zotero/storage/KMJQ3SF5/.zotero-ft-cache

AFAL-TR-88-031 AD:
AD-A 197 537
Final Report 21st Century Propulsion
for the period
7 July 1987 toCo c p
7 January 1988 DTIC
AUG 1 9 18
April 1988 Author: Veritay Technology, Inc. R. L. Talley P.O. Box 305 4845 Millersport Highway East Amherst, NY 14051
F04-88-1 F04611-87-C-0058
Approved for Public Release
Distribution is unlimited. The AFAL Technical Services Office has reviewed this report, and it is
releasable to the National Technical Info-mation Service, where it will be available to the general public, including foreign nationals.
. Preparedfor the: Air Force
Astronautics
Laboratory
V
.Air Force Space Technology Center
"SpaceDivision, Air Force Systems Command
Edwards Air Force Base, California 93523-5000
• %" bb.d
•v" t- 'r


NOTICE
When U.S. Government drawings, specifications, or other data are used for
any purpose other than a definitely related Government procurement operation,
the fact that the Government may have formulated, furnished, or in any way
supplied the said drawings, specifications, ur other data, is not to be
regarded by implication or otherwise, or in any way licensing the holder or
any other person or corporation, or conveying any rights or permission to
manufacture, use, or sell any patented invention that may be related thereto.
FOREWORD
This final report was submitted by Veritay Technology, Inc., East
Amherst, NY on completion of Small Business Innovitive Research contract
F04611-37-C-0058 with the Air Force Astronautics Laboratory (AFAL), Edwards
AFB, CA. AFAL Project Manager was Dr Frank Mead.
This report has been reviewed and is approved for release and
distribution in accordance with the distribution statement on the cover and on
the DD Forn 1473.
FR LIN B. M'AD, JR. WILLIAM A. SU LL, CAPT, USAF
Project Manager Chief, Advanced Concepts Branch
FOR THE COMMANDER
ROBERT L. GEISLER
Deputy Chief, Astronautical Sciences
Division
0 . . .. . • .., . . •


UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE
REPORT DOCUMENTATION PAGE
Is. REPORT SECURITY CLASSIFICATION lb RESTRICTIVE MARKINGS UNCLASSIFIED 2a. SECURITY CLASSIFICATION AUTHORITY J. DISTRIBUTION/AVAILABILITY OF REPORr b_DECLASSIFICATION IDOWNGRADING SCHEDULE Approved for public release, distribution
AI is unlimited.
4 PERFORMING ORGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION REPORT NUMBER(S)
F04-88-1 AFAL-TR-68-U31
65 NAME OF PERFORMING ORGANIZATION 61b OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION
(If applicable) Air Force Astronautics
Veritay Technology, Inc. 7V715 Laboratory
k. ADDRESS (City, State, arid ZIP Code) 7b. ADDRESS (City. Statoe, and ZIP Code)
4845 Millersport Highway, PO Box 305 LKCT East Amherst, New york 14051 Edwards AFB, CA 93523-5000
.aN.AME OF FUNDING iSPONSORING 1b OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
ORGANIZATION (If applicable) F04611-87-C-0058
SC.ADDRFSS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS PROGRAM PROJECT TASK IWORK UNIT
ELEMENT NO. INO. NO. ACCESSION NO.
1_._TT ________________u_____C___.________n__. __ 65502F 3058 00 4M
11. TITLE (include Security Classificationi
21st Century Propulsion Concept (U)
12. PEISONAL AUTHOR(S)
Talley, Robert L.
13a TYPE OF REPORT 13b. rTMECOVE9 D 14. DATE OF REPORT (Year. Month. Day) rPSAG LUNT
iFinal I FROMI8 7 (/I/ TO 88/I/7 88/4 "
16. SUPPLEMENTARY NOTATION
"-'1L/ *'
17 COSATI CODES B8.SUBJECT TERMS (Continue on reverse if necessary and identif by block number)
FIELD GROUP SUB-GROUP Biefield-Brown Effect >Electrostatic Force Genera
22 01 •lectrostatic Field tion.'PQOJa5 'kcq
,,e N
-Propu 1sion) Advanced Propulsion Technique.
19.ABTAT(ontinue on reverse if necessary an identify by block numbewr) - ---------
--rhis Phase Iý SBIR contract-'*f fort was intended to explore the
Biefield-Brown effect, which allegedly converts electrostatic energy
into a propulsive force.
Activities under this program emphasized the experimental
exploration of this electrostatic thrust-generation concept to verify
its existence, to verify its operation in a vacuum, and to establish
*- the magnitude of its thl-ust.
To meet these goals an overall laboratory test configuration was
designed and developed for quantifying the electrostatically induced
'A•. propulsive forces on selected experimental devices. This
configuration utilized a vacuum chamber with a torsion fiber type
measurement system for direct assessment of propulsive forces.
20 DISTRIBUTION /AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION
0 UNCLASSIFIEDIUNLIMITED tR SAME AS RPT 0 DTIC USERS Unclassified 22. NAME OF RESPONSIBLE INDIVIDUAL 'gb TELEPHONE (include Area Code) 22c. OFFICE SYMBOL rFcnkLtn S. Mead, Jr. (S05)275-5540
DD FCRM 1473, 84 MAR 83 APR edition may be used until exhausted. SECURITY CLASSIFICATION OF THIS PAGE
All other editions are obsolete I


19. ABSTRACr (CONT.)
veometrical symmetries were incorporated in the design to minimize the influence of reaction forces which can arise from nearby bodies, including the walls of the vacuum chamber itself. Tests were conducted at atmospheric pressure and over a range of partial vacuum conditions. Direct experimental results indicate that when an electrostatic potential difference is applied between asymmetrical electrodes of an all metal test device, a residual propulsive force is generated and acts on the device. This residual force acts in the opposite direction to electrical wind forces and to the forces claimed to have been measured in a vacuum by T.T. Brown.<tL.IU
p."
S
RCK


f_
TABLE OF CONTENTS
Section Page
INTRODUCTION 1
INVESTIGATIVE PROGRAM 3
LITERATURE SURVEY AND BACKGROUND 4
Literature Survey 4
Background and Review of Selected Observations 4
EXPERIMENTAL CONFIGURATION 9
General 9
Main Test Configuration 13
Torsion Fiber Subsystem 15
Optical Readout and Data Acquisition Subsystems 21
* Electrical Subsystem 24
Vacuum Subsystem 27
ITet Devices 29
TEST RESULTS 31
Test Approach 31
Calibration 32
Device Tests 35
Auxiliary Tests and Considerations 52
System Errors 53
Evaluation 54
CONCLUSIONS 58
RECOI'MENDATIONS 61
REFERENCES 62
* BIBLIOGRAPHY 64
0, % •':•.J:dJ coeds
iDist


SLIST OF FIGURS
1 Block Diaqram of Overall Test Layout 10
2 Room Layout 12
3 Test Configuration Schematic 14
4 Figure-Eight Device Support 16
5 Details of Rotating Device Support and Fixed
Standpipe in Electrical Connection Region 26
6 Variation of Measured Force With Air Pressure 42
7 Variation of Measured Force With Applied
Potential Difference 45
8 Variation of Measured Force With Applied
Potential Difference 47
iv


LIST OF TABLES
_Table Til PAU,
1 Selected Mechanical Properties of Metal Fibers 17
2 Parameters for Calibration Cylinde= and Mirror
Holder 21
3 Brown Effect Test Device No. 1 30
4 Brown Effect Test Device No. 2 30
5 Cone for Brown Effect Test Device No. 3 30
6 Calibration: Torsion Fiber No. 1 33
7 Asymmetrical Devices Test Data Summary 38
8 Symmetrical Device Data Summary 49
9 Electrical Condition-, for Symmetrical Device Tests 50
10 Array for Comparison of Forces Measured Using
Symmetrical Device 51
V-V?
A.
a


INTRODUCTION
The objective of this contract effort was to experimentally
and theoretically explore the Biefield-Brown effect, which allegedly converts electrostatic en~ergy into a propulsive force.
This force-generation scheme, originally suggested by Dr. Paul Biefield and subsequently discovered by Townsend T. Brown in the late 1920s, arisesi when a large DC electrical potential
differvnce is applied between shaped electrodes fixed with
respect to !'ne another by a dielectric. Under these conditions a significant net force is generated, which acts on the entire
electrode/dielectric body and typically causes it to move in the direction of the positive electrode.
This concept atay well represent a direct field-field, or
field-vacuum interaction scheme with the potential for producing thrust without the conventional action-reaction type of momentum
transfer brought about by ejective expenditure of an onboard
fuel. The significance of this propulsion concept to launching
and/or maneuvering payloads in space is potentially very great.
maniy years, but at the outset of the present effort it was still
inadequately explored and remained without confirmed operation in
a vacuum, without adequate quanftitative characterization, and
V ~without an adequate theoretical basis for its operation.
Therefore, activities under the current program emphaxsized the experimental exploration of this electrostatic, thrust
generation concept to verify its very existence, to verify its
operation in a vacuum and to establish the magnitude of its
thrust. To conduct this program, an overall laboratory test
design and configuration was developed that was suitable for quantifying the resultant propulsive force. Instrumentation
* schemes and techniques were also developed for measuring the


propulsive force and key physical parameters. Candidate propulsive devices were developed and tested. The overall
experimental effort centered on making direct measurements of electrostatically induced propulsive forces on these test devices, inasmuch as details of the nature of the Brown effect were too sketchy to provide a reliable base for interpreting indirect measurements. Theoretical activities centered largely on examining contributions from known phenomena, which could influence or confound measured force values.
This report discusses the various features of the investigation, including a search and review of the available literature relevant to the Brown effect, the experimental configuration, the test techniques and results, the evaluations made, and the findings and recommendation. to further explore and develop the potential of the Brown effect.
The experimental results of this investigation give a
0 preliminary indication that a force phenomena does arise in conjunction with the presence of a non-linear electrostatic field. At present these findings are based on very limited results. Additional tests and efforts will be required to confirm or refute this finding. We are cautiously optimistic that this finding will remain valid when the results of other possible minor effects are explored and evaluated.
II


INVESTIGATIVE PROGRAM
An experimental-theoretical approach was used to explore the
existence and nature of the Biefield-Brown effect. Particular
attention was given to developing an experimental configuration
and technique to provide a sound basis for quantifying electrostatically induced propulsive forces on selected test
devices, and to overcome shortcomings of previous experiments that have led to criticism of Brown's work. This program was carried out .n the following three phases:
1. Lite•ature search and review---This included a search and review of the available literature
related to the Biefield-Brown effect, its nature, and its use for propulsion.
2. ExDerimental evaluation---This involved developing an overall laboratory test design and configuration suitable for quantifying the
propulsive force; advancing instrumentation schemes for measuring the propulsive force and relevent physical parameters; designing,
fabricating and testing candidate propulsive devices; and evaluating the test results.
3. Theoretical studies---This primarily involved
finding key parameters and recognrOed phenomena on
which the Biefield-Brown effect may depend.
I. Secondarily, considerations were given to
developing a preliminary theoretical model for the propulsion concept, mainly for use in designing
the experiments and evaluating the test results.
IV
* 3&


LZTERATURE SURVEY AND BACKROUND
Literature Survey
Information directly related to the Biefield-Brown effect
(sometimes just called the Brown effect) is rather linited, often obscure, and sometimes hard to acquire. A reasonable sample of past and current material written about the Brown effect has been acquired, or at least identified. A bibliography of these and other references believed to be relevant is presented later in this report.
It is noted that a portion of the most definitive work of T.T. Brown has been reported in the patent literature. Other published articles written directly by Brown are scarce. His findings and views have occasionally been summarized by others. The six-volume set entitled, "The Scientific Notebooks of T.T. Brown," (currently being published) is worthy of note, but
information related to Brown's propulsion work seems to be
limited to Volume 1 and part of Volume 2.
Background and Review of Selected Observations
The impetus for T.T. Brown to conduct electrostatic- type
propulsion investigations apparently received a considerable boost from Dr. Biefield's query as to whether an electrical capacitor, hung by a thread, would have a tendency to move when
it was given a heavy electrical charge. "Yes," was the answer
found by Brown. This finding formed the essential basis for the Biefield-Brown effect.( 1 ) Subsequently, Brown broadened the
phenomenological basis of his investigations to consider possible couplings between the forces of electricity and gravity, as
analogs to known couplings between those of electricity and
magnetism. Whereas the coil is a key link in e!ectromagneric phenomena, the capacitor is the analogous link for the
electrogravitational case. This, in turn, may account for
4
kv


Brown's emphasis on the use of capacitor-type propulsion devices, 4,
including the use of dielectrics.
Five factors were indicated by Brown as determining the intensity of the effect:(2)
1. The separation of the plates of the capaoitor
(closer plates give a greater effect);
2. The magnitude of the dielectric constant K (larger K gives a greater effect);
3. The area of the capacitor plates (larger plates give a greater effect);
4. The potential difference applied to charge the
plates (larger potential difference gives a larger
effect); and
5. The mass of the dielectric between the plates (greater mass gives a larger effect).
Brown claimed that it was the last factor which is inexplicable
from the electromagnetic viewpoint, and which provides the
connection with gravity.
Essentially, the force produced in Brown's experiments was nearly always in the same direction as that from the negative to
the positive potential within the test device. It is important to note, however, that in his early diolectric-type devices the force and motion sometimes would also occur in the reverse
* direction, when the so-called "saturation voltage" of the
N dielectric was exceeded.(3)
Later--- when Brown used asymmetrical d3vices---apparently no ambiguity in the direction of the propulsion force was observed.
5


In these devices, a pair of different aixed electrodes was
attached to opposite ends of a dielectric element. When a high
electrostatic potential difference was applied ac.iossa these
electrodes, the electric field lines from one electrode converged....steeply to the other. The force tending to propel the device ....
was, in this case, in a direction from the region of high flux
density toward the region of low flux density, and generally in
the direction through the axis of the electrodes.(4 ) Brown further claimed: (4)
" The thrust produced by such a device is present if the electrostatic field gradient between the two
electrodes is non-linear. This non-linearity of
gradient may result from a difference in configuration of the electrodes, from the electrical potential and/or
polarity of adjacent bodies, from the shape of the
dielectric member, from a gradient in the density, electrical conductivity, electric permittivity and
magnetic permeability of the dielectric member or a
combination of these factors".
In another set of experiments,P() Brown observed an impulse (or
time-dependent) action with a type of test device made from a
solid block of massive dielectric. Such a test device was
immersed in oil, but suspended wita electrical leads to act as a pendulum and swing along the line of its elements (i.e.,
lengthwise). When a DC potential in the range of 75 to 300 kilovolts was applied, the pendulum swung up the arc and stopped
when the vertical component of the propulsive force balanced the
gravitational force on the pendulum. But the pendulum did not remain there; it gradually returned to the starting position even
whiles the potential was maintained. The time for the pendulum to
reach the maximum amplitude of swing was less than five seconds, but from 30 to 80 seconds were required for it to return to zero.
It was necessary to remove the electrical potential fo:.: several
minutes to allow the system to regain its normal condition and to
enable the cycle to be repeated.
6
-- - - - - -- - - _ - - - 


I7 . .. . ..........
The observations made by Brown can likely be accepted at . .........
face value, but his claims and explanations are another matter.
Unfortunately, despite Brown's efforts to the contrary, his ........
results and explanations have been repeatedly reviewed,
evaluated, criticized and often discredited( 6 ) with little
attempt to adequately explore his observations---largely because
the reviewers found it impossible to accept the possibility that
the Biefield-Brown effect could represent a "new force."
Thus, at the outset of this program, the literature discussing Brown's work reported no confirmation of the existence
of the affect in a vacuum, no adequate characterization of the effect in quantitative terms, and no theoretical basis for the -Lserved generation of the effect.
From the space-propulsion point of view, the issue of
operation in a vacuum is paramount. A known phenomenon called
"electrical (or ion) wind"( 7 ) has frequently been invoked as the
mechanistic basis of thrust on electrostatic-driven propulsion devices similar to, and including, those explored by Brown. In fact, electrical wind does contribute to the thrust on such
devices when they operate in air.* However, Brown( 4 ) has made
the claim:
" In a vacuum, the reaction forces appear on solid environmental bodies, such as the walls of the vacuum
chamber. The propelling force, however, is not reduced
to zero when all environmental bodies are removed
beyond the apparent range of the electrical field."
* The electrical or ion wind eff.act typically arises in the
vicinity of coronas and results from the momentum given to air
molecules by impact or drag of the i,.ns and electrons as they
move out from the high field regions.
7
L
A.MMA LA.fA 16 M f.lfn.XARUMV~~ AAAAlAXL~AJAM W.IIIRd
L4~L1 NA J %AAA.&j&MA~xAxaA mfrM MMM


Further, in a 1973 letter written to Rho Siga,** Brown
commented on his propulsion investigations in a vacuum
environment: (2)
"The experiments in a vacuum were conducted... in 195556...in 1957-58 and...in 1959 .... We were aware that the thrust on the electrode structures was caused largely by ambient ion momentum transfer when the experiments
werp conducted in air .... In the Paris test miniature
saucer-type airfoils were operated in a vacuum
exceeding 10-6 mmHg. Bursts of thrust (towards the positive) were observed every time there was a vacuum spark within the bell jar....The result which was most significant from the standpoint of the Biefield-Brown effect was that thrust continued, When there was no
vacuum sgark, causing the rotor to accelerate in the
negative to positive direction .... In short, it appears there is strong evidence that the Biefield-Brown effect does exist in the negative to positive in a vacuum of
at least 10-6 Torr. The residual thrust is several orders of magnitude larger than the remaining ambient
ionization can account for." j
Thus, we focused the activities of the current program on
experimentally determining the existence or nonexistence of the
phenomenon, confirming its operation in a vacuum, and
establishing the magnitude of its thrust in quantitative terms.
The following section describes the experimental configuration
designed and built by Veritay for this purpose.
II
** Rho Sigma is a pseudonym
8


EXPERIMENTAL CONFIGURATION
General
A principal step in investigating the Biefield-Brown effect is the development of a test configuration for quantifying this electrostatically induced propulsive force on candidate propulsion devices. The configuration design advanced and implemented for this project uses a vacuum chamber as the central element. This permits tests to be conducted either at atmospheric pressure or over a range of partial vacuum conditions. Further, the test set-up incorporates geometrical symmetries to minimize the influence of reaction forces, which may arise from nearby bodies, including the walls of the vacuum chamber itself.
These features reflect the importance of investigating the Brown effect over a range of pressure conditions, and especially in a vacuum. Under normal atmospheric conditions, electric3l wind effects are known to be significant, and can be sufficient to account for some of the results often attributed to the Brown
effect.
A block diagranl of the overall test layout is given in Figure 1. This figure shows schematically the key components,
their grouping into functional subsystems, and their interconnections within the subsystems.
The subsystems include the following:
1. Main test configuration, which encompasses the vacuum chamber and the components within its interior.
2. Torsion fiber subsystem, which is the central elenent for measuring propulsive forces, is located
9
L~rkvx"Ml V: AA 1^10RAA W A)4A'k)AXIAAX'AMM AYAfAPJVWN '.NN1. W\UI N


u~wm=
Mcw
A.mm
-cul
Kima
FIgure 1. B ocDiakgn=am.DLQ&HTestI&Mtjy
10


inside the ch&u'ber, and is noted specifically
because of its importance.
3. Optical readout and data Acquisition subsystem,
which is critical in extracting and recording measurement information. It interfaces with the chamber through the telescope.
4. Electrical subsystem, which includes both the high
voltage source to activate the test devices and the
instrumentation to quantify several electrical
quantities.-
5. Vacuum subsystem, which includes the pump, gauges
0 and components related to achieving and quantifying partial vacuum conditions in the chamber.
Wo 6. Test devices, which are critical to meaningful
exploration of the Brown effect.
The actual implementation of this test layout, shown by the
physical arrangement of various components, is given in the room
layout of Figure 2. The room itself is located in the corner of
a concrete bunker, formerly a part of the Nike missile launch
site. The concrete floor is approximately 4.8 m (16 feet) below grade, and the concrete ceiling is covered with about 1 m (3
feet) of earth and another 0.2 m (8 inches) of concrete at the
surface. The short-b-erm temperature stability of this room
location is good without imposing controls, and the mechanical M vibration stability during test runs is adequate to maintain a
low mechanical noise level in the experimental measurements.
M Details of these experimental features are discussed in
SI subsequent subsections.
Lk 11
DK


3t.LUMN~lo
r 1.LSAurrm
4. VACUUM PUMP IL. VIDEOCAMERA
S. VACUUM 11083 13. W=DE
L. VACUUM GAUGE 14. ONTORF
7. VACUUM MEIER I&. OIClLLcmCW'
L. POWER SUPPLY AIL POlR ChILU
FS LM
I
Fime Bm 2 lyg12
0~i


Main T-est configuration
The main test configuration is shown schematically in Figure 3. It consists of an AISI type 316 stainless stoel vacuum
chamber approximately 1.04 m (40,8 inches) in diameter by 1.52 m
(60 inches) overall height with a 1.17 m (46 inches) straight
side arid a welded, dished top and bottom. Access to the chamber is provided through a 0.51 m (20 inch) top manway and smallerI
ports. The latter are used as instrumentation, illumination and vacuum port3. The chamber is mounted on legs (not shown) with
leveling screws for adjustment of the vertical axis of the
chamber.
Other general features associated with the test * configuration are also indicated in Figure 3. Candidate devices
for the Brown effect tests are mounted in tandem on a figure
eight type device support. This unit with an attached mirror is
suspended from the top of the chamber by a torsion fiber. The
latter is fastened to a vacuum-sealed adjustable rod which can be
raised, lowered or rotated, to achieve a suitable device height
or angular zero position for tests.
The height adjustment for the fiber support rod was
implemented by mounting the rod on a precision rack and pinion
slide mechanism. In use, this unit was coupled with a dial
indicator, and height settings and resettings to a precision of
+ 0.0013 cm (± 0.0005 inlches) were achievable. This precision
conduct of tests under this program.
The electrical input to the test devices are fed from a high
voltage DC power supply into the bottom of the chamber through a standpipe, and are connected to concentric liquid mercury
contacts at the top of the standpipe. Electricity is then fed to the test devices via conductors with points which touch the
p. mercury. This arrangement provides nearly frictionless


-TEST DEVICES
SCALE
DIAL INDICATOR RAC AINO
TORSION FIBER
MIRROR
OBSERVATION PORT HIGHI VOLTAGE FENW
IIIIIIIIIIIIIIIIIIIý ijSTAND PIPE
HIGH VOLTAG.E SUPPLY
Flgure 3. Ih~nlufi~hui
9 14
118i


elentrical contacts, and permits electricity to be conducted to the test devices while a force measurement is being made.
A force measurement uonsists of determining the angular twist of the fiber caused by action of the force. This together with the fiber calibration constant and moment arms of the test devices permits the force acting on the devices to be evaluated.
The angular twist of the torsion fiber during a test run is measured using a telescope, a mirror mounted to the test device ý_apport, and a scale fastened to the chamber wall.
Torsion Fiber Subsystem
The measurement of forces generated by electrostatically driven test devices is carried out using a single torsion fiber system which supports the devices. The torque to twist the fiber
is provided by device generated forces which are directed
perpendicularly to radial moment arms and to the fiber axis. Two
such devices are directed in tandem and mounted on the figure
'eight device support at "equal" radial distances from the fiber.
The figure-eight device support is shown in Figure 4. The actual
moment arm distances for tests conducted in this program ware 0.1024 m and 0.1018 m, with an average value of 0.1021 m.
The selection of a torsion fiber involved making a tradeoff an~on-' a number of factors such as:
o elastic after-effect and hysteresis
o relative strengLh in tension and torsion
o elastic behavior almost up to the b~reaking point
o non-corrosive
o electrical insulator-ý or conductors
o freedom from internal strains which can cause zero
drift
o cost and availability.
15


NALL
MRASS TUBE
______ AneHYvi~t~ * ma"k
OLECTR*ICAL, MECHANICAL
SUPPORT AING
viEW A-A
FIgur 4. t DamhgI~!mice oz
16


II
Metal fibers were of particular interest because of their
electrical conductivity. A conducting fiber was expected to be used as an electrical lead for measuring the leakage current of various test devices; these tests have not yet been conducted.
Metal fibers, however, are known to have long term drift problems. In this application we planned to make zero angular position readings of the fiber system both before, and after, each individual test, although zero drift was not expected to be a problem. In retrospect, drift has been observed occasionally, but has not been a serious problew. The drift observed was always linear over a period greater than the duration of a test and showed an average magnitude of about 1.7 x 10-4 radians per minute when it occurred. The drift continued at the same value when power was applied during a test run; deflections were * measured directly between drift lines.
The types of metal fibers considered for use were copper, tungsten, platinum and silver. Selected mechanical properties of these fibers are given in Table 1.(8)(9) Typically, the fiber strength becomes the limiting mechanical factor for each material and determines the corresponding minimum fiber diameter.
TABLE 1. Selected Mechanical Properties of Metal Fibers
MATERIAL YOUNGS RIGIDITY TENSION YIELD DENSITY
MODULUS MODULUS E(N/m 2 ) G(N/m 2 ) N/m2(psi) g/cm3
Copper 12.1-12.8x10 1 0 4.6x10 1 0 5.24x10 8 (76x10 3 ) 8.3-8.93
(hard) Tungsten 34.x 1010 13.5x10 1 0 6.89x10 8 (100x10 3 ) 18.6-19.1
Platinum 16.7x10 1 0 6.4x101 0 4.14x10 8 (60x10 3 ) 21.4
(hard)
Silver 7.5x10 1 0 2.7x101 0 3.58x10 8 (52x10 3 ) 10.4
(hard)
17
--------- - 


For reasons of availability and cost, together with features
of strength, corrosion resistance and conductivity, the fiber
type selected for test use was tin plated copper. Copper without
the tin plating was sought, but was not readily available. The
size of fiber used is B. and S. No.34 with a diameter of I6G#m
(0.0063 inches). This will support a load of 0.538 kg with a
safety factor of two. The fiber length used is 0.927 a (36.5
inches).
Force measurement with a single torsion fiber systea depends
fundamentally on how the fiber twists when it is subjected to a
torque. The total torque Q needed to twist a torsion fiber is
given as the sum of three contributions( 1 0 ):
Q - s + Qz + Qb (1)
0 where
Qs - torque associated with the shear stress of twist (this
torque constitutes the major part of the resistance of
an elastic fiber to torsion)
Qz - torque associated with the longitudinal stresses of
twist (when an elastic member is twisted, its
longitudinal fibers are forced to take up a helical
configuration around the axis of the twist. The ones
farthest from the axis are stretched and put in
tension; this causes those near the axis to be put into
compre:sion).
Qb= torque associated with the application of an axial
force. (Since the tensile stress acts along helices,
the force on each elementary area of the elastic member
cross section has a component at right angles to the
axis of twist. This component produces a torque dQb
tending to untwist the member).
This can be written
UQ Qs( I+QZ/Qs + Qb/Qs
18


41
and for solid circular cross sections: I
Q Q5s ( 1 + 0.2 q2 /G 2 + p/0 ) '(21 _
where p " plongitudinal tensile or compressive stress due to applied axial force
q - maximum shear stress of twist
G - shear (or torsion) modulus, also called modulus of ---- . rigidity.
in practice the ratio q/G is approximately .001 when the shear stress has its maximum allowable value. Qz is therefore
negligibly small compared to Qs for typical amounts of twist. The quantity Qb/Qs or p/G is not negligible when heavy fiber loads are used.
The shear torque Qs for this case is given by (10)
QS _ (A2 G0)/(2wL), (3)
with A- fiber cross sectional area
L- fiber length
e- angle of twist in torsi,.#n connection. J!
Likewise, the torque Qb is given by
Qb (PAe)/(2nL) (4)
with P- axial tensile (+) (or compressive (-)) force applied to
a torsion connection (and hence to stretch a twisted fiber in tension).
These tension relations are included for completeness, and
19
-~ - - - - - - -- -- -- - - - - - - - - - - 


are useful in the delddtift of tibet Sttertilso thd Ineva~t the rigidity modulus G of materials.
Direct calibration of the torSion fiber used for toots Was done by measuring the swing periods 6f the fiber as part of a torsional pendulum, In this ca&e the torsional stiffness 8 o••fa
fiber is given by (i.
S
=ye - (4f2 Ic)/T 2
where Q = total torque N--
e = angle of twist of a torsion fiber;radian.
Ic = moment of inertia of the suspended magslkq-m2
T = oscillation period;sec.
The connection between the torsional stiffness, S, and the
rigidity modulus follows from equations (3) and (5)p
G (21tL)/A 2 . (6)
The fiber calibration is most conveniently carried out using a cylindrical block of known mass and dimensions, for which the moment of inertia can be determined with considerable accuracy.
A particularly important consideration noted nearly a century ago by Limb (11) concerns the shape of the otcillating calibration mass to be suspended from the torsion fiber under test. He noted that a solid cylinder with the length l, which is C3' times its radius R, i.e. I= R6, has the same moment of
inertia about all axes through its centroid (the ellipsoid of
inertia for the cylinder becomes a sphere). Therefore, by using
a suspended mass of this shape for calibration, an error is not
0 introduced if the point of attachment of the fiber to the
cylinder does not lie exactly on the axis of the cylinder.
20


Such a calibration mass was employed, but the overall shape criterion was only approximately satisfied, since the
contribution of the mass and moment of inertia of the mirror
holder used to attach the calibration mass to the fiber were neglected. Values of key parameters for the calibration cylinder
and mirror holder are given in Table 2.
TABLE 2.
Parameters For Calibration Cylinder and Mirror Holder
CYLINDER MIRROR TOTAL HOLDER
Length, m(inch) .02730(1.075)
Diameter, m(inch) .03152(1.241)
Mass, kg .16768 .01904 .18672
Moment of 2.1012xi0- 5 3.3528x10- 7 2.1347x10-5
inertia, kg-m2
Optical Readout and Data Acquisition Subsystems
A readout subsystem consisting of a simple telescope, mirror
and scale was used to determine the angular twist of the torsion fiber. In this subsystem an alignment telescope equipped with a
crosshair reticle is directed so its line of sight intersects,
ml near center, a mirror attached to the lower end of the torsion
fiber. The telescope and mirror are further positioned so this
linear scale fastened to the inner wall of the vacuum tank. The scale is mounted horizontally and is curved to conform to the
K
'shape of the wall.
21
0m


The scale can then be viewed through the telescope and a definite reading determined at the intersection of the croeshair with the scale image. A change in the angular direction S of the mirror causes a deflection of the light path through an angle 20. Concurrently, this causes an apparent movement of the scale image relative to the crosshair. For a scale deflection (x-,X), relative to some zero reading xo, the angular motion of the mirror is e = (x-xo)/2r.
In the arrangement used, the radial distance, r, from the mirror to the scale is 0.5175 m (20.38 inches). The scale,
itself, is a stainless steel, flexible machinist's scale,
0.6096 m (24 inches) long with subdivisions in English units of 1/50 and 1/100 inches on two separate rulings.
The scale is typically read to the nearest
1.27x10- 4 m (0.005 inches), and deflections of about twice this
value are representative of the threshhold of measurement, taking into account reading uncertainties and random pendulum oscillations due to noise. This results in a minimum deflection
angle of the mirror
emin = 2.45x10-4 radian per 0.010 in. apparent motion
of the scale.
This value of emin corresponds to about 50 seconds of arc, which is not of great sensitivity, but has proven to be a useful
level for tests conducted under this program.
During propulsion device tests some observations were made visually through the telescope. A preferred technique involved
using a video camera and monitor, with the camera lens positioned to see through the telescope eyepiece. Deflection readings
during tests were then made directly from the video monitor and entered in a laboratory notebook manually.
22
""


A General Electric VHS movie Video System SE 9-9608 with video tape recorder was used for a number of the test runs.*Use of the recorder permitted a real time record of the actual test
run to be made. This recor-d is helpful since it permits 7
replaying the test run to confirm certain deflection readings or
ascertain missing readings after-the-fact.
Readout of the torsion pendulum oscillations for calibration purposes required a different approach, since the swing time
periods cannot be determined visually with sufficient accuracy.
The same telescope and mirror system is again used, but.an active signal in the form of a laser beam is passed through a beamsplitter, through the telescope, and onto the mirror. Twice each
period as the mirror oscillates, the laser beam is reflected back
S through the telescope, into the beam splitter and deflected into
a fiber optic cable ani then onto a photodetector and amplifier. Real time electrical signals from the detector-amplifiar are sent
to a digital oscilloscope and recorder, where the signal pulse
can be displayed and timed accurately.
For calibration runs, the zero angular position of the
torsion fiber and mirror was adjusted so the laser beam wasI
reflected into the detector when the pendulum passed its
librium position. Near this position the angular deflection
rate of the mirror is the greatest, and the duration of the optical pulse at the photodetector will be rolose to a minimum.
This facilitates making accurate measurements of the instant when
th-e pulse occurs (usually reckoned at the maximum ampiitýuC'e of
t.'Ie pulse).
The actual system assembled uses a Spectra Physics type
155A, 0.5 mW, helium-neon laser; a 10 mm x 10 mm prism beam splitter; a plastic fiber optic light guide; a photodiode
detector with an IC instrumentation amplifier designed and built
at Veritay; and a Nicolet Digital oscilloscope Model 2090 with
* disk recorder.
23


Typical digital sampling times used were 0.05 seconds per
point. At this sampling rate a total sample of 32 x 103 points (26.7 minutes) could be recorded directly. The total time used for calibration runs normally encompassed about six to eight
complete oscillation periods.
This same data recording system employed for oscillation timing was also used to track early pump down runs in the vacuum chamber. In these vacuum checks, gauge readings were taken every 20 seconds, which provided a continuous 4000 point record that spanned approximately 22 hours.
Electrical Subsystem
High voltage was supplied to the devices under test with a Spellman regulated, 0-60 kV, DC power supply Model UHR60PN30,
which has a maximum power output of 30 Watts. The voltage
polarity of this supply can be reversed internally, but is not directly switchable for safety reasons.
The output vcltage and current from this supply can be read
directly on a panel mounted kilovoltmeter and microammeter,
respectively. At low voltages and currents the sensitivity of these meters is inadequate. Input voltages up to about 2 kV were
measured with a Fluke digital voltmeter model 77. A ± 5 A full scale instrument to measure supply current was designed and
fabricated at Veritay. It consists of a high grade differential
amplifier with programmable gain to accuratly measure the voltage
drop across a precision resistor of known value. This unit has a minimum current resolution of about 100 A.
The high voltage is fed to the test devices inside the test
chamber using Belden No. 8866-80C high voltage cables, which have a DC breakdown voltage of approximately 80 kilovolts. These
cables penetrate, and are sealed with a high vacuum leak sealant,
424
Sim


to an acrylic plug at the base of the feed-through standpipe. The cables, themselves, are enclosed inside additional acrylic
tubes inside the standpipe to help keep the cables straight and
away from the grounded stainless steel walls of the standpipe.
The conductors of these cables are'attached to platinum wires at the top of the standpipe. These wires are sealed into the
insulating acrylic block located there. This dielectric block,
itself, is threaded into the interior of the standpipe and vacuum sealed to it with an 0-ring and vacuum grease. The platinum
wires then pass through this block to a shallow trough and well,
each. of which contains liquid mercury. The separated mercury
regions are symmetric about the axis of the standpipe. Details of the electrical conductor location at the top of the standpipe
are shown in Figure 5.
When the mercury is touched by conducting platinum points
protruding from the base of the figure-eight device support,
nearly frictionless electrical contacts are formed. These
contacts permit the supply voltage ro be fed directly to the test
devices.
The standpipe column and dielectric block at its top were originally designed with provision to cool the liquid mercury.
This was deemed necessary, since the liquid mercury was to be
* subjected to the full vacuum in the chamber, and mercury will boil at room temperature (200 C) under a vacuum of 0.160 Pascal
(1.2x10-3 torr).( 12 ) At a temperature of -5.60 C the: vapcr
pressure of mercury is 1.33x10-2 pa (10-4 torr), which is the lower limit which could be expected with the vacuum system used.
I
These concerns were legitimate, but unfulfilled as ot~her vacuum
sealing problems (believed to be largely associated with the
* standpipe) precluded achieving vacuum levels in this system lower
than about 1.3 Pascal (10-2 torr). Efforts to cool the mercury
during this program were thus abandoned.
25


F....M
swmvm RUIw
Fiue .D3ILARWILIAhUIMMn3
0.O.T=V IITT 0
26R 6ZCUt
L
inM~~ M M A~ KkX M A CAU B JJJLO
MM~ Lr , t A X W U M 3 f


The surface tension and viscosity of the mercury did present problems of causing excessive friction for the contact points
protruding from the device support. This was resolved by
adjusting the fiber height so the contact points just touched the surface of the liquid mercury.
As part of the electrical considerations, an instrument for measuring the leakage current from the tezt devices was also developed. This leakage current meter is similar to the output
current meter noted earlier, except this unit has four ranges of operation with the most sensitive being 0 to 500 nA. In this
case the minimum resolution is 1 nA.
The collector for the leakage (diffusion) current from the test devices was a .406 m (16 inch) diameter stationary, hollow
aluminum sphere, mounted inside and electrically insulated from
the externally grounded vacuum tank, except via its connection
through the leakage current instrument. This hollow sphere
surrounded the figure-eight device support and test devices; the devices were free to rotate inside the sphere without contacting it.
In principal, the entire vacuum chamber could have been used as the leakage current detector. This would have involved isolating the entire chamber and associated equipment from
electrical earth ground, and then connecting the chamber to earth ground through the leakage current meter. For safety reasons the
chamber was not isolated from earth ground in this maizner during the program.
Vacuum Subsystem
The vacuum achieved for test purposes was developed with a Boekel Cenco HYVAC 45, two stage rotary gear pump with gas ballast and 500 liter-per-minute (17.7 CFM) capacity. The vacuum
N27


U.
capability of this pump is 1.3 x 10-2 Pa(10" 4 torr). While a
lows, vacuum level would have been desireable, the chamber size
involved, the range of partial vacuum levels desired for test,
the initial uncertainty as to the vacuum level needed to overcome
electric wind effects, and pump costs precluded employing a
pumping system in this program to achieve lower vacuum levels.
A Boeko TKO-19 vacuum pump fluid with a vapor pressure of
1.3 X 10-4 pa (10-6 torr) at 25 0 C was used in the pump.
The pump was connected to the chamber with a reinforced,
flexible vacuum hose with capability of holding a vacuum of about 1.3x10- 4 pa (10-6 torr).
All new flexible seals for the vacuum chamber ports were used, together with Dow Corning 976 vacuum grease and a high vacuum aerosol leak sealant, as required.
An Edwards Pirani gauge head PRM 10K with a model 503 controller was used as the primary m3ans for routinely determining vacuum levels in this chamber. This controller unit has both a direct meter as well as an analog voltage signal for readout. This voltage signal was typically read out using a Fluke model 77 digital multimeter.
A Labconco McLeod gauge was available for calibration checks of the Pirani instrument.
Although strictly not a part of the vacuum system, an air inlet unit was installed on the vacuum tank. This unit consisted of Union Carbide type 4A molecular sieve dessicant and an EPM 2000 high volume air sampling filter to trap incoming aerosol particles larger than about 0.3 1um. Full utilization of this unit was not possible, since many of the test preparations required complete entry into the chamber.
28
m


Test Devices
The candidate devices selected for initial propulsion tests
had all metal, but few or no dielectric components. This avoided
the inherent problems associated with assessing initial and final
electrical conditions when testing dielectrics (13) (14) and
improved both the reproducibility and interpretability of results from a small number of tests.
One simple candidate device, which is patterned after a test device used by Brown (4), is the all metal ball and disk unit shown in Figure 4. The construction materials and device dimensions are given in Table 3. A key feature of this test unit is that it will develop a non-linear electric field gradient between the ball and disk.
A second test device is a pair of identical metal toroids, each with a flat disk which fills in the normally open central region of the toroid. The construction materials and dimensions of this device are gi-en in Table 4. This test device is expected to develop a relatively uniform field between the two
toroids.
Test device No. 3 was a modification of device No. 1, in which a truncated cone of dielectric was placed between the ball and disk. The material and cone dimensions are given in Table 5. Device No. 3 is expected to form a non-linear electric field
gradient within the dielectric.
29


TABLE 3. Brown Effect Test Device No.1
Configuration: Ball and disk, separated normal to disk by 0.0400 m (1.57 inch)
MATERIAL DIAMETER THICKNESS
Ball: Aluminum 0.0127 Z(0.50 inch) -----
Disk: Brass 0.0793 9(3.125 inch) .00038 a (0.Ol5inh)
TABLE 4. Brown Effect Test Device No.2
Configuration: Two filled in metal toroids coaxially located and
separated by 0.0338m (1.33 inch).
0
MATERIAL DIAMETER WEB THICKNESS OUTSIDE INSIDE
Toroid Aluminum .01905m .01270m .00122m
(0.748inch) (0.500inch) (0.O048inch)
TABLE 5. Cone for Brown Effect Test Device No. 3
Configuration: Truncated dielectric cone to fit between the ball
and disk of device No. 1.
MATERIAL DIAMETER HEIGHT
BASE TOP
Truncated Acrylic 0.0508m 0.014m 0.0455a
0Cone (2.000inch) (0.55inch) (1.790inch)
• 30


TEST RESULTS
Test Approach
The overall test approach used here is directed toward measuring the propulsive forces over a range of partial vacuum conditions, and extrapolating these measures to a limit
appropriate for a full vacuum. The purpose of this schene, of course, in to overcome any residual force effects due to electrical wind. The propulsive force limit so obtained
represents the sum of the Brown effect and the ion propulsion effect. The contribution due to ion propulsion can be estimatedj and the existence and magnitude of the Brown effect can be evaluated.
At the outset it was anticipated that separate, moveable walls might need to be installed in the chamber, and then tests run with the walls located at various distances from the devices to enable estimates to be made of the influence on the test devices of induced surface charge on the walls. Preliminary tests with boundary plates and devices outside the chamber led us to believe that such boundary effects were probably negligible, except perhaps at test conditions in excess of 20 kilovolts.. Subsequent test results obtained in the chamber appear to confirm this expectation for the torsion fiber system, at least at relatively low voltage test conditions.
As noted earlier, a stationary hollow aluminum sphere (two joined hemispheres) was placed around the test device support as
a collector for leakage currents from the associated devices. In effect, this sphere acted as a separate wall of the type just
noted. The ability to neglect its influence as a boundary is
important, in order to enable it to be used as a current
col lector.
31


Calibration
The calibration of torsion fibers was carried out by
measuring the period of a torsion pendulum, consisting of the
fiber, calibration cylinder and mirror holder.
Results for a first fiber are given in Table 6 over a range of vacuum levels. The torsional stiffness, S, of the fiber was calculated using equation (5) given previously. A regression
analysis of the torsional stiffness as a function of vacuum level
indic:ated essentially no change over the range of conditions tested. Thus a simple mean value of the torsional stiffness S was used. The corresponding rigidity modulus was calculated using equation (6); this giAve a mean value G-4.737 x 101ON/m2 . -I This compares favorably with the handbook value GH - 4.6 x 1010 N/m2 for rolled copper. This first fiber was
used for propulsion device shakedown test run numbers I through
8A, before it was accidentally broken while replacing a test
device.
Thus a second similar wire, called fiber No.2, was
calibrated at one atmosphere and at a temperature of 15.2 0 C(59.4 0 F). The value obtained for the torsional stiffness was S= 3.2033x10- 6 N-m2 , and a corresponding value for the
rigidity modulus was G- 4.599x10 1 0 N/m2 . Fiber No.2 was used for propulsion device test runs after No. SA, starting with No.9.
The effect of temperature on the fiber calibration values
has not been measured directly, but was determined indirectly from data on the temperature variation of Youngs modulus for
copper given by McGregor Tegart. (15) At 20 0 C (68 0 F) this coefficient is approximately -0.040 percent per degree Celsius
increase (-0.022t/°F). This same coefficient applies
approximately to the rigidity modulus, G, Lnd to the torsional stiffness, S, of the copper fiber used. The small value of this
coefficient indicates that temperature variations of
* 32
C


0.
4h 0 04
0 4jo
z a V4
m-0 -i
Pist
00 M~ ~p
bOGfI-I
EE
P4 %Or4r4 M
0 oG 4 0 
-o- 0 0
M%
41 Z 0
0-4C44
cc N0 0%
.1 .
In!C x
iN N4%
N V-4 t- 0 Ch c"
Ch~
0%t 0 0 mU N t
4l t'ci4
* ON N
a4 0
t0 RU 4 Ua01.,4
lo 4) Ie)Euca
q3e 0 t 01. 4F
FAr,-ON4 
4 r 0O 6 3
wE-4 E-4
33


t 2.40 C (± 4.3 0 F) experienced during the tests conducted under this program should alter the fiber calibration and resulting force values measured by no more than about t 0.1 percent.
Next, the figure-eight support with two test devices was run as a torsion pendulum in order to determine the moment of inertia, Is, of the system. Again equation (5) is used, but this
time it is solved for Is, using the known value of S for the fiber. The values obtained (using fiber No.2) for the pendulum with test devices No.1 and No.3 are:
Is No.1 type devices 4.6416 x 10-4 kg-m2 No.3 type devices 1.6871 x 10-3 kg-m2
The symmetric test device, type No.2, was not calibrated (but
could be) since its main use up to the present has been in
evaluating features involving symmetry of the device support and electric wind.
Once these values of Is are known, each pendulum with a
particular type of device can be used to check the calibration of
torsion fiber directly, without the need to reuse the calibration cylinder. This is particularly useful once the devices are
closed in the test chamber and a vacuum is established. No great use of this feature has been made to date, but it allows
recalibration to be carried out in a straight forward manner during a test series, if desired.
It should be noted that the moment of inertia Is only needs
to be known in order to recalibrate the fiber with the test device pendulum. The original fiber calibration value of S can be used to determine the force generated by any particular device
set-up, provided the temperature or resulting axial tension does
not otherwise alter the value of S. Therefore, the fact that the
moment Is was not determined for the No.2 type devices (symmetric
34
~WuVVV1AnM-W NJLWPW


toroids) above, does not preclude determining force values from _
deflections observed during tests with this unit.
The force, F, on each test device is obtained directly using
the following expressions:
e= (x-xo)/2r (7)
Q= se (8)
F= Q/2R (9)
where
xo= zero position on linear scale(no force applied)
x= final position on linear scale(after force applied)
r- mirror to scale distance e= angular deflection of mirror and devices, and angle
of twist of the torsion fiber; radians S= torsional stiffness; N-m2
Q= total torque on fiber; N-m
R= moment arm of each device about fiber axis; m
F- force on each device; N.
These relations can, of course, be combined into a single
expression for' evaluating F. It has been used in reducing test
data under this program.
Device Tests
The propulsion device tests conducted under the program fall with two major groups according to the asymmetrical or symmetrical nature of the device. Secondarily, tests were
4 conducted with the test devices mounted to swing in the complete
35
4A


chamber (a condition referred to "open chamber"), or to swing inside a stationary hollow conducting ball (a condition called ..
"inside ball") The test device, Vere located at the same
position with respect to the chamber in the two cases; the ball was simply mounted to surround the devices in the one case. As
noted previously, the ball was introduced in order to measure diffusion current from the devices uniter test.
Initially several shakedown type tests were conducted in the
open chamber at atmospheric pressure using a modified form of the
asymmetrical test device No. 1. These tests were to examine the
nature and behavior of the overall test configuration. The
devices tested had a small dielectric ring placed around the edge of the plate to help prevent the edge region from becoming a source for electrical discharge. This ring had a cross-sectional diameter of about 0.00416 m (0.188 inch), centered at the edge of
0 the disk. These tests indicated that even this small amount of
k dielectric caused inconsistencies in device performance, much as if the electrical characteristics of the dielectric changed successively during the first few tests. Whether this behavior
would stabilize after many tests is unknown. As a result, the shakedown tests' results are greatly suspect and have been
omitted from this report.
After shakedown, the dielectric rims were removed so the
test devices reverted to the all metal ball and disk form of
device No. 1.
Test conditions and resulting forces measured with
asymmetrical devices numbers 1 and 3 are given in Table 7. The
tests cover a range of pressures from atmospheric to 1.33 Pascal
(10-2 torr). At atmospheric pressure, driving voltages from 0.5 to 6.04 kilovolts were used. Under partial vacuum conditions of
V7 1.33 x 103 Pascal (10 torr) or less, these applied voltages were
generally limited to 0.5, 1.0 or 1.5 kilovolts. This was so the applied voltage would remain under a Paschen type voltage
* 36


breakdown curve, where the applied voltage would not cause arc or
spark breakdown to occur. As the results in Table 7 indicate,
arcing did occur occassionally anyway. It was especially
difficult to obtain force measurements at pressures of
13.3 Pa (10-1 torr) and 133 ra (1 torr), which lie in the region §
where the voltage breakdown is a minimum for the test configuration used. This minimum breakdown voltage was just
above 500 volts. This pressure region, on the other hand, is most important in examining electrical wind which may be present.
It will be noted that Table 7 includes data for only one
successful test (No. 42) with device No. 3, which included the
truncated dielectric cone. Several abortive attempts were made to conduct further tests with this device, but after the first
run the dielectric had changed, and it continued to change. The
total force became less and less with successive runs, until
essentially no deflections were observed.
The final group of tests, numbers 43 through 50, in Table 7 were run under the open chamber condition. The results
of these tests were compared with those made earlier inside the
sphere to determine whether the sphere boundary-affected the
total force measurements. This comparison indicated that the
influence of the boundary on the total force measurement is not appreciably different whether the sphere is present or absei., at
least for the relatively low voltage tlst conditions used. This finding will need further verification for conditions of higher
driving voltages.
Measured values of the average total force acting on the
asymmetrical test device No. 1 are presented in Figure 6 as a
function of vacuum level, with the voltage applied across the device as a parameter. The number alongside each point
designates the corresponding test number from which the data for the point originated.
37
VI


040
in 0 g- in
t 1 N 01 c, 4
04 MU 4 cft a4
0 0 0 0a
0~~ ~ ~ ~ .
r4 9 r - D4f P
%a t
N N M M - -. V. i
%. a f
90 an 00D 05% 0 0 0 0 a@.N
0- ~ i in
V4 0. 0.4
H.A 4 r- .4 .-4 r- H 4 SA m
@ ~~ ~ .-~ r Mn
I 0n ~ o .
- r4 N M H M M *4 *4 *I v4 H r4 r4
.44.
4.3r 0 in 0 0 In N 0 N 0 a el 0 0 -4~I 0 0a H0 n
4 v v 0n in
w 4-1 "1 H
r4 N
H -1 4
Hn H4 M4 H 45
4J 0 P N m 0 0 %0 0 0 m 0 aq
?A~ ~4~ 4Hr ~ i 4H N N M
In I n i n0
*DZ44J~0 0 0 0
.- - ' .4 .4 . . 4 438.4 .


ii in
1! 6
M 00 0 0 0ý 0 40
in In
E- m a t. 0 CI In
.4 H .
r4 C4 m m' Hn 'A .4 H 4
P4 a a 0 0 0 a a 0a
Ii aN1'
rllC;
4
fn m r4 a
4r ~
00 0 to 4 N o 0 9 0 t 0 M-,N r4 AN H H'
fgnn ~%* 0 m 4 en m 60 Sn M m e
z 4 A 4 C4. P4 .H44 r4 H4 Hn m M r4
0f0 0 v mn 0N mag in 4 nC0 in
0a ON N 00r N N 0 0C 0
'4 0 4 a v 0 o a% m a - al H m4 (q .
EP4 e 4 H 0Ar4 0 H
ON N N N N .4
1 0 NNN 000 000000 H4 H4
N 4 4 . 4 .4 . 4 .4 .4
N -410 '-C' - - '4 N- -~ - CN' N
.4 .4 N- N N4 N H4 .4 .l .4 .4 r4 .4 .4 H4
.0 In %4 rý c 0 N .4 Nn m9 IV in 0o In
.4 (0 N 0 0 N4 MO N i0n in, f" m0
t ~O 0 . 4 M .4 .4 . 4 . 4 .
E4' 39


MCI
A4 M V4N
54 i
to +
ii
* c4 0 0 0
MW ~0 %
S• n
m irlN I, en eN e
•4 "0 i n . . N
00
0 0 00 0 0
.4 -4 r- F4 V4 .4M in in i
lo Uo in in kA in
Iv
a0 0
0
0 :.ý f4r M. C----f
V4S A.4 4 1- r4 M.
cc co In en V
Nn qN H
IJ I1.0 r4t
q, q l l P 4 M, 1q4 P.
40O
V'* r 4 r-4 r-4 C4 . 4
z ~ ~~ ~N N
4
CO 0% 0 C40n Mr in w 
04


1i
I
41 .
4 43
E.4U $
"4Vý I
41
0-4 0
E44
.40)
0 .40 4
PN4' 4.1
> 4* 4.1
~~.Mo0 .
p. '0j 00 04.
E-.. a i 0 -A
00 r4 m~ 41 In ~ 00
414


PRESSURE - PASCAL
10"I 1 101 102 103 10 4 105
+2xl. +.020
APPLIED POTENTIAL DIFFERENCE 39
I0J dod.ts
• IAS kilovoits
+lxl67 _ +.010
40'0
o i-'o--"-Z2-21 1 101..
. .. .. io _
2
29 50 4
PRESSURE - TORR
Figure6 aito fMntd oc ihArPesr
942
Ill .1
L4 - 211


The positive direction of force indicated in Figure 6 is
from the ball towards the disk for device No. 1. For the test
results shown, the disk was always at a positive potential with resp~act to the ball, and the ball was at the same earth ground as
the chamber walls. The stationary hollow aluminum sphere placed
around the test device support was held at earth ground potential when leakage currents were not being measured. It was connected
through a 10k ohm resistor to ground during current measurements. Since the observed leakage currents were less that about 1 nA, the potential of the sphere was essentially zero also.
The region on Figure 6 labelled "measurement threshold" is bounded by the approximate positive and negative force levels
that correspond to the combined effect of oscillation tinoisenl in
* the torsion pendulum and readout uncertainties. Forces measured
Xn with the current torsion fiber system are not reliable when their absolute values lie within this threshold region. 1n this region
the forces are considered to be approximately zero.
The average force values given iii Figure 6 include the t contribution of the electrical wind, the ion propulsion effect,
and presumably the Brown effect. It appears that the electrical wind force is appreciable at and near atmospheric pressure, but
falls off to zero by the time the pressure is reduced to about
133 Pa (1 Torr). This pressure limit may be slightly lower for
higher applied potentials. This behavior needs to be explored
further in future investigations. The electric wind region
indicated, however, is consistent with investigations noted in
the literature;( 1 6 )-(1 9 ) apparently this phenomenon has received
little attention in regions of higher vacuum.
The net force on the device in the region where electrical wind predominates is in the directio~n toward the more positive
potential, as indicated by Brown. In the vacuum region, however,
our measurements show this force to be in the opposite direction.
0 Brown ulaimed (as noted in the "Background and Review of Selected
I
43


observations") that under vacuum conditions the force was "in the negative to positive direction."( 4) This discrepancy in the
direction of force has not yet been resolved. If our results
remain valid when further tests and other possible minor effects
are taken into account, then it seems that the force measured
here is not the same one Brown claimed to have investigated.
In the pressure region bel'ow about 133 Pa (1 Torr) the total force level associated with each potential appears to be
independent of pressure. Further, the lack of significant boundary influence in tests run inside and outside the current collecting sphere is apparent. Test numbers 47 and 50 were run with no sphere present, whereas test numbers 9 and 34 were run with devices mounted inside the sphere.
The pressure independence also implies that any specific force level should exist unchanged even at atmospheric pressure, and should add (algebraically) to the electric wind force to give
the total force actually measured. Near atmospheric pressure the
electric wind force, in turn, would be larger than the total measured force shown in Figure 6.
The variation of total force with applied potential difference is shown in Figure 7 for forces on test device No. 1
at atmospheric pressure. The line through the data points is a
regression line fitted to the points. On a logarithmic basis the equation of this line is
Slog F = a + b log V, (10)
or in terms of physical parameters,. it corresponds to the simple
functional form
F = A Vb, (11)
44


10-5 
Air Pressure:
1.01 x 105 Pascal (760 Torr)
"is
10-6'17 " 10-1
2m
0-16
>4
039
41
10-8 1 101
APPLIED POTENTIAL DIFFERENCE - KILOVOLTS
Fgure 7. Viraton of 1murd Force With "MW
45


whore F- measured force v- applied potential difference A- constant - 10a
b- line slope (an log-log plot).
For the data of Figure 6, the regression line becomes approximately
F - 5.21 x 10-8 VI'888, (12)
where F is expressed in newtons and V in in kcilovolts.
Forces in the pressure region at and below 1.33 Pa (I !rorr), which do not seem to vary with pressure, are combined at selected levels of applied voltage and are plotted in Figure 8. Error bars about the mean value points at 0.5 and 1.5 kcilovolts represent the positive and negative values of one standard deviation estimated from experimental data. These values
correspond to ± 20.6 percent and ± 17.4 percent of the means, respectively. Corresponding error estimates for the single points at 0.575, 1.0 and 5.0 kilovolts are each assumed to be +21% of the respective measured force values.
The regression line in this case becomes
F - 3.55 x 10-8 V 0.722, (13)
using the same units as before.
The exponents in each of the preceding functional forms are rather sensitive to the actual data, and hence, to any minor systematic errors in the measured values. The values given,
therefore, should be considered preliminary.
46


Air Pressure:
S1.33 Pascal (1 Torr)
10-0
0F
0
7I
10- W2
10-'1
APPLED
PTENTAL
IFFEENCE KILOOLT
Flp L xdo@Lo~m md am M "ie
47


I•-
Results for tests conducted inside the hollow sphere using
the symmetrical device No.2 are given in Table S. The purposzeof
these tests was to determine if the figure-eight device support
produced an asymmetrical force, and whether such a force, if present, was due to electrical wind. In view of Brown's work, the symmetrical device, itself, should produce no net force due to either the electrical wind or to the Brown effect. These tests were conducted using different combinations of electrical potentials on the two toroids and different grounding arrangements. The electrical conditions for each test run are shown In Table 9.
For all cases the electrical connections between the terminals at the bottom of the figure-eight support (or equally at the top of the stand pipe) and the device elements remains
fixed. The center terminal indicated in Figure 9, is always
connected to the ball end of the asymmetrical device, or to the
toroids of the symmetrical device which fastens onto the same support arm. Likewise, the outside terminal fastens to the disk of the asymmetrical device, or the other toroid in the symmetrical case.
r reference, all the tests noted previously for asymmetrical devices were run using electrical condi 4 -n #1, given in Table 9.
There is a slight asymmetry in the figure-eight device support shown earlier in Figure 4, consisting of the mechanical
support ring for the brass tubes that hold the disks of the
asymmetrical device. This same ring also serves as an electrical U- connection between the two disks and the vertical brass tube
which ultimately connects to the outside terminal at the top of
the stand pipe, as shown in Figure 5.
* 48
S


o 00 04
p4~ ..9,...
9..4
"4 "4,
0 0i @0
0 c:
Go 0o *1
10a 0'
0 0In %a 0 wtG a C * Um en4 in 0 M 4 . + + + I+ + I+ + I+
*4 0 m A M *4 *
mn a0 '0 ' 0
%49.4 .. 9- 9 1 1 9 . 9. 9 97
a W4 4 94 4- '4 A4 4 94 C4 .4. . 4
r40 0 0 0 0 0 0 0 0 0 0 0 0
P40 0 0 0 0 0 0 0 0 0 a 0 0 0
?0 4m0 m40 m0 4 a0 0 a0 0- 0 .
mO Il r4 Mn In In Mn In In H A
0 0 0 0000 0 0000 0
EA A4 .4 A4 .4 H4 H4 m 4 .4 H4 00 0 0 0 0 0 a 0 0 04 0. 0 04
c. r44 A4 m4 H4 .H4 4 r4 4 .4 .4 4n
0N N m N N m N NNN N
M M 60 V .4 M H i5n.4
M q 0 M4 w M4 m 4 N In .4
A
0 mn in V q 4 v4 In In In w0
.4 H4 .4 .4 .4 .4 4 . .4 4 .4 4
04 .4 .
4 r 4 .4
49


Table 9. Electrical Conditions for Symmetrical Device Tests Signs of Electrical Potentials on Terminals at Top ofStape
CONDITION CENTER OUTSIDE GROUD
TERMINAL TERMINAL TIRKINAL
#1 + ~Center(
#2 + -Center (.
#3 + -Outside(
# 4 + Outside (.
TEST NO. ELECTRICAL TEST NO.* ELECTRICAL (TABLE 8) CONDITION (TABLE 8) CONDITION
1iT 8ST 4
2T 3 9T 2
3 T 3 10OT 2
4 T 1 11 T 4 ~
5 T 2 12 T 1
6 T 4 13 T
7 T 4 14 T 3
It is perhaps easiest to grasp the symmetry test results by
examining Table 10, where the measured total force values are indicated, together with test run number, in an array
corresponding to given electrical conditions and potential differences applied to the device.
First, it is noted that the measurement threshold corresponds to a force of about .385 x 10-8 newton. Hence, the
force values under condition #2 at atmospheric pressure are
essentially zero. The values for conditions #1 and #4 combine to
50


Table 10. Array for Comparison of Forces Measured
Using Symmetrical Device
Entry: Measured force, newtons(symmetrical device test number)
ELECTRICAL CONDITION APPLIED POTENTIAL DIFFERENCE
1500 VOLTS 1000 VOLTS 500 VOLTS
"Pressure:
j1.01xl05pa (760 torr)
# 1 - .866xI0-S(lT)
# 2 - .385xi0-8(5T) + .385x10- 8 (9T)
# 3 +3.657xl10 8 (2T)
0 +3. 368x108 (3T)
# 4 +5.197x10-8(6T) +4.138x10-8(7T) 1.059x10-8 (8T)7
Pressure:
1.l9x103Pa(9 torr)
# 1 +.289xi0-8(12T)
# 2 +.577xi0-8(10T)
# 3 ---- +.289xi0-8(14T)
# 4 ---- -. 481x10 8 (liT)
indicate a bias in the force of approximately +2.1 x 10-8 newton.
This corresponds to the polarity used in the asymmetric tests.
The bias causes the force measured under the #1 condition to be more positive (less negative) than would be the case without the bias. This bias is attributed to the figure-eight device support
hardware (probably associated with the asymmetrical support ring mentioned above), since the toroids of the No.3 devices used are quite symmetric and their force contribution should be zero.
51


A similar bias, with the same sign but with smaller
magnitude (about +1.6x10- 8 newton) also exists for the electrical
conditions #2 and #3. In this case the electrical polarity at
the device terminals are reversed.
The most important feature, however, is that at the pressure of 1.19x10 3 Pa(9 torr), each of the forces measured at atmospheric pressure has essentially disappeared. The force
values for conditions #2 and #4 which are slightly in excess of
the measurement threshold, are still considered to be a noise
deflection. This disappearanLe of measured forces implies that the forces observed with the symmetrical devices at atmospheric pressure are caused by electrical wind. In turn, this source of
wind interaction is most likely the figure-eight support, and particularly the metal ring. This slight asymmetry does not seem
to be operative at reduced pressures, so the previous test results with the asymmetrical test devices should be unaffected.
Auxiliary Tests and Considerations
A few auxiliary tests were conducted in attempt to further
define or estimate the importance of factors which could influence the results obtained during tests of asymmetrical
device No.l.
Perhaps the most important of the auxiliary tests were the
ones run to assess the effect of lighting within the chamber on
the torsion fiber, the test devices, and the residual air in the chamber.
Heating of the torsion fiber by incident radiation (especially under vacuum conditions) was examined briefly via
comparison runs for a normal device load with the fiber shielded,
with the radiation turned off, and with radiation turned on as used during device test runs. These several conditions appeared
to have no effect on fiber performance or drift when they were
52
Li


individually applied, or juxtaposed,' over a period of about 30 minutes.
Direct application of radiation to the test devices ina the open chamber indicated no measurable effects of radiation pressure. Inasmuch as the devices were inside the sphere during many of the test runs, radiation pressure effects were not expected to be influential during those particular test runs.4
Radiant heating of air (or gas) in the chamber does cause mild convection currents to appear when the air pressure is near
atmospheric. At reduced pressured of about 1.33x10 3 Pa(l0 torr)I
or less no convection currents strong enough to influence force readings were observed.
Magnetic effects were considered, but were essentially
negated by the symmetrical design of the figure-eight deviceI
support and the current carrying electrical connections and
The figure-eight support design was chosen and irplemented for two reasons; first, to eliminate the need to place a specific dielectric material between the electrodes of the test devices; and second, to balance any residual electric vind forces between support arms so that no net torque would act on the fiber.
Other auxiliary considerations, such as boundary effects, t~m electrical breakdown conditions, use of symmetrical devices, changing electrical polarities, and use of different grounding points have been noted previously and are relevant.
System Errors
Any measurement system is subject to errors, both random and systematic, and the torsion fiber system used in this effort was * no exception.
53


The principal random errors encountered were a slow zero
drift of the fiber and a fast shift of the apparent zero position
of the fiber, usually when an electrical potential was applied to
the devices under test. The slow zero drift was straight forward
to assess, since the zero positions were evaluated both before
and after each test run. The fast shift proved more difficult,
and was apparently not directly associated with the fiber behavior, but with that of the mercury when an electrical potential was applied. This can likely be circumvented by eliminating the mercury contacts from the system.
The main sytematic error encountered was the mechanical drag of the mercury on the alectrical contacts mounted to the oscillating device pendulum. This drag tended to reduce the deflection of an electrically driven device, thereby indicating . that a smaller total force was causing the device to deflect.
This is another key'reason for eliminating the use of mercury for electrical contacts.
A second systematic error is unconfirmed, but is apt to be associated with a changing value of the torsional stiffness, S, of the copper fiber used, with temperature and with applied load
on the fiber. Copper is probably not strong enough or sufficiently stable to serve as a trouble free fiber for this
application; tungsten is believed to be a better choice.
Evaluation
The total force measurements need to ba compared to
estimates of the magnitudes of the electric wind and ion
propulsion effect to determine if any residual force exists.
In this program, asymmetrical devices were used for testing
purposes to emphasize the Brown effect rather than electric wind. Given this selection, the burden of accounting for the magnitude
54


of the electric wind during this effort was placed on
measurements of the total forces generated electrostatically
rather than on calculated results. The geometry of the
asymmetric ball and disk are such that direct calculation of the
electric wind ef'ect for this type of device becomes a
significant three-dimensional axisymmetric boundary value
problem. The major difficulty arises because the electric field
and induced air flow (electric wind) are coupled and are
generally not in the same direction at any point in the
longitudinal plane, which includes the symmetry axis of the
device. While such calculations can be made, they were not
considered to be within the scope of the Phase I effort.
Tt is of interest to note, however, that an analytical model
has been advanced by Chang( 2 0 ) for a simple one-dimensional case of electric wind generation by a device consisting of closely spaced parallel planar electrodes constructed of light wire
meshes. When these electrodes are driven with a DC potential difference of several kilovolts, a thrust is generated that consists of electric pressure and electric wind. The electric
pressure arises from a nonuniform elec'%. i field energy density between the electrodes (the nonuniformity results from space charge effects); the electric wind arises from the induced flow of neutral air molecules. For the sake of simplicity, thp energy
density term has not been separately called out in this report;
instead, it has been included as part of the electric wind effect
itself. Both electric pressure and electric wind cause a force on the device described by Cheng, which acts in a direction towards
the positive electrode, just as observed here and as observed by
Brown.
It is considered desirable in any follow-on effort to
incorporate a device configuration that will allow direct
comparison of test results with Cheng's model. This should help corroborate experimental findings with theory and strengthen the
interpretation of any residual force effects observed.
55
I_- - - - - - * . . * . -- - - - - - - - - - - - -- - - -
0- - - - - - 


The magnitude of the ion propulsion effect :iepends on the size of the diffusion current not collected by the disk or ball of the asymmetrical device No.1, but which passes to and is collected by the conducting sphere surrounding the test devices. The limit of sensitivity of the metering system used to evaluate the leakage current to the sphere was 1 nA. At no time (except during electrical breakdown) during the test runs under a vacuum was a measurable value observed in excess of this current sensitivity limit. Thus lx10- 9 ampere represents an upper bound on the current expected to contribute to the ion propulsion effect.
A general expression for the force, F, expected from such an electrostatic thruster is given by Sutton and Ross (21):
F - 1 (14)
where F= accelerating force, newton
I= propelling current flow, coulomb/sec V= accelerating potential difference, volts A= mass of accelerated particle, kg
e= charge per particle, coulomb.
For a threshold current of 1= 10-9 amp, V= 1000 volts and e= 1.60x10- 1 9 coulomb, equation (14) gives
F= 111.84 newtons.
If the accelerated particles were all electrons, protons, or aluminum ions (single charge) the corresponding maximum forces would be:
56
II


electrons: ,u 9.11 x 10"31 kg
F- 1.07 x 10-13 newton - 1.07 x 10-8 dyne
protons: u= 1.672 x 10-27 kg
F- 4.57 x 10-12 newton - 4.57 x 10-7 dyne
aluminum a- 26.T8xl.66x10- 2 7 kg/AMU = 44.8X10-27 kg
ions: F- 2.37 x 10-11 newton = 2.37 x 10-6 dyne.
The force per device would be one-half of each of these values.
These estimates assume that all the ion current would act
collectively to propel each device in one direction. These force
values per device are at most less than 1/1000 of the force measurement threshold value of .385xi0-8 newton (0.000385 dyne).
Here these estimates for ion propulsion effects are negligible, and the extrapolation procedure suggested earlier is unneccessary. The measured total force values for pressures less than about 133 Pa (1 torr), as shown in Figure 6, are the forces sought.
These electrostatically generated interaction forces are in the opposite direction of the forces claimed to have been measured by T.T. Brown. Hence, these interaction forces will not
be referred to as due to the Brown effect, but will be called
residual forces.
57


CONCLUSIONS
The following conclusions have been reached based on the investigations of the Biefield-Brown effect conducted on this
project:
1. Direct experimental results show that when an electrostatic potential difference is applied between
asymmetrical electrodes of an all metal teat device, a
propulsive force is generated and it acts on this
device.
2. This electrostatically induced propulsive force consists of at least three components: electrical wind, ion propulsion, and a significant residual force.
a.The electrical wind acts in the direction from the
negative to the positive electrode and occurs only for air pressures greater than about 133 Pascal (1 torr),
at least for applied potentials in the low kilovolt
range.
b.The ion propulsion effect (estimated on a theoretical
basis) is completely negligible for the tests
c.The residual force acts (for the tests conducted and
the test device used) in the direction from the
positive to the negative electrode, i.e., opposite to the direction of the electrical wind force. This
W 0 residual force was observed directly and remained
independent of the partial vacuum level over the
approximate range of 133 Pascal (1 torr) to 1.33 Pascal
(10-2 torr). Observations further indicate that this residual force remained constant up to atmospheric
58


pressure and subtracted from the electrical wind to yield the total force actually measured.
3. The electrostatically generated residual forces
measured here act in the opposite direction to the
forces claimed to have been measured in a vacuum by
T.T. Brown. As a result these forces are referred to
as residual forces, and not as forces caused by the
Brown effect.
4. The residual force appears to vary approximately as the
0.72 power of the potential difference applied to the asymmetrical propulsion device tested. This finding is
based on only a few datapoints, and may need revision when more data become available.
5. The measured total force at atmospheric pressure, due to contributions from electrical wind and (presumably)
the residual force, varies approximately as the 1.9 power of the potential difference applied to the asymmetrical propulsion device tested.
6. The magnitude of the residual force appears to be
rather small, but the size, shape and configuration of
the. device tested are not necessarily optimal for
residual force generation, and it may be possible to
generate larger forces with devices similar in overall
size.
1~7. only cursory attention was given to the exploration of
electrostatically induced propulsive forces using devices which incorporate dielectrics in their design.
The few tests which were conducted at atmospheric
pressure using such devices, exhibited problems with
reproducibility.
59


8. The torsion fiber type measurement system employed in
this program needs a few modifications to improve
performanco, but the overall measurement scheme appears
suitable for investigating the fundamental aspects of
electrostatically induced propulsive forces.
60
Ik L2


RECOMMENDATIONS
As a result of this investigation, it is recommended thatmeasurements of propulsive forces generated on test devices by--application of applied electrostatic potentials or fields be
continued. The purpose of this activity would be to further verify the existence of the residual force noted in this report,,, and to develop a more extensive data base which can be used to-more thoroughly explore and characterize its nature. Particular . attention needs to be given to extending the range of test conditions to greater vacuum levels and to higher applied electrostatic potentials. Selected improvements in the overall measurement and test configuration need to be incorporated to facilitate test reproduciblity, more efficient data collection,
* and improved accuracy of measurements.
A
<
N
ri.•
0.


REFERENCES
1. G. Burridge, Townsend Brown and His Anti-Gravity DiscsFats __ pp 40-46, 1956.
2. Rho Sigma, Ether Technology: A Rational Approach to
Gravity-Control, Private Publication, Clayton; -GA-1977,-pp. 27-28, 39, 44-49.
3. T.T. Binwn, "A Method of and an Apparatus or Machine For
Producing Force or Motion," British Patent #300, 311, Nov. 15, 1928. p. 4, line 46.
4. T.T. Brown, Electrokinetic Apparatus, U.S. Patent 3,187,206_
June 1, 1965.
5. T.T. Brown, How I Control Gravitation, Science and Invention. August 1929, p. 374.
6. Office of Naval Research, The Townsend Brown ElectroGravity Device: A Comprehensive Evaluation by the Office of
Naval Research, with Accompanying Documents, W.M. Moore
Publications, Prescott, Az. Sept. 15, 1952.
7. L.B. Loeb, Electrical Coronas, University of California Press, Berkeley, 1965, pp. 402-406.
8. D.E. Gray (ed.) American Institute of Physics Hardbook, McGraw-Hill Book Co., New York, 1957, pp. 2-61, 3-78- 3-80.
9. A. Elliott and J.M. Dickson, Laboratory Instruments: Their Design and Application, Chemical Publishing Co., New York, 1960.
10. P.J. Geary, Torsion Devices, British Scientific Instrument Research Association Report R249, 1960.
11. C. Limb, Sur la determination du moment du couple de torsion d'une susrension unifilaire, Compte Rendus Vol. 114, pp.
1057-1060, May 9, 1892.
12. R.C. Weast (ed.), CRC Handbook of Chemistry and Physics,
55th Edition, CRC Press, Cleveland, 1975.
13. H.A. Pohl, Dielectrophoresis, Cambridge University Press, New York, 1978.
14. A.D. Moore, Electrostatics and Its Applications, John Wiley
& Sons, Inc., New York 1973.
15. W.J. McGregor Tegart, Elements of Mechanical Metallurgy, MacMillan Co., New York, 1966, p. 97.
62


F...... . . . . . .- .. ....
REFERENCES (CONT)
16. A.P. Chattock, Philosophical Magazine, Vol.48, p 401, 1899.
17. A.P. Chattock, Philosophical Magazine, Vol.1, p 79, 1901.
18. A.P. Chattock and A.M. Tyndall, Philosophical Magazine, Vol .17, p 543, 1909.
19. S. Rattner, Philosophical Magazine, Vol. 32, p 442, 1916.
20. S.I. Cheng, Glow Discharge as an Advanced Propulsion Device, ARS Journal Vol. 32, No. 12, pp 1910-1916, December 1962. 
21. G.P. Sutton and D.M. Ross, Rocket Propulsion Elements, 4th
ed., John Wiley and Sons, New York, 1976, p 481.
63


BIBMLIOGRAPHY
AIMA Subcommittee 1971 "UFO Encounter I -- Samle case selected
by the UFO Subcommittee of the A~A,
Astrnaic A&rni en p6-*
Bearden, T.E. and A. Michrovski,
1985 The Emerging Energy science, Planetary
Association for Clean Energy Inc.,
Ottawa, CA
Brown, T.T., 1928 "A Method of and an Apatsor-Machine .
for Producing Force or Nlotion,"0 ritish 
Patent #300,311. Nov. 15, 1928.
Brown, T.T., 1929 "How I Control Gravitation," scinceand~
Invention. August. pp.312-313, 373-375.
Brown, T.T., 1934 "Electrostatic Motor," U.S. Patent * #1,974,483. September 25, 1934.
Brown, T.T., 1956 "Electrical Self-Potential in Rocks," Psychic observer and Chimes. 37(l) Jan.- 
Brown, T.T., 1956 "Electrokinetic Apparatus," The Psycic
Observer and Chimes. 37(l)i Jan. -Mar. pp.
34-40.
Brown, T.T., 1956 "The Fluid Pump," The Psychic Obgney.er
and Chimags. 37(1) Jan.-Mar. pp.54-59.
Brown, T.T., 1956 "How I Control Gravitation," The sychic obsrvyer and Chimes. 37(.l2) Jan.-Mar.
pp.14-19.
Brown, T.T., 1960 "Electrokcinetic Apparatus," U.S. Patent #2,949,550. August 16, 1960.
Brown, T.T., 1962 "Electrokinetic Generator," U.S. Patent #3,022,430. February 20, 1962.
Brown, T.T., 1962 "Electrokinetic Transducer," U.S. Patent
#3,018,394, January 23, 1962.
Brown, T.T., 1965 "Electric Generator," U.S. Patent
#3,196,296. July 20, 1965.
Brown, T.T., 1965 "Electrokinetic Apparatus, U.S. "Patent #3,187,206. June 1, 1965.
* 64


... .--i.
BT.T, 1973 L"Research in Shackles," Information and
Brown,Ltter from Brown
Brown, T.T., 1986 'The Scientific Notebooks of T.T. Brown" William L. Moore Publications & Research.
Vol.l., Vol. 2.
Burridge, Gaston, 1956 "Townsend Brown and His Anti-Gravity
Discs," TE, pp.40-46.
Chattock, A.P. 1899 Philosophical Magazine. Vol 48, p. 401.
Chattock, A.P. 1901 Philoso2hical Magazine. Vol 1, p. 79.
Chattock, A.P. and A.M. Tyndall
1909 Philosophical Magazine. Vol 17, p. 543.
Childress,D.H. 1985 The Anti-Gravity Handbook, Publishers
Network/Adventurer Unlimited Press, Stelle, IL, 1985.
Christenson, Edward and Paul Moller
1967 "Ion-Neutral Propulsion in Atmospheric Media," AIAA Journal. Vol.5(10)., pp.
1768-1773.
Cleaver, A.V., 1957 "Electro-Gravitics: What it is--or Might j
Be," Journal of the British j
Interplanetary Society. VoI.16, pp.84-94.
Coll. Univ. Wisdom 1966 "The Biefield-Brown Effect,"
"Proceedings" College of Universal
Wisdom. Yucca Valley, Calif., Volume 8, Aug.-Oct. pp. 1-4, 4-6, 36-41, 87.
DeSeverskyA.P. 1964 "Ionocraft," U.S. Patent #3,130,945, April 28, 1964.
Gravity Rand, Ltd. 1956 "The Gravitics Situation." pp.3-30.
Hagan, D.E. 1964 "Flying Apparatus," U.S. Patent #3,120,363. February 4, 1964.
Intel, 1956 "Towards Flight without Stress or
Strain...or Weight," ILnterAyi. Vol.11(5), pp.373-375.
65


Intel, 1956 "Towards Flight without Stress or
Strain...or Veiqhtew The Psyahei Observer
and Chimes. 37(11 Jan.-Nar. pp.10-13.
Mead, F.B.,Jr., and J. E. Cox
1976 "Laboratory of T. Townsend, Sunnyvale,
CA," Trip Report, January 24 & 25.
Moore, W.L. 1979 "The Force Fields of Townsetnid-rm." The
Philadelphia Experiment. Fawcett Crest
Books: New York. pp.207-225.
Naval Research Off. 1952 "The Townsend Brown Electra-Gravity
Device," A Comprehensive Evaluation by the Office of Naval Research, with
Accompanying Documents. W.M. Moore Publications, Prescott, AZ, September.
15.
Rattner, S. 1916 PhilosoDhical Magazine. Vol 32, p. 442.
Rose, Mason, 1956 "The Flying Saucer," The Psychic Observer
andCimes 371l) Jan.-Mar. pp.20-27.
Schaffranke, R 1975 "Letter to the Editor--Energy Research:
is America Losing the Ball?" Astronautics
& Aeronautics. Vol.37(l). pp.46-47, 69.
Sigma Rho 1972 Forschung in Fesseln, Ventla-Verlag,
Originalausgabe 1972, VENTLA-Verlag, D
6200 Wiesbaden-Schierstein
Sigma, Rho, 1972 "Research in Bondage, The Riddle of the
Electro-Gravitation." pp. 11-22 (translation)
Sigma, Rho, 1977 Ether-Technology: A Rational Approach to
Gravity-Control
Thomas, W.R. and J. L. Martin
1956 "Evaluation of T. Townsend Brown's
Electrokinetic Apparatus," Psychic
Observer and Chimes, 37(1). Jan-Mar.,
pp. 42-45
66