that must be supplied by the corrected orifice.
The orifice size to give this AP can be read directly from the ordinate on the left.
If the net AP should be negative, removal of the 78.1 mm
orifice would not provide complete
correction. It would then be necessary to add a restriction in the alcohol side. This can be determined
by locating the oxygen pump pressure and the net AP, on the oxygen pump performance curve (Fig. 53),
and reading the new mixing ratio error at their intersection. This error is then taken to the alcohol pump
performance curve (Fig. 54). On this curve the intersection of the new mixing ratio error and alcohol
pump pressure is found. Directly across from this point read the AP and the corresponding size of
orifice to be installed in the alcohol line.

The above procedure for selecting ratio orifices is based on the assumption that the test-stand
orifices used to simulate combustion pressure are perfect and give exactly the correct pressure drop. In
practice this was not entirely true and it was therefore necessary to transfer that pressure - drop error
to the ratio orifice when calculating the size of that orifice for flight. As experience was gained, the com-
bustion pressure simulating orifices were made more nearly correct and the above correction was thereby
minimized. The changes in simulating orifice size are shown in the table of calibration data, Table II.

B.4.5.5.2 Flow Rate Correction

It was required that the mixing ratio and total flow be within five percent of the standard values

during a final calibration test. However, the final run seldom gave exactly the right ratio or exactly the
right flow. Therefore, it was necessary to change the flow orifice (as described above) to obtain the
correct mixing ratio for flight. This change, in turn, had an effect on the flow rate. The correction for
.

flow rate was composed of two factors: (1) correction for the change in mixing ratio and (2) correction
for the original flow error.

Flow rate was adjusted by varying the setting of the pressure regulator which supplied gas to
pressurize the steam plant. The corrections are given in terms of change in regulator setting, psi. The
correction for mixing ratio change is shown on the pressure regulator correction curves (Fig. 55). The
total flow correction is shown on the pressure regulator correction curve (Fig. 56). The use of these
two curves should be evident from the two examples which follow.

B.4.5.5.3 Sample Calculations

a. Example 1

1. Standard starting values

Regulator setting 450 psi

Alcohol orifice 125 mm
Oxygen orifice 78.1 mm
st data

Corrected alcohol pump pressure 310.0 psi

Corrected oxygen pump pressure 255.0 psi
Alcohol flow 127.8 lb per sec
Oxygen flow 155.2 lb per sec
Total flow 283.0 lb per sec
Mixing ratio 0.823

110
Test No. 57-19 34 95-45 96-46 97-47 98-48 99-48
I
100-49
Rocket No. BU-3 )
61 54 57 55 55 52
Date 8-3-48 >-49 9h 26-50 10-18-50 12-28-50 2-21-51 2-21-51
( 5-15-51

Orifices Used in Place of

Alcohol Burner (in.) 1.8305 1.8209 1.8209
Oxygen Burner (in.) 1.9108 1.9210 1.9210
Alcohol Flow (mm) 125 125 125 125
Oxygen Flow (mm) 80 78.1 78.1 78.1

Low Air Bled to (psi) 470 55 455 462 448 N.G. 458 452
Low Air Held (psi) 450 40 451 450 431 445 442
Low Air Bled - Held (psi) 20 15 4 12 17 13 10
H2O2 Tank (psi) 443 22 435 441 418 434 430
Steam (psi) 370 54 356 355 340 N.G. 355
Steam Temp (°C) N.G. 80 N.G. 367 350 402
Turbine Speed (rpm) 3780 40 3750
3 3840 3812 3816 3810

Test Flow (lb per sec)

Alcohol - H2O 146.4 146.6 139.7 142.9 140.8
Alcohol - Alcohol 125.9 126.0 120.2 122.9 121.1
O2 - H2O 132.5 129.9 129.6 128.4 129.0
O2-O2 151.0 148.0 147.7 146.4 147.0

Mixing Ratio (%) 83 85.1 81.4 83.9 82.3

Total Flow (lb per sec) 276.9 274.1 267.9 269.4 268

Pressures (Corrected to 123.5 lb per sec

Alcohol Pump (psi) 309 P 304.8 281.5 317 295 309
Alcohol A.O. (psi) 102.5 2 89.8 84.3 98 87 92.7
AP (psi) 206.5 2 215.0 197.2 219 208 216.3
Cooling Jacket (psi) 99 I 88.9 82.6 97 87.8 92.6
Injection (psi) 40.8 j> 33.2 28.9 43 33.4 33
Cooling Jacket-Injection(psi) 58.2 6 55.7 53.7 54 54.4 59.6

02 Pump (psi) 244 I 241.5 252.9 247 254 255

A.O. (psi) 44.3 5 45.5 51.0
2 63 53 _
AP (psi) 199.7 2 196.0 201.9 184 201 _
Injection (psi) 22.7 23.0 25.8 25 30.8 29.6

Corrected Values for Firing

Turbine Speed (H20Flow)(rpm) 3878 91 3739 3867 3910 3910 3908
Low Air (psi) 465 50 460 465 465 470 470
O2 Orifice (mm) 86 Ale. 80 Ale. 80.5 94 Ale 89
Turbine Speed( Ale. Flow) (rpm) 880 28 3680 3765 3918 3835 3885

ION D
Turbine Speed (H20Flow)(rpm) 3B7B 3896 3884
3. Mixing Ratio Correction

AM.R. - 0.81 - 0.823 = -0.013

Going to the oxygen pump performance curve (Fig. 53) and following the pump pressure
line of 255 psi to the AM.R. line of 0.013 gives a pressure drop (AP) of - 9.6 psi. Adding this to the
orifice (78.1 mm) drop of 17.7 gives a net AP of 8.1 psi and a required orifice size of 88 mm.

Total Flow Correction

To obtain the change in pressure regulator setting required by the above change in orifice,
use the pressure regulator correction curve (Fig. 55). On the curve for "Orifice in Oxygen Line," (Fig. 55,
bottom) locate the point for a ratio of 0.823 and directly across to the left find the value of - 7.5 psi.

To obtain the change in pressure regulator setting required by the excess flow shown in the
test data (283 vs 276 lb per sec), go to the pressure regulator correction curve for total flow error (Fig. 56).
Locate the 283 lb per sec point on the curve and directly across to the left find the value of - 26 psi.

The regulator setting for flight is 450 - 7.5 - 26.0 = 416.5 psi.

b. Example 2

1. Standard starting values

Regulator setting 450 psi

Alcohol orifice 125 mm
Oxygen orifice 78.1 mm
2. Assumed test data

Corrected alcohol pump pressure 295.0 psi

Corrected oxygen pump pressure 255.0 psi
Alcohol flow 124.1 lb per sec
Oxygen flow 147.9 lb per sec
Total flow 272.0 lb per sec
Mixing ratio 0.839
3. Mixing Ratio Correction

AM.R. = 0.81 = 0.839 = - 0.029

Going to the oxygen pump performance curve (Fig. 53) and following the pump pressure
line of 255 psi to the AM.R. line of 0.029 gives a pressure drop (AP) of 21 psi (this value is negative when
AM.R. is negative). Adding this to the installed orifice drop of 17.7 psi gives a net AP of - 3.3 psi. Since
this value is negative, the remaining correction must be made on the alcohol side. Transferring from one
performance curve to the other (oxygen, Fig. 53 to alcohol, Fig. 54) must be accomplished on the basis of
AM.R. only. The 3.3 psi on the oxygen curve gives a AM.R. of 0.0045, which must be corrected on the
alcohol side. From the alcohol pump performance curve (Fig. 54), a pump pressure of 295 psi and a
AM.R. of 0.0045 gives a AP of 3.6 psi, which requires an orifice of 97 mm
in the alcohol line.

4. Total Flow Correction

To obtain the change in pressure. regulator setting required by this change in orifices, it

is necessary to use both sections of the pressure regulator curve (for change in mixing ratio, Fig. 55).

From the curve for "Orifice in Oxygen Line," (Fig. 55, bottom) read the correction cor-
responding to that portion of the ratio correction accomplished by removing the oxygen orifice. This is
0.029 - 0.0045 = 0.0245. Add this value to the desired ratio of 0.81 to obtain 0.8345, the value to be used
with the "Oxygen" curve (Fig. 55, bottom). From that curve the correction is read as - 14 psi.

From the curve for "Orifice in Alcohol Line," (Fig. 55, top) read the value corresponding
to that portion of the ratio correctionaccomplished by the addition of the orifice in the alcohol line. The
mixing ratio used in this case is the desired ratio plus that portion of the error which was not corrected
by removal of the oxygen orifice. This is 0.81 + 0.0045 = 0.8145. From the curve (Fig. 55, top) the cor-
responding correction is read as + 1.3 psi.

Ill
The total pressure regulator correction for the change in mixing ratio is then the algebraic
sum of the above, or: -14.0 + 1.3 = - 12.7 psi.

From the pressure regulator curve for total flow error (Fig. 56), read the correction re-
quired by the low flow as shown in the test data (272 lb per sec). This correction is + 13 psi.
The final pressure regulator setting for firing is then:

regulator setting
Initial 450 psi
Correction for change in mixing ratio -12.7 psi
Correction for low test flow + 13.0 psi

Regulator setting for firing 450.3 psi

+ 40

+ 20
CO
Q_

O
UJ
tr
-20
tr
o
o
CL -40
<

-60
260 270 280 290
TOTAL FLOW, LB PER SEC

Fig. 56 Pressure Regulator Correction Curve for Total Flow Error

112
+20

+ 10

a o
tr.

6
° -10
ORIFICE IN
ALCOHOL LINE

20 1 1 1

.77 .78 .79 .80 .81 .82 .83 ,84 .85

MIXING RATIO

.77 78 .79 .80 .81 .82 .83 .84 .85

MIXING RATIO

Fig. 55 Pressure Regulator Correction Curves for Mixing Ratio Change

113
B.5 CALIBRATION AND FLIGHT DATA
B.5.1 Average Values from Calibration Data
The following table gives average, maximum and minimum values obtained by 19 calibration runs
with final calibration equipment and procedures. In compiling these figures, test data from runs 79-33
through 100-49 (omitting runs 84-37, 85-37 and 94-44) were used. Complete data on these test runs will
be found in Table II.

MEASUREMENT UNITS MINIMUM AVERAGE MAXIMUM

Air pressure (bled) psi 440 452 462
Air pressure (held) psi 431 442 452
Tank pressure - hydrogen peroxide psi 418 428 441
Steam pressure psi 340 353 360
Steam temperature oc 350 378 402
Turbine speed rpm 3750 3857 3996
Alcohol flow (water) lb per sec 138.0 143.4 147.9
Equivalent flow using alcohol lb per sec 118.7 123.3 127.2
Oxygen system flow using water lb per sec 128.4 131.7 136.5
Equivalent flow using oxygen lb per sec 146.4 150.2 155.5
Mixing ratio 0.78 0.821 0.854
Alcohol pump outlet psi 281.5 307.1 333.0
Alcohol pressure below combustion orifice psi 84.3 96.7 116.0
Pressure drop across combustion orifice psi 197.2 210.4 222.6
Alcohol pressure at burner inlet psi 82.6 96.1 113.0
Alcohol pressure before injection nozzles psi 28.9 37.1 48.8
Pressure drop across cooling jacket psi 52.1 59.3 69.9
Oxygen pump outlet pressure psi 234.0 246.8 255.0
Oxygen pressure below combustion orifice psi 45.0 51.6 63.0
Pressure drop across combustion orifice psi 184.0 194.6 201.9
Oxygen pressure before injection nozzles psi 18.6 27.7 36.0

B.5.2 Static Firing Data

Three propulsion units were tested by static firing after the normal "cold" calibration had been
made. The total flow and the mixing ratio, as determined by the static firings, is tabulated below:
TEST DATE CALIBRATION TOTAL FLOW MIXING RATIO
TEST
NO. NO. LB PER SEC
1 10/21/48 58-20 290 0.791
2 10/25/48 58-20 272 0.802
3 7/19/50 91-43 288 0.790
4 7/29/50 94-44 292 0.778

Tests 1 and 2 were made on the same propulsion unit. The ratio orifices used in test 3 were the same
as those used in the "cold" calibration of that unit. The mixing ratio shown by the "cold" calibration was
0.807, which indicated that no change of orifice was required.

Although the above data show some departure from the intended flow and ratio, the errors are not un-
reasonable when the accuracy of the measurements is considered. Flow rate measurements during static
firings were not completely satisfactory, particularly in the oxygen system. It is believed that appreciable
errors were introduced by the low temperature (-183°C) of the liquid oxygen.

B.5.3 Flight Data

Flight data pertinent to this section is included in Table III, p. 116.

114
B.6 SPECIAL V-2 PROPULSION UNIT TESTS
B.6.1 Introduction

Special tests were run on the V-2 propulsion unit to aid in determining the cause of rocket motor
failures in flight. By observing the effect of controlled malfunctioning of the propulsion unit, it is possible
that the accuracy of analyzing rocket motor failures in flight may be greatly increased. These tests were
arranged to determine the effect of the actions noted below on rocket motor performance:
a. Applying control pressure to the oxygen main valve,

b. Applying control pressure to the alcohol main valve,

c. Applying control pressure to both the alcohol and oxygen main valves,

d. Closing the preliminary alcohol valve.

B.6. 2 Test Procedure

a. All tests were run on the calibration stand with the following changes from normal calibration pro-
cedure.

1. Combustion pressure for the alcohol system was simulated by four 0.952-inch diameter orifices
machined in a sleeve inserted in the combustion chamber dome (around the main alcohol valve).
Combustion pressure for the oxygen system was simulated by eighteen 0.4823-inch diameter
2.
normal oxygen injection nozzle position. The oxygen injection nozzles were
orifices located in the
mounted on a seven-inch pipe extension.
3. Conventional main alcohol and main oxygen valves were used in place of the special valves nor-
mally used for calibration (special valves are necessary for calibration because the combustion pressure
is simulated by an orifice in the pump discharge of the alcohol and oxygen system).

4. The alcohol and oxygen main valve positions were recorded on a photoelectric recorder by means
of a slide wire potentiometer mechanically connected to the valves.
5. Provisions were made in tests 2 and 3 (Tables IV, V and VI) to measure the alcohol by-pass flow.

This was accomplished by connecting the by-pass line to a separate tank and weighing the water. The as-
sumption was made that the gain by a lower by-pass discharge pressure would be approximately offset by
a reduction in pump inlet pressure. Data from tests 2 and 6 and tests 3 and 4 indicate the assumption
was not valid.

6. The test method was the same as that used in calibration runs with control pressure applied to
the alcohol and oxygen main valves (under valve control pressure) for the time intervals indicated.
b. Discrepancies between test and flight conditions and the effect on test' results are outlined below.

1. Water used in place of alcohol and oxygen:

(a) Under normal operating procedures the effect will be negligible because the reduction in
horsepower required by the oxygen pump is offset by the increase in horsepower required by the alcohol
pump.
(b) When control pressure is applied to the main valves, the amount of valve opening depends
solely on the pump discharge pressure and not on the volume flow. Therefore, under flight condition the
oxygen valve would tend to close less and the alcohol valve more. This would make slight changes in
test results.

2. Combustion pressure simulated by orifices:

Reduction of flow in one system (by applying control pressure to the main valve) will not
(a)
reduce the combustion pressure in the other system as it would in flight. This means that in flight there
would be a greater increase in turbine speed with corresponding changes in flows and pressures when the
oxygen valve is operated. In addition, turbine speed would probably increase (instead of decrease as
shown by test data) when the alcohol valve is operated.

115
B/O TS EXP B/O
Radio Radio

B/O
Radio

B/O
ooooo o ooSo a O O O Oi
55 3 5 B/0 PQHHCQCQ KOHIxO 00 GQ i O «

o o m o t- o o o o to ,
O CO
o
o ooooo voomo oJ?o*e
I — -H O 00 o eg t- in cm ,- in *h

o eg in
CO
o CO 6 rH t- S5S 2S552S S?"8
o o to
O OOO o c- in o co
eg m eg -. cn
oeot-ma in o o w o>
co o M co t-
I

»»ooe«
m>r «*» ,.»»,.
in <o ee oo oo
*oi»ic» «o
oi«mo
^ •»
o
» «o
en en —

o (o gj O in OiflidOid OflOO)
||S 55 8

I £
u o.
—i.
_o a -r- o
oco o o
m o
m
I 8.? too o
e« , en , co

* C.T? oo ooo
5<3

I 8-1
O CJ OOiD«i

o .'0(d to o o o *r ooooo V Ol (D N O)
co cm •» co
00 cn t- -. m oo oo
-. eg o
o rr to -r
co
hc-»-
oo oo co o» a coaeg —
a oo
co

o to oo o o o cm o o
o cn oooo o I

oo o oinmoo §§§§§ 22§§§

CO CO 00 CO CO CO OO CO 00 CO
<! s CO 00 oo CO CO 0O oo OO 00 CO CO CO CO oo CO CO CO.CO CO

§ CO CO
o r-
t-
^ CM cm CM ^1 o -r in
CO
o J_ T o ocn
eg o
o o ^
CO
j_ CM CO e- cn CO
O
TT cn
co a a. CM
s
o o to o s o
s
w |
— 00 00
CO
cn OS 00 OS CD cn en cn cn oo cn 5 cn cn cn 00 cn en cn co cn cn cn cn 00 a> cn en
S
m to
o o m o o in o o cm in
V O O CO CO CO CM o oo
o o
1=
| s CO en
CM in
CO t- cn CO o CO to cn to CO
in
t- cn to
CO CM
a.
e_
r-
*p S •" *" ** ~* CM

CO to
35 ^
CO to CO CO CO CO CO to to to to SS t- j. r- J. j. c- r- ti- j. CO 00
V o 'T
o 'T eo
o o t-
O o m Oa i
T
s in to en in cn cm oo CM cn en
Q S
cn
O o CN
to
co V m CO cfl to r- t- * co oo "" "" "* ** ""' rt H
" "* CM CO "* ** t- * -< — " *"* CM CO <r O1

73 d ' ' • • ' ' ,. o a « to co CM

Z ' ' ' ' • ' ' ' ' ' ' ' ' > ' ' ' ' ' •

s
8 I -~ „ - m CO t- oo cn o -22 3 in CO r- OO cn
8 s M eg CM s ao Kg CM
in CO
CO co CO CM

116
"3 u6O ||oft
uo 2 I-. u, ca co H K CQ 2 CQ in u< m 2 nH to m5 S ca ti m u w «

bp 8 OOOOO
k 1 " 5 -«S

o in o
gog
ooot-co
mcomooo 000m
tomcotN S§ •'

O CO O C- CN
— v „ g to
*> <o o *• c- en a *r co
:

in rr ^. W CN «r CO •*

"S 00 — 1

Sg» «. in o
to co t- en t-; o in

CO Sh« co"
CD
>
O in"
<£,
a «
rj cd
0«NCO N i-' f'oivod 1

o o in

OOOOO
m mmom
SScoSS
0000 co in co

PS — 1 tt en in t- c- r-
wm •» *
9 X in »
Si eno-*eg«m
co en ej t-
co 00 to en »N« c-tDl-t-t-
o t-eoffco — N -. Ol rt CO

H t»
in in in m inininintn w m o in
n 1 2. % § a
<o,s -1 rt ^ r-
in in in in in in to in in
tlMNON
in in in

-j S
£•6 3600 3650 3650 3700 3900 3878 3941 3816 3884 3797 3711 3650 3789 3791 3867 3692 3797 3943 3791 3829 3816 3887 3750 3956 3988 3680 3867 3910
—1
en

m Tf O*O C-
|S* V V V «in
in to in
§5§53 v
gg° t- in in to
to in
4" •*vmv
in
•* •*
m
V» «•
O O
r- c-
in in
co to CO

in ,-.
o» CO
27938 30076 29826 29607 28658 27447 28037 27417 27470 28523 28106 29450 29217 28297 29636 28973 27936 28940 30171 28839 28911 29315 29695 29433 27937 28245 30183 28180 29227 30067 r- •*
00 en
3*
•h in
9995 9725
en «
10000 10450 10450 10350 10450 10150 10400 10150 10150 10300 10150 10700 10350 10350 10450 10228 10355 10630 10584 10475 10290 11045 10385 10590 10440 10296 10694 10548

So
000 §1 iliii OOOOO
OOOOO 00000 O O
OO 00
00
CO
O O
O *•
1

ggggg
*f& gss
oo in co 00 t- c- i

CO 00 CO CO 00

>S CO -* -* W CO
5, te-r- e- — to -r a
nMoneo o
co
f-H
in
O
om t-
co m co t- co t-
c S
6 » O£ m *r — co co
teooaco co co 00 S co o 00 cn o
o>
o> o o co co
coSS8°
0)000)0

53? A in co c- co
ss Of eo>H Ok E» CO ififfl G-y O COCO
ca •» 00 co -h t- t- co co co cointococo mco
III
G> TT CO OO
TcoT 1
?"?
i CO H CO
i -H HHMMtl ««iotoc« e».
I I

NNNNN OJOfflrtO
CM CO M CO CO
»
co
in co 00
co co co «r
o en -< in to

to 00
o Q
1 in cv r>

CO CD CD
35Si* to
in o *
1 c- eg
•* co
CO CO -H CO
tt in co in d
I

117
§ „
21-32
2750 3575 3600
503 485 395

f ^
05 13
6
o 0-20 2850
502 482 395
3750 3720
_ O
o> CO
oo >_
'
CO «• to
o o o
Z nn<H
.
t>H — — O
.
I I — — OO I

5 o in o
^ o m + en — en
1
21-46
2100
445 432 360
3900 4002
S I- -J 5 CB of § g
"».
VCMCOCMCOCOt-TfZ . C- CO CO
co co co in
O CO CO
co — — CO'
I

, I

M 13
S © t- eg o co en
in O O 8
#
cm o
CM * CO CO t-
S3 cm t- t- o e> op
I O 1

CM «• •* CO CO
CO
CO
o> t-
. t- •» <o
OHMC-HriHOCO en .
I
'

noNHHOoa
co co co in co co

*« OS O °. CO I o
**
31-38 I

442 430 360

NG
CM CM CM —— O5O
a
a
26-30

442 430 360

3840 i«' r- t- ". in
CJ> t- — m«
CO ^- CO — —o
NG

c
W S O °. o 3
in co co ". 1- a
CO CM — $S w
2o^
11-25 -
2050
442 430 360 NG
3960
.
incMcocMCMcoc-^*inO
S. ' TJ- CO O— —
CO CO CM CO
O Ol
00 *«
§
a •o
a * m
w z 0-10 2150 3790
°. SS—- In U U h—
cm— ZZooiU
hS 1 442 428 360
NG
co —
IV „«5 O CM CM
SPECIAL
in o» o § »O °. r^ »' h in

TABLE
PROPULSION
26-45

4050 4092
•»NinN<rr-»«zO OMN-M-rtl-fZ
442 430 360

CO o
21-25 in _ in co co _ £ co co
FROM
" 2400 441 429 360
3975 3960
"». S r- ». «* «' V SSA -f co en co
-2 1
r
Q
•a

i
s
0-20 2500
440 428 360
3850 3810 o ,r m
BHHHrt-HOOlU
""^ CJ»
t- — °.
O
CM CO •» CO
CO t-
co>- -•
CM
J)
S
——oo co

ON ^ CO
u i oooo in co
«. s § 2 ii us
HNHHHHriOZU
o m t- in §
P CM
I

fe (

r cm to en 5 ——————oZ
CM •* T CO CO CO CM CJ cocoincM— — oooo
o en o
I" slislss *". 8DS '
a S3 S I sss «-• s s* a

ooo g
§*«n»«HHOOo
co g> co in co
oo

f! - If!
•a v
a la «*-^ P 01 CO

i t ^~ a. a. j- =• s § r-sas-gsss*
I I islfllliil
flfiiil ffilfsillli UflljiiSII

118
TABLE V

'NORMAL' PORTION OF SPECIAL V-2 PROPULSION UNIT TESTS

Test No. Avg

Oxygen
Pump inlet (psi) 2.5 3 2.5 3.2 2.9
Pump (psi) 197 198 197 199 202 199
Before orifice (psi) 176 177 177 179 178 177
Injection (psi) 12.5 12.2 11.5 7.1" 13.1 12.3
AP-B.O. (Inj.psi) 163 165 165 172* 165 164.7
Flow (lb/sec. 2 OJ 141.7 142.8 142.6 141.4 145.1 142.7

Alcohol
Pump Inlet (psi) 1.5 3 4.7 5.6 5.6 4.1

Pump (psi) 337 340 342 343 340 340

Before Orifice (psi) 334 334 337 339 338 336
Injection (psi) 51.5 52+ 52+ 56+ 56.8 54+
AP-B.O. (Inj.psi) 282.5 282- 285- 283- 281.2 282-
Flow (lb sec.H 2 0) 155.4 156.1 154 156.4 155.5 155.5

* Values not included in averages.

+ Gage linkage hit limit.
-O2 Injection pressure is approximately 6 psi low.

TABLE VI

"SPECIAL" PORTION OF SPECIAL V-2 PROPULSION UNIT TESTS

Test data on runs 2 through 6 corrected to average normal speed of 3840 rpm;
control pressure on valves as indicated.

Valve Cont. Pre s. Normal O2 On 2 & Ale. On Ale. On

Av.
Test No. 2-6 5 3&5 3 4 2

Turbine Speed 3840 3960 3930 3712

Oxygen
Pump inlet (psi) 2.9 4.6 4.5 4.5C 4.4 5.3 1.3 2.8
Pump (psi) 199 254 256 255 264 246 202 185
Before orifice (psi) 177 37.5 37.4 37.4 51 30 180
Injection (psi) 12.3 2.5 2.1 2.3 2.5 2 11
AP-B.O. (Inj.psi) 164.7 35 35.3 35.2 48.5 28 169 157.4
Flow (H z O lb per sec) 142.7 75.3 73.3 74.3 79.5 62.5 144.2 138.3
Valve pos. 1/64 (in.) 91 6* 5

Alcohol
Pump inlet (psi) 4.1 3.5 2.5 0.3 7.7 7.4
Pump (psi) 340 374 371 288 344 280 310
Before orifice (psi) 336 363 369 366 284 338 273 304
Injection (psi) 54+ 53 57 55 13.2 20.6 12.5 16.5
AP-B.O.(Ihj.psi) 262- 310 312 311 270.8 317.2 260.5 287.5
Flow (H 2 Olbpersec) 155.5 160 169.6 165 165.3 104.3 161.9 94.3
Valve pos. 1/64 (in.) 86 12 12 12

Taken from test No. 1.

Ale. by pass line open in tests 2 and 3; connected in tests 4 and 6. This accounts for the
discrepancy in change of turbine speed and flow with air pressure on the alcohol valve.

119
(b) Combustion pressure during flight varies directly as the flow; during calibration test, this
pressure varies as the square of the flow and is correct at normal operating conditions only. Therefore,
changes in combustion pressure due to changes in flow, would not be as great for flight tests as for
calibration tests.

Only two of the above discrepancies are of prime importance: (1) reduction of flow in one system
does not reduce the combustion pressure in the other and (2) combustion pressure during test varies as
the square of the flow instead of directly with the flow as in flight.

The effect of these two discrepancies tend to cancel each other when either the alcohol or oxygen
valve is operated separately. This is not the case, however, when both valves are operated simultaneously
since only the second condition is (in effect) resulting in a greater change than would be expected under
flight conditions. A comparison of test data supports this condition. Using turbine speed as a criterion,
the following changes resulted from applying control pressure.

CONTROL PRESSURE TURBINE

APPLIED TO SPEED
Lox main valve increased 165 rpm
Alcohol main valve decreased 128 rpm
Lox and alcohol main valves increased 120 rpm

From
these data it can be expected that test results are, in general, representative of flight con-
ditions except when both valves are operated simultaneously. In this case the data should be corrected to
a 37-rpm increase in turbine speed. A more detailed analysis of the effects of the discrepancies would
be of questionable value since there is only one constant - the power to the turbine and several variables -
turbine speed, flow rates, pressures, valve positions, division of horsepower between the two pumps and
combustion pressure.
B.6.3 Results

B.6.3.1 General Performance

Propulsion unit operation was very smooth with no pulsations occurring during the application of
control pressure on the main valves. It is very likely that the combustion-pressure simulating orifices
did not give a true picture in this respect. In the flight it is reasonable to expect that changes in com-
bustion pressure are not in phase with the upstream changes; this may give rise to the severe oscillations
the Germans reported.
B.6.3. 2 Test Data

All test data are itemized in Table IV. Data during the normal portion of each run, corrected to a
3840-rpm turbine speed, are presented in Table V. Table VI is a compilation of data from the "special"
portion of each run. Data in table VI are corrected in proportion to the average normal turbine speed
(3840 rpm). All test results are in terms of water flow. Results of test 1 were disregarded since it was
obviously in error.

B.6.4 Conclusions

a. Application of control pressure to either the oxygen main valve or the alcohol main valve or both
valves will not cut off the propulsion unit through turbine overspeed. Under any of these conditions, the
turbine speed will change only about 100 rpm.

b. The thrust under any one of the three test conditions will be approximately one half.

Closing the preliminary alcohol valve will result in immediate cutoff by turbine overspeed as soon
c.
as the valve actually is closed. It requires approximately seven seconds to bleed the control air from the
preliminary alcohol valve.

d. The increased pressures which result from operating either or both valves should not in themselves
be sufficient to damage parts or burst lines; however, not included are the effects of pulsations.

120
APPENDIX C
STEERING SYSTEM COMPONENTS

C.l GYROS
References: Backfire, Vol. II, p. 136.

During the early part of the V-2 program, the availability of usable control gyros was a serious problem.
There were two main types Anschutz and LGW (Fig. 57) available from Germany. However, most of these
units were in such poor condition that a general rebuilding was necessary. Both types were used to some ex-
tent at WSPG, at times with a pitch gyro of one type and a roll yaw of the other type on the same plate.

Fig. 57 German and American-made Gyros Used in the V-2 Project at WSPG
121
C.l.l Anschutz Type
construction of the Anschutz gyro was not as neat nor as sturdy as the LGW. The gimbal
In general, the
system consisted of a die-cast gimbal on which was mounted a heavy, two-phase a-c torque motor. The
gimbal was pivoted on a cantilever support also of die-cast material. The cantilever support was quite weak
considering the weight of the gimbal and motor system. There were a few cases of bending of this canti-
lever during shipping and handling.

The Anschutz pitch gyro contained an unusual erection device. The gyro axis was erected vertically by
two torque motors which were powered through a sensing device such that the torques applied caused the
gyro to precess to the vertical. The sensing device consisted of a small dish-shaped, sealed container
which was mounted on the bottom of the gyro motor. Within the container was a small amount of conduct-
ing fluid and a set of four contacts. These contacts stuck into the container in such a manner that if the
unit was not held level, contact was made from one or more of the contacts through the fluid to the dish
or common contact, thereby energizing appropriate torque motors.

The program was set in by a d-c timing motor which rotated a porcelain cylinder. The cylinder was
coated with a number of silver contact strips on which contact fingers rode, making a circuit through cali-
brated resistors. This action caused the torque motor to precess the gyro in pitch. Consistent program
angles were hard to obtain with this system because the: (1) time duration was affected by variations in
d-c supply voltage, (2) torque motor current was directly affected by the a-c inverter voltage and (3) gyro
precession rate was affected by the frequency of the a-c inverter.

Some of the most frequent troubles encountered with the Anschutz gyros were as follows:
a. Wiring faults: small size, solid wire was used with insufficient bundling and anchoring.

b. Contacts: the button-type rotatable contacts used for wire lead-ins often opened up either com-
pletely or intermittently.

c. Pick-off potentiometers: very inaccessible, making thorough cleaning almost impossible. Con-
tact wipers frequently rode off the pots.

d. Torque motors: two phase a-c type with very small clearance between rotor and stator caused
considerable binding if a small amount of dirt entered the air gap.

e. Program timer: silver on porcelain cylinder pitted due to arcing, causing bad contact and oc-
casional snagging of contact fingers.

f. Program resistors did not actually cause much trouble but a large amount of time was consumed
in calibration.

C.1.2 LGW Type (German)

The LGW type of gyro was of much sturdier design and allowed easier access to all parts. The gim-
bal was machined from metal with a much larger cross section than the die cast gimbal of the Anschutz
type. Torque motors were much lighter than those of the Anschutz, causing the load per unit cross-sectional
area of the gimbal to be much less than the loading in the Anschutz. The gimbal was supported on a canti-
lever of considerably larger cross-section and with better web reinforcing than Anschutz.

Erection in the vertical axis was accomplished by use of a small pendulum mounted on the gimbal.
The pendulum carried a pair of contacts which energized the torque motor to erect the gimbal in a vertical
position. Erection of the roll axis on the pitch gyro was effected by a pair of silver contact surfaces re-
cessed into a glass plate. A contact finger riding over the plate energized the torque motor when the gyro
axis was not perpendicular to the gimbal axis. A small space between the silver strips represented the
perpendicular position. The roll axis of the roll and yaw gyro was erected by using the roll output signal
to operate a relay. This relay would energize the torque motor driving the roll signal to zero.

122
Torque motors on this gyro consisted of a permanent magnet rotor inside two coils, one of which was
energized with d-c when a torque was desired. This type of torque motor was limited in the angle over which
it would operate but was satisfactory for the LGW (German) gyro; mechanical stops prevented its turning too

far.

The programmer on the LGW (German) potentiometer gyro was superior to the Anschutz design in that
itturned the gyro potentiometer through the program angle instead of precessing the gyro. It consisted of a
program motor driving a cam to which was fastened a metal band. The band passed around a drum on which
the pickoff jpot was mounted. The total program was adjustable (to some extent) by changing the portion of
the cam used and the total turning angle. To reset the device after a test run, the driving arm was disen-
gaged from the program cam thus allowing the cam to snap back to zero position (when the cam snapped
back, the metal band occasionally broke). Since the program was originaLy intended to be approximately
45 degrees on this cam, a lever reduction had to be inserted between the pot drum and the cam for use at
WSPG. This was necessary to bring the program angle down to approximately 10 degrees. The program
motor was a step type unit which received pulsating d-c from a vibrator in the time switch. Each pulse
caused the program motor to take one step. Pulses were at a frequency of 45 per second. Natural fre-
quency of the clapper arrangement was checked on one program motor and found to be approximately 45-47
cps. It is possible that the device was designed this way to make it considerably less sensitive to low bat-
tery voltage. At the time this apparent "Tuning" was noticed only one original German assembly was
available for test, therefore it was not possible to decide if all units were "tuned" in this manner.

C.1.3 LGW Type (American)

As LGW type gyros became depleted it was decided to produce the
the supply of both Anschutz and
LGW type in America. Waldorf and Kearns Incorporated were selected to build the domestic units similar
to the original German models with a modified program device. The American-made units seemed to be
at least as good as the German models in almost every respect. Typical drift rates were of the order of
0.02 degree per minute while the specifications allowed as much as 0.1 degree per minute.

The chief modification between American and German LGW gyros was the program device. The
American model used the same program step motor but instead of transmitting motion to the pot with a
metal band, the domestic unit utilized a cam following lever. The cam was designed with changing slope
to obtain program angles of about 3.7 to 11.0 degrees. An additional modification was made on a few
gyros to allow program angles of 72 and 91 degrees for certain missiles. These units used a cam follower
with a step-up gear sector and pinion to multiply the rotation obtainable from a cam of reasonable size.
The domestic -designed program device was easy to adjust and gave very good repetition of program angles.
The US-program motor would not operate on as low a voltage as the German unit. This may have
been due to the fact that the German unit was "tuned," that it had a better iron path or a number of other
reasons. When the 72 and 91 -degree programmers were made, special attention was given to this opera-
ting difficulty. A wider gear was used and the motor was "beefed-up" in general to enable the American
model to operate at lower voltage than before but still not as low as the German-made motors.
Alnico magnets used in the American torque motors were stronger than the alloy used in the Ger-
man models. This made the US-torque motors stronger and able to erect the gyro faster. The plastic
forms on which the torque motors were wound did not have as much heat resistance as the German forms.
This allowed deforming of the torque motor and binding of the gyro if the torque motor was overheated.

In the German model, the erection contacts that caused the pitch gyro to erect perpendicular to the
gimbal (roll axis) consisted of strips of silver recessed in glass. The surface on which the contact-wiper
rode was flat. On the American device, the silver was cemented to the glass leaving a depression down
to the glass between the silver pieces. The silver was beveled to allow the wiper to ride up the edge
easily. This left the edge quite thin and subject to damage by arcing. A filter circuit was installed to re-
duce the amount of arcing at this point. However, the arcing was still sufficient to cause snagging of the
contact wiper where the silver had been roughened. This trouble did not cause very large drift errors
in the gyro but was probably one of the most consistent troubles encountered with the American-made gyros.

123
Originally, the test instructions for the American gyros called for the contact adjustment on the gim-
bal erection pendulum to be such that the "dead space" (space between erection from CW position com-
pared to erection from CCW position) should be less than 0.1 volt, or approximately 0.04 degrees. Since
there was no minimum specified, a few gyros were built with zero dead space. A minimum value of dead
space was established to avoid having both torque motors energized simultaneously.

Sparking at the contacts of the pendulum device occasionally caused interference with radio equip-
ment in the missile during ground tests. This was not a critical problem, however, since erection power
was cut off at lift.

C.1.4 Test Proc edures and Specifications

The following is an outline of the procedures used in tests of LGW (American) gyros:

a. Inspect all solder joints, lead-in contacts, potentiometer windings and wipers.

b. Check continuity and circuit resistances through gyro motor, potentiometers and torque motors.
c. Megger each circuit to ground. Minimum allowable, 10 megohms.
Test motor starting and running current, time to full speed and direction of rotation. Starting cur-
d.
rent should be approximately 2.1 amperes. Running current should be 0.25 to 0.50 amperes. Time to full
speed should be less than 2,5 minutes. Motor should turn CCW looking from lead-in end.

e. Test vertical erection (pendulum) from both CW and CCW directions. Dead space should be more
than 0.04 and less than 0.06 degrees.

The following tests apply to the roll-yaw gyro only:

Set gyro on vibrator-oscillator test table (Fig. 58) with the connector plug toward the west (i.e. gyro
f.

axis in a north-south plane when erected). Level the test table and turn on gyro and erection power. When
gyro has erected, note the readings on both roll and yaw output signal instruments on gyro test panel (Fig.
58). These zero readings should be less than 0.2 volt (0.08 degrees). Turn off erection voltage for two
minutes and read the "wander" for this interval on signal output voltmeters. Yaw "wander" should not
exceed 0.2 degree for any two-minute period on three consecutive tests. "Wander" in roll should not ex-
ceed 0.4 degree under same conditions.

g. Repeat test f with vibrator and oscillator running during the two-minute intervals. The limits given
in f apply.

The following tests apply to the pitch gyro only:

h. Perform both the static and vibrator-oscillator tests of (f) and (g) above, reading the "wander" on the

pitch output voltmeter. Wander should not exceed 0.2 degree for any three consecutive tests of two-minute
duration.

Energize the time switch from the gyro test panel to provide vibrator supply (45 pulses per second)
i.

for the gyro program motor, After four seconds the program motor should start operating and run con-
tinuously for 48 ± 1 seconds. The program signal, which is fed to a photoelectric recorder, should increase
along the selected angle-time curve until equal to the desired program at 52 seconds (total time). In the
case of the seven degree program (the program used most frequently for the V-2 at WSPG), the angle-time
curve is essentially a straight line. At the end of the program test, the output voltage to the recorder should
stay constant (17.5 volts equals 7 degrees) until the program motor starts to recycle to zero. The final
program should be within + 0.2 degree of the desired value for three consecutive tests.
j. The coast time for all gyros should be more than 12 and less than 30 minutes.
The following applies to the location of both gyros on the gyro plate:
Accurately level the tilt stand (Fig. 59) with the level plate. Install the flight gyro-mounting-plate
k.
on the tilt stand, using the machined index fixture to obtain accurate orientation of the plate with respect

124
Fig. 58 Gyro-test Table and Panel

Fig. 59 Gyroscope Tilt Stand

125
to the stand. Secure the pitch gyro to the plate (the machining of the plate will determine the location of
the pitch gyro with sufficient accuracy). Install the roll-yaw gyro with its axis displaced approximately
90 degrees from the pitch gyro. Energize both gyros, including their erection circuits, through the gyro
test panel. After both gyros are up to speed and erect, disconnect the erection voltage and turn the tilt
stand through the program angle. Note any output signal in either roll or yaw. If either signal is pre-
sent, adjust the roll-yaw gyro orientation, re-erect the gyro and repeat the tilt test. Continue this
test-and-try adjustment until the roll-yaw gyro orientation is such that it produces no signal when the
tilt stand is rotated through the entire program angle.

C.1.5 Special Vibration Test

Normally, the gyros were not given a vibration test because it was felt that more trouble would be
introduced than would be detected by such tests. However, one US-built gyro was given a rather thorough
vibration test.

The gyro was mounted on a standard plate which was attached to the vibration table without shock
mounts of any kind. The direction of vibration was approximately 45 degrees with respect to the gyro
and gimbal axes.
Preliminary "wander" tests were made to provide a comparative basis for the "wander" rates to be
determined after each vibration cycle. The vibration amplitude was changed in steps of 0.01 inch from
0.01 to 0.10 inch. At each step, the frequency was varied from about 10 to 60 cps.

During each test, oscillograms were made of motor current, signal output of the pitch pick-off and
vibration table frequency. After each run "wander" rates were checked. The roll rate was measured by
an optical system and the pitch rate by the voltage from the pitch potentiometer. The "wander" in a period
of five minutes is given below:

TEST PITCH ROLL

degrees degrees
Before vibration 0.00 -0.11
After 0.01 inch vibration -0.08 -0.17
After 0.02 inch vibration 0.04 0.11
After 0.03 inch vibration 0.02 -0.06
After 0.04 inch vibration 0.02 -0.11
After 0.05 inch vibration -0.16 -0.50
After 0.06 inch vibration -0.12 -0.28
After 0.07 inch vibration 0.02 0.17
After 0.08 inch vibration 0.08 0.06
After 0.09 inch vibration -0.64 0.06
After 0.10 inch vibration -0.29 0.23
After return to room
0.64 -3.45
temperature
It will be noted that (with one minor exception) the wander rate remained within specification limits

(0.5degrees maximum in five minutes) throughout the tests. After the gyro had cooled down, the static
balance along the motor axis shifted enough to create a large drift.

The oscillogram showed that the lead-in leaf contacts opened momentarily at various times in the
range of frequencies from 48 to 55 cps. This occurred, in varying degree, at all amplitudes above 0.01
inch.

C.2 MIX COMPUTER (Autopilot servo amplifier)

Reference: Backfire, Vol. II, p. 138.

The name of this device, mix computer, was developed from a translation of the German name, Mischger'at.
In this country it would be called an autopilot servo amplifier.

126
The mix computer as used on this project was a servo amplifier with inputs from the pitch gyro, the roll-
yaw gyro, and the jet vanes 2 and 4 synchronizing circuit. There was also a provision for a guide beam (LS)
input into the yaw circuit. This input was used with a German radio guidance system to correct the azimuth
of the missile during burning. The German guide beam system was never used at WSPG.

From these inputs, the mix computer developed output control current for the four servos attached to jet
vanes 1, 2, 3 and 4. It contained the necessary rate networks to make the complete servo system stable.

Both German and American mix computers were utilized at WSPG. The American units were built by the
General Electric Company. Electrical circuits were essentially the same in both units; however, the mechan-
ical construction was entirely different.

The German computers were 13 x 9-3/4 x 7 inches over -all, weighed 32 pounds and were constructed with
components of a good quality. However, the wire was generally plastic-insulated, solid-copper. This solid
wire gave some trouble in test (particularly vibration tests). The wire would break occasionally, probably
due to; (1) handling the wire improperly during manufacture, insulation may not have been removed correctly
from the wire (and the copper was nicked or weakened) and (2) age (units were old and had been reworked so
many times some of the copper wire had been fatigued). It is felt that stranded wire would have been more
satisfactory in this application.

From discussions with German personnel and from the appearance of different computers, it was ap-
parent that these units were tested and adjusted by laboratory personnel in Germany. In many units there
were wires which had been added to replace wires in the regular cable harness.
For the discussion below, the V-2 rockets launched will be separated into three groups. The chronological
I, page 6
listing is included in Table .

Group I: In general, missiles 2 through 20 and missiles 25 and 38 (missiles 25 and 38 were chronologically
launched during Group III below, but because of the type of mix computers used will be considered part of Group
I). Missile 19 had a non-standard V-2 steering system; therefore it is not included.

These missiles all had standard German mix computers with no modifications. Of the 20 missiles in this
group, the steering systems of eight failed and gave erratic flights.

The specif ic components which gave trouble in the German mix computers were wire (as mentioned above),
vacuum tubes, dry-disk rectifiers and electrolytic condensers. Vacuum tubes are mentioned here notonly be-
cause of the number of failures, but also because the supply was short. American-equivalent vacuum tubes were
tried in one computer This computer worked satisfactorily but it was difficult to replace the German tube sockets
.

with American. It never became necessary to make this modification on more than one unit. Dry-disk rectifiers
and the electrolytic condensers were replaced with an American equivalent whenever a failure occurred.

Group II: Missiles 21 through 27 (after missile 24, they were not fired in numerical order). Of the eight
missiles in this group, the steering system in five failed and gave erratic flights.

During this time one missile was launched by Operation Sandy and is not included here. This missile
carried a German mix computer as in Group I but there were modifications to the steering system; the steer-
ing system of this missile failed.

Because of the number of steering -system failures, it was decided to mount the German computer in a
container designed to keep the unit at atmospheric pressure during flight. This was done to prevent voltage
breakdown in the power supply as the altitude of the missile increased. The possibility of voltage breakdown
was discussed in General Electric Company - Project Hermes report 45779. In addition, the subject was
discussed with German personnel. Their comments were that such a breakdown had been considered in Germany
but that it was not believed serious. By enclosing the computer in a pressure tight container, the possibility of
voltage breakdown was eliminated; however, the problem of heat dissipation from this container was increased.
No method of ventilation (such as circulating air) was used; therefore, high temperature rises were possible.
Group HI: Missiles "Special" through 52. Missile 36 had a non-standard V-2 steering system and it is
not included.

The mix computers in all missiles of this group were American-made.

127
When the number of missiles to be launched was increased to 100, it was apparent that there would not be
enough German-made mix computers. Therefore, 80 American-made mix computers were built, 60 for the V-2
program and 20 for other programs. These 80 units were to duplicate essentially the German electrical design.
To obtain electrical-design data, several German units were disassembled; a new mechanical design was made
to utilize standard American components.

The American computers were 13 x 9-5/8 x 7-1/8-inches over-all and weighed 28 lbs. The only difference
in the electricaldesign between the German and American units was in the power supply. The power supply in
the German unit was a full-wave, center-tapped transformer type using a vacuum tube rectifier. The power
supply in the American version was a voltage -doubler type using a dry-disk rectifier. This power-supply change
reduced the secondary voltage of the power transformer one half, which minimized the chance of voltage break-
down.

Six of the American-designed mix computers were built as production samples. Tests performed on these
units included: (1) electrical, (2) vibration, (3) cold and (4) hot.

In electrical tests it was found that the gain was low. By adding condensers to tune some of the transformers
to 500 cycles the gain was increased to a value equivalent to the German unit. In vibration tests it was found
that some components, brackets and shelves were not mounted properly. The components were relocated and the
brackets were strengthened. The shelves had additional supports added changing the cantilever type construc-
tion to box type. No troubles were encountered in the hot tests.

In general, it was learned that some mechanical changes were necessary to facilitate maintenance. It was
also found that the stranded wire being used, although better than solid wire, had poor insulation; this was
changed in later units. The exposed shielding on some of the cables gave trouble in the grounding of exposed
terminals. This was corrected by adding insulation where necessary.
All missiles (36) in Group III used American-made mix computers. The steering systems of three mis-
siles failed and gave erratic flights. Two of these failures, Bumper No. 7 and 8, were not due to the mix
computer.
The first missile of this group "Special" was launched primarily to test operation of new equipment.
This rocket was the first in which an American-made mix computer was used. The test was considered
satisfactory as a steering system test; however, there were other failures which did not make the flight
completely successful.

Starting with the next missile (28) there was some oscillation in vane 2 during test. By replacing the mix
computer, this trouble was eliminated. During the launching of missile 25, it was found that there was oscil-
lation of vanes 2 and 4. This oscillation occurred so late in the launching sequence that it was decided to re-
place the unit with a German computer and proceed with the launching. The steering system performed
satisfactorily.

After the launching of missile 25, the cause of the oscillation of vanes 2 and 4 was investigated. It was
found that the gain of the vane 2 and 4 synchronizing circuit was sufficient to cause this oscillation. The gain
of this circuit was reduced. The vane 2 and 4 synchronizing circuit, of which the mix computer is part,
causes the vanes to follow each other. These vanes control pitch only and follow each other to minimize the
roll introduced to the missile by the vanes themselves. A switch was added to the Vane 2 and 4 synchronizing
circuit to remove the synchronizing signal during final vane balance. This facilitated the vane balance and
minimized the effects of a failure in the synchronizing circuit.
During the preflight tests on missile 38 (the next rocket launched after rocket 25) it was found that the
30 mfd rate condensers were breaking down. The condensers are small in size (2 1/4 x 2 1/4 x 2 inches) for
their capacity (30 mfd) and voltage rating (100 volts d-c working); they are constructed using a metalized
paper process. These condensers had one bad characteristic as far as trouble shooting was concerned;
after voltage breakdown caused a portion of the condenser to short out, it would reheal and the condenser
would appear perfectly normal. This problem of voltage breakdown with the 30 mfd rate condensers was
referred to the manufacturer.
Because of the condenser trouble, missile 38 was launched with a German computer. The steering system
of this missile failed.

128
In view of the trouble with the 30 mfd rate condensers and to give the manufacturer time for tests, it was
decided to use the American-made mix computer with German rate condensers. Missiles Bumper 1 through
V-2 45, (12 missiles) used this equipment.
Voltage breakdown of the 30 mfd rate condensers was overcome by a heat treating process found to be
very effective. Of the 120 condensers subjected to this heat treating process only one could not be used.

At the time Bumper 1 was launched, a considerable amount of electrical pick-up experienced in the mix
computer was traced to the guide beam receiver input. This circuit was shorted out to prevent further pick-up
troubles; all remaining missiles also had this circuit shorted out.

During the pre-launching tests of Bumper 3, there was a considerable amount of random motion of the
command current for vanes 1 and 3. This effect was traced to electrical pick-ups on the leads to the block-
house. These leads (part of the roll erecting circuit for the roll yaw gyro) were connected directly, to the
roll output circuits of this gyro and therefore, were the inputs to the roll circuit of the mix computer. The
pick-up gave false signals to the rate network in the mix computer and caused command current fluctuations.

A similar result could also be seen in the pitch and yaw circuits when gusty winds moved the missile.
The small motion between the missile structure and gyro pick-offs gave signals to the rate networks and
appeared as random variations of command current.
Fluctuations of the roll command current caused by the pick-up were not considered serious. However,
since they were undesirable some attempts were made to eliminate them. The two erection leads to the
blockhouse were shielded and a condenser was added directly across the leads at the ground stotz plugs
(missile drop-away plugs). This method was discontinued because the hazard involved was not considered
worth its value.

During later missile firings it was noted that American-made mix computers adjusted in the laboratory
would not give the same results when mounted in the missiles. Investigation showed that this trouble was
being caused by the output transformers. It was noted that two differences between the German-built and
American-built transformers existed: (1) the windings in the German transformers were random-wound
while the American were layer wound and (2) the laminations of the American transformers were a better
grade of transformer steel than the German.
During investigation of the transformers it was found that if a one microfarad condenser was connected
between each side of the computer output to the common lead, the trouble was reduced to a negligible value.
Therefore, because of the small number of missiles involved it was decided that the condenser solution was
more economical than a transformer re-design. If, however, a large number of these units were to be built
for future use, it is recommended that these transformers be redesigned.

The remaining V-2 missiles (48 through. 52) were flown with condensers in the output circuits and with
more careful laboratory adjustment.

During the launching of Bumper 6, the servo balancing potentiometers opened circuit and caused incorrect
command currents. It was found that in adjusting the potentiometer, the arm was turned past the stop and
bent.When it was turned back to the resistance card it would cut the resistance wire. Turning the potenti-
ometer past the stop was a result of improper adjusting of the lock nuts. During this investigation, it was
noted that the electrical connection to the ends of the resistance card of the potentiometer were made by a
friction fit. For these reasons, a larger potentiometer having better mechanical construction and better
electrical connections was used.

C.2.1 Investigation of the American Version of the German Computer

The object was to determine the reasons
of this investigation for the variation in operating charac-
teristics between the German and American computers.
It had been found, that a German computer could be tested and adjusted in the laboratory, put in the

rocket and pass all pre-firing tests satisfactorily. The American computer would be tested and adjusted
to give duplicating laboratory measurements but when placed in the rocket it would not fill the require-
ments of pre-firing tests.

129
For the computer investigation, the procedure outlined below was followed. Both a-c and d-c input
voltages, variable from zero upward were available. The d-c voltage could be varied from a plus to
a negative value and the a-c varied 180 degrees out of phase. Zero-center milliammeters in the center
leg of each servo circuit were used to observe the plus or minus d-c current.

When tests for yaw, pitch and roll were made, a signal of + 6 volts d-c was applied to each input in
turn and the output currents recorded. A current of 9 i 2 ma with an input of 6 volts would have been
satisfactory. Equal outputs with both plus and minus inputs were desired. To arrive at equal outputs it
was necessary to tune various transformer windings with capacitors and adjust potentiometers in balanc-
ing circuits. When satisfactory balance was secured the synchronizing circuit was tested.

During the above tests, the loads were a set of servo-valve coils similar to those used in the missile.
It was foundthat when the American computer was adjusted and balanced for correct operation (9 ± 2 ma
with a 6 volt signal) on the laboratory loads, duplicating values of output current could not be secured
with missile servo loads. Test results are included in Table VII.

TABLE VII

LABORATORY AND ROCKET TESTS, GERMAN AND AMERICAN COMPUTERS

Roman numerals in heading refer to fins
Arabic numerals in Table are milliamperes

Computer Sig Pitch Yaw Roll Sync Ls Balance Pot Current 750 Ohms Across Output
- II IV I ni i ni n iv I III II IV I III II IV I in
German 322 -1 -1 +1 +1 -8+6 -9+4 +6-6 +7-5 +6-8 +5-8 +9-10 +10-8
in Lab* +6 +8 +8 +10 +10 +10 -8 +6 -10 +13 +13
-6 -9 -10 -8 -9 -10 +10 -8 +6 -12 -12
in Rocket -6+7 -5+6 -10+8 -10+8 -7+7 -8+8 -6+4 +6-5
+6 +8 +9 +8 -8 +7 -7 +12 +12
-6 -8 -8 -9 +9 -7 +8 -13 -13

American 24
in Lab +6-6 +6-6 -4+4 -5+4 +8-8 +8-8 +10-10 +10-10
+6 +10 +10 +10 +10 +10 -10 +4 -4 +16 +16
-6 -10 -10 -10 -10 -10 +10 -4 +4 -18 -18
in Rocket +4-6 +4-3 -2+3 -3+2 +5-10 +9-6 +7-5 -5-6
+6 +14 +12 +8 +12 +11 -5 +18 +23
-6 -15 -11 -8 -13 -13 +5 -20 -25

* This computer received no tuning or adjusting for months.

130
Considered most important were the unbalance and load tests. In the former, the balance pot shown
in Fig. 60 was turned to the limits on each side. The current unbalance indicated on the meter should
have been 5 ± 2 ma. In the load test, a 750-ohm resistor across one leg of the three control-current
leads should have resulted in 9 + 2 ma. During laboratory tests the computers were tuned to meet these
conditions. Results obtained on American computers prepared for firing are included in Table VIII.
The differences shown in Table VIII were evident in all American computers tested. Results of the
750 ohm tests were just as variable between the laboratory and rocket installation.
TABLE VIII

LABORATORY AND ROCKET TESTS - AMERICAN-MADE V-2 COMPUTERS

Numerical values in Table are milliamperes

COMPUTER FIN
NO. II IV I in

17 LAB +6 -8 +8 -6 -4 +5 +4 -4
ROCKET +10 -11 +12.5 -11 -7 +7 +5 -12.5

13 LAB +4 -5 +4 -4 -4 +4 +4 -4
ROCKET +7.5 -3 +6 -6 -6 +7 +7 -8

26 LAB* +3 -3 +3 -4 -2 +3 +2 -2
ROCKET +2 -0.5 +3 -0.5 -6 +3.5 +2 -3.5

24 LAB +6 -6 +6 -6 -4 +4 +4 -5
ROCKET +7 -11 +8 -10 -5 +5 +6 -7

This computer was not worked-over before test. Adjustments were

made on gain and balance pots only.

ANV\
dt M « fa-

1.6 K

750 750

Fig. 60 Computer Output Windings

131
From the beginning of this investigation it was felt that the main difference was in the output trans-
former (Fig. 61). Therefore, two German-made and two American-made transformers were tested.
Inductance measurements were made of the windings; a General Radio type 650 A inductance bridge was
used for measurements at 1000 cycles. The test data are included in Table EX.

It should be noted from Table EX, that winding 1-2 of the American transformer decreased in inductance
when other windings were shorted; in the German transformer this winding increased in inductance.

Figure 62 is a log-log plot of the inductance of control winding 1-2 versus the resistance on windings
5 to 7 and 8 to 10. Attention should be directed to the slope of the curves labeled American KV1921 and
German, over the operating region. This section is considered the operating range since the load of the
servo valve coils fall between these points. The change of inductance in the American transformer is
quite large as compared with the German when the loading varies between 220 to 280 ohms.

TABLE IX

TEST RESULTS, GERMAN

AND AMERICAN TRANSFORMERS
GERMAN AMERICAN
WINDING A B A B
1-2 1.07 1.26 8.0 8.2
3-4 2.2 2.25 1.6 1.14
5-7 0.7 0.7 0.55 0.63
8-10 0.58 0.65 0.56 0.67

All windings open except winding being tested.

Numerical values are in henrys.

OTHER WINDINGS OTHER WINDINGS

SHORTED OPEN
AMERICAN GERMAN AMERICAN GERMAN
WINDING 1000 500 1000 500 1000 500 1000 500
1-2 1.7h 1.25h 3.65h 2.3h 7.8h 6.1 8h 1.04h 0.73h
3-4 8.5mh 4mh 1.07h 0.65h 1.95h 1.50h
5-7 4.7mh 2.5mh 0.56h 0.47h 0.68h 0.50h
8-10 4.4mh 4.3mh 2.5mh 7.4mh 0.56h 0.45h 0.68h 0.55h

Tests at 1000 and 500 cycles.

::i
10

Fig. 61 Schematic Wiring Diagram of Output Transformer

132
3 Q i

oo
o
2
<

2-1 9NIQNIM*3DNV10naNI AHN3H

Fig. 62 Inductance in Windings 1 and 2 Versus Resistance in Windings 5-7 and 8-10
133
To determine the variation in valve coil characteristics a number of servo pumps were measured, The
data are indicated below:

INDUCTANCE RESISTANCE INDUCTANCE RESISTANCE

mh ohms m> ohms
720 241 575 252
790 242 570 250
625 255 500 252
600 253 505 245
598 245 686 249
The maximum variation was 290 mh (790 to 500 mh) inductance and 14 ohms (255 to 241 ohms) resistance.
From Fig. 60can be seen that if an unbalance of 10 ma existed in the circuit due to a control signal
it

applied from the computer through windings A and B, a difference of 14 ohms should not cause a change in
the current of more than one-half of one percent. Yet, in the actual operation of a computer, changing the
servo had caused a variation in current up to 100 percent.
With the above knowledge, it was felt that the unbalance was due to the reactive effect caused by the
variation in inductance of the servo coils. To counter this effect, one-microfarad capacitors were placed
across the output leads from the computer. After rebalancing the computer it was found that a change in
load caused no appreciable change in current.

To prove that the output transformer in the American computer was the cause of most of the trouble,
a set of German transformers was substituted for the KV1921 type. Test results from this modified
American computer were comparable to results from the German units.

In an effort to pinpoint the trouble, a KV1921 and a German transformer was disassembled. The
German core was inserted in the American windings. Test results are shown in Fig. 62. In addition,
laminations from a Stancore Input Transformer, Type A-53-C was found to fit the windings from a KV1921.
Test data are included in Fig. 62.
Another set of curves were made on control winding 1 and 2. Initially a volt-ampere test with all
other windings open was made (Fig. 63). Then a volt-ampere curve with an excitation voltage of 40 volts,
500 cycles on windings 3 and 4, with the normal load (rectifier and servo pump) on windings 5 to 7 and
8 to 10 was completed (Fig. 64). In operation, the voltage and current in winding 1 to 2 varied between
2 volts at 0.2 ma to 40 volts at 7.6 ma.

In general, it was felt that the type KV1921 transformer was the offending unit and if made to con-
form more closely with the German output transformer, better operation would result.

3 4 5
MILLIAMPERES

Fig. 64 Volt Ampere Curve With 40 Volts 500 Cycles on Windings

3-4 and Normal Load on Windings 5-7 and 8-10
134
300

KVI92I WINDINGS

S "fTANCORE
COKE 6U

/
200 /s
K VI9 21 .
//
/ '
/
/, /
CO /g.ERM AN
Li //
y
o
> /,•

/
f

/
100 /
//
/
//
//
/
y
//
Lr
r\ V 10 15 20 25
MILLIAMPERES
Fig. 63 Volt Ampere Curves for Control Windings 1-2 With Other Windings Open
135
C.3 SERVO
Each jet vane was driven by a self-contained servo-mechanism'^'. Its basic components were a double-
gear pump driven by an electric motor, solenoid-operated control valves, an individual oil reservoir and a
hydraulic piston.

An analysis of all in-flight steering troubles indicates that the servo was seldom, if ever, the cause of
steering failure. It was, however, one of the weaker components of the V-2 and considerable effort was ex-
pended in providing servos which were considered adequate for flight service.

One of the reasons for the marginal character of the servo was the fact that it was not designed for the
V-2 but was originally produced for aircraft use. It appears that its power output was somewhat less than
would normally be specified for a servo specifically designed for the V-2. Nevertheless, the performance
of the servos when new, must have been reasonably acceptable since they were used on some thousands of
successful flights in Germany. This assumption is substantiated by the fact that a limited percentage of the
servos received at WSPG met the test specifications without difficulty.
Another factor contributing to the servo difficulties at WSPG was wear, particularly in the gear pumps.
A number of different types of gears were found in the pumps. Some were hardened while others were of
relatively soft steel. Some had a smooth finish while others showed tool marks very clearly. As might be
expected under such conditions, some showed the effects of wear in a short time while others were much
less susceptible. With a fairly extensive amount of testing at WSPG, added to an unknown amount of running
in Germany, many of the servos fell below the specified performance.

Some servos would pass test specifications with cold oil but fail when the lubricant reached a higher
temperature. The following tables show the failure point for 132 servos received in one shipment.

TEMPERATURE NUMBER OF SERVOS NUMBER OF

°C EXCEEDING SIX- SERVOS
SECOND LIMIT STALLING
35 13 9
40 33 9
45 18 6
50 26 2
55 9 2
60 3
65 1

70 1

104 28

The following table shows the average time required to produce a rise of 5°C with the servo running at
no load.

TEMPERATURE AVERAGE TIME NUMBER OF SERVOS

°C REQUIRED IN AVERAGE
Minutes
35-40 4.2 108
40-45 4.3 66
45-50 4.4 45
50-55 3.7 14
55-60 6.0 5

136
Fromthe above data it was evident that it would be necessary to improve the performance of the weaker
servos if trouble was to be avoided in the latter stages of the program. It was hoped that sufficient improve-

ment could be realized through the use of some hydraulic oil having a smaller change of viscosity with tem-
perature. An investigation showed that the Germal oil was approximately equal, in this respect, to the hydraulic
oils available in this country. This meant that any appreciable improvement in viscosity index would have to be
obtained through the use of a silicone oil. Tests demonstrated'* 5 ) that silicone oil could not be used in the
servo gear pump without modification because of the high probability of seizure. Modification of the pumps
did not appear justified because the improvement in performance was not great and was partly neutralized by
poor circulation of the fluid in the servo case.

The German pump motor was known to have exceptionally poor speed regulation. Application of full load
would reduce the motor speed to approximately 50 percent of its no-load value. It was evident that great im-
provement could be expected by substituting a motor which would hold a reasonably constant speed over the
normal load range.
No standard motor was found which would satisfy the requirements of this application without modifica-
tion.Eventually it was necessary to rewind a motor originally designed for aircraft service. The perfor-
mance of the motor was very satisfactory, holding speed within about three percent from no load to full load.
However, the improved motor performance was obtained at the cost of weight and power requirements.
The individual German servo weighed 17 pounds. The servo with the new motor weighed 25 pounds. No-load
current was doubled, increasing from an average of four amperes for the German motor to an average of
eight amperes for the new motor. To meet the added current requirements, a third main-power battery was
added to the two normally carried.

At the same time that the motor was replaced, each servo was thoroughly reconditioned. Pump gears and
pistons were replaced and cylinders were re-lined if required. After this overhaul each servo was given a
complete set of tests at the factory. This was followed by similar tests at WSPG prior to installation in the
missile. Sample test sheets are shown in Fig. 65 and 66.

The result of the motor change and the overhaul was an outstanding increase in performance. Average
time required to complete the full travel was reduced to about 2.3 seconds. Previously, the time varied from
6+ seconds to a complete stall.
Neither the motor substitution nor the overhaul introduced serious troubles. There were three minor dif-
ficulties, all concerned with oil leakage. Initially a lead gasket was used at the adapter plate between the new
motor and the servo case; considerable leakage resulted. A fiber gasket was tried but no appreciable improve-
ment was obtained; a neoprene gasket proved satisfactory. There was also some leakage past the oil seal on
the shaft of the motor. The problem was to obtain an oil-tight seal without placing unnecessary load on the
motor or causing heating of the shaft at 4000 rpm. A change in the type of seal produced acceptable results.
The third trouble was the "wicking" of oil from the servo case by the control wiring. A number of interim
remedies were tried with some improvement. None were entirely successful. Selection of wire with suitable
insulation avoided this trouble.

Low power output, as previously discussed, was the defect responsible for the majority of rejected servos
at WSPG. Two other defects, drift and low sensitivity, accounted for most of the remaining rejections.

Excessive drift was the result of improper adjustment of the zero position of the two control valves.
Test specifications required that with no mechanical load and with zero signal current, the time of movement
from one extreme to the other must be greater than 30 seconds. The servo design included adequate pro-
visions for adjusting the position of the valves, but the actual process was rather difficult. This was due
(in part) to the fact that a change in adjustment to improve drift was very likely to upset the adjustment with
respect to sensitivity. The valves were delicate and were sensitive to very small changes in adjustment.
Special test facilities would be required to do the job properly. The rejections for excessive drift were not
considered large enough to justify the construction of such facilities at WSPG. Consequently, adjustments
for drift normally were not made at WSPG.

Test specifications for sensitivity stated that the servo must move a load of one meter-milogram very
slowly with a differential control current of ± 4 ma. The foregoing comments regarding valve adjustments
apply to sensitivity as well as drift; such adjustments seldom were made at WSPG.

137
d
s
H „ CO CO T)< m IM CO
" ™ ™ "» "* *
3 "
o
rt
Z '

o 1
1

l-J
CO r-
! "J
CO
o!o|
CO CO
1
CO CO
i

in
* u ft< "U.H'-Ip; pi
CO

lATION

I.Y.

IS
SERIAL
40

R
80

R
o

OS
M
d
PS
S
-!

PS
*
d
PS co
7~ i Mj «
01
j

a.
CORPOI
a <
York,
a
< co oj
H r r
j

New
O m « 2 a
38.5
208
80 ti
f 31311
JMENT
d J d < < j
:

St.,

PS2 J j J J J J J PS J PS J K J PS
H ?
INS
Lafaye
£ C € £ °o € €
A-4)
RF
1 ,

§s I . c
$ s
a U CJ
u
53
!
w 1
o CD CD 4>
in
0)
m
•
\

** « ; * *
k a ft
as-- £"" «- ^• »* ^ -
, ~ i
s in S a

8 - 1 £ s £ 2 H
w
os
gjssg^aa -^sq ^sq ^sq
H U Z co co co to co

H ( !
_' cm co' » in" co' i-'

ago rt Z Z
ho
pa

< ° O O
CJ
H SJ
b § j j 1
J OPS
Z
1 K 5 1
sir soils
13 *
JI
=
-
3
«
OS g S ps
£ S - 1
8 u.
CO N O «S '

h 63 3S|
H
CM W o S PS
i H z

1 III-
jps i
so
^| BJ
alt rt

> §e «
a
c
a o«°
i
1 1 1 s
.£_ a

Fig. 65 and 66 Sample V-2 Servo Test Sheets

138
The German test specification called for the servo (with 28 volts applied and a control current of 30 ma)
to move from one extreme to the other, without load, in less than 1.6 seconds. With a 30 meter -kilogram
load, the maximum time should be less than six seconds. A special test panel and load fixture(^) (Fig. 67)
provided facilities for accurately measuring the travel time. This type of test was used exclusively during
the earlier part of the program. In 1947 facilities were arranged for measuring the stall-torque for both
CW and CCW motion.

'
Hi '
1 ' I
'
!

"
tutni !
u ^^^ m
nJfit^ i'

bHJp J( t

^B
.

g*

fffiffifr 71 \
Uw \M
W* V

--': "

^EStE jEr^'fr

W .•--,-"
H9

Fig. 67 V-2 Servo Being Tested With Load Test Panel

139
CA TIME SWITCH
Reference: Backfire, Vol. II, p. 132.

The German time switch was satisfactory and was used with only a few modifications at WSPG. These
modifications usually consisted of slight wiring changes and some changes in time of operation of certain
cams. The switch consisted of a regulated-speed d-c motor which, through a gear train, turned a cam shaft
mounting 14 cams. These cams actuated switches at various times depending on the calibration of the cams.
The vibrator that furnished power for the gyro program motor was also mounted in the time switch cam.
This vibrator consisted of a vibrating metal reed tuned to vibrate at 45 cps. It was mounted between two
contacts such that when it hit one contact it gave a pulse of current to the program motor in the gyro and
when it hit the other contact it energized an electromagnet which kept the reed vibrating. Dirty contacts
and misalignment of the reed constituted most of the troubles encountered with the time switch. The test
panel and time switch are shown in Fig. 68.

The following is an outline of the test procedure followed in preparing the time switches for missile use.

a. Megger each circuit to ground. Should be more than 10 megohms.

b. After making necessary cam adjustments and associated wiring changes, check timing of all cams to
be used in proposed missile flight. All should be within + 0.1 second of desired time for any three consecu-
tive tests.

c. Check time for a complete cycle of the time switch. Should be 90 * 2 seconds over a 25 to 30 volt range.

d. Be sure the vibrator will start repeatedly and will continue to run with the time switch mis-oriented
from its prescribed missile position up to ± 30 degrees.
e.Check time switch in conjunction with the pitch gyro proposed for use and the spare. The gyro program
motor must operate reliably over a range of 26 to 30 volts, when supplied by the vibrator of the time switch
tested.

Fig. 68 V-2 Time Switch and Test Panel

140
C.5 INVERTERS AND REGULATORS
Reference: Backfire, Vol. II, p. 129

The inverter and regulator set proved to be a very dependable unit. It supplied 500 cycle a-c power for
operation of the gyro motors and the missile servo amplifier. The inverter unit consisted of a d-c adjustable-
speed motor driving an alternator with a six-pole permanent-magnet rotor and a wound stator. The regulator
was a rather unique unit with no moving parts. The frequency sensitive portion of the circuit consisted of two
inductance-capacitance tuned circuits, one tuned slightly above 500 cps and the other tuned slightly below 500
cps. In series with each of these tuned circuits was a bridge rectifier circuit. Direct-current from these
rectifiers flowed through two saturable reactors which controlled flow of a-c current into another rectifier, the
output of which operated the control field of the d-c motor driving the inverter. Each tuned circuit tended to
make the inverter operate at its frequency. This system held the frequency to about one-tenth of one percent
with as much as 10 percent change in line voltage. Adjustment was obtained by changing one of the adjustable
inductors. Difficulties with the inverters and regulators consisted mostly of sparking at the brushes of the
inverter and poor soldering or broken connections in the regulator. Radio interference was encountered a few
times but seemed limited to individual cases where brush sparking was a little worse than usual. Since there
was a large supply of these inverters, no attempt was made to repair an occasional troublesome unit. The
original German wiring system used a List 14 pole plug. Because the supply of the matching plug was limited,
the plugs were removed and an American terminal board substituted. The inverter, regulator and test panel
are shown in Fig. 69.

roR

Fig. 69 Inverter Regulator and Test Panel

141
The following is an outline of the test procedure used for the inverter and regulator.

a. Remove German List 14 pole plug and replace with a Jones-type strip. Connect regulator to inverter
terminals and anchor new connector cable to inverter case.

b. Remove regulator case; visually inspect all solder joints and connections. Check for capacitor grounds
to base.

c. Apply d-c line voltage (with load switch off) and observe test panel instruments for any abnormalities.
Check regulator adjustment to be sure it will cover the desired range (500 cycles).

d. With supply voltage set at 29.5 volts, turn load switch on; after inverter has taken up the load, adjust
the regulator for 500 cycles (the load is a bank of two gyro motors in parallel).

e. Turn voltage down to 27.5 volts, frequency should remain at 500 + 1 cps.

f. Turn voltage down to the point at which the regulator can no longer control the frequency. This voltage
must be less than 25 volts.

g. Megger each point on terminal strip to ground. Must be more than 10 megohms.

142
APPENDIX D
MISSILE FAILURES

D.l SUMMARY
MISSILE PROBABLE CAUSE
NO. OF FAILURE
2 Broken jet vane
8 Pump bearing seized
10 Open circuit in wiring to servo
11 Pickup in guide-beam circuit
14 Failure in computer or roll-yaw gyro
16 Open circuit in pickoff potentiometer of pitch gyro
18 Failure of power supply in computer
20 Failure of tube in computer
24 Failure in computer
26 Failure in yaw gyro
27 Failure of tube in computer
29 Pickup in guide-beam circuit
30 Open circuit in wiring to servo
32 Break in alcohol piping
37 Open in control circuit to alcohol preliminary valve
38 Open in computer or roll-yaw gyro circuit plus open in synchronizing pot
39 Hydrogen peroxide tank not fully loaded
40 Open in control circuit of alcohol main valve
42 Computer failure or open circuit in control wiring
45 Poor regulator operation plus high winds
46 Cutoff relay operated through undetermined failure in control system
50 Intermittent contact in control circuits for both main propellant valves
52 Leak in alcohol piping
54 Break in lox tank pressurizing system
55 Separation explosives detonated
57 Leak in alcohol piping
Bumper 2 Failure in control circuit of alcohol preliminary valve
Bumper 4 Break in alcohol piping
Bumper 6 Cutoff relay operated by undetermined failure in control system
Bumper 7 Sneak circuit erection system of pitch gyro
in
Bumper 8 Sneak circuit erection system of pitch gyro
in
Special Failure in control system for alcohol preliminary valve

D.2 INDIVIDUAL REPORTS ON FAILURES

The following pages contain detailed description of each failure with comments concerning the probable
cause.

143
MISSILE 2

Performance
The motion of the missile was erratic from lift; thrust was terminated by radio at 19 seconds.
Data
No telemetry equipment was carried on this missile. The only sources of information on missile
behavior were photographs and visual observations.

Remarks
The type of gyrations executed by the missile could have been caused by breakage of one of the carbon
jet vanes. There area number of possible causes of this breakage such as inclusions, voids or other de-
fects within the vane itself. There is also the possibility that the vane was struck by one of the ignitors as
it was blown from the motor. On later missiles an elaborate series of tests and precautions were taken to
prevent vane breakage. The vanes were X-rayed and given a mechanical load test. The mounting threads
(tapped in the carbon) were given a torque test. Protection against damage by the ignitors was provided by
thick cardboard covers for the vanes. These measures apparently were effective since no other missile
acted in a way as to suggest vane breakage.

Probable Cause

A broken jet vane.

MISSILE 8

Performance
Missile operation appeared entirely normal up to 28.25 seconds. The velocity was about five percent
above the general average and the steering was good. At 22 seconds (time of the last good trajectory data)
the pitch program was developing normally and the east-west deviation was only 40 feet. At 28.25 seconds
there was an explosion which broke up the missile.

Data
Optical instruments provided the following information:

a. Missile velocity was about five percent above the general average

b. Pitch program was developing normally

c. East-west deviation at 22 seconds was 40 feet.

Telemetry data was somewhat questionable due to the loss of certain calibration equipment^). The
time of the explosion was recorded as 28.25 seconds.

Recovery after impact was particularly valuable. The bearing of the oxygen pump showed clearly that
it had overheated and seized.

Remarks
Seizure of the oxygen pump bearing could be expected to wreck the oxygen pump, the alcohol pump,
and the steam turbine. An explosion in the tail would certainly be expected to follow.

Probable Cause
Rust on the oxygen pump shaft, resulting in failure of the oxygen pump bearing.
MISSILE 10
Performance
The flight appeared normal up to 13.9 seconds. Steering and propulsion were both good. At 13.9
seconds the missile went into a spiral motion. The motor was cut off by radio at 20 seconds.

144
Data
Optical instruments provided the following information:

a. Missile velocity was about 10 percent above the general average.

b. At 13.9 seconds the deviation from the north-south line was about 40 feet west.

c. There was some north movement, indicating that the program was starting.

Telemetry results were good and showed that:

d. Vane 3 moved rapidly to its extreme position at 13.9 seconds.

Shortly after 13.9 seconds, vane

e. 1 moved in a direction to counter the roll-producing action of
vane 3 as described in d.

Remarks
Missile movements were those that v/ould be expected if vane 3 moved to an extreme position. The
action of vane 3 can best be explained by an open circuit in the control wiring to that servo. Output of the
computer is fed to the servo over a three-wire circuit. The middle wire is connected to the common point
of two control solenoids. The outer wires connect the opposite ends of the solenoids to the computer. The
solenoids are arranged to oppose each other. Under balanced conditions the solenoids receive equal
amounts of current, but no movement takes place because of their opposing action. If one of the outer
wires is opened, one solenoid is de-energized and the other causes the jet vane to move to an extreme
position. The action of the vane strongly suggests that such an open circuit occurred. Since the steering
was good for over 13 seconds, it seems probable that the open circuit was caused by vibration.

Probable Cause

An open circuit in one of the control wires connecting the computer to the servo.

MISSILE 11

Performance
Immediately after lift the missile turned to the east. After about four seconds it had reached an angle
of approximately 70 degrees from the vertical. At about this time the missile rolled so that fin 1 was up.
The rocket continued in fairly level flight about 300 feet above the ground with slight turning toward the
north. Radio cutoff was given at about 6.5 seconds.

Data

Telemetry records were good. Starting from the instant of lift, vanes 1 and 3 moved rapidly to a de-
flection of about seven degrees in such a direction as to turn the missile east.

Remarks
Calculations indicate that a seven-degree deflection of the vanes would produce movement in
reasonable agreement with that observed. It therefore became a question of what caused the vanes to
act improperly.

Prior to launching it had been necessary to adjust the vane-balance potentiometers twice to stop vane
drift. At the time, this was attributed to temperature changes. Experience in Germany had shown that
this might be expected if the missile stood for over one hour after oxygen loading (stand time in this case
had been 100 minutes). However, the subsequent action of the missile could not be explained by tempera-
ture change alone. Calculations indicated that the observed vane action could be expected if the original
cause of unbalance, which had made readjustment necessary, were removed at lift. One such possibility
would be the opening (due to vibration) of a circuit within the computer. It is not reasonable to believe
that the open occurred beyond the computer, because both vanes started to move in the same sense at the
same instant and their last common junction is within the computer.

145
Later experience indicated another possibility. The original computer was designed to accept guide-
beam command in yaw. This was a relatively high-gain circuit and it was necessary to short the input
terminals of the computer when guide-beam command was not used. There is no certain knowledge that
these terminals were shorted on this particular computer. If the short was not present, it is possible that
the yaw channel was being influenced by pickup from the ground control cables. This could explain the
need for readjustment of the vane balance. It would also explain the disappearance of the spurious signal
when the ground-control cables were dropped and the missile lifted.
Probable Cause

A spurious yaw signal resulting from induced potential in the guide-beam circuit of the computer.

MISSILE 14

Performance
Takeoff appeared normal and the missile flew normally for about two seconds. At this time it ap-
peared to go into a combined roll and yaw. After a few gyrations it headed south, flying approximately
level and having a periodic oscillation from south to west. Cutoff was given by radio at about 31 seconds.

Data
Telemetry records were satisfactory. From these records, it was found that during the first two
seconds of flight the rocket corrected two minor deviations and held on course in a normal manner. After
two seconds, a combined roll and yaw movement of the vanes took place, but no vane moved to zero or to
an extreme position. Vanes 2 and 4 remained synchronized.

Remarks
The fact that the missile corrected successfully for early deviations indicates that all polarities were
correct and that the gyros, the computer and the servos were all functioning normally at the start. Since
the telemetry record shows that the servos continued to operate, it appears reasonably certain that the
servos were receiving spurious signals in both roll and yaw. An attempt was made to find some circuit
in the computer that would duplicate the observed effects, however, this attempt was unsuccessful. It is
reasonably certain that the trouble originated in either the gyro or the computer, but there is no evidence
to suggest which was at fault.

Probable Cause

A failure in either the roll-yaw gyro or in the computer.

MISSILE 16
Performance
The missile tipped north rapidly during the first three seconds after lift. At the end of three
seconds the angle with the vertical was in the order of 15 degrees. This angle increased gradually
to approximately 21 degrees at the end of powered flight. Azimuth angle,
from the north-south line,
earlier
varied in an erratic manner but was less than two degrees at the end of burning. During the
part of the powered flight the velocity ran a few percent below the general average, but the
maximum
velocity was much higher than normal due to an unusually long burning time.

Data
The following data were obtained from the tracking instruments:

146
a. According to Doppler data the missile velocity was as follows:

TIME VELOCITY PERCENT

(seconds after lift) feet per second OF AVERAGES
10 325 -7.1
20 772 -3.5
30 1303 -3.6
40 1914 -2.2
50 2755 -1.6
60 3971 -0.7
69 5204 (max) *

* Only two missiles launched during this program exceeded 5204 fps.

b. East-west movement was erratic during the powered flight:

East-west movement
Built up graduallyfrom 3 fps at 3 seconds
to westward movement of 44 fps at 20 seconds
Dropped abruptly to 10 fps at 21 seconds
Continued gradual decrease to 1 fps at 31 seconds
Reversed direction at 31 seconds
Slight eastward movement to 38 seconds
Gradually increased (west) to 46 fps at 54 seconds
Averaged 41 fps to 62 seconds
Reversed direction abruptly at 63 seconds

Deviation from north-south line was 1058 feet west at

end of burning.

c. Photographs of the launching show that the missile began to tip north within one second after lift.
V/ithin two seconds, the angle with the vertical was approximately 10 degrees. Thereafter, the apparent
pitch angle (calculated from increments of range and altitude) ran as follows:

PERIOD AVERAGE PITCH ANGLE

degrees
4 to 8 seconds 19.2
8 to 12 seconds 19.0
12 to 16 seconds 19.8
16 to 20 seconds 19.3
20 to 24 seconds 17.2
24 to 28 seconds 17.8
28 to 32 seconds 17.2
32 to 36 seconds 18.5
36 to 40 seconds 19.0
40 to 44 seconds 18.9
44 to 48 seconds 18.8
48 to 52 seconds 18.5
52 to 56 seconds 19.2
56 to 60 seconds 19.8
60 to 64 seconds 20.4
64 to 68 seconds 21.2

Except for the first three seconds after lift, the telemetry records were good. The action of the
jet vanes during the first three seconds would have been of much interest, but the record was completely
lost during that period. Beyond three seconds, the data showed the following:

147
d. The turbine speed action appeared normal throughout the powered flight, with a steady-state value
of approximately 3800 rpm. From about 80 to 100 seconds, the turbine was turning slowly (about 100 rpm).

e. The combustion chamber pressure remained essentially constant at about 190 psi during the
powered flight. After burn-out the readings failed to drop in the proper manner. At 71.5 seconds there
was an abrupt drop to approximately 80 psi. Starting at 75.5 seconds the pressure reading showed a
gradual rise to about 125 psi. This value continued to at least 110 seconds,

All four vane positions were recorded and showed unusually large deflections. Vanes 2 and 4
f.

showed excellent synchronization throughout the record. Starting with a pitch-north deflection of five de-
grees at five seconds, they moved steadily to a pitch-south deflection of nine degrees at 22 seconds. Sub-
sequent action wasrecorded as follows:

Fairly steady decline to 3.0 degrees pitch south at 26.5 seconds

Fairly steady rise to 8.0 degrees pitch south at 31.0 seconds
Fairly steady decline to 3.0 degrees pitch south at 34.5 seconds
Fairly steady rise to 7.0 degrees pitch south at 39.0 seconds
Fairly steady decline to 5.0 degrees pitch north at 53.0 seconds

There was no additional change of any appreciable magnitude to the end of burning.

The function of vanes 1 and 3 was the control of both roll and yaw. The recorded positions of
these vanes have been resolved into equivalent roll and yaw position as follows:

constant at about 4 degrees (yaw east) to 15 seconds. A gradual movement

Yaw approximately
to about 7.6degrees (east) at 26 seconds. A return movement to about 3.0 degrees (east) at 42 seconds.
A gradual increase to 5.6 degrees (east) at 47 seconds. From this time to the end of the record, the vane
angle varied from 3.1 to 3.9 degrees (east).

Roll showed minor fluctuations around the zero position up to 8 seconds. From 8 to 15 seconds
the angular deflection of the vanes was approximately 1.5 degrees in such a direction as to produce clock-
wise roll. Starting at 15 seconds there was a gradual and fairly constant movement which continued to a
deflection of 14 degrees in such a direction as to produce CCW roll. From 28 to 44 seconds this deflection
averaged about 11 degrees. Starting at 44 seconds, the angle gradually decreased until the roll correction
reached approximately zero at about 62 seconds. Thereafter, there was no appreciable change in the posi-
tion of the vanes.

All vanes showed one peculiar change. Normally, the trace of vane position is not completely
smooth but shows ripples, or minor fluctuations. This was true of all four vanes up to 50.5 seconds. At
that time, the trace of all vanes suddenly became exceptionally smooth. After that time, additional vane
movements occurred but the trace remained unusually clean.
Remarks
Evidence indicates that the missile held a pitch angle of 19 ±2 degrees. The pitch-angle plot leaves
some question as to whether the normal pitch program of 5 degrees was superimposed on a fixed angle of
approximately 15 degrees. There is a suggestion that it did, but the data are not conclusive.

Much evidence points to the pitch gyro as the origin of the abnormal pitch angle. In general,
of the
the failure of any steering component, other than the gyro, would result in a continuing turn or in a purely
random motion. In this case the missile appears to have been steered along a more-or-less fixed angle.
In addition, vanes 2 and 4 were exceptionally well synchronized. This establishes, beyond reasonable
doubt, that: (1) all power sources, (2) a part of the computer and (3) all steering components beyond the
computer, were operative. If the preceding it-ems are eliminated, only the gyro, the command battery and
the input circuits of the computer remain.

The pitch gyro in this missile was of the Anschutz type. The pickoff potentiometer consisted of
two segments, each wound in an arc of approximately 175 degrees. However, the full arc was not used.
Each potentiometer was tapped at points 20 degrees (plus and minus) from its center position; the com-
mand potential was brought in at these taps. The most probable location for an open circuit in one of
these potentiometers would be near one of the taps.

148
If one of the pickoff potentiometers should open at any given angular location, the steering
system
would be satisfied only when the missile had assumed a corresponding angle. For example, assume that
one potentiometer opened at a point 10 degrees from its center. If the pickoff wipers were on center, one
half the full command potential would appear as a gyro output signal. The polarity of the signal would be
such as to drive the missile, and the wiper, in the direction of the open. The signal would not become
zero until the open was reached. If the wiper was driven across this open, a strong signal of reverse
polarity would appear; this would operate to drive the missile and wiper back toward the open. Thus,
the steering system would try to hold the missile at an angle corresponding to the angle of the open.

Assuming an open circuit in the pickoff potentiometer near the 20-degree tap, the above sequence
would appear to offer a reasonable explanation of the observed trajectory. Unfortunately, the telemetry
record of vane positions offers nothing to confirm this theory and (in some respects) suggests the presence
of some other type of fault. The discrepancies are most pronounced in the position of the pitch vanes. Ac-
cording to the data, the pitch vanes were in a pitch-south position from about 13 to 48 seconds, while the
missile was certainly pitched north. If this actually was the case, then these vanes were opposing some
other force. This point should not be given too much weight, however, because there is the possibility
that the records are in error as to polarity.

an extraneous force is the fact that the pitch vanes began to move toward
Another suggestion of
their zero position at about thesame time that "Q" (1/2 PV 2 ) started to decrease. It should also be noted
that the pitch vanes had an average displacement of approximately five degrees from 17 to 43 seconds.
During this 26-second period, the apparent missile pitch angle showed remarkably little change; starting
at about 19 degrees, dipping to about 17 degrees and returning to about 19 degrees. During the same
period the roll component of vanes 1 and 3 was unusually large, averaging about 10 degrees.

The action of the vanes suggests the possibility of unusual aerodynamic forces. It seems possible
that some structural distortion may have been produced by the rapid turn at the start of the flight.
Probable Cause

An open circuit, near the 20-degree tap, in the pickoff potentiometer of the pitch gyro.

MISSILE 18
Performance
Missile performance was satisfactory for about 38 seconds. Velocity was approximately 15 perceht
above the general average and steering was adequate. At 38 seconds the deviation from the target line
was about 200 feet west. The pitch program angle was 4.85 degrees compared to a desired value of 4.95
degrees. At about 40 seconds the missile started to roll. Since there was no danger of the missile leaving
the range, radio cutoff was not required. Burn-out occurred at about 60 seconds.

Data
Optics and doppler provided the following information:

a. Missile velocity was approximately 15 percent above the general average

b. Deviation from the target line was about 200 feet at 38 seconds

c. The pitch program was developing properly (4.85 degrees compared to 4.95 degrees desired at
38 seconds).

d. At about 40 seconds there was a pronounced change in the deviation rate and in the program angle.
e. Doppler data indicated that a roll started at about 40 seconds.
Telemetry records were adequate but their usefulness was impaired because of a lack of voltage
calibrations. It was necessary to fly a transmitter without calibrations due to technical difficulties with

the telemetry. Some useful information (noted below) was obtained on the basis of relative values.

f. Vanes 1 and 3 showed oscillation, with a period of approximately 0.6 second, from lift.

g. The movements of vanes 1 and 3 indicated correction for a combination of both roll and yaw.
149
h. At about 38 seconds, all four jet vanes moved toward their center positions. Vane 4 steadied
almost exactly at center while the others appeared to settle a few degrees off center. No appreciable
movement of vanes 1 and 3 took place until burning ended. Vanes 2 and 4 remained steady to about
50 seconds. From this time to the end of burning the vanes showed some activity.

i. Starting at 40 seconds, noise appeared periodically on the telemetry record.

Remarks
There is no suggestion of a malfunction in the propulsion system other than the fact that the thrust
was somewhat higher than normal. This was probably the result of poor performance of the air-pressure
regulator. The reduced burning time is in reasonable agreement with the observed increase in thrust.

It appears that the steering system was working steadily from lift to maintain roll-yaw control. All
evidence indicates that as long as the steering system was operative, adequate control was maintained. In
fact, pitch control appears to have been exceptionally good.

All four vanes started toward center at the same time indicating the failure of some component com-
mon to all vanes. Among such common components are: (1) inverter supplying the computer, (2) power
supply within the computer, (3) command battery and (4) power leads to the servos.

If any of the above are lost, the vanes are driven toward center by the jet.

As well as can be determined from the telemetry record, the change in conditions appears to have
taken place abruptly, possibly within 0.1 second. This tends to eliminate the command battery since a
decline in battery voltage would be expected to be more gradual. A complete loss of command voltage,
by open circuit, would be possible, but this seems unlikely since vanes 2 and 4 show appreciable motion
between 50 seconds and burn-out. The same objection applies to the loss of a common power lead to the
servos and to the complete loss of the inverter; it is difficult to conceive of a partial loss of an inverter.
Available information suggests an abrupt drop in all servo signals to a value which would allow the
jet to drive the vanes back to approximate center, but not a complete loss of signal. Under this assump-
tion, a very strong signal from the pitch gyro might produce the observed motions of vanes 2 and 4 after
50 seconds (section h of Data above). Vanes 1 and 3 would not be equally susceptible to gyro signals
because they were mechanically coupled to their respective air vajies and would therefore offer greater
resistance to movement away from center.

It is suggested that some fault within the computer, perhaps the partial loss of effectiveness of the
power tube, caused the internal power supply of the computer to drop to a critical value.
Probable Cause
Partial failure of the internal power supply of the computer.
MISSILE 20
Performance
For about 27 seconds missile performance was excellent. Velocity was approximately 15 percent
above the general average. Steering was good, with a deviation from the north-south line of only 133 feet
at 27 seconds. The pitch program was developing normally. At about 27 seconds the east-west motion
reversed direction and the pitch motion showed a disturbance. At 37.3 seconds the missile began to roll.
At the end of burning the roll rate had reached one rps. Combustion chamber pressure began to drop at
55.5 seconds and burning ended at about 58 seconds.

Data
Telemetry records were satisfactory and showed the following:
a. Aside from a momentary disturbance at lift, the jet vanes showed very little activity up to 20
seconds.

b. At 20 seconds vane 3 began to show an oscillation of increasing amplitude.

150
c. At 24.3 seconds vane 3 moved rapidly to its center position and remained there to the end of the
record.

d. From 20 to 24.3 seconds vane 1 acted as if it were working to correct a combination of roll and
yaw. When vane 3 ceased to function, vane 1 immediately went into an oscillating motion similar to that
previously seen on vane 3.

e. Vane 1 became relatively inactive from 30 to 33 seconds. At 33 seconds it started a slow move-
ment toward an end position. At 36.5 seconds the end position was reached and at 37.3 seconds a con-
tinuous roll started.
Remarks
It is clear that the final cause of trouble was the loss of jet vane 3. The subsequent action of the other
vane, and of the missile itself, were those that would be expected under the circumstances.

There are many different faults which would cause a vane to go to its zero position, but a few of these
can be eliminated. First, the roll-yaw gyro did not appear to be at fault because vane 1 reacted in a
proper manner when vane 3 was lost. Second, the command battery and the computer inverter were not
suspected since the other vanes continued to function normally. For the same reason, the power supply
within the computer was not suspected.

The following possibilities remained:

a. An open wire in the vane 3 circuits of the computer

b. Loss of the vane 3 electronic tube in the computer

c. An open circuit in the middle leg of the command circuit to the servo

d. Loss of drive power to the servo pump

e. Some type of mechanical failure within the servo

Of the above, the loss of the electronic tube seems most probable. Old German tubes were used;
these had shown a fairly high failure rate when subjected to tests on a vibration table. All telemetry
records indicate considerable vibration near the speed of sound. This failure occurred within two
seconds of the time at which the missile reached that speed.

Probable Cause

Failure of an electronic tube in the computer.

MISSILE 24
Performance
Missile performance was normal up to 57.5 seconds. Velocity was within five percent of the general
average and the steering appeared to be normal. Deviation from the north-south line was less than 100
feet at 40 seconds and there was no significant change of angle prior to 57 seconds. The pitch program
was fair. Starting at 57.5 seconds, the missile began to roll.
Data
The experimental equipment aboard this missile required a maximum number of telemetry channels
leaving only two channels for monitoring missile performance. These two channels were assigned to
turbine speed and combustion chamber pressure. There was no telemetry data on the steering system.

Remarks
In the absence of any information on the steering system, there is no way to eliminate any of the
possible sources of roll. Among the many possible sources were the following:
a. Fault in roll-yaw gyro

151
b. Fault in computer

c. Fault in servo

d. Failure of command battery

e. Failure of inverter supplying the computer

f. Fault in wiring anywhere in steering system

In general, the computers showed less reliability than the other devices listed above. For that reason
only it is considered the most probable source of the trouble.

Probable Cause
Computer failure.

MISSILE 26
Performance
Starting at lift, or very close, the missile flew a remarkably straight course approximately 40 de-
grees east of north. The pitch program was near normal and the velocity was very close to the general
average.

Data

The telemetry channels assigned to missile performance did not produce data due to some failure
within the telemetry transmitter. The only information available for analysis of the steering trouble con-
sisted of trajectory data obtained by optical and doppler means. The trajectory data appeared to establish
the following:

a. Missile velocity was normal

b. Pitch movement was near normal

c. The missile did not roll prior to burn-out

d. East movement was not proportional to the missile velocity

e. In general, the east movement held a fixed relation to the north movement.
Remarks
Since the pitch movement was approximately normal, it can be assumed that servos 2 and 4 were
operative. Since no roll occurred, it can be assumed that servos 1 and 3 were operative. The same
evidence would indicate that the command battery and the computer power supply were not at fault. The
fact that theyaw angle appeared to increase gradually from zero to about six degrees, over a period of
60 seconds, does not suggest the breaking of a wire. It is difficult to conceive of a fault in the computer
which would produce such a gradual increase in yaw angle. By elimination, the yaw portion of the roll-yaw
gyro is suggested as the most probable source of the faulty steering.
Probable Cause

A fault in the yaw portion of the roll-yaw gyro.

MISSILE 27
Performance
Throughout the powered flight, the missile velocity was approximately 11 percent above the general
average. Up to 48.4 seconds the azimuth steering was exceptionally good, the total deviation at that time
being only 90 feet (west). At 48.4 seconds there was an abrupt increase in the west movement; shortly
thereafter the missile began to roll. Since the roll prevented any further increase in the azimuth angle,
the missile was not cut off.

152
Data
Trajectory data, from optics and doppler, indicated the following:

a. The missile velocity was approximately 11 percent above average.

b. At 50 seconds the pitch angle of the trajectory was 5.2 degrees.

At 48.4 seconds there was an abrupt increase in west movement. For the preceding 10 seconds,
c.
the average west velocity was less than 2 fps. In one second this value increased to 17 fps. In 10 seconds
the west velocity had increased to 257 fps.

d. The start of roll is not precisely indicated. It appears to have been after 52 seconds but before
57.4 seconds.

Telemetry data indicated the following:

e. The turbine started to slow down at about 62 seconds.

f. showed no motion throughout the flight. This undoubtedly was

Jet vane channels the fault of the
telemetry system since the vanes must have been operating.

Remarks
In the absence of information regarding jet vane actions, any analysis must be based on trajectory in-
formation. If the west movement had developed at 30 or 40 seconds, it might have been possible to deter-

mine whether the movement was due to roll or yaw. Since the movement started within about three seconds
of the end of the program, a selection between roll and yaw is not clear. Since the west movement may
have been caused by a slow-starting roll, all elements of the steering system were subject to suspicion.
The only remaining basis for the choice of a particular element is the general reliability of the various
elements. On this basis it seems most likely that the fault originated in the mix-computer.

Probable Cause

Failure of tube in mix-computer.

MISSILE 29

Performance
Within the first two seconds after lift the missile showed movement to the east. During the first 13
seconds very little, if any, north movement was apparent. This resulted in a large azimuth angle east of
north. At about 13 seconds north movement started. This movement increased fairly rapidly; the azimuth
angle began to decrease. At about 31.5 seconds the missile reached the safety limit and was cut off by
radio.

Data
Trajectory data, obtained by optical means, indicated the following:

a. East motion occurred within the first two seconds.

The average eastward velocity was a fixed percentage of the missile's total velocity.
b. The ap-
parent yaw angle was about 2.5 degrees.

c. There was no north movement during the first 13 seconds of flight. After 13 seconds the north
movement was fairly close to that of a seven-degree program.

Telemetry data were available for turbine speed, jet vane position (four vanes) and missile position
with respect toits roll, yaw and pitch axes. These data indicated the following:

d. The missile did not roll during powered flight.

Immediately after lift, a yaw angle appeared and remained relatively constant
e. to 17 seconds.
From 17 seconds to the end of the powered flight a larger yaw angle is indicated.

153
f. The pitch angle showed no significant change during powered flight.

g. general, the vane movements were of small magnitude.

In Maximum departure from zero posi-
tion (for any vane) was about three degrees.

h. All vanes showed enough activity to indicate that they were operative.
i. There were definite indications that vanes 2 and 4 were synchronized.
j. All vanes showed increased activity near the speed of sound.

Remarks
Since there is evidence (section d. above) that the missile did not roll, from (b) above the missile
apparently maintained a reasonably constant yaw angle from lift (section b. above). This suggests that a
fixed bias appeared in the yaw steering system at lift. If such bias had been present prior to lift, it
would have been seen in the steering command to servos 1 and 3.
The fact that north movement did not develop as early as expected does not appear to have any connec-
tion with the yaw The lack of early north movement might be attributed to: (1) a slight delay in
trouble.
the start of the program, (2) a program of less than the expected seven-degree maximum or (3) temporary
cancellation of the north movement by the wind. The trajectory data was not smooth enough to allow a
definite selection among these alternatives. It appears, however, that the velocity and direction of the
wind could explain fully the temporary delay in north movement.

While there was no apparent connection between the delay in north movement and the yaw failure, it
should be noted that this delay did exaggerate the azimuth angle produced by a relatively small yaw angle.
In this respect it was a contributing factor in making cutoff necessary.

Since the missile appeared to be under control, although flying a fixed yaw angle, there is little cause
to suspect the command battery or the servos. The gyro or the computer might be suspected, but the fact
that there was no evidence of roll tends to weaken this suspicion. The computer, in particular, seems un-
likely since the same tubes are used for both roll and yaw control.

There is another possibility which appears to fit all the known conditions. The guide-beam part of the
yaw circuit was not being used. If there was pickup on this input, it could have been balanced out by the
vane balance potentiometers (an angle of 2.5 degrees was within the range of these pots). If the pickup
originated in the ground controls, it would disappear at lift, leaving a fixed bias in yaw. To satisfy
this bias, it would be necessary for the missile to fly a corresponding yaw angle.

Probable Cause

Pickup in the guide -beam input to the yaw control.

MISSILE 30
Performance
Steering in azimuth was satisfactory, with a deviation of 400 feet (west) at burn-out (62.5 seconds).
at burn-out was 3970 feet, which was about one half that expected. Propulsion was
Range above normal,
with missile velocity about 18 percent above the general average. After burn-out the missile moved
south at a low rate and west at a somewhat higher rate. Impact was approximately 3.25 miles west and
1.0 miles south of the launcher.

Data
Optics, doppler and telemetry all gave very useful information. Trajectory data, from optics and
doppler, indicated the following:

a. Missile velocity was about 18 percent above the general average.

b. The missile pitched about 16 degrees to the north immediately after lift, followed by a pitch south
of about 10 degrees.

154
c. After a few oscillations (resulting from the initial disturbance) the missile began to move north at
a rate corresponding to a normal seven-degree pitch program. This continued up to about 27 seconds.
d. At approximately 27 seconds the apparent pitch program began to decrease slowly (from about
three degrees at 27 seconds to about 2.2 degrees at 54 seconds.

e. At about 54 seconds the apparent pitch program began to decrease rapidly, reaching zero degrees
at about burn-out.

f. The north position of the missile remained approximately constant from burn-out to about 120
seconds. After that time the missile had a slow south movement to impact.

g. At burn-out the missile had moved 422 feet west at an average rate of about seven feet per
second. After burn-out the west velocity increased to approximately 60 fps at 120 seconds. From this
time the west velocity remained essentially constant to impact.

Telemetry data indicated the following:

h. There is some evidence that the "desensitizer" functioned during the first three seconds of flight
and that it was at least partially removed at approximately three seconds.

i. All four jet vanes were operating normally up to 27 seconds.

j. The pitch vanes were very closely synchronized up to 27 seconds.

k. The roll-yaw vanes began typical oscillations as the speed of sound was approached.
1. Jet vane 4 ceased to function at approximately 27 seconds. It moved to its zero position in about
0.3 second and remained there to the end of burning.

m. Jet vane 2 began a mild oscillation at 30 seconds. This disappeared at about 46 seconds. It is

clearly shown that the roll-yaw vanes were responding to movement of vane 2 during this period.
n. When burning ended, vanes 1, 3 and 4 moved to a position other than zero.

General information is as follows:

o. This missile was launched with a tilt of 1.5 degrees to the north.

p. A "desensitizer" was used to reduce the servo response to gyro signals during the first three
seconds of flight, after which it was supposed to be removed automatically from the circuit.

q. The center of gravity (missile not fueled) was 226 inches. This was several inches below that
normally desired. The normal lox load was reduced by 1000 pounds to increase the initial loaded center
of gravity.

Remarks
In view impact location, it would be reasonable to suspect that the program did not start or did
of the
not function properly. Trajectory data, however, clearly establishes that the program was developing
normally up to 27 seconds.
There was also the possibility that the "desensitizer" was not removed from the circuit as planned.
Telemetary data gave a fairly strong indication that it was at least partially removed. The subsequent
action of the vane tends to confirm this. This is particularly evident for the roll-yaw vanes (they were
able to prevent roll after vane 4 ceased to operate).
It is quite clear from the telemetry data that vane 4 ceased to operate at or near 27 seconds. Since
the performance had been normal up to that time, it is reasonable to assume that this failure was the
origin of the unusual trajectory.

The fact that a single pitch vane was unable to maintain a normal pitch program does not seem un-
reasonable. The pitch thrust for a given gyro signal would be reduced and the synchronizing circuit would
attempt to hold the active vane in the same position (center) as the inactive vane. Under these conditions
any sort of stability in pitch could be considered remarkable. It is also suprising that the steering system
was able to counteract the roll-producing effects of the single pitch vane.

155
It seems clear that the roll-yaw system continued to operate in a very effective manner and that the

unusual trajectory can be explained by the loss of vane 4. If this is true, all common components of the
steering system are above suspicion. The servo, wiring and one small part of the computer remained as
questionable.

There is very little among these alternatives. Perhaps the least likely is the servo.
basis for a choice
A mechanical servo would probably show up as a gradual decline in its activity. No such de-
fault in the
cline was noted. If the motor should stall due to mechanical binding, the load on the power battery would
be so great that there would be little probability that the steering system could remain effective for 35
seconds longer.
Ordinarily a computer fault would be selected in preference to a wiring fault on the basis of general
reliability. In this particular case the situation is modified by the fact that additional equipment (the
"desensitizer") was connected in the servo circuit. Although this device was tested thoroughly, its re-
liability in flight had not been established in the same degree as the normal V-2 components. In any
event, its presence introduced additional connections which increased the possibility of open circuits in
the command line to the servo.

Probable Cause
An open circuit in the command line to the servo.

MISSILE 32
Performance
Missile velocity was approximately 15 percent below the general average. The steering appeared to
be satisfactory, although, normally, little can be determined about the steering during the first ten seconds
of flight. At 10.7 seconds there was a fairly violent explosion in the tail. After the explosion, the missile
began to roll and to veer to the west. Thrust continued to about 24.7 seconds when a second explosion took
place.

Data

Optical records indicated the following:

a. Missile velocity was approximately 15 percent below the general average.

b. There was very little north-south or east-west movement prior to the first explosion.

c. Photos show clearly a fairly violent explosion in the tail at 10.7 seconds.

d. The jet continued strong after the explosion.

e. There was no photographic evidence of any abnormal condition prior to the explosion.

f. A second explosion occurred at about 24.7 seconds at which time the jet stopped abruptly.

Telemetry offered no useful information because of a fault within the telemetry system which pro-
duced excessive jitter.
Recovery provided the following information:

g. Various tail hatch covers were torn off leaving many of the holding screws still in place.

h. The burner was recovered one piece although one side was mashed flat. The burner was cut
in
open for inspection but no evidence of trouble was found either inside or outside. The nuts which secured
the alcohol and oxygen pipes to the burner were all still securely in place.

i. There was no evidence of appreciable burning within the tail. Many pieces of tail skin were re-
covered but they did not show signs of burning. Many tail cables remained clean.

156
Remarks
In the absence of useful telemetry data, there is very little that can be learned concerning the cause of
the explosion. Good telemetry data would have indicated whether there was any connection between the low
thrust and the explosion.

Certainly, there is the possibility that the cause of low thrust was also the cause of the explosion. One
objection to this theory is the time element. An alcohol leak of sufficient volume to produce the observed
reduction in thrust would be expected to produce an explosive concentration in the tail in less than 10 sec-
onds (trajectory data indicated that the thrust was low from lift or shortly thereafter).

One possible explanation for a delayed explosion could be the absence of a source of ignition during
the first few seconds. On all missiles elaborate precautions were taken to eliminate all potential sources
of ignition. Assuming that there was no source within the tail, ignition might have resulted later if there
was a flow of alcohol from the tail to the jet.
It is highly probable that the explosion resulted from alcohol vapor. The only other source of such a
violent explosion would be hydrogen peroxide. In this particular case it seems unlikely that there was any
appreciable break in the peroxide system since the turbine continued to run for approximately 14 seconds
after the first explosion.

Probable Cause

A break of appreciable size in the alcohol system.

MISSILE 37
Performance
The missile velocity was about 18 percent below the general average. Steering was good with the
program very close to the intended value and a deviation of about 250 feet west at burn-out. Burn-out
occurred prematurely at 57.5 seconds.

Data
Trajectory data, obtained from optics and doppler, showed the following:

a. Missile velocity was about 18 percent below the general average.

b. The pitch program was slightly above the intended seven degrees at 52 seconds.
c. West velocity averaged about four fps with a total deviation at burn-out of 250 feet.

Telemetry data showed the following:

d. Low-air pressure was 395 psi (about 17 percent low).
e. Combustion chamber pressure was 175 psi (about 17 percent low).

f. Turbine speed was 3550 rpm (about nine percent low). The turbine started to overspeed at 56.4
seconds and reached its peak of 5370 rpm at 57.3 seconds. There is telemetry evidence that the missile
was cut off at about this time. This was to be expected since the overspeed trip was tested for 5150 rpm.
Remarks
The low air pressure, low turbine speed, low combustion pressure and low missile velocity were
probably the result of a poor pressure regulator. This regulator (in general) proved to be inconsistent
and unreliable in its operation.

The turbine overspeed, resulting in missile cutoff, was almost certainly caused by unloading of one
or both of the pumps. Tests have proved that the main oxygen and alcohol valves cannot be closed fully
against pump pressure and that such partial closure will not result in any appreciable increase in turbine
speed. This leaves three major possibilities for unloading the pump or pumps: a large rupture in either
the alcohol or oxygen system, the exhausting of alcohol or oxygen or the closing of the alcohol preliminary
valve.

157
rupture in the propellant piping system cannot be dismissed as highly
improbable. The piping sys-
A
to severe vibration and to extreme temperature changes. It
would be expected, how-
tem was subjected
an explosion. Since two telescopes
ever that such a rupture would be followed immediately by a fire or
the probability of piping trouble
tracked the missile for 170 seconds and reported no such occurrence,
was diminished.
There is little probability that the alcohol supply was exhausted. The quantity loaded was checked by
two independent means; a meter on the alcohol pump and float switches in the alcohol tank. In addition,
the color of the explosion at impact suggested the presence of considerable alcohol; also there were strong
alcohol fumes at the crater the following day.

The exhaustion of oxygen might be suspected because the missile remained at the launching site for
three hours and ten minutes after oxygen was topped. Simple calculations show this to be highly im-
probable. From a number of topping operations, oxygen boil-off rates have been established to an ac-
curacy of about 95 percent. The established rate is 6.5 pounds -per -minute for the first hour and 5.7
pounds per minute thereafter. Initial oxygen loading was completed at 9:35 PM and topping at 12:15 AM.
Therefore, the boil-off after topping should be based on a rate of 5.7 pounds per minute; the loss for 190
minutes should be about 1080 pounds. For a typical launching the loss would be about 560 pounds and the
burning time would be 65 seconds or more. The loss for this missile exceeded the typical figure by 520
pounds, which would equal 3.4 seconds of burning time at the normal rate. Thus under normal conditions,
less than half of the lost burning time is explained. In this particular case the propellant consumption
rate was at least 10 percent below normal. This increases the expected burning time to at least 71 sec-
onds and thereby widens the discrepancy.

There is the possibility that one propellant was exhausted prematurely because of some error in the
selection of the orifice which controlled the mixing ratio. This is highly improbable. To account for the
lost burning time in this way would require a very large error in mixing ratio. The alcohol orifice was
125 mm diameter and the oxygen orifice 92 mm. Calculations show that (for this particular missile) the
mixing ratio would be changed only 1.6 percent if the two orifices were reversed. This would not be
true for all missiles. For some missiles, a reversal of this kind could result in a change of perhaps 15
percent in mixing ratio. But in this particular case the characteristics of the propulsion system were
such that a large change in orifice dimensions would have little effect on the mixing ratio.

The most probable cause of the turbine overspeed is premature closure of the alcohol preliminary
valve. doubly susceptible to accidental closure because the loss of either air pressure or
This valve is
electrical power will cause it to close. The circuit to hold this valve open includes one relay contact,
one connector and a number of wiring junctions plus one auxiliary control valve. The opening, or failure,
might be anywhere within this system. There is no data to aid in locating the source more closely.

Probable Cause

An open in the wiring of the control circuit for the alcohol preliminary valve.

MISSILE 38
Performance
Between 13 and 29 seconds the missile was observed to roll 40 to 50 degrees counterclockwise and
back to normal four times. At approximately 29 seconds a continuous roll in a clockwise direction
started. At the moment the continuous roll started, the pitch program was about 10.5 degrees (three
times normal). After roll started, the missile settled on a course about 20 degrees west of north.
Missile velocity was close to the general average. At about 57 seconds the missile was cut off by radio.

Data

Telescopic data indicated the following:

At about 13 seconds the missile rolled about 40 degrees CCW and then rolled back to its normal
a.
position. Similar rolls started at about 16.5, 20.5 and 24.0 seconds. The magnitude of the roll appeared
to increase slightly with each cycle. After the return from the last cycle, the missile held its normal posi-
tion for about one second and then started a continuous CW roll. The first full roll was completed in about
eight seconds.

158
Trajectory data, from optics and doppler, indicated the following:

b. Missile velocity was close to the general average.

c.The pitch program showed a steady increase at about three times the normal rate. At 28 seconds
the pitch angle was approximately 10.5 degrees.

d. From 18 to 28 seconds the west velocity was nearly constant at about 30 feet per second. Total
deviation at 28 seconds was about 475 feet (west). At about 29.5 seconds the west velocity started to in-
crease rapidly.

Telemetry data indicated the following:

e. Vanes 1 and 3 showed no movement from their center position.

Vanes 2 and 4 showed a most unexpected action. For approximately five seconds they showed no
f.

movement from their center position. From five to 12 seconds there were possible minor excursions
from center. From 12 to 27 seconds, the pitch vanes showed a rather consistent cycle of about five sec-
onds duration. During one half of this cycle, vane 4 remained near center while vane 2 moved about five
degrees in such a direction as to produce CW roll. During the other half-cycle vane 2 remained near
center while vane 4 moved about five degrees in a direction to produce CCW roll. The complete cycle had
a remarkably consistent period of about four seconds. In general, a given vane moved from center in one
direction only. The only appreciable exception occurred a fraction of a second before continuous roll
started. At that time, vane 4 showed about three degrees of movement to the opposite side of center.
There were brief instances when both vanes appeared to be synchronized in pitch-producing movements.
It should be noted that the vane-position channels were commutated and information on vane position was

available for only one third of each second.

g. The missile carried a roll-measuring gyro which was telemetered. The total deflection of the
recording pen was 3/8 inch for 360 degrees of roll. Consequently the accuracy of roll readings was
limited. Within this limitation, the telemetry record confirms the telescopic data on roll as given in a.
above. There is a suggestion on the telemetry record that one or two smaller rolls occurred prior to
the first observed by the telescopes, but this is subject to question.

Remarks
The telemetry showed a constant reading of approximately 2.5 volts for vane 1 and vane 3 for the
duration of the record. It is hard to conceive of a fault in the telemetry which would produce this type of
record. It is equally difficult to believe that both position potentiometers became disengaged simul-
taneously after lift. It is therefore accepted that neither vane 1 nor vane 3 moved.

Two independent means have established that the missile rolled CCW and returned to its normal posi-
tion at least four times. Without roll control, the missile would be expected to develop a continuous roll
in one direction because of imperfect positioning of the fins. But to return. to normal position, after having
rolled away, requires jet vane action. Since vanes 1 and 3 were inactive, the roll force in at least one di-
rection must have come from vanes 2 and 4.

If vanes 2 one servo must be much faster than the other or the synchronizing cir-
and 4 produce roll,
cuit must be abnormal in some
Servo trouble is almost out of reason in this particular case.
respect.
Test records on the two servos show that their speed was reasonably close. Of more importance, the roll
period was entirely too long to be explained by differences in servo speed. Further, the telemetry record
of vane movements does not support this possibility.

The symmetry of the vane movements, as shown by the telemetry, strongly suggests that the synchro-
nizing circuitwas operative, although the manner of operation was abnormal. Also, there were brief
periods when the vanes appeared to be synchronized in a normal manner. This directed suspicion to the
synchronizing potentiometers which were attached to the vanes.

An open circuit near the center of one potentiometer would explain many of the recorded actions.
This, together with the previous assumption of vanes 1 and 3 being inactive, would appear to satisfy all
the available data.

159
To explain
the apparent fact that each vane usually moved on only one side of center, it is necessary
open a few degrees to one side of center of the potentiometer. The maximum movement of the
to locate the
vanes suggests that this be in the order of five degrees from center.

It seems probable that the open occurred shortly after lift. Otherwise, the strong synchronizing com-
mands should have been noted in test and on the steering desk. Since the record shows that the vanes re-
mained at center for approximately five seconds, it is suggested that the open occurred at the time the
program started.
The following explanation of the missile's action is offered: When the synchronizing potentiometer
(sync pot) opened, it appeared to the sync circuit that the particular vanes had moved to an extreme posi-
tion; it is assumed that the pot on vane 4 opened. The sync circuit started to move vane 4 away from
center toward the open. At the same time, the sync circuit started to move vane 2 away from center in
the opposite sense. During this time the pitch program was applying a signal to the vanes. The pitch
program added to the movement of vane 4 while it opposed that of vane 2. This tended to drive vane 4
faster than vane 2. Thus, vane 2 remained near center while vane 4 moved in a direction to produce
CCW roll. This movement continued until vane 4 reached the open and thus removed the sync signal.
If the pitch pickoff were satisfied with the missile angle, all servo signals would disappear and the jet

would drive the vanes back toward center. When the sync pot wiper again made contact with the pot, the
above cycle would start again. This cycle would repeat until further pitch-north signal was received.

When
further pitch-north signal was received, the sync pot wiper would be driven across the open in
the pot and make contact to reverse the sync signal. This would cause reverse motion of both vanes but
by then the program and sync signals were opposed for vane 4 but added for vane 2. Thus, vane 2 would
move faster than 4 producing CW roll.
should be noted that when the missile roll (CCW) reached approximately 30 degrees, the roll gimbal
It

of the pitch gyro came up against a stop which caused the pitch gyro to precess in a direction to increase
the pitch-north signal. This in turn increased the vane action to produce CW roll and return the missile
to its normal roll position. The excessive pitch program can be explained also in this manner.

When the CW roll had pulled the gyro gimbal away from the stop, the precession ceased and the CW
force decreased. Apparently there was a type of balance between the normal pitch action and the inherent
roll tendency of the missile. This seemed to operate to bring the CW force to zero or the missile stability
caused the vanes and the missile to reach their center positions at about the same time. Thus, conditions
were set up for the start of a new cycle.
The question remains as to why the missile finally took off on a continuous CW roll. This could have
been caused by some disturbance in pitch (perhaps a wind gust) which gave a strong pitch-north signal
just before the missile started a new CCW roll. It should be noted that the final roll started in the speed
of -sound zone where disturbances are to be expected.

The cause of the inactivity of vanes 1 and 3 remains in question. In view of the activity of vanes 2
and 4, it seems certain that no common component of the steering system was at fault. It seems equally
improbable that the fault was in the two servos or their individual components, since this would imply
simultaneous and like failures in the time between final test and lift. This points to the probability of a
fault in equipment or wiring which was common to the roll-and-yaw system only. These would include
the roll-yaw gyro, a limited number of components of the mix-computer and the associated wiring. The
computer seems improbable since it has no tube common to both servos and very few parts of any sort
that are common to servos 1 and 3 only. Normally, the gyro would be among the last devices to suspect
because of their remarkable record of reliability. In this case, they are even less liable to suspicion
because both roll and yaw are involved. It is difficult to conceive of a mechanical fault which would make
both axes inactive and not produce some kind of spurious signals. From the above, it seems most probable
that there was an open circuit in the wiring associated with the computer or the gyro.

Probable Cause
An open circuit in the wiring associated with the computer, or the roll-yaw gyro, plus an open in the
synchronizing potentiometer for vane 4.

160
MISSILE 39
Performance
Preliminary stage appeared normal but main stage thrust failed to develop in the usual manner. The
missile lifted after about six seconds of partial main stage. During the first few seconds acceleration was
low but it increased smoothly to near normal at about 17 seconds. For the following six seconds the
missile performance appeared normal. At 23 seconds the thrust started to decrease and was very low by
27 seconds. Steering was remarkably good during the powered flight with no evidence of instability until
the start of the final decrease in thrust.

Data
Trajectory data, obtained from optics, indicated the following:

a. Missile velocity increased slowly and smoothly to a maximum of about 710 fps at 25 seconds.

b. Pitch program was present but was developing at nearly double the normal rate.

c. Between 6.6 and 23.6 seconds there was east movement at a relatively constant rate of about
4 fps.

Telemetry data showed the following:

d. Air Pressure

Started up at X - 8 seconds
Rose to 125 psi in approximately 0.5 second
Held level for about 0.5 second
Rose to 150 psi at X - 4 seconds
Rose on smooth curve to 460 psi at 17 seconds
Remained approximately constant at 460 psi to 23 seconds
Broke sharply at 23 seconds
Reached 340 psi at 25 seconds
Smooth curve to 25 psi at 65 seconds
e. Turbine Speed

Started up at X - 6.5 seconds

Rose slowly on smooth curve to 3900 rpm at 17 seconds
Held essentially constant to 23 seconds
Broke sharply at 23 seconds
Reached 3000 rpm at 24 seconds
Reached 1100 rpm at 38 seconds
Held 1100 rpm to 50 seconds
Broke to rpm at 57 seconds
f. Combustion Chamber Pressure
Started up at about X - 5 seconds
Rose slowly on smooth curve to 220 psi at 17 seconds
Held at approximately 220 psi to 23 seconds
Sharp break at 23 seconds
Reached 140 psi at 24 seconds
Smooth curve to psi at 36 seconds
Remarks
From the above data it is evident that the combustion chamber pressure increased in an abnormally
slow manner because of the slow buildup in air pressure. In considering the unusual behavior of the air
pressure, the following points should be noted.

a. It can be assumed that an adequate supply of high-pressure air was available at the start since
the low-pressure air held the correct value some 31 seconds later.

161
b. It can be assumed that the pressure regulator was not directly at fault. The regulator was re-
covered in operative condition and actually passed a normal test after recovery.

c. It seems reasonable to assume that the abrupt drop at 23 seconds was caused by the same trouble
that produced the slow buildup. This is supported by evidence that the cutoff relay (a9z) was not energized
until X plus 55 seconds.

It can be assumed that the bleeder valves were at least partially closed.
d. Tests have shown that if
both bleeder valves are fully open, with no flow from the peroxide or permanganate tanks, the air pressure
will hold at about 50 psi. In the case of this flight, the pressure built up to approximately 460 psi while
peroxide and permanganate were being delivered.

Among the possible causes of abnormal air pressure are the following:
Foreign material acting as a restriction in the air system ahead of the tanks: this might account
a.
for a slow buildup, but it is hard to visualize a complete stoppage as indicated by the rapid drop at 23
seconds.

b. Bleeders not completely closed: a small opening might account for a slow buildup but would not
cause a loss of peroxide. The steam generator would continue to operate until the air supply was ex-
hausted. This would imply a gradual decline in air pressure, not an abrupt drop.

c. Leakage due to a mechanical break in the pressurizing or bleeder system: same comments as
for b. above.
d. Leakage in peroxide tank or feed lines: this could account for a slow buildup. It would also in-
volve the loss of peroxide which could produce an abrupt drop of both air pressure and turbine speed when
the peroxide was exhausted. The most serious objection to this possibility is the fact that such a spray of
peroxide in the tail could be expected to produce visible evidence. If it is assumed that the tank had a nor-
mal load of 126 litres and that the flow rate was normal at 1.4 litres for 31 seconds, then 2.6 litres per
second must be sprayed into the tail to exhaust the peroxide at X plus 23 seconds. It would seem reason-
able to expect, at the very least, a cloud of steam from the tail, starting before the missile lifted.

e. Peroxide tank not fully loaded. It appears that this condition accounts for both slow build up and

abrupt end of thrust. The calibration curve for this steam plant shows that it required about 15 seconds to
reach normal air pressure when the tank contained 72 litres of peroxide. For this launching the peroxide
loading detail was made up largely of inexperienced men, so that the possibility of a personnel error
existed.

Probable Cause
Peroxide tank not fully loaded.

MISSILE 40
Performance
The missile velocity was very close to the general average, although the missile was 1142 pounds
over design weight. Steering was satisfactory but showed a west movement somewhat larger than normal.
The pitch program was good. Up to 45 seconds the propulsion system was performing about 5 percent
above normal. From 45 to 50 seconds there was a series of disturbances in the propulsion system. At
59.4 seconds the turbine started to overspeed and cut off the missile.

Data
Trajectory data, from optics and doppler, indicated the following:

a. Missile velocity was very close to the general average.

b. Pitch program was good

c. West
velocity built up gradually to about 48 fps at 33 seconds. This velocity remained fairly
constant to 45 seconds when propulsion disturbances started. From 46 to 51 seconds the velocity was to
the east. Between 51 seconds and cutoff the velocity was again to the west, building up to a value of about
90 fps.

162

i
d. Film from telescope 3 showed that the jet flame decreased to about 30 percent of its normal
length at intervals starting at about 45, 46, 47, 48, 49 and 50 seconds.

Telemetry data indicated the following:

e. There was no evidence of improper propulsion unit operation prior to 45 seconds. Approximate
values are as follows:

Combustion chamber pressure 235 psi - about nine percent high

Low-air pressure 500 psi - about four percent high
Turbine speed 4070 rpm- about seven percent high

f. At 45.2 seconds a disturbance, lasting about 0.2 second, appeared on combustion chamber pres-
sure. Almost simultaneously, disturbances appeared on low-air pressure and on turbine speed. It was
not possible to determine the sequence in which the disturbance showed on the various channels due to
reading errors and variations in end-organ response time. These disturbances were repeated on all three
channels at about 45.2, 46.0, 47.1, 48.0, 49.0 and 50.0 seconds and consisted of two to four oscillations, at
a frequency of 10 to 20 cps. The oscillations appeared to be centered about a value somewhat below the
preceding steady-state value.

g. There was no indication of abnormal performance from 50 to 56 seconds.

At about 56 seconds the combustion chamber pressure dropped sharply from 235 to 157
h. psi. It

remained at this level (with oscillation at about 16 cps) to 59.4 seconds when it started a decline to zero.
Zero was reached at about 60.6 seconds.
i. During the period from 56 to 59.4 seconds, the turbine speed dropped only 110 rpm. At 59.4
seconds the turbine speed started up, reaching a peak of about 5200 rpm at 60.2 seconds. It commenced
to drop rapidly shortly after 60.2 seconds.

j. From 68 to 206 seconds the turbine speed showed very unusual actions. The variations were too
numerous to describe in detail, but the most interesting features were: (1) it stayed at about 4000 rpm
from 127 to 195 seconds, (2) it reached a maximum of about 6000 rpm at about 180 seconds and (3) the
start of its final decline (at 181 seconds) corresponds closely with the peak of the trajectory.

Remarks
The tendency of a turbine to run after cutoff has been noted occasionally. In some cases this could be
attributed to leaky control valves. In this case the sharp decline at cutoff suggests that the 8- and 25-ton
valves seated properly. The only other path for peroxide would be through the pressurizing line, and its
associated check valve, to the permanganate tank. Steam generated there could flow through the generator
to the turbine. The check valve in this line has been known to stick open in test. This appears to be the
most probable reason for the erratic turbine action after cutoff.
Since telescopic data and telemetry data are in very close agreement, there is no cause to doubt that:
(1) the propulsion unit performed as described above and (2) the overspeed trip on the turbine operated to
cut off the missile.

Since the combustion pressure started to drop as the turbine speed started to rise, it is reasonable to
assume that one or both of the pumps lost load. In view of the high turbine speed and the high combustion
pressure, propellant exhaustion would be expected at about 60 seconds. The high-performance figures are
supported by the fact that the missile maintained average velocity although overloaded about 1140 pounds.

The disturbances between 45 and 50 seconds and the reduced chamber pressure after 56 seconds re-
main to be explained. It seems reasonably certain that this trouble was caused by a loose connection to the
auxiliary valve controlling the main alcohol valve.

Tests have shown that the alcohol valve can be closed part way against pump pressure if control air is
applied to its piston. A short movement of the alcohol valve opens a by-pass line around the alcohol pump.
Thus, the combustion chamber can be robbed of alcohol without a corresponding change in pump load.
Tests have shown that under these conditions, the pump speed will decrease only about 128 rpm or about
three percent. This corresponds remarkably well with the observed performance from 56 to 59.4 seconds,
when the combustion pressure dropped 33 percent while the turbine speed dropped only 2.7 percent.

163
It is probable that some connection to the auxiliary control valve for the main alcohol valve was loose

and was making intermittent contact to produce the disturbances noted from 45 to 50 seconds. The pre-
sumption is that this circuit opened up permanently at 56 seconds.
Probable Cause
An open circuit in the wiring to the auxiliary control valve for the main alcohol valve.
MISSILE 42
Performance
A casual inspection of the trajectory data would lead to the conclusion that the missile's performance
was near normal. Velocity was about 5 percent below the general average. The pitch program angle was
close to that desired. Azimuth steering to 22 seconds was exceptionally good with a total deviation at that
time of 25 feet (east). From 22 to 53 seconds the azimuth angle remained rather constant at about four
degrees east of north. Total deviation at 53 seconds was 400 feet. After 53 seconds the azimuth angle in-
creased rather rapidly but did not become large enough to require cutoff by radio. Cutoff was by time
switch in two steps. Reduced thrust occurred at 60.6 seconds and complete cutoff at 64.0 seconds. An
inspection of telemetry data showed that vane 3 went to its center position at about 22 seconds.

Data
Trajectory data, obtained from optics, indicated the following:

a. Missile velocity was about 5 percent below the general average.

b. Pitch program angle was close to normal.

c. Up azimuth steering was excellent, with a total deviation at that time of 25 feet.
to 22 seconds,
From 22 to 53 seconds the azimuth angle remained nearly constant at about four degrees east of north.
After 53 seconds the angle increased rapidly to approximately 20 degrees at burn-out.

d. Vanes 1 and 3 showed a moderate oscillation from lift. This appeared to be mostly roll with per-
haps a little yaw. Vanes 2 and 4 were nearly steady.
e. At 21.5 second vane 3 was active. At 22.2 seconds, the next commutation point, vane 3 had re-
turned to its center point where it remained for the rest of the record. Vane 1 continued to oscillate with
possibly a slight increase in amplitude.

f. Vanes 2 and 4 were synchronized and showed some activity throughout the powered flight. They
did not show any excessive or abrupt motions.
Remarks
It is that vane 1 was able to hold the missile in roll after the loss of vane 3.
somewhat remarkable
There however, a little cause to doubt that vane 3 was lost: (1) there was a trajectory disturbance at
is,
about the right time and (2) the telemetry channel showed a steady value of about 2.5 volts. If this value
had been or 5 volts, the telemetry would be suspected. But there are few ways in which the telemeter
can fail and still show a middle voltage. Such a voltage could appear if the position-measuring potenti-
ometer became disengaged from the vane. Recovery after impact showed the potentiometer still coupled
to the vane. Therefore, it is reasonable to assume that vane 3 was actually lost.

Since vane continued to operate in a very effective manner, it seems certain that the fault was not in
1
a component which was common to vanes 1 and 3. Thus, the gyro and the command battery are eliminated.
This leaves the mix-computer, the vane 3 servo and their associated wiring.

The servo was recovered after impact. The motor was broken away from the servo mechanism. All
windings of the motor measured properly. The commutator looked good and there was no evidence of
overheating. The brush holders were broken but this probably occurred at impact. The wire linkage be-
tween the control magnet and the control valve was broken but the nature of the break suggested damage
at impact. The control magnet windings showed proper resistance and insulation values and the magnet
action was good. The gear pump rotated with moderate freedom. In view of this inspection, it seemed
unlikely that the servo was at fault.

164
Probable Cause

A computer failure or an open circuit in the control wiring.

MISSILE 45
Performance
The missile velocity was consistently about 17 percent below the general average. Otherwise, per-
formance was good to 18 seconds. Up to this time the average east velocity was one fps. From 18 to 25
seconds the east velocity averaged seven fps. Starting at 25 seconds there was a sharp increase in east
velocity reading a maximum of about 105 fps at about 51 seconds. The missile was cut off by radio at
about 56 seconds.

Data
Trajectory data from Askanias indicated the following:

a. Missile velocity was consistently about 17 percent below the general average.

b.At 30 seconds the pitch program angle was very close to the desired value. After 30 seconds
the pitch angle continued to increase but at a rate below normal. At the end of the program period, the
pitch angle was 4.7 degrees compared to a nominal value of 7.0 degrees.

c. Up to 18seconds the east velocity averaged one fps. From 18 to 25 seconds the east velocity
averaged seven fps. Starting at 25 seconds, there was a sharp increase in east velocity which reached a
maximum of about 105 fps at about 51 seconds.

Telemetry data showed the following:

d. Turbine speed and combustion pressure were about normal; low-air pressure was a few percent
low.

e. Vane movements were small, up to about 18 seconds. Thereafter, some very fast movements
took place. The amplitude of these movements was large. Apparently both roll and yaw correction were
involved.

f. The telemetry commutator began to slow down at about 30 seconds. There was a gradual change
to about 75 percent of its original speed.

g. Telemetry signals ceased at 48 seconds.

165
Wind data taken at 0600 Mountain Standard Time.

ALTITUDE APPROXIMATE VELOCITY DIRECTION

TIME AFTER
LAUNCHING
feet above WSPG seconds fps degrees
9 270
1000 9 20 260
4470 18 -

5100 19 20 250
8440 24 67 270
11950 28 66 280
17070 32 67 280
19440 34 75 270
21270 36 92 270
23340 37 135 270
24550 38 154 270
25540 39 153 270
27340 40 111 260
29440 41 89 260
31240 42 69 260
33380 43 53 260
35250 44 44 250
43130 48 91 280
44660 49 91 280

Remarks
There is a lack of agreement between the missile velocity (by trajectory data) and the propulsion unit
measurements (by telemetry). The latter shows turbine speed and combustion pressure near normal,
although the missile velocity appears lower than the average. Some reduction in velocity would be ex-
pected because the missile was 737 pounds overweight. Also a part of the discrepancy can be explained
by probable errors in the telemetry readings.

The most probable cause of low thrust would be poor performance on the part of the pressure regu-
lator. These devices have proved inconsistent in their operation.
The fact that the commutator began to slow down at 30 seconds suggests that the voltage of the power
battery was dropping. This does not necessarily follow since mechanical binding may have developed in
the commutator. The low-voltage possibility is strengthened, however, by the fact that the pitch program
rate changed at about 30 seconds. The program device is known to be sensitive to low voltage and begins
to operate erratically at about 24 volts. On the other hand, the servo motors were also sensitive to low
voltage and they showed very fast movements up to 48 seconds (the end of the telemetry record).

It appears that a large part, if not all, of the east movement can be explained by the high winds above

5100 feet. It will be noted that there was very little deviation from the north line up to 18 seconds when
the missile encountered the high wind velocities. From that time on, the missile velocity very obviously
follows the wind velocity closely. There is, of course, a lag between changes in wind velocity and the
missile response, but this is to be expected. The down-wind movement implies that the missile had very
stiff control in yaw, otherwise it would tend to turn into the wind. This contradicts the low-voltage theory.

Probable Cause
Low thrust, due to poor regulator performance, plus high winds.

166
MISSILE 46
Performance
Missile velocity was about 5 percent below the general average. Movement of the missile to the north
and to the east was above normal but can be easily accounted for by wind forces. Burning ceased at about
26 seconds.

Data
Trajectory data, from optics, indicated the following:
a. The missile velocity ran about 5 percent below the general average.

Movement to the north was above normal. At

b. 25 seconds the apparent pitch angle was 5.5 degrees
as compared to a normal value of 3.1 degrees.
c. Movement to the east was above normal. The east velocity gradually built up to about 25 feet per
second at 25 seconds.

d. Telescopes reported that after burn-out, the rocket took about 1/4 revolution CW and then re-
versed, making about two turns CCW by the time it reached its peak.

Telemetry data indicated the following:

e. There was no evidence of any abnormal condition in the propulsion system for the first 25
seconds.

f. Low-air pressure, turbine speed and combustion pressure were near normal.
At 25.7 seconds the turbine speed started down in a typical cutoff decay.
g. The earlier record was
studied (frame by frame). and there was no sign of overspeed.

h. The pilot valves for the main valves were de-energized somewhere between 25.0 and 25.7 seconds
(commutated channel - no record available for 2/3 of each second).
i. Pressure appeared in the control chamber of both main valves between 25.0 and 25.7 seconds
(commutated channel).
j. The alcohol preliminary valve and its pilot (slh) were in normal flight conditions at 25 seconds.
By 25.6 seconds slh was de-energized.

k. After cutoff, vanes 2 3 and 4 remained in their center positions. Vane 1 moved to an extreme po-
sition and remained there.
Remarks
The slightly low missile velocity was probably the result of the pressure regulator action. This device
was not very consistent in operation and variations of 5 percent were common. No malfunction was indi-
cated.

The fact that the north and east velocities were greater than normal does not prove that the steering
system was at fault. The east movement in particular shows a very obvious relation to the wind velocity.
From lift to 25 seconds, the east component of the wind velocity averaged 31 fps. During the same period
the missiles' east velocity averaged about 10 fps, with a maximum of about 25 fps at 25 seconds.

There is little cause to suspect steering trouble. On the contrary, the down-wind drift indicates very
stiff control in yaw.

In relation to premature cutoff, an important point is the fact that the turbine did not overspeed. This,
together with the fact that all the main valves were operated to close at the same time the turbine speed
started down, constitutes strong evidence that the cutoff was initiated by electrical command rather than a
malfunction or abnormal condition of the main propulsion system. This could have been brought about by
the pickup of A9z or the dropout of A6x (see Backfire, Vol II for detailed identification of A9z and A6x).

167
There is evidence indicating that the cutoff was by A9z rather than A6x. Operation of A6x cuts off the
propulsion system but leaves the steering system operative. Operation of A9z cuts out both propulsion and
steering. If cutoff had been by A6x alone, vanes 1 and 3 would be expected to show considerable movement
since the missile rolled as described in d. This did not take place; therefore, it is probable that cutoff was
by A9z.
In this missile there were three sources for energizing A9z: (1) overspeed trip circuit, (2) radio cut-
off receiver and (3) ground cutoff relay A90z.

Since the turbine did not overspeed, cutoff from this source would come only through a mechanical or
electrical failure. With reference to the overspeed device itself, an extensive series of tests were made
to determine its susceptibility to vibration. A special test fixture was designed to allow both rotation and
vibration simultaneously and a high-speed camera recorded its operation. From these tests it was ap-
parent that the overspeed device was not sensitive to vibration. This does not, of course, eliminate the
possibility of a mechanical failure.

The ARW-37 type receiver had been subjected to thorough tests with reference to both vibra-
of cutoff
tion and interference. These tests indicated that this receiver was highly reliable. There was no evidence
suggesting that the receiver was the source of cutoff.

The contacts of ground cutoff relay, A90z, were fed through contacts of take-off relay A7y. If A7y
functioned properly, there should have been no voltage on the contacts of A90z during flight. This means
that both A7y and A90z would have had to close contacts simultaneously to energize A9z. These two relays
are of different types and the probability of simultaneous closure due to vibration is slight.

Probable Cause

The energizing of cutoff relay A9z by some unknown mechanical or electrical failure within the control
system.

MISSILE 50
Performance
Missile performance appeared entirely normal up to 43.4 seconds; missile velocity was within about
1 percent of the general average and the steering was good. The pitch program was very close to the de-
sired value and deviation from the north-south line was not great. At 43.4 seconds the combustion
chamber pressure dipped to about 69 percent of its previous value for approximately one second. A second
dip occurred at 48.4 seconds. At 51.7 seconds this pressure dropped to about 66 percent of its original
value and remained at the lower figure to the end of burning (62.5 seconds). The steering remained good
to burn-out and there was no need to cut off the motor by radio.

Data
Trajectory data, obtained from optical instruments, indicated the following:
a. From 10 to 43.4 seconds the missile velocity remained within about 1 percent of the general
average.

b. The pitch program was unusually good.

c. built up gradually to about 15 fps at 26 seconds and then gradually decreased to
West velocity
about three fps at the time of the first disturbance. The deviation at 43.4 seconds was 271 feet (west) and
at 62.6 seconds (burn-out) was 158 feet (west).

168
d. Telescopic photographs show the following:

SECONDS OBSERVATION OF JET

43.35 short for one second
48.50 fades for one second
49.5 to 51.7 detached by 1/2 missile length
51.8 grows very dim
58.3 back but weak
58.35 to 59.2 not visible on film
59.3 back but weak
62.3 very intense
62.55 no longer visible on film

Telemetry records were reasonably good and provided much valuable information. Channel 27 failed
to record for reasons unknown. This was particularly unfortunate since this channel monitored the control
commands to the two pilot valves controlling the two main propellant valves, the probable source of the
propulsion difficulty. Telemetry records showed the following:

e. The turbine speed reached a steady state value of about 3900 rpm at about one second after lift
and held this value without appreciable change to about 58 seconds. From 58 to 61.6 seconds the speed
oscillated slightly. The turbine started to overspeed at 61.6 seconds, reaching a maximum of about 5450
rpm at 62.15 seconds. The speed had dropped to zero by 65 seconds.
f. From lift to 34 seconds, the low-air pressure remained constant at about 474 psi, approximately
three percent above the intended value. Between 34 and 35 seconds, the pressure rose to about 506 psi
(approximately 10 percent high) and remained fairly steady at this value to the end of burning.

g. Combustion chamber pressure varied as follows:

TIME AFTER OBSERVATION PRESS!

LAUNCHING
seconds psi

3 reaches steady state of about 270

16 completes a gradual rise to 290
43.4 dip starts (minimum value) 200
44.5 back up to 290
48.4 dip starts (minimum value) 200
49.7 back up to 290
51.7 dips to 192
57.8 still at 192
57.85 severe oscillation starts
61.65 starts final drop
61.90 has completed drop to about

h. The main power supply voltage remained essentially constant at approximately 27 volts through-
out the powered flight.

169
i. Summary of major events
TIME AFTER OBSERVATION
LAUNCHING
seconds

3.0 Essentially steady-state conditions reached

43.4 CCP (combustion chamber pressure) dips 31 percent

(change in turbine speed barely perceptible)

43.5 Pressure appears on oxygen main valve (95 psi or more)

(was all right at 41.8 seconds which was last commutation)

44.2 CCP starts back up, fast oscillation

44.43 Pressure on oxygen main valve gone (below 95 psi)

(note: possibly earlier, this, was next commutation)

48.31 Pressure on alcohol main valve (170 psi or more)

(was below 170 psi at 48.30 seconds)

48.45 CCP starts dip of 31 percent (change in turbine speed

barely perceptible)

49.36 Pressure still on alcohol main valve (end of commutator

segment)
49.5 CCP starts up, severe bump in oxygen pump inlet pressure
50.1 Pressure off both main valves

50.4 Oxygen pump pressure at zero, remains at zero, oxygen

tank pressure satisfactory

51.7 CCP drops about 34 percent and remains at this value

51.93 Pressure on alcohol main valve

52.27 Oxygen tank pressure down to zero, remains down thereafter
57.85 CCP starts severe oscillation, turbine speed indicates
oscillation

58.50 Pressure is now on both main valves (pressure on alcohol

valve only at 57.8 seconds)

61.65 CCP starts final" drop, turbine speed starts to increase

61.90 CCP is essentially zero

62.15 Turbine speed reaches maximum of about 5450 rpm

65.00 Turbine speed is essentially zero

170
Remarks
The steady-state combustion chamber pressure was recorded as being approximately 32 percent
above the desired value. There is reason to believe that the actual value is in error. In the first place,
this is not in agreement with the low-air pressure and the turbine speed (in reasonable agreement with
each other, both being from 5 to 10 percent high). In the second place, the missile velocity was very
close to the general average, indicating that there was no large increase in chamber pressure or thrust.
In the third place, such a chamber pressure would have required a large increase in propellant flow rate,
resulting in very early exhaustion of at least one propellant.

Although the actual magnitude of the combustion chamber pressure is believed to be in error, the
record is very useful since it does show the changes which took place. There is considerable reason to
trust the relative values since they are confirmed by the telescopic records.

Important information is available from the record of channel 28 (used to record the presence or
absence of control pressure on the main propellant valves). If full control pressure is applied to the con-
trol chambers, the valves will move toward the closed position against pump pressure, but will not close
completely. Calibration stand tests with water instead of the usual propellants, were made to obtain in-
formation on the effects of applying control air to the main valves. Since the combustion chamber pres-
sure was simulated and the effects of oxygen evaporation were not present, the tests were not wholly
representative of flight conditions. It is believed, however, that, with certain corrections, they present
a fair approximation of the effects to be expected in flight. The tests indicate that the application of
control pressure to either valve individually, or to both valves simultaneously, will result in a change of
turbine speed of less than 5 percent. Under the same conditions, the drop in combustion pressure will
be somewhere between 20 and 50 percent.

From the summary of major events, part i. above, it will be noted that each drop in combustion pres-
sure was accompanied by the appearance of pressure on one, or both, of the main valves. It is conceivable
that the pressure on the oxygen main valve might have been produced by the leakage of lox past its seal.
There are, however, several reasons for discounting this possibility. One reason is that it would require
a tremendous leakage (a magnitude never experienced before) to build up substantial pressure against a
special vent of very large capacity. Second, if 43 seconds were required for a leak to build up a pres-
sure of 95 psi, it would hardly be expected to double or triple that pressure in one or two seconds: it is
reasonably certain that a few hundred psi would be required to produce the observed effect on combustion
chamber pressure.
At 48.31 seconds, pressure appeared on the alcohol main valve alone. This could hardly have been
caused by anything other than the operation of its associated pilot valve, S2h. This pilot valve, when de-
energized, applies air pressure to force the main valve toward the closed position. It is highly probable
that an intermittent contact in the circuit of the pilot valve was responsible for the partial closure of the
alcohol main valve, with a resulting drop in combustion pressure. The oxygen main valve is controlled
by a similar pilot valve, 03h, located on the same block with S2h. The control wires for both pilot valves
are brought out of the main distributor on the same plug (V). Likewise they are brought out of the sec-
ondary distributor on a common plug (29). If either of these plugs should be loose or defective, inter-
mittent contact might be expected in both valve circuits.

It is V or plug 29 was loose. Special measures were taken to insure against

unlikely that either plug
this type of trouble. Plug V was secured in place by a strip of rubber backed by a heavy steel strap.
Plug 29 was held in place by a spring-loaded latch, which was further secured by safety wire. The
probability of a defective plug is greater. This is particularly true in the case of plug V. Little trouble
was experienced with this type of plug as made in Germany, but the US -made copy showed a tendency to
break. The breaks followed a fixed pattern. Almost invariably the break ran through three adjacent
pins, such as 3-13-23 or 6-16-26, located parallel to the short dimension of the plug. The probability of
such a break is suggested by the fact that the control circuits for 03h and S2h used adjacent pins 9 and 19.
Probable Cause
Intermittent contact in the control circuits for pilot valves 03h and S2h (possibly a broken plug V in
main distributor).

171
MISSILE 52
P erformance
Missile performance appeared normal up to 6.0 seconds. The velocity was slightly above the general
average and the steering appeared to be satisfactory. Unless there is some serious steering trouble, it is
not possible to evaluate the steering performance during the first few seconds. It can be said, however, that
no steering difficulty was apparent. At 8.0 seconds there was a violent explosion in the tail. This caused
trouble in the steering system but showed very little effect on the propulsion system. Thrust continued to
approximately 22 seconds when the propulsion unit was shut down by radio.

Data
Optical equipment provided information as follows:

a. The missile velocity was about three percent above the general average.

b. As well as could be determined, the steering was satisfactory.

c. Excellent pictures of the explosion were obtained. A 96-frame-per-second film shows that the ex-
plosion started near the bottom of fin 3. One frame shows the missile in normal flight. The next frame
shows a bright spot near the bottom of fin 3. The next frame shows gas, vapor or smoke being blown from
a number of openings in the tail. Three frames later a large hole is visible in fin 3 at the point where the
bright spot appeared four frames earlier.

An excellent telemetry record was obtained but its usefulness was somewhat impaired by the fact that
were no calibration marks on the record. This made it
the calibrating device failed to function and there
impossible to obtain accurate values but approximate values were obtained. They showed that all monitored
quantities were very close to the desired values up to the instant of the explosion. No abnormal conditions
could be detected. An accurate value of turbine speed was obtained from a geared contact-making device
which was independent of the calibrating device. Turbine speed was recorded as 3918 rpm, within three
percent of the desired value.

Telemetry offered the following additional information:

d. There was no detectable change in turbine speed after the explosion.

e. The record shows combustion chamber pressure dropped at the instant of the explosion to
that the
about 76 percent of its previous value. This may have been the result of a damaged pressure gage, since the
outlet pressure of the alcohol and the oxygen pumps show no detectable change from their previous values.

Remarks
The two probable sources of a violent explosion in the tail are hydrogen peroxide and alcohol. The fact
that the steam plant continued to operate near capacity for 14 seconds after the explosion means that no large
amount of hydrogen peroxide could have been discharged into the tail. It is therefore highly probable that
the explosion was caused by an alcohol leak. Since the propulsion unit performance was normal, or slightly
above, it is unlikely that there was a large rupture in the alcohol system. It appears probable that there
was a relatively small alcohol leak which gradually built up an explosive concentration in the tail. There is
no further information to. aid in locating the point of the leak or the source of the ignition.

Pr obable Cause
A relatively small alcohol leak, resulting in an explosive concentration in the tail, ignited from some un-
known source.
MISSILE 54
P erformance
Preliminary stage was very good. When main stage was energized, the turbine started and thrust in-
creased but the missile did not lift. The appearance of the main-stage flame was not normal. An unsuccessful
attempt was made to cut off the motor. After about ten seconds of burning, the missile lifted. Acceleration
was very low. Fortunately the missile was stable. The powered flight continued up to 44 seconds when the
combustion chamber pressure started to drop reaching zero at about 50 seconds.

172
Data
Optical instruments provided information as follows:
o
Acceleration was low throughout the powered flight with a
a. maximum of about 16 ft per sec at about

43 seconds.

b. There was definite evidence that the pitch-north movement of the missile was increasing, indicating
that the pitch program was operative.
c. The missile was remarkably stable in yaw, with a reported maximum excursion of about one degree.

Telemetry records were good to about 56 seconds. Unfortunately, the commutator failed and valuable
information was lost Good records were obtained in alcohol pump outlet pressure, turbine speed, combus-
tion chamber pressure, alcohol flow, oxygen flow and vane positions.

Telemetry provided information as follows:

d. Alcohol pump outlet pressure started up at -8 seconds, reached 234 psi at -6 seconds and remained
constant to 50 seconds (normally this pressure should be about 300 psi).

e. Turbine speed (by counter, should be accurate) started up at -9 seconds, reached 3978 rpm at -5
seconds, and remained constant to 42 seconds, at which time it started to overspeed. From 45.2 to 53.35
seconds the value was 5886 rpm.
f. Combustion chamber pressure started up at -8 seconds, reached 112 psi at -6 seconds and remained
constant from -6 to 44 seconds. It started down at 44 seconds and reached zero at 50 seconds.

g. Alcohol flow was approximately 170 pounds per second (37.7 percent above the nominal value).

h. Oxygen flow had an average value of about 60 pounds per second (39.4 percent of the nominal value).

i. and 4 were synchronized and showed remarkably little motion during the flight. There was
Vanes 2
a gradual movement of these vanes away from their zero position. This amounted to about 10 degrees in 39
seconds.

Recovery was exceptionally good. The entire propulsion unit and much of the midsection were returned
virtually intact. Information from recovered equipment was as follows:

j. The main distributer was recovered in fair condition. A careful wire check indicated that the wiring
to relay A9z was correct and continuous.

k. The oxygen tank was recovered with the oxygen-flow float still intact on its guide wires.

I. The orifice in the oxygen line was rechecked and found to be 125 mm, which was the diameter of the
oxygen pipe.
m. The entire oxygen system was checked for any obstruction to flow and none was found. The search
was very complete, even to the extent of removing the oxygen rosettes from the motor head. They were ab-
solutely clean and open.

n. The oxygen pump was recovered in good condition and there was no evidence of anything wrong.

o. The heat exchanger check valve was recovered in good condition. After impact, a check showed
that it would open at 38 psi.
The main oxygen valve was recovered in excellent condition. It was tested before disassembly.
p.
The showed the valve would start to open at 80 psi and would be fully open at 1-00 psi (at room tempera-
test
ture). The valve was then taken apart and inspected; no abnormalities were found. The valve was reas-
sembled and placed in a cold-test fixture; lox was applied to the top for two hours. Under this condition,
the valve would start to open at 100 psi and would be completely open at 150 psi. Next, about 1/2 cup of water
was dumped on top of the valve before lox was applied. case the valve opened fully at about 200 psi.
In this
Next, moisture was allowed to condense on all surfaces of the cold valve before lox was applied. In this case
the valve opened at about 250 psi.

173
q. The switch battery was recovered in good condition. Both valves operated satisfactorily (electri-
cally and pneumatically).

Remarks
It is clearly established that the flow from the lox tank was only about 40 percent of its nominal value.
The telemetry record of lox flow is supported by the combustion chamber pressure and the alcohol flow.
Since the flow was measured in the tank, there is no cause to suspect that the low flow at the motor was due
to a break in the lox piping.

Assuming no break in the lox piping, the low flow must have been due to incorrect pump operation or to
some restriction in the lox system.
It has been established that: (1) the pump was rotating at a value slightly above normal and (2) there was
no visible defect in the pump. The only remaining probability of pump trouble would be cavitation. It is
probable that the pump would cavitate if the lox tank were not pressurized. It is known, however, that the
tank was pressurized prior to preliminary stage. This means that the tank would have had to lose pressure
at, or near, main stage. This could result from the loss of the ground air supply or from a malfunction of
its control valves. This is highly improbable, however, because in such a case improvement would be ex-
pected when the heat exchanger became operative.
Loss of lox tank pressure would result if the lox vent valve opened. The valve was observed to close
by two witnesses. Further, the time required to pressurize the tank originally was normal. It is therefore
established that the vent operated normally prior to main stage. There remains the possibility that some
malfunction in the main stage switching reapplied air to open the vent. This is improbable because, in such
a case, the vent should have been reclosed by spring pressure at lift. Thereafter improvement should have
been noted.
Pump cavitation could have been caused by a break at any point in the tank pressurizing system. The
result would be the same as the opening of the vent valve. This is one of the more probable causes of the
failure.

There were three major possibilities of a restriction in the lox system: (1) foreign material (2) incor-
rect lox orifice or (3) a partially opened lox valve. A careful search revealed no evidence of any foreign
material. A recheck after impact showed that the lox orifice was of the same diameter as the lox pipe.
This eliminated the possibility of an error in calibration calculations. In view of these points, the lox
valve remains as the only probable source of restriction.

The
lox valve could produce a restriction to normal flow in two ways: (1) it could fail to open fully or
(2) itcould be operated to partially reclose. Tests described above showed the valve to be in excellent
operating condition even after impact. The last two tests were made under extreme conditions that could
hardly be expected to exist at the time of launching. Failure to open fully could be considered as possible
but improbable.

A malfunction of the pilot valve for the main lox valve could apply air to the control chamber of the
main lox valve. This would tend to close the lox valve part way and would cause a large reduction in flow.
The pilot valve, a part of the switch battery, was recovered and tested. Even after impact it operated in
a satisfactory manner. This does not prove, however, that there was not a malfunction inthe electrical
circuit which controls that valve.

There is another factor which raises considerable doubt that the lox valve tried to close. In the original
V-2 system the control chamber of the lox valve was vented during the powered flight. Under certain cir-
cumstances the original vent was too small and a second vent (many times the capacity of the original) was
added. For reasons of timing the pilot valve for the second vent was not connected in parallel with the pilot
valve for the lox valve. Instead, it was in parallel with the pilot for the alcohol valve. Thus the added vent
should be open when the alcohol valve was open. Since there is ample evidence that the alcohol valve was
wide open, it is probable that the added vent was open. If the added vent was open, application of control
air to the lox valve would not be expected to produce appreciable movement of the lox valve.

In addition to the low lox flow, there were three failures, apparently electrical, to be explained: (1) the
telemetry commutator failed to operate, (2) the ground cut-off circuit failed to cut off motor operation and
(3) the over speed device failed to shut down the propulsion system.

174
The failure of the commutator may be unrelated to the other failures. There had been considerable
bearing trouble with these commutators and further trouble of this nature would not be surprising.

The ground cut-off failure also may be unrelated. After the launching, the fin cut-off plug was found
to be out of its holder. It was in place prior to launching and it is probable that it was blown out after the
missile lifted, but this is not a certainty.

As stated in section j. the wiring to cut-off relay A9z was re-checked after impact and appeared to be
in order. Some other failure must have taken place since A9z should have been energized when the turbine
speed exceeded 5200 rpm.
There is one possible explanation for all the observed failures. The various devices in the missile
showed considerable variation in the minimum voltage required for operation. It is conceivable that there
was a particular value of resistance between the power battery and the bus which lowered the bus voltage
to a critical point where some devices would operate and others would not. Such a voltage might be sufficient
to hold the alcohol side of the switch battery but not enough to hold open the lox valve and the added vent. It
would be assumed that the voltage was too low to run the commutator or to pick up A9z at overspeed. Low
battery voltage could produce the same result but a battery that had dropped that low could hardly be expected
to carry the missile load for an added 55 seconds. Although the resistance idea could explain all the observed
effects, it is difficult to accept because of the very narrow range of resistance which would satisfy the con-
ditions.

Probable Cause
Ah opening in the pressurizing system of the lox tank, resulting in the loss of tank pressure and cavita-
tion of the lox pump.

MISSILE 55
Performance
Preliminary stage developed normally. Main stage was energized and the plugs dropped. Shortly there-
after there was a
fairly violent explosion .in, or above, the control chamber. After an instant the
missile
toppled to the ground. At, or shortly after impact, there was a violent explosion followed
by several other
explosions of varying intensity.

Data

There was very good camera coverage. One high-speed camera, located 1000 feet to the east, provided
particularly useful information. A study of various films disclosed the following:

a. The missile definitely lifted, perhaps as much as six inches.

b. It was not possible todetermine whether the explosion occurred just before or just after lift. It is
clear, however, that it occurred within a few hundredths of a second of lift.
c. One 96-frame per second camera caught the very start of the explosion. In one frame the missile
is absolutely normal. In the next frame bright flashes are present. Two of these flashes are
certainly at
TNT locations. This frame also shows a bright streak where prima cord ran from the TNT to the
nose cone.
Remarks
There is the possibility of an alcohol explosion, but this seems highly improbable. First,
great care was
„.
taken (as always) to insure that no alcohol was spilled in the missile.
Second, an alcohol explosion would be
expected to produce visible damage to the midsection. Excellent photographs
show absolutely no distortion
of the midsection. Third, it would be necessary to assume
that the alcohol explosion set off the TNT since
photos clearly show that the TNT was detonated. Fourth, it would be a
most remarkable coincidence for an
alcohol explosion to occur at the precise instant that the protective short
was removed from the TNT

175
K it is assumed that the explosion was initiated by electrical means, there is a choice between induced
potential and the application of voltage by direct contact.

It is highly improbable that the explosion resulted from induced The short-circuiting wires
potential.
were the only wires of any appreciable length. These were made up
a separate cable. Its principal ex-
in
posure was a run of about 20 feet parallel to: (1) the 500 cycle potentiometer supply or (2) the 28 volt servo
supply. The former was an unshielded pair carrying about 0.14 amperes at 40 volts, 500 cycles. The latter
was an unshielded pair carrying the servo power at 28 volts d-c. Probably the worst surge on the 28-volt
line would come from an open in the servo feed, resulting in a drop from about 30 to amperes. Evidence
indicates that this did not happen. It does not seem probable that induction from either of the above sources
.

would set off a squib of five ohms requiring a minimum of 0.050 amperes. The normal squib requires 0.10
ampere so the 0.05 figure is conservative. In addition two such squibs are connected in parallel. The
probability of pick-up from some "high-frequency source seems small.

In considering direct contact, two possibilities exist. One is that movement of the rocket removed the
protective short circuit and allowed the squib to fire. The other is that movement of the rocket produced
a contact which caused the squib to fire.

In either of the above cases there is the remote possibility that rocket vibration disturbed the wiring in
such a manner as to produce direct contact. This would require a double fault that would place positive
polarity on one wire and negative on the other. This is too freakish to warrant much consideration.

There is always the possibility of the existence of some permanent-type connection which could apply
voltage to the squib circuit. Under this classification would come sneak circuits, wiring errors and per-
manent short-circuits between wires. The chief objection to this type of fault is the fact that it would have
to be timed perfectly. If such a connection did exist, it would have to be so located that it was effective for
only a short period during some specific switching sequence. Otherwise it would be detected during tests
conducted for that purpose. If the short-duration feature is accepted, then the time-of -occurrence is limited
to a fraction of a second before or after lift. This limitation does not remove the possibility of such a connec-
tion, but it certainly reduces the probability.

One other possibility has been considered. The protective short-circuit was completed through the pull-
away plug. This plug has butt-type contacts. The pins are individually spring loaded and have a follow-up of
about an eighth of an inch. If the missile had moved horizontally about 1/2 inch before it lifted 1/8 inch, the
short-circuiting pins could have touched certain energized contacts. These contacts were so arranged that
the missile would have to move in one specific direction (within narrow limits) to apply potential to the short-
ing pins. In a clean, fast lift the probability would be low. K, however, the missile bobbled about on the stand
a few times prior to final lift, the probability increases.

None of the above possibilities is particularly attractive and none is supported by additional evidence. In
such a case, the least complicated has the most appeal.

Probable Cause
The squib short-circuiting pins made contact with energized points on the ground plug as the missile thrust
fluctuated a bit prior to final lift.

MISSILE 57
Performance
Performance appeared normal up to 15.5 seconds. The missile velocity was about 10 percent below the
general average but was very close to the average for missiles of comparable weight and contour. Steering
was good, with a total deviation at 15.5 seconds of five feet. Trajectory data indicates that the pitch pro-
gram was developing. At 15.56 seconds there was a large explosion in the tail section and small pieces were
seen to leave the missile. This explosion appears to have damaged the steering system since the missile had
begun to tip south by 16 seconds. Although the jet showed signs of being disturbed, thrust continued up to about
18.5 seconds. The jet flared up at 18.25 seconds and ceased abruptly at 18.5 seconds. At that time there may
have been a second explosion in the tail. There was another explosion of considerable magnitude at 19.5
seconds, after which the missile was enveloped in flame.

176
Data
Trajectory data provided the following information:

a. The missile velocity was normal for a missile of its type, with a value of 501 fps at 15 seconds.

b. The pitch program was developing at about the normal rate.

c. The east-west deviation was small, five feet at 15 seconds.

Good telemetry records were obtained at all three ground stations. Unfortunately, their value was greatly
reduced by the fact that the ground reference (zero volts) channel was erratic throughout the entire flight.
For this reason the telemetry data is not considered reliable as far as actual values are concerned. The
records were useful, however, in that they indicated that all propulsion unit measurements were near nor-
mal and that no disturbances took place prior to the first explosion. The turbine speed was monitored by
a revolution counter in addition to the usual tachometer. Since the counter produced pips, rather than a
proportional voltage, an accurate determination of turbine speed could be obtained. This indicated that the
average speed was 3900 rpm compared to a desired value of 3910 rpm.
Remarks
Neither trajectory data nor telemetry data gave an indication of any abnormality in the propulsion system
prior to the explosion. Certainly there was no rupture of sufficient size to have any appreciable effect on
the thrust. It therefore appears that there must have been a small alcohol leak which eventually built up an
explosive concentration in the tail. There is no evidence to aid in determining the location of such a leak or
the source of ignition. The xx explosions and other events which followed the first explosion are to be ex-
pected under such circumstances and do not appear to offer any evidence concerning the origin of the explosion.

Probable Cause
A relatively small leak in the alcohol system, resulting in an explosive concentration in the tail, ignited
from some unknown source.
BUMPER 2

Performance
The Missile velocity was about 10 percent low compared to the general average for missiles of standard
V-2 contour. Steering was good up to about 25 seconds when some type of trajectory disturbance was indi-
cated. From 25 seconds to cut-off the steering appeared to be somewhat erratic but acceptable. The motor
was cut off by turbine over speed at about 33 seconds.
Data
Trajectory information, from Askania records, indicated the following:

a. Missile velocity was about 10 percent low compared to the general average for missiles of standard
V-2 contour.
b.Pitch program was developing in a normal manner to 24 seconds when the pitch angle started to de-
crease. The angle reached 0.3 degree at 26 seconds, after which it started a rapid increase toward normal.
At cut-off the angle was about 2.4 degrees.
c. Azimuth steering was good with an average east velocity of about 2.5 feet per second up to 25 seconds.
From then to cut-off there was little change in velocity.

Telemetry data indicated the following:

d. The steering system was working hard but the missile was under control.
e. Up to 28 seconds the propulsion system performance was completely normal with measured values
as follows:

Turbine speed 3840 rpm

Combustion Pressure 213 psi
Low-air pressure 473 psi

177
f. Combustion pressure and low-air pressure readings were lost at 28 seconds due to the failure of a
potentiometer.

g. The turbine started to overspeed at about 33 seconds. It reached a peak speed of 4800 rpm a few
tenths of a second later. The speed then decreased in the manner typical of overspeed trip.
Remarks
All available evidence indicates the loss of certain telemetry readings had no connection with the turbine
overspeed. There is much evidence to indicate that the telemetry failure was caused by an overloaded potenti-
ometer.
Four possible causes of turbine overspeed are suggested:
a. a large rupture in the propellant piping system,

b. the exhaustion of one or both propellants,

c. cavitation of the oxygen pump due to loss of pressure in the oxygen tank,
d. the closing of the preliminary alcohol valve.

It would be expected that a large rupture in the propellant piping would result immediately in an explosion,

a fire or at least a large cloud of vapor. Optical conditions were favorable and four telescopes had the mis-
sile in view at the time of cut-off. Since no evidence of a break was noted, that prospect seems poor.

Assurance that the alcohol tank was full at lift was checked by two independent means. The oxygen tank
was overflow occurred. The oxygen boil-off time was only 95 minutes. It is reasonably sure that
filled until
both tanks were nearly full at lift. It is hardly conceivable that one of the propellants could have been pumped
at twice the normal rate. It is therefore highly unlikely that propellant exhaustion took place at 33 seconds.

It is unlikely that the oxygen tank lost any appreciable pressure through the vent valve since this valve is

held closed by adequate spring pressure. This still leaves the possibility of a leak in the pressurizing piping
or the failure of the heat exchanger. Since this system consists mainly of simple, rugged piping, the proba-
bility of failure is relatively low.

A more probable cause of turbine overspeed is premature closure of the preliminary alcohol valve. Since
both air pressure and electrical power are required to keep this valve closed, it is one of the more vulnerable
components of the propulsion system. The control circuit for this valve includes one relay contact, one con-
nector, one pilot valve and a number of wiring junctions. The opening, or failure, might be anyplace within
this system. No data is available to aid in a closer determination of the source.

Probable Cause
Closure of the alcohol preliminary valve caused by a failure in the circuit controlling this valve.

BUMPER 4

Performance
The flight appeared normal in every respect to 28.5 seconds. Missile velocity was very close to the
general average and the steering was good. At 28.5 seconds it appeared that a tail explosion caused the jet
to broaden, the telemetry record to go bad, the beacon signal to disappear and the servo
signals to increase
to near maximum. These spurious signals drove the jet vanes hard over, causing the
missile to execute a
fast turning motion. Impact was at approximately 130 seconds.

Data
Trajectory data indicated the following:

a. Missile velocity was normal, sonic velocity was reached at about 25 seconds.

Steering was good up to 28.5 seconds. East deviation at that time was only 72 feet and the pitch
b.
program was developing normally. The telescope data indicated that the missile showed no sign of roll.

178
c. The jet flame showed a disturbance starting at 28.5 seconds. At that time the jet broadened until
it appeared wider than the fins.
d. In the first telescope picture showing a jet change, two bright specks are visible at, or near, the
junction of the tail and the midsection. These do not appear in the next frame (16 frames per second) al-
though the jet becomes longer and broader.

e. In the third frame the jet begins to decrease and appears normal by the tenth frame. The telescope
record also shows that the missile had reached the vertical by the tenth frame. In frame 17 there is clear
evidence that something is streaming from the forward part of the tail. By this time the rocket showed
considerable pitch south and the tail appeared to be burning. By frame 35 the missile is about horizontal.

Telemetry data showed the following:

f. Up to 28.5 seconds all monitored quantities gave good records with the exception of combustion
chamber pressure which ceased to give useful data shortly after lift. All values appeared absolutely nor-
mal up to 28.5 seconds.
g. The telemeter itself survived the disturbance at 28.5 seconds but it appeared that all pressure-
measuring end-organs were damaged by the disturbance. Signals continued to be recorded occasionally
from the more rugged end-organs such as pressure contacts, vane-position potentiometers and the turbine
speed tachometer, up to 30.25 seconds when all signals disappeared.

h. The turbine speed remained essentially constant to the end of the record.

At 28.5 seconds all four servo signals increased in 0.03 second

i. to near maximum values and the
jet vanes immediately started to move toward their extreme positions.

j. There was no evidence that any pressure contact operated, or that any solenoid contact operated
or that any solenoid valve was improperly energized.
Remarks
All the evidence points to a tail explosion. First, the appearance of trouble was sudden and occurred
simultaneously in a number of independent forms, such as the change in the jet, the loss of the beacon sig-
nal, the disturbance in the telemetry and the increase in the servo signals. All these suggest a severe
shock. The momentary appearance of bright spots at the junction of the midsection and the tail supports
the idea of a tail explosion. The photographic evidence of a liquid or gas flowing from the tail implies a
broken propellant line. The fact that the trouble appeared in the trans-sonic region suggests that a break
may have been caused by the vibration which is expected at that time. In view of the above, it appears
probable that there was a break in the alcohol piping. There is no data to aid in the location of the precise
point of the break or the source of ignition.

A break in the alcohol piping which resulted in an explosive mixture in the tail, ignited by some un-
known source.

BUMPER 6

Performance
Missile performance was normal up to 47.5 seconds. The velocity was very close to the general average.
Although there was no north movement until about 19 seconds, this can be explained readily by the fact that
there was an appreciable wind component to the south while the average pitch angle to the north during this
period was about 0.2 degree. Westward deviation reached a maximum of about 30 fps at about 44 seconds,
but this can also be explained by the west component of the wind which reached a maximum of about 90 fps
at approximately 40 seconds. At about 47.5 seconds some malfunction apparently shut down the propulsion
unit.

Data
Optical data indicated the following:

a. Missile velocity ran very close to the general average up to the point of shut off.

179
b. North movement was apparent at about 19 seconds. There is evidence that the pitch program was
developing.

c. West deviation reached a maximum of about 30 fps at about 44 seconds. Total deviation was about
220 feet at shut-down.

Telemetry records showed the following:

d. Prior to shut-down there was no evidence of any type of malfunction. All monitored quantities were
normal.
e. At 47.56 seconds, the auxiliary control valves (03h and S2h) for the main propellant valves were
energized.

f. At 47.66 seconds, pressure appeared in the control chambers of the main propellant valves.

At 47.75 seconds the turbine speed started to drop to zero. An examination of the record showed
g.
no overspeed prior to the drop.
Remarks
The action described in e, f and g. above could have been caused by the drop-out of relay A6x. The
telemetry record showed, however, that the inverters were de-energized at the same time that the propul-
sion unit was shut down. This indicates that at least one "back" contact of relay A9z opened at this time.
It is therefore reasonable to assume that the drop out of A6x was caused by the opening of another back contact

of A9z.

The opening of the A9z contacts (the cut-off relay) may have been caused by vibration. There is no direct
evidence to prove or disprove this possibility. This relay does, however, have a good record with respect
to vibration. Further, it gave no trouble during the trans-sonic period when greater vibration might be ex-
pected.

Aside from the possibility of vibration effects on A9z, there were four circuits connected to energize
A9z under certain circumstances: (1) turbine overspeed switch, (2) radio cut-off receiver, (3) ground cut-off
relay A90z and (4) a circuit concerned with experimental equipment. Since the turbine did not overspeed,
circuit (1) was not responsible unless some mechanical or electrical failure occurred. Extensive tests of
the overspeed device itself have indicated that it is not sensitive to vibration.

It was reported that the radio cut-off transmitter was not actuated. The cut-off receiver (ARW-37) has
demonstrated in tests that it is not sensitive to vibration. No other evidence points to the cut-off receiver.
The contacts of ground cut-off relay, A90z, are energized through contacts of take-off relay A7y. H A7y
functioned properly, there should have been no voltage on the contacts of A90z during flight. This means
that both A7y and A90z would have had to close contacts simultaneously to energize A9z. Since these two
relays are of entirely different types, the probability of simultaneous closure due to vibration appears to be
slight.

It is highly improbable that circuit (4) caused the trouble. There is clear telemetry information that it

was not at fault.

Probable Cause
The energizing of cut-off relay A9z by some unknown mechanical or electrical fault within the control
system.

BUMPER 7 AND 8

Performance
The experiments carried on these missiles called for a relatively low trajectory, with a separation angle
of approximately 20 degrees from the horizontal. This trajectory required a relatively rapid turn during the
powered flight of the V-2. Both missiles made the turn successfully and the general performance appeared
good. A closer examination of the trajectory data showed, however, that the program had been greater than
desired.

180
Data
Trajectory data showed the separation angle for Bumper 7 to be approximately 10 degrees and that for
Bumper 8 to be about 13 degrees.

Remarks
The fact that the two trajectories showed the same type of discrepancy indicated a systematic rather than
a random fault. Since it seemed highly improbable that the pitch device itself would fail in such a fashion as
to increase the program, precession of the pitch gyro was suspected. Since the pitch gyro circuits had been
modified to obtain a much larger than normal program, these circuits were among the first investigated.
This investigation showed up a "sneak circuit" which caused the erecting motors of the pitch gyro to be
energized after take-off. This in turn caused a precession which operated to increase the program angle.
This fault appeared to answer fully the observed discrepancy.
Probable Cause
A "sneak circuit" which caused the erecting motors of the pitch gyro to be energized after take-off.

MISSILE SPECIAL
Performance
missile performance was normal up to about 36 seconds. Up to the end of burning, the velocity
In general,
was very close up to 18 seconds the azimuth steering was exceptionally good with a
to the general average;
deflection of only 11 feet (west), beyond that time there was east movement but this was probably caused by
fairly high winds. Motor pressure started down at 36.6 seconds and reached zero at 37.8 seconds.

Data
Trajectory data indicated the following:

a. Missile velocity was very close to the general average during powered flight.

b. Azimuth steering was exceptionally good up to 18 seconds. From 18 to 28 seconds east movement
rose to about 37 fps. This movement remained fairly constant to the end of burning.

c. Pitch program was near normal at 18 seconds but did not show a normal increase thereafter.
Telemetry data indicated the following:
d. Motor pressure started down at 36 seconds and reached zero at 37.8 seconds.
e. The turbine started to overspeed at 36.6 seconds, reaching a maximum at 37.3 seconds. This maxi-
mum was approximately 130 percent of the average during normal operation. After remaining at the peak
for about 0.2 second, the turbine speed dropped off rapidly to zero.

f. The steering system may or may not have been cut off at the time the turbine speed started down.
The evidence is not conclusive.
Remarks
An important fact was that the combustion pressure started down before the turbine speed started to in-
crease. This immediately limits the fault to one which would unload the turbine. Among such possibilities
are: (1) large rupture in the alcohol as oxygen piping, (2) closure of the main alcohol valve, (3) closure of
the main oxygen valve and (4) closure of the preliminary alcohol valve.

It seems highly probable that a large rupture in the alcohol system would result in a fire or an explo-

sion that would be visible to the telescopes. A large rupture in the oxygen system probably would produce
visible evidence.

181
Tests have demonstrated that the main valves can not be closed fully against pump pressure. When air
is applied to close either or both of these valves, the combustion pressure should not drop over 50 percent
and the turbine speed should not change over five percent. Since the combustion pressure went to zero and
the turbine speed increased by about 30 percent, it seems clear that the main valves did not produce the
shut-down.

Closure of the alcohol preliminary valve will produce a complete loss of combustion pressure in a short
time. The evidence therefore points to this valve as the cause of the trouble. In addition, it should be noted
that this valve is more susceptible to accidental closure than are the main valves since both electrical power
and air are required to keep it open. A break in either the electrical system or the air system will cause
it to close.

The presumption is that the failure was caused by vibration. Under this assumption, a break in the
electrical wiring is more probable than a break in an air line. The electrical circuit included one relay
contact (a6x), one connector (k) and a number of junction points. Any one of these could be the location of
the break. There appears to be no data to allow a closer determination of the probable source of the trouble.

Probable Cause
A break in the control wiring to the alcohol preliminary valve causing this valve to close, thus cutting
off alcohol to the motor.

182
REFERENCES

1. General Electric Company - Project Hermes report DF-71369 "The Missile A-4, Series B," February
1, 1945
2. British Special Projectiles Operation Group, "Report on Operation Backfire," November 7, 1945
3. Ibid
4. Ibid, Volume IV, p 52
5. Ibid, Volume IV, p 57
6. Ibid, Volume IV, p 40
7. Ibid, Volume IV, p 65
8. Ibid, Volume IV, p 47
9. Ibid, Volume IV, p 39
10. Ibid, Volume II

11. Ibid, Volume II, p 72

12. Ibid, Volume II

13. Corrections are recorded in Archives 56.7, 57.16, and 57.22. In addition, Archive 57.19 contains the
references for combustion and injection pressure
14. British Special Projectiles Operation Group, "Report on Operation Backfire," Volume II, November 7,
1945 and General Electric Company - Project Hermes report 45770, "German A-4 Electric Hydraulic
Servo," Broome, J. W.
15. General Electric Company - Project Hermes report DF-78155, "Report on Servo Test - A-4 Rocket,"
Campbell, E. J. and Boynton, E. R.
16. British Special Projectiles Operation Group, "Report on Operation Backfire," Volume H, p 112,
November 7, 1945
17. General Electric Company - Project Hermes report DF-71395, "Telemetering Data, Rocket No. 4,
A-4 Serial B," Cunningham, H. A.

183
DISTRIBUTION

External distribution in accordance with parts A, B, and C of the Research and Development Board
Guided Missile Technical Information Distribution List, MML 200/1, List No. 1. In addition, copies dis-
tributed externally as follows:

Armed Services Technical Information Agency

U. B. Building, Dayton, Ohio (15 copies)

Commanding General, Redstone Arsenal, Huntsville, Alabama

Attn: Technical Library (5 copies)

Internal distribution as follows:

Dr. R. W. Porter, CAP-1, Schenectady, N. Y.

R. A. Duff, CAP-1, Schenectady, N. Y.
H. G. Kingham, CAP-1, Schenectady, N. Y.
C. C. Botkin, CAP-1, Schenectady, N. Y.
L. B. Cowles, CAP-1, Schenectady, N. Y.
G. R. Fawkes, CAP-1, Schenectady, N. Y.
E. I. Finger, Malta Test Station
R. M. Fritz, Washington, D. C.
Dr. C. F. Green, 28-408, Schenectady, N. Y.
J. R. Haeger, Huntsville, Alabama
Capt. A. E. Hansen, CAP-1, Schenectady, N. Y.
R. P. Haviland, CAP-1, Schenectady, N. Y ,

E. H. Hull, Malta Test Station

L. T. Huntington, 28-508, Schenectady, N. Y.
M. J. Jang, Los Angeles, California
Library, A & OS Division, Schenectady, N. Y.
Library, Guided Missiles Department, Schenectady, N. Y.
Dr. H. A. Liebhafsky, Research Laboratory, Schenectady, N. Y.
R. W. Mayer, CAP-1, Schenectady, N. Y.
L.J. Neelands, Syracuse, N. Y.
R. H. Norris, CAP-1, Schenectady, N. Y.
Lt. Col. F. F. Poppenburg, 23-238, Schenectady, N. Y.
A. W. Robinson, CAP-1, Schenectady, N. Y.
Major H. B. Sloan, CAP-1, Schenectady, N. Y.
T. S. Teague, Malta Test Station
H. E. Vigour, CAP-1, Schenectady, N. Y.
L. D. White, Las Cruces, New Mexico (6 copies)

184
3 9088 01477 6371