High voltage flashover of insulating barriers
by Paul Uhlrich
A THESIS Submitted to the Graduate Committee in partial fulfillment of the requirements for the
degree of Master of Science in Electrical Engineering
Montana State University
© Copyright by Paul Uhlrich (1952)
Abstract:
Many investigation of high-voltage flashover in air have been recorded in the literature. Studies of
various standard configuration such as rod gaps, needle gape, sphere gape, needle point to plane, sphere
to plane, etc. are numerous. Studies of corona discharge, insulation creep age properties, and the
potential breakdown of gases have been made under various conditions of pressure, temperature, and
humidity. This paper is offered as an extension of these data to include the study of air breakdown over
insulating barriers.
This paper presents information concerning the use of insulating barriers between adjacent live
terminals or between a live terminal and ground. Curves are included showing the effect of various
barriers upon d-c flashover voltages for altitudes up to 80,000 feet. Similar curves show the effect of
60-cycle a-c flashover voltages for altitudes up to 100,000 feet. High altitude conditions were
simulated in the laboratory by use of a vacuum chamber. All tests were performed using the same
insulating material (black laminated micarta) for the test blocks. This report is intended to determine
the comparative effectiveness of various barriers; no attempt is made here to study the properties of
insulating materials.
The necessity for adequate spacing between terminals of electrical equipment is fundamental.
Sparkover between terminals or to ground normally causes the arcing part to fail with resulting circuit
malfunction. Protection against sparkover is especially important in the design of aircraft electrical
units because circuit failure in flight may be costly. The use of insulating barriers between points at
different potential provides protection against sparking with small sacrifice of space and weight. HIQH TOLTAQX ILASHOVEB Ol INSULATING BABBIEBS
hy
PAUL UHLfiICH
A THESIS
Submitted to the Graduate Committee
In
p a rtia l fulfillm ent of the requiremente
fo r the degree of
Master of Science in E lectrical Engineering
at
Montana State College
Approved:
Iyf Z / J
Head, Major Department
Dean, Graduate/Division
Bozeman, Montana
June, 1952
H37f
Cl h (p h
' V
’
-2TABLB OB OOHTBtiTS
Acknowledgment.........................
3
A b stra c t.............................................................................
4
In tro d u c tio n ...............................................
5
Test Bqulpment............................................................................................ .....
Test P rocedure............................................
15
Preliminary D iscussion.................................................................................. 21
B arriers of Similar Thickness
...............................................................
23
B arriers of Similar Oreepage D is ta n c e .................................................... 34
Various Configurations Using Similar B arriers
..................................
36
Effect of Electrode S h a p e ................................................
43
Effect of Humidity ................................................
44
Effect of Ozone . ............... . . . . . . . . . . . . . . . . . . .
45
High Voltage Tests of A ircraft Equipment. . . . .
47
Conclusions
.............................
............................................................... . . . . . . . . .
49
Appendices
I Reduction of Laboratory Data to Standard Atmosphere . . . .
51
II L ist of Equipment............................................................................. 52
I II Description of Test Electrodes andTest Blocks
....................... 53
IV Photographs........................................................................................ 55
L iterature Cited and Consulted ...............................................................
103028
62
-SAOKSOWLSDOMMT
This thesis was prepared xinder the supervision of Mr. Don V.
Bxner and Mr. Bruce Morgan of the Boeing A ircraft Company. The author
wishes to express appreciation to these men for th e ir invaluable help
and suggestions and to the Boeing A ircraft Company fo r the use of
equipment and technical f a c ilitie s .
ABSTRACT
Uany investigations of high-voltage flashover in a ir have been
recorded in the lite ra tu re . Studies of various standard configurations
such as rod gaps, needle gaps, sphere gaps, needle point to plane, sphere
to plane, etc. are numerous. Studies of corona discharge, insulation
creep age properties, and the potential breakdown of gases have been made
under various conditions of pressure, temperature, and humidity. This
paper is offered as an extension of these data to include the study of
a ir breakdown over insulating b arriers.
This paper presents information concerning the use of insula­
tin g b arriers between adjacent live terminals or between a liv e terminal
and ground. Curves are included showing the effect of various b arriers
upon d-c flashover voltages f o r a ltitu d e s up to 80,000 fe e t. Similar
curves show the effect of 60-cycle a-e flashover voltages for altitu d es
up to 100,000 fe e t. High altitu d e conditions were simulated in the
laboratory by use of a vacuum chamber. All te sts were performed using
the same insulating material (black laminated micarta) for the te st
blocks. This report is intended to determine the comparative effective­
ness of various barriers; no attempt is made nere to study the proper­
tie s of insulating m aterials.
The necessity fo r adequate spacing between terminals of elec­
tr ic a l equipment is fundamental. Sparkover between terminals or to
ground normally causes the arcing p art to f a il with resu ltin g circ u it
malfunction. Protection against sparkover is especially important in
the design of a irc ra ft e le c tric a l u n its because c irc u it fa ilu re in flig h t
may be costly. The use of insulating b arriers between points at d iffer­
ent potential provides protection against sparking with small sacrifice
of space and weight.
-5 -
IHTflODUCTIOH
The major portion of th is paper deals with the breakdown of
dry a ir .
The effect of humidity upon a ir breakdown has been reported in
previous lite r a tu r e . 1 ,8
Some te sts were made with saturated a ir (with
and without moisture condensation on the te s t block) and the resu lts
seem to verify the conclusions of these w riters.
A q u alitativ e study of
the influence of ozone upon a ir breakdown is included.
fo r convenience, breakdown voltages are plo tted here as a
function of a ltitu d e .
All curves are based on the United States standard
atmosphere as established by the National Advisory Oommittee for Aero­
nautics . 3 *4
As seen in figure I , the temperature of the standard atmos­
phere decreases lin early with altitu d e from sea level to 35,000 feet and
is isothermal &t-67°f. from 36,000 feet to 105,000 fe e t.
Since the data
used here were taken at room temperature, correction is required in order
to correlate these data with standard temperature conditions.
At each
T.
Berberlch, L .J., Uosest &.L., S tile s, A.M., Veinott, C.O. STFBCT Of
ALTITUDE ON ELECTRIC BREAKDOWN AND TLASHOVSH Of AIBCBAfT INSULATION.
AISX Transactions, Volume 63, 1944, pages 345-53.
2.
Delerno, M.J, POTENTIAL HBBAKD01 N Of SMALL GAPS UNDER SIMULATED
HIGH-ALTITUDE CONDITIONS. AIES Transactions, Volume 63, 1344,
pages 109-12.
3.
National Advisory Committee for Aeronautics, Technical Report No.
638, 1935. ALTITUDE-PRESSURE TABLES BASED ON THE UNITED STATES
STANDARD ATMOSPHERE.
4.
NACA Technical Report No. 83? and Supersonic Aerodynamics Handbook.
STANDARD ALTITUDE TABLE.
6-
#
# titff#
TjtjS1 tit
SffIStffit . . . . BBSBfPf ffiffiffiffiffiffim
ffiffi ffi S
W\ Stt
ffjp # ffit ffiffiSffi ffi #
:
rs
#
5
SS
Ef
i Tttt Sffi # t
#
ffi 0 ffi k 1 Sg P I
0
ffitffiffi0
m
I
i
:::
# rtlnwl I #-H±h0 I ti::T
P pfj0s
frit
m i .. :I j ;S ffiffi
# 0Trrr# I I f I ffi 0 m # I
tj+
ffi::
j-p
|tit
H
ff
m
±
I
g
E Se M1I s
Iff 0 0 0 # Ititi I
i #
I Sffi # S f
#
I p
B
#
i±a
$
i
I
$
#
=
I
W
I
i
T
tr#
#
#
+tirffi mt I
T
n
T
0 :i
0
S
I M
I
#P
m tit M
HffiI
:'
0 4 I I 0 0 ffiffi
# -M # 0 0 Btt
TTTT
# # tti}
0 S Tti
0 #
I B 1
m# . m tit
P
ffi
0
tti I # # -H::::st Sff 0
pp
.
I
sm S i B I m #
K
ffi
tiff rn I
T
ffir
H
S
ffi
tit
p
I
#
S
m I #
B ffs I 0 I f # 0 0 Sffi
ftiff # iS # ffi ffi 0ffi ffi I
#
#
tH t
mm p # 0
pp
I
ffi
0
0
0
titi
1 1 Iw I
ffi
f&tt
W 0 0 pgp i rn 1 ffi ffi 0 P # I
0 m I g g i # I i BB
#
1 0 0 I I I 1 1 i l I I sSieB#BI 0 0 0 # 0 # B. :s:Bp PB• 0fa SS
I
I l Pti0 # 0
1 p ■ Itti0 pp
I SI I I [ I I
I f I
PBI
tit I
tit
E:B
.
K
g
ttx#
m
i
m
S
e
H
•
P
#
0
#
#
P
0 I
0 tit- titi
#
# # :::: 1
I -tti
0
m 0
+r
titi
Bti
H
f:
Sjf
l
i
ffitp P 0 m SBffi ffi
A# -H-H-#
#
# I # i P #
m
#
g
titi
m
m
I
1
1
20
# 0 P#
0 #1 #0
p
4
0
C
U
i
i
tit
B
ti
Iff*
B
fft
BB
# Iti U.
I
titt I p #
i s B # pti# tffi # 0 # 0 0 0 S :St 0 H
B
L
I
ti
H
B
ti
#
tip: 1 0 m i tititi
BBSI M # # 0 m 0 m
0
TTTT 0 BB--isffi
@ 0ffi H
T
tift p JBf p ffi
tit ttitrn Sffi • SS W Th+S
ffiffiS
ft
ffi
ffiS
&
Sr
ffii ffiffi # 0
ffu
Sffitit # HitTtffHffi
rfTtiB TiBis 111}a
tB
f
Hff
ffiffi
#-HtrffitWHff
I
0 0 ffi ffi 0 I
B i S # ffift 6LlBtiff J: 0 Btr0
I
0 $
I
P
ffi
S
p
&
#5
# I
s#
0
I
0
0
0
0
tit # #0 I
m H W f tp tHm m DS0 BF ffi# is 0 A # ffi # ffi I
#
nI
Sfl gg S i ffi# af
S
ti;“ 0
U
B
: E B t 00
E tip
B
#
titi
T
I
U
y
tir
Br
p Ir:
ffi r:
# 0 0 # #
titi p
% PfHttSffiffi 0 # 0
1
C37ffi
PE
tip
m
pp
p
#
Si=S
A
m
0 0
0 tip
0 ffi
0 -M
Sr X
JffffT
I I Jti t i #
\ titi
0 0 ffi ffi I 0 BB
SS ii #l 0 0
tit 3fcLiiffE tti
p P 1 S ffi # ffi0
i
ffi is
# Bi
Trff
f
gtit I
P
m
!
I
mHE1 0ffir 0 I m 0 1:• I
0 i i ffi I
40 t i t i s
&
W
0 B
Ir;1:
BA
f# # 0 I
i
p
ffi ffi I
f
f
i
1
\
sff
tti
P
0
0#
0 I
# r tit JtH
tIi t i
m
0
ffiffi
Ir
M
P
ffiffi
tit tip 0 0 0 0 T Ptt t i l # # Sft m SS
l+ 4-ltiiff ff ii BHf f i
0 0 ffi S
Er
Sff
ffiff I-P
ffiffi S
ffiSS Hffi P PYIffiffffiffiHSfFffiffi
as Si
HEi+tip
#
rn Ss is
tti 35
-Sfpp
SffiSfftn
Tffi-ffi TTttffi ffi
fffffffff
SSft
p :: '
Sp- j-ti 0 0 0 M m S i # 0 titi
ffi
ffi
#
ffi
$
ffi
I
#
# p j r r ffi ffi ffi ffi II
# ' ffiffffiffi f m
# ffiffititititi
55 fff:
SjtffI S ffiffi
# tir
-tj-H
rnt
ffi
T
E
pp
•
+
ffiffiti
Tff 1
TiM
ffiBtiff4p
#
0 I pTrlt I ffff tit titi' -Hti- TiffIjtJ- P
IT
PBB
B
E
B
ffi
ffi
tti
Bti
ffi I
# m m tir W # I # i a
ffiffi
Titr rn *
ffiffi
ffi
rn
I *
S I 0
ffi I
#
0 ffiffi1 # i m I 4
i i # 0 # Bp
tWi
I p 0 I # I
I
I S I1 1 # 0 I P P # # p
o.v.
I
I SffiS
::’# # W
0
0
#
s
1
I
SS
ffi ffi # I
#
#
H
0
0
0 W
1
01ffi
#
■
I BtiftHE
B 0 is
tin
. . . tiQ
ixt
IG
m # I 0 E# I
ffi I # f f i tf iftii # I I
B tip
tHttiti # 0m ffiffi
r tf Tmr
TTT
# 0 B
titif
a
ttib
Bpf f i BJl t i t i S
ffiS
tti
tf
Bti
B
T
-i
Tt
Sff
#
pp
BBiS URYE
Sff
m
0 1 I f; TFis >ESffi v« r# # -M
# Ss
#
"ffi" p t RAJElfR
r n #
I
1
)
n
-:
M
p # ::: I
ffl
I
S P M # +rnm ffi EH # p 0
:: :
M
TTTT
0::: :
1
1
m tit
p
pB
f
f
i
BAs
m W IB:: $ 0
0
tiff
:J ;
0 mi 0BB
# 1 0 W titi titiBBf
#
ij£
r
. E-jTJ SEl
g
•A fg-C a
g m#
#
tiff
ti
p
■:
~~
TB
tin:
t
i
p
ffff
£
ffiffi
SEffi jiff
ffirr
#
f
f
i
SSS e P m
P p
S e Era 5
# m
5S
S
#
I I
I
....
I
A
m
I
-7 -
data point, the density of the a ir in the evacuation chamber was calcu­
lated from temperature and pressure readings according to the formula!
t o n . l t , of Dr, Air -
^
Density in pounds per cubic feet
3? in inches of mercury
t in degrees centigrade
The a ir density was converted to the corresponding altitu d e by
use of the NACA density-altitude relationship (figure 2) which is calcu­
la ted from the standard atmosphere ta b le s.^
Therefore, at each set of
readings, the density of the a ir in the evacuation chamber is equivalent
to the density of the a ir at the specified altitu d e under standard atmos­
pheric conditions.
See Appendix I for a sample reduction of laboratory
data to the standard atmosphere.
This method of temperature correction is based on the assump­
tion that the potential breakdown of a ir at a given pressure is dependent
upon a ir temperature only insofar as the temperature influences the a ir
density.
The v alid ity of th is assumption may be proved from Paschen1s
law which is fundamental fo r a ll work rela ted to gaseous breakdown.
Paschen 1s Law states that the sparking potential of a gas is a function
1% National Advisory Committee for Aeronautics, Technical Report No.
638, 1935. ALTITUDB-PRESSURE TABLES BASED ON THE UNITED STATES
STANDARD ATMOSPHERE.
2.
NACA Technical Report Ho. 837 and Supersonic Aerodynamics Handbook.
STANDARD ALTITUDE TABLE.
8-
ES±#AAAA A AAAAA
W i fn
StrEmp MP ffffm HFffff
A SEAAA 55
ffff
ffff
H
X
jff
I
A
rS
ffff
ffl
ffl
E
#
I.I
#
#
P
P
P
#
# A A AE
ArfflX
P
B
B fjAtiff
TprFxf:
•#Hr±! ffffffff
I
X
T
R
T
ffff
ffff
#
itt S
±
p
I
A A A SB i A
SS # Afl]A
P SE B HXH i A
#
fffr
T
m
ffl
p
ff=
ffl
fflI
ffl
# aIT P P A
#
#
S
tf
ItP# #
I P Sr A
I Es
I AI A
X
E
2
rr
ittSit #' :■E
X
I P
trSs#
P# i PT
ff-LtfflA A
I A I I fflffl
IjfffflI Im:pA I I I A S IA
mx m
M
I
I
II
#
m
ffl
i
P
ffl
P# I
TiTT#
P TFTP A A
AI
B ffl
B
iff:
±
ttt
m #
P
#
S trt
F
M
M
ffl P P ifffI A P A I A kfflI
M
—
I ffl ffl
::‘ ttB
HaII P 0 m pffl ffl i I ffl
ffl
T
p
r
S
S
ffl
I
AAI
PfrrffffA
frrn
Ip#W
I P ffl ffffffl
S
AI
P T
I
a
ftpffffEIPft# P M
•ttSA B B I APfhtffl A Ta A ST-A
AIiII
# tn SS tW
A A
# GEff#N■ffff
TSIf # # I
i AB
IBP A A A# 55
# iA
P
P
P
AI
p i V
r:
m P Am
B
I
K
*
r
:
A
I
#
P
PP#
I P AM
MB MA I R: A
TF
jT
f ILflf g
M
#
T
te
I
P
A
I
S
ff
W
#
#
B
: A
ffff
B
-Pffff- P A
ti±# M m ‘tiff
Ii
i
P
P
P
P
PP
P I P P AA A I A
# H# Tfft \ P P I S P I I i P I P I A I I A AfP # IR
S fflP A I A fflI FP i | s I
# # ifflP
oI3
nS:E: I P
i fflP P SI I B B fflS
sS
P
I B A I AA i P I
@§ m
SI
PEANI B I ffl P ffl P
P SaP S
M
M
pHis ±fftP XVL ffff
Ef I I
&# pXZft Tfn<-£FTtB
A
fflm
ffl
E
E
tiffffl I B I A
fflaixTxffff
A A AA B
Si
B
I fflI
:
i:
•
A
IL
•
P
S M si g
#
T
A
P
i
:
I
ffl
ffl
A
P
#
iR
i
B
B
A S
# BB
fflm ffl
ffff ::::-4 ffx
t ffl
is# a
ffff
ffl
I
I
ffl
A
P
ffl
A
A
P
P
P
P
B
i
B
B
ffff: : m I # P fflI # # M:ffl I ffl I I M
:TM#
P A AP B I
P# B
B
M
MB
I
ffff
I ffff
I MI : 'XRT : ffffl
I
I
ffl
P
ffl
I P A II I
I ff
-T
Pm I
ifif
P
#
{#
If m I
P B B B B # ES
#
A P
SmI -A
I # Sb A p S"iti■
I
I
I
#
ffXA P A P PIf-P A II
P 0 m s # S ■ P B B B B B IA #P T
A fflSEE
p
BrfL
A P
S
S
P
ffl
ffffl:ffff
M
P
as
A
ffl
M
A
P
A
r
E
tc
r
T
4
E
*
S
A
-fffffffff
AffffA
EsTI
t
jSx
B sP I A ## P A ffl■:Atar
:Iffff
•:: H
fflp
3
T
F
T
I
m
I
fflffl
P
A
A
I
P
R
L
I EM
I
M
NffA
I
ffl
Tffi P I S
I I
is
ffff
A
fflA # Pffff A
#
I
#
##
I
I
I
B
B
ffffffff
H
A
ffl
m
ff
#
i
ffff
SeP Se
P A aTB fflB I P P A S ffffA A I
i-jx# ## # P: :
ffff
p ffff
ffff P I P fflI AI ffli #ffffffli ffff
ffffA ffl i
M#
I PSEP B
±
f
f
±
P
±
f
f
±
tffx
ffff
ffl
#
ffl ffl ffffA B ffffP B I t # ff
ffl ffffP ffff
W
# ffl
# fflffl
ffff
ffix:#
m P p
I :I A fflm
fflffffA P m A fflfflB
P
f
fff
P
rm
B
0
N# p I
# P # ffl Iff: P B I P A P # $ P ffl S #THrA f
§
P
$I p I I I I i: A A B it # I I P I ffl : P I p # M
I s#
# # I -mRTr # :: ffff
ffffP P P S # ffP $ I ffl A a fffflI AfflI P P #
IW #
#
# P
I I
# # iff
I
K
A
#
#
#
P
S #
#
P
B
E
0
P
#
#
S
E
#
P
P
P
P
P
A
S
B
iff:
ffl S:P P B # fflB fflP A I I A P m P I
i #
##
B
#
#
P
p
A P P if BT I AA P $I AfflA AAA fflffl
fflffff
S # I # P TnTffffm #P
SP1 iw ; B fflffl $• m ffl m i
#
I
PP A
#I
#
SI
# I
E
n
## # P
P
Pfflg
A
0
P IO drrti A A A I
a #
m
.
.........
.
f f f f
f f l
ti
f
f f l
.
•
f f l
f f l
f f l
f f l
f f l
f f l
f f f f
f f f f
t
f f l
f f l
f f l
f f f f
f f l
f f l
f f l
-,:3
f f l
.
::
f f l
f f l
f f l f f l
f f l
f f l
f f l
f f l :
f f f f
f f l
f f l
f f l
. _.
-4-4-
.
f f l
f f l
f f l
f f l
BE
f f l
Ertr-
M
MM
:
.ffff
mi
•
:
ffr:
S ff
: r::
••
'
f f iT
ffrT
:' .
E rR
XCff
.....
:.
TF T
;:::
S
MM
ffff
.....
f#
MM
I
m
tm
E ffl
f f l f f l
. XM
f f l f fl
ffl f f l
MM.
A
TRRF-H r m:
T mT
P
P
Th t
a
ffff
f f l
ffff
B A s e
X lff
q ff
R T n frIi'
£f f l
H E f f l
f f l
.! R lf f l T f
A
•
A
f f l Ei
x ffi
EE-R
ffff B B ffff
ffff
T rf
O f f l
X nx f f f f
B
E fff
f f l ffff
flu .
....:
f f l
f f l ffl
f f l :
m
.
:.
:
X jf f
m
E s
A
M
rm ± f f #
tj r ffff
irtt
f S T ttt tif f
#
m
-LUf
I
XE T ±
TRX
ffff
MM
ff fff lf f f l t f f r f f l
jx b - MM
f f l ffl
Mt
f f l
I
.:
fj
■9-
only of the product of the sparking distance and the pressure of the gas
(constant temperature) or, in other words, the mass of the gas between
the electrodes.^
That is , for constant temperature.
Breakdown voltage =
(pressure x electrode spacing)
where the constant Ki is dependent upon the u n its chosen.
The actual
mass of the gas between terminals is proportional to the density of the
gas which is a function of temperature as well as pressure.
A more con­
venient statement of Paechen1s Law would involve sparking distance and
density of the gas as independent variables rather than sparking distance
and gas pressure:
Breakdown voltage = Kg (density x electrode spacing)
Therefore, for a given electrode-barrier configuration, the sparking
potential of a gas is a function only of the density of the gas.
I.
Knowlton, A.S. STANDARD HANDBOOK IOR ELECTRICAL ENGINEERS.
Edition, 1941, Section 4-654, page 405.
Seventh
-1 0 -
TEST EQUIPMENT
The Tacvum equipment used In these te s ts consisted of a bell
ja r , rotary vacuum pump, mercury manometer, drying tra in , and thermometer.
The b e ll-ja r base was made of one-inch black laminated micarta with
six terminal etude sealed into the base.
See Photograph 123030. Higi-
voltage te ste showed the leakage current between etude to be negligible.
The b ell ja r was sealed to the micarta base with a vacuum wax.
The a ir introduced into the b e ll ja r was dried by passing i t
slowly through a two-stage s ilic a gel drying tra in .
drying the a ir was quite effective.
This method of
To prove th is, the en tire evacu­
ation chamber, highly evacuated, was placed inside a cold chamber.
The
chamber was cooled to -?5°Y and allowed to soak for several hours. When
dried a ir at room temperature (TO0F) was released into the chamber, no
v isib le condensation resulted in the b ell ja r .
Thus, the indicated
re la tiv e humidity was le ss than one percent.
▲ standard U-tube mercury manometer was used for most of the
d-c te s ts .
One tube of the manometer was expoeed to the atmosphere and
required correction for barometric pressure.
An accuracy of plus or
minus 2,000 feet is expected for the U-tube manometer fo r altitu d es up
to 80,000 fe e t.
A Wallace and Tleman reservoir-type manometer (type
TA 135) became available for the a-c te s ts .
See Photograph 122029.
With th is instrument, a ltitu d e s may be determined to an accuracy of
plus-or-minus 1,000 feet up to 100,000 fe e t.
barometer with one o u tlet sealed.
I t was set up as a mercury
Consequently, th is manometer system
-1 1 .
vae Independent of the barometric pressure.
A Takk Corporation model 56 high-voltage Insulation te ste r was
need to determine d-o breakdown voltages.
See Photograph 123038.
The
te ste r w ill supply up to 15 kilovolts d-o and Ie provided with a 1.5 and
a 16 kilovolt scale.
Corrected voltmeter readings are considered to be
within plus-or-alnue 3 percent of fu ll-sc a le value.
An ammeter with 100
and 500 microampere ranges is Included to indicate the current supplied
by the instrument.
The output of the te s te r was connected to the te st
terminals without series resistance.
The high regulation of the in stru ­
ment prevented destruction of the te a t blocks except for a condition of
prolonged arcing.
work,
The Takk te ste r has two lim itations fo r th is type of
f i r s t , i t is unable to supply top-scale voltages in cases where
300 to 500 microamperes of corona current flow before arcing occurs.
Second, the d-c voltage waveform supplied when using the 1.5-kilovolt
scale has a strong a-c component.
The peak of the waveform was found to
be 13 percent higher then the voltmeter indication in the upper half of
th is low scale.
Since i t is the peak voltage which determines the point
of breakdown, a ll readings obtained on the 1.5-kilovolt scale are cor­
rected with a multiplying factor of 1.13.
The readings obtained from
th is scale may be distinguished from high-scale readings because only
low-scale readings are recorded with two decimal places.
The circ u it of figure 3 was used fo r the a-c breakdown te s ts .
See also Photograph 122029.
'Two 115/11500-volt potential transformers
were used with the high-voltage windings in series.
A continuously-
-1 2 -
l KA
HO
VOLTS
500
MA
I KA
LR.
L .
INDUCTION
REGULATOR
INDI CATOR
LAMP
s, a S2
a T2
T,
VOLTMETER,
M,
CREST
M2
VACUUM
TUBE
VOLTMETER,
0 - 5
M3
VACUUM
TUBE
VOLTMETER,
0 -0 3 -1 .0
M4
VO L T MET ER ,
Mg
VACUUM
O - ISO
TUBE
0 - 3 0
K V.
VOLTS
TOGGLE
SWITCH
POTENTIAL
TRANSFORMER
PEAK
K V.
PEAK
(WHEN
M A.
(USED
M A.
( USED
USED
WITH
WITH
R2
AS
VOLTMETER,
O- 0 . 3 - I.0
WI TH
R4
AND
C
USED
WITH
METER
M,
<n
T ERMI NAL S
A
AND
B
USED
WITH
METER
M2
(S2
122029
FIG.
CIRCUIT
DIAGRAM
FOR
AND
APPENDIX
II
CM
A
PHOTOGRAPH
SHUNT)
RMS
TERMINALS
SEE
Rj )
FOR
AS
C L OS E D)
OPEN)
DETAILS.
3
A-C
TEST
SHUNT )
EQUIPMENT
-1 3 -
varlable range of voltage up to 30 k ilo v o lts was obtained by using an
Induction ragalator In the primary c irc u it of the transformers.
A 30-
PAgohR rerlee resistance (Ri) was used to prevent destructive arcing at
the te st terminals afte r breakdown.
This required a very low-current
measuring device In order that the drop across the current lim iting re­
sistance be a minlnraa.
A General E lectric Type A3 Orest voltmeter (M^) was used to
measure the alternating voltage applied to the terminals.
46869-B for panel view of th is meter.
See Photograph
The maximum voltage which mey be
measured by th is instrument is 30 k ilo v o lts and its nominal Input imped­
ance at 50 cycles Ie about 250 megohms. L0Cal calibration, of the meter
was not possible, but an accuracy of plus-or-minua 3 percent of fu llscale value le claimed by the manufacturer.
The meter le battery-operated
which permitted the "ground" terrains! (end case of the meter) to have a
high potential with respect to actual ground p o ten tial.
Tha meter was
insulated from ground and isolated physically to avoid human contact
while in use.
Readings below 5 kilovolts on th is raetsr were not considered to
be su fficien tly accurate for these te ste , so a second metering circu it
was prepared to obtain more accurate readings* in th is range.
The te st
c irc u it was arranged so that one or both of the transformers might be
used.
Transformer Tg could be removed from the circ u it by opening switch
@2 sn<J grounding one terminal of the teet block. With one test terminal
grounded, a potential divider circ u it (S3 ) and vecuvra tube voltmeter («2 )
-1 4 -
oould be calibrated to meaeure voltage.
The divider c irc u it consisted of
52 10-megohm re sisto rs in series and effected a 1000 to I step-dovn ra tio .
After calibration th is arrangement was known to be accurate within plusor-minus 3 percent of i t s fu ll-sc a le value.
leads and divider resistances were shielded.
The vacuum tube voltmeter
The current, supplied by the
transformers was determined from the reading of voltmeter M3 across Rg.
The current flow to the te s t block was measured by voltmeter Mp across £4 .
Appendix II contains a l i s t of the equipment used.
The te s t blocks were made of one-inch black laminated mlcerta
with from two to four insulating b arriers milled into each block.
See
Photographs 122031, 122032, end 122033. With th is method, the b arriers
are an Integral part of the block.
Each b arrier was designed ao that the
breakdown path around the end of the b a rrie r was longer than the break­
down path over the top of the b arrier.
over the top of the b a rrie r.
Therefore, a ll arcing occurred
A l i s t of the te st blocks used is given in
Appendix I I I .
The brass electrodes were chosen as a compromise of the various
configurations used in a irc ra ft e le c tric a l equipment.
cut from one-elg^th-inch brass stock p la te s.
one-fourth inch wide and four inches long.
were one-fourth inch wide and one inch long.
All terminals were
The ground terminals were
The high voltage terminals
On most of th e te s t blocks,
flat-head screws and round-head screws were used at opposite ends of each
b a rrie r.
All terminals were cleaned with fine-grain emery paper before
they were mounted on the te s t block.
The terminals and b arriers were
washed with lacquer thinner before they were used in a te s t.
TEST PBDCEDDBE
The following procedure was used in preparing the vacuum cham­
ber and te s t block fo r each te s t.
The chamber was evacuated for several
minutes u n til the absolute pressure was about 3mm. of mercury.
Then dry
a ir was slowly released into the chamber fo r a time, shut off for re­
evacuation, and released into the chamber a second time.
After a third
evacuation of the chamber, a valve in the pump line was closed and the
vacuum pump turned o ff.
Beadings were taken f i r s t at low pressure (hi
a ltitu d e simulated) and repeated in steps for higher pressures (lower
altitu d e s) by le ttin g in additional dry a i r a t Intervals.
This procedure
reduced any error due to leakage of the vacuum system by firmly seating
the b e ll ja r before data is taken.
I t also assured the presence of some
fresh a ir , not yet ionized by breakdown, a t each data point.
The method used fo r saturating the a ir in the chamber was more
d iffic u lt to control.
water.
The drying tra in was replaced by a b o ttle of warm
With minimum pressure in the chamber, a ir was released slowly
through the water into the chamber.
The incoming moist a ir was warmer
than the b e ll ja r so that some condensation would occur on the Inside sur­
face of the ja r .
The system was then allowed to reach equilibrium at
room temperature.
The resulting humidity was considered to be near
saturation provided that some condensed moisture was s t i l l v isib le in the
b e ll ja r.
By proper control of th is operation, moisture could be con­
densed on the b arrier and terminals under te s t or the te s t block could
be l e f t v isib ly dry with saturated a ir in the bell ja r .
Bo experimental
16-
determination of the re la tiv e humidity of the a ir was attempted here.
I t is readily conceded that the procedure need for the saturated a ir
te sta could he refined considerably, but the resu lts desired here were
qualitative in nature and intended only ae corroboration of sim ilar work
already published.
In a l l te sta, the breakdown voltage was considered to be the
minimum potential required for a v isib le arc to jump repeatedly from
terminal to terminal.
However, corona-starting voltages were recorded in
many cases in addition to the voltage required for complete a ir break­
down.
Breakdown sometimes occurred at voltages 6 to 10 percent lower
than the value recorded, but these were considered to be false points.
These points may be accounted for in p a rt by an analysis of the Takk
te s te r c irc u its .
The d irect voltage is obtained from a 0.02 microfarad
condenser which acquires additional charge each half-cycle from the
secondary winding of a high-voltage transformer. Beatifying action of
two type 60.8 cold-cathode tubes prevents discharge of the condenser back
through the transformer winding.
I f the te ste r voltage is held at a
certain value, the output d-c voltage w ill be steady because the condenser
w ill be charged to the peak value of the ar-c wave supplied by the trans­
former.
While the output voltage Ie being increased to obtain a point
of data, the output waveform w ill include some a-o rip p le.
The peaks of
the a-c component cause breakdown, while the d-o meter indicates an
average voltage.
To obtain a reading, i t was necessary to increase the
voltage to the point of arcing, note the approximate breakdown voltage.
-1 7 .
and then approach the point more slowly to obtain a more accurate reading,
fo r th is reason, the breakdown voltage was chosen as the lowest voltage
which would cause repeated arcing a t the te s t block with no change in
the voltage adjustment.
%ch point was determined several times to ver­
ify the fin a l breakdown point.
The rate of increase of voltage was not accurately controlled
in th is work.
The d-c voltage was controlled by manual adjustment.
The
rate of increase of voltage for fin al adjustment was no more than 100
v o lts per second.
The &-c voltage was controlled by manual adjustment
of the induction regulator.
As an estimate, the rata of increase of a-c
voltage was 300 to 500 v o lts per second.
The relative importance of th is
factor was investigated fo r the &-e equipment by "soaking" the test con­
figuration at a voltage slig h tly under the value required for breakdown,
Hesults indicate that the effect of the time of voltage application is
too small to be significant in these te s ts .
In addition, i t is consid­
ered here that voltages high enough to cause breakdown in a irc ra ft
systems w ill most lik e ly be transients or peaks of re la tiv e ly short
duration.
Howell*- has shown that the surface condition of tlie electrodes
is an important factor in the study of a i r breakdown.
Houghened elec­
trodes lower the breakdown voltage considerably in comparison to highlypolished smooth electrodes.
I.
The breakdown voltages obtained with smooth
Howell, A.H. BREAKDOWN STUDIES OZ COMPRESSED OASES. AIEB Trans­
actions, Volume 58, 1939, May Section, pages 193-206.
polished electrodes are too high fo r application to p ractica l electrodes.
In addition, Howell point* out that nor© consistent re su lts can he ob­
tained a fte r the electrodes arc conditioned.
The conditioning process
consists of repeated arcing between electrodes to bum off small points
or ridges which may influence the point of breakdown.
As mentioned pre­
viously, the electrodes used in this investigation were cleaned with
fine-grain energy paper and washed In lacquer thinner to remove any sur­
face film ,
'fhe electrodes used here were not highly polished and screws
were used in the electrodes to obtain a more p ractical electrode con­
figuration.
Hach configuration was arced several times before actual
te s t in an attempt to condition the electrodes.
I t is believed that the
re su lts obtained here w ill be d irectly applicable to many practical
electrode-barrier configurations.
Before b arrier studies were begun, preliminary te sts were made
using teat block B-I shown in Photograph 122032. These te s ts afforded
the author a period of fam iliarization with the te st equipment and pro­
cedure while attempting to correlate th is data with corresponding values
recorded in the lite ra tu re .
le s t Block B-I was assembled to compare with
the electrode configuration used by Berberich, Moses, S tile s, and Veinott.^
The comparison of the two sets of data is shewn in Figure 4.
The break­
down values obtained here were slig h tly lower than those reproduced from
the lite r a tu r e , but since the curves were on the safe side and reasonably
U
Berberich, L .J., Moses, O.L., S tile s, A.M., Veinott, C.G. smOT OF
ALTITUDE UK ELECTRIC BREAKDOWN AND FLASKOTER OF AIRCRAFT INSULATION.
AIES Transactions, Volume 63, 1944, pages 345-53.
-1 9 -
-2 0 -
con ei ate n t, the etudy of h arriers was in itia te d without farth er pre­
liminary work.
-2 1 -
y m ,IMiaAlY DI SOU33103
In the discuseion w>iloh follows, the toraa "oreepa^e distance"
«md "clearance dletaaoe" w ill be used according to the following defi­
n itio n s.
Oreepage distance 1 » the shortest distance from electrode to
electrode neaeured along the surface of the insulating m aterial.
Glear-
anco distance Is the sliortest path la a ir from electrode to electrode . 1
See Table I fo r diagrams.
Kaay factors are known to influence the breakdown of gases under
d ie le c tric stre ss.
In addition to the electrode spacing and the gas den­
sity , the breakdown voltage is dependent upon the gas I ts e lf; the elec­
trode metal; the surface condition end the configuration of the electrodes
and b arrier ( i f any); the frequency, waveshape, and rote of increase of
the voltage used.
I t Ie lik e ly that the material of the Insulating bar­
r ie r w ill also have some effect upon the breakdown of the gas surrounding
the b a rrie r.
All of these factors are controlled by the choice of te at
equipment, te s t procedure, and the design of the te s t blocks.
Slie te s t blocks were designed such that four conditions of
electrode and b arrier configuration might be studied on a comparative
basis*
(1) A series of b a rrie rs having the sazno thickness but different
height.
BCLeotrodee were mounted flush against the b arriers.
(2) A series of b a rrie rs having the same calculated creepago dis­
tance but d ifferen t thickness and height.
I.
Slectrodes were
A fZ lIGAB STAJDAkD DJBiliii'ZIOiS 0? SILZOfilIOAL TSiRKS, AIKS.
2 0 .1 0 .4 1 6 -4 2 0 , page 8 5,
S e c tio n "
-32mounted flu s h a g a in st fcho h s r r lo r ,
(3 ) A Bariea o f b a r r ie r s Imviiig tha aaiae th ick n e ss and h e ig h t.
I le o -
trod es were mounted a t various p o s it io n s p a r a lle l to th e h a r r ie r .
(4 ) A study o f th e e f f e c t o f d if fe r e n t e le c tr o d e shapes (th e use o f
ronad-head and. fla t -h e a d machine screws) u sin g e given h a r r ie r .
-3 3 -
BABBIERS OE SIMILAR THICKRESS
All Tiarriere of the f ir s t series were l / l 6 inch thick with
height varying from 1/8 inch to 3/4 inch by increments of 1/16 inch.
The brass term inals were mounted tig h t against the b a rrie r in each case.
JPour of the te s t blocks of th is series are shown in Photograph 122031.
figure 6 shows the d-c breakdown voltage vs. altitu d e curves obtained
from these b arriers.
In order to compare creepage distance over b arriers
with plane-surface creepage distance, d-c breakdown te s ts were made using
block B-7 shown in Photograph 123032.
This block is equipped with a
movable ground electrode so that the surface creepage distance could be
adjusted as desired.
The d-c breakdown voltage vs. a ltitu d e curves from
th is data are shown in figure 6 .
These plane-surface breakdown curves
compare favorably with the corresponding d-c curves seen above for the
f i r s t series of b arriers (figure 6 ).
The maximum difference between
corresponding curves is apparent at 3/16 inch creepage distance.
The d-c
breakdown voltages fo r the 3/16 inch b a rrie r are from 6 to 26 percent
lower than the corresponding d-c breakdown voltages for 3/16 inch planesurface
creepage.
fo r longer creepage distances, comparative breakdown
voltages over the b arriers are slig h tly higher at low altitu d e s and some­
what lower a t high altitu d e s than those for plane surface creepage.
It
is well to note here that creepage distance and clearance distance are
identical for these configurations.
Thus a creepage (or clearance) dis­
tance measured over a b a rrie r (with electrodes close to the barrier) w ill
require nearly the same direct voltage fo r breakdown as a sim ilar plane-
-2 4 -
25-
26 -
-2 7 -
surface creepage (or clearance) distance.
Jor example, terminals ( l /8
Inch high), which must he separated 7/16 inch to meet breakdown require­
ments, w ill he adequately insulated by a 1/16 Inch h arrier which is
5/16 inch high.
Comparison of the d-c and a-c breakdown voltage vs. altitu d e
curves in Jigore 6 and 7 shows that the 60-cycle a-c peak voltage re­
quired to cause arcing over a given h a rrie r is nearly the same as the
d-c voltage necessary fo r sparkover under the same conditions.
Therefore,
i t is concluded that d-c and a-c breakdown voltages are sim ilar for the
corresponding clearance distances used in th is f i r s t series of h arrier
configurations.
The breakdown voltage vs. clearance distance curves shown in
Jigure 8 , 9, and 10 were p lo tted from the breakdown voltage vs. altitu d e
curves in Jigures 5, 6 , and 7 respectively.
The breakdown voltage vs.
clearance distance curves present the same data as the original curves,
but lend themselves more readily to detailed analysis of the relative
advantages of increasing clearance distance to meet breakdown voltage
requirements.
Examination of these curves reveals that as altitude in­
creases, the slope of the curves decreases.
This means, for example,
that a given increase in the clearance distance in an attempt to meet
voltage requirements is much less effective at higher a ltitu d e s.
This
phenomenon may be demonstrated most effectiv ely with the curve in Figure
11.
In th is curve, the slopes of the respective curves in Figure 10 are
p lo tted against altitu d e.
This curve shows the increase of breakdown
-2 8 -
8 0 OOO
-
OiSTANC t H-
•4
3D 0 )0 0
50 000
6 0 0'
-3 0 -
CLEARANCE ."!dietahci[!m g 5 E S a w #
VOLTAGE
INCH
ijRAN^E
-Z l-
REPRESENT
SLOIhE
FllSUR
SLOPE ^
OF
AN
INCREASE
-3 2 -
voltage which may be expected for each additional l / l 6 inch of clearance
distance more than the f i r s t 1/16 inch,
from design considerations, i t
is the higher-altitude portion of th is curve that is most significant,
'Ihe knee of the curve at 50,000 feat marks the lower lim it of th is
c r itic a l region.
To illu s tr a te th is point, consider the following example. At
100,000 fe e t, tne breakdown voltage is 700 volts peak for 1/lG inch
clearance and 1200 volts peak for 15/16 inch clearance (figure 10).
This
represents a 500-volt increase of breakdown voltage for a 14/16 inch in ­
crease of clearance distance or 35 volts per sixteenth inch increase of
clearance distance.
'This la tte r value may be read d irec tly from the
slope vs. altitu d e curve (figure 11).
In th is example, the voltage re­
quired for breakdown is raised only 71 percent by increasing the clear­
ance distance to 15 times i t s original value,
As the operating altitu d e of modern a irc ra ft steadily increases,
the design of a irc ra ft e le c tric a l equipment becomes more d iffic u lt.
Un­
fortunately, the insulating properties of a ir become dangerously ineffec­
tive at the low densities found at high altitu d e as illu stra te d in the
preceding example.
The necessary increase of creepage distance to meet
breakdown voltage specifications may become impractical because of space
lim itations.
The use of hermetically-sealed equipment solves th is problem
inside the seal, but outside connections may be surrounded by the ra rifie d
a ir .
In any case, the use of insulating b arriers seems to be the most
feasible method of Increasing creepage distance with negligible loss of
space
-3 3 -
I t is known, however, th e t th e in e u le tin g q u alitie s o f a ir
never disappear e n t ir e ly .
For uniform f i e l d s , a minimum breakdown v o lta g e
o f approxim ately 550 v o lt s peak e x i s t s fo r a ir a t come c r i t i c a l product
o f a ir d e n s ity and e le c tr o d e sp a cin g .^
For any product o f d e n s ity and
spacing o th er than the c r i t i c a l product, th e breakdown v o lta g e w i l l be
g r e a te r then 550 v o lt* peak:.
For uniform f i e l d c o n d itio n s , then, t h is
c r i t i c a l product d e fin e s th e worst p o s s ib le co n d itio n o f p r o te c tio n from
breakdown in a ir .
For non-uniform f i e l d s , i t i s probable th a t a p o t e n tia l
somewhat l e s s then 350 v o l t s peak would overs tr e s s a ir along a c e r ta in
p ath such th a t breakdown might occur.
However, i t i s w e ll to r e c a ll that
a minimum breakdown p o t e n t ia l does e x i s t , fo r any g iv en e le c tr o d e con­
fig u r a t io n , which i s independent o f the a ir d e n s ity .
F ortu n ately, the in s u la tin g p r o p e r tie s o f s o lid in s u la tio n are
not m a te r ia lly a ffe c te d by a lt i t u d e .
I f term in als norm ally exposed to
th e a ir were sea le d in p l a s t i c or were p a in te d h e a v ily w ith an in s u la tin g
m a te r ia l a f t e r i n s t a ll a t io n , adequate p r o te c tio n a g a in st breakdown a t
extreme a lt it u d e might be obtained w ith r e la t iv e l y short creepage p a th s,
C
Knowlton, A.S.
STANDARD HANDBOOK FOH ELECTRICAL ENGIHEEHSl
M it lo n , 1941, S ectio n 4-554, page 405,
Seventh
-3 4 -
BARHIERS OE SIMILAR CREEPAOB DISTANCE
A ll t e s t s S lecu ssed eo fa r mere made with, h a r r ie r s o f uniform
th ic k n e ss .
is
At t h is p o in t the q u estion a r is e s whether breakdown v o lta g e
a fu n c tio n o f the th ick n e ss o f a h a r r ie r .
h a r r ie r s was designed to study t h is q u e stio n .
The second a e r ie s o f
See Photograph 122033.
The e le c tr o d e s were mounted t ig h t a g a in st the h a r r ie r s a s fo r the f i r s t
s e r ie s so the terms Mcreepego d istan ce* m l "clearance d istan ce" me;*
again he used in terch an geab ly.
ware chosen
COHLion
so
Tiie s p e c i f i c dim ensions o f the h a r r ie r a
th at each e le c tr o d e -h a r r ie r co n fig u ra tio n
clea ra n ce d istan ce o f 15/16 in ch .
dim ensions.
w o u ld
have a
See Apoandiz I I I fo r b a rrier
Data obtained from t h is b lo ck could than be compared w ith
the d ata recorded fo r th e l / l 6 inch wide by 9 /1 6 in ch -h ig h b a r r ier which
hao
ix
sim ila r clearan ce d is ta n c e .
The s im ila r it y o f clea ra n ce d ista n ce
fo r th ese c o n fig u r a tio n s ia dependant upon the use o f term inal a 1 /8 Inch
high counted flu s h a g a in st the b a r r ie r .
The c lo s e resemblance o f the breakdown v s . a lt it u d e curves fo r
th e s e b a r r ie r s i s shown in Figure 12.
I t i s c le a r ly asen th a t d-c break­
down v o lta g e i s not a fu n c tio n o f b a r r ie r th ick n ess fo r t h is co n fig u ra tio n .
Tlio corresponding curve o f 5 0 -c y c le a -c v o lta g e v s . a lt it u d e fo r the 1 /1 6
x 9 /1 6 in ch b a r r ie r i s shown aa a dashed li n e in Figure 1 2 .
I t means
reasonab le to assume from p revious r e s u lt s th a t 6 0 -c y c le a -c breakdown
v o lta g e are a ls o independent o f b a r r ie r th ick n ess fo r the co n fig u ra tio n
u se d .
35-
-3 6 -
VARIOUS OONTIQURATIOSS USING SIMILAR BARRIERS
The third series of b arriers were a l l 1/16 inch thick and l /3
inch higfc.
The purpose here was to determine the most accurate method
for calculating the actual length of the breakdown path for electrodeb arrier configurations in which the electrodes were set back from tne
b a rrie r.
The nine configurations used in th is series are illu stra te d and
labelled in Table I .
(Block B-8A in Photograph 122032 illu s tra te s four
of these configurations.)
Three different methods of determining the
length of the breakdown path between electrodes are lis te d in Table I.
The creepage and clearance distances have been defined previously.
It
is evident that the creepage distance is greater than the clearance dis­
tance for any configuration in which one or both electrodes are set back
from the b a rrie r.
The method using the shortest perpendicular distance
is included as a compromise between the creepage distance and the clear­
ance distance.
Rach distance lis te d in Table I is referred to the creep-
age distance of configuration No. I on a percentage basis to f a c ilita te
comparison.
The curves for the nine configurations in Table I are shown in
Figures 13 and 14.
Curve Mo. I is drawn in both figures to provide a
common reference for the two sets of curves.
At a given a ltitu d e , the
breakdown voltages of the nine configurations should vary with the actual
length of the breakdown path.
The breakdown voltage fo r a l l configura­
tions i s within 20 percent of the breakdown voltage of configuration
No. I a t any altitu d e up to 50,000 fe e t.
Tor some a ltitu d e s above 50,000
fe e t, the breakdown voltage from curve 9 is as much as 145 percent of the
-3 7 SHORTEST
CONFIGURATION
NUMBER
CREEfj AGE
CLEARAN CE
DISTANCE
DISTANCE
PERCENT
INCHES
OF NO.
I
m
e
INCHES
PERCENT
OF
NO. I
0 .8 1 3
100
0 .8 1 3
100
1 .0 6 3
131
0 .8 3 2
102
0 .9 3 8
I 15
I
146
0 .8 8 8
109
1 .0 6 3
131
1 .0 6 3
131
0 .8 3 2
102
0 .9 3 8
1 15
1 .3 13
162
0 .8 5 2
105
1 .0 6 3
131
1 .4 3 8
177
0 .9 1 7
11 3
1 .1 8 8
146
1 .1 8 8
146
0 .8 8 8
109
1 .0 6 3
131
1 .4 3 8
177
0 .9 1 7
113
1 .188
146
1 .5 6 3
192
0 .9 6 3
11.8
1 .3 1 3
162
m
3
m
NO. I
- I
I +
2
PERCENT
OF
DISTANCE
100
0 .8 1 3
I
INCHES
I
PERPENDICULAR
188
i
4
s rm
m
I
5
m
e
e m
4
a m
6
m
I
7
4
n r
- 1 00
I
4
4
T
8
m
E
m
9
m
METHOD
m
USED
C A L C U L A T IN G
DISTANCE
n—I
--L — —
—
----- 1
I----- -L--------
T A B L E
C O M P A R IS O N
THE
/
FOR
OF
E F F E C T IV E
THREE
[ V -L - - J
I -----1
L - V l_ _ _ _
_ r : r d
[
METHODS
D IS T A N C E
JI
OF
BETW EEN
C A L C U L A T IN G
ELECTRODES
>
:4 m m .
Aim
39-
-
40-
correttpondin^ voltage from curve I ,
The variation of clearance distance in column 2 of Table I is
seen to correspond well with the variation of breakdown voltage for the
major portion of the a ltitu d e range.
The use of the creep age distance
or the shortest perpendicular distance cannot account fo r the relativ e
proximity of the nine curves.
Therefore, the actual breakdown path over
a b arrier may be expressed with reasonable accuracy as the clearance
distance.
The p ractical advantage of moving the electrodes away from the
b arriers is slig h t.
The resulting increase of clearance distance ie
negligible compared to the loss of space in the design.
I t is evident
then, th at the most p ractical design of high-altitude a irc ra ft elec tric a l
equipment, in terms of space and weight considerations, ie the use of
b a rrie r configurations with terminals mounted close to the b arriers.
One additional observation was made during th is series of te s ts
which may be of in te re st here.
was being tested.
The configuration illu s tra te d in Table II
I t was discovered that arcing would occur at the same
breakdown voltage from the high voltage (+) terminal to eith er ground (-)
terminal.
This phenomenon was observed fo r a range of altitu d e from sea
level to 50,000 feet.
The arc from A to 0 terminated on the outside cor­
ner of C as shown by path n in Table I I .
Since the breakdown voltage is
the same, the actual length of the breakdown path must be sim ilar over
eith er b a rrie r.
The clearance distance from A to C (path m) does not
apply here because the arc did not follow th is path.
The distance along
t
-4 1
SHORTEST
BREAKDOWN
PATH
FROM
A
TO
B
FROM
A
TO
C
ALONG
PATH
"M "
FROM
A
TO
C
ALONG
PATH
"n"
FROM
A
TO
C
ALONG
PATH
IS
NOT
t THlS
FOR
A
COMPARISON
CREEPAGE
CLEARANCE
PERPENDICULAR
DISTANCE
DISTANCE
DISTANCE
INCHES
PERCENT
OF A-B
DISTANCE
INCHES
PERCENT
OF A - B
DISTANCE
INCHES
PERCENT
OF A - B
DISTANCE
1.688
I OO
1.082
I OO
I .438
IOO
L I 88
70
0.888
82
1.063
74
0.963*
89
1.082*
IOO
..p .I
CLEARANCE
WITH
D IST ANCE.
CLEARANCE
TABLE
COMPARISON
FOR
A
OF
VARIOUS
CERTAIN
TEST
IT
D IS T A N C E
IS
IN C L U D E D
FROM
A
TO
Il
BREAKDOWN
PATHS
CONFIGURATION
HERE
B.
-
42-
path n is 11 percent shorter than the clearance distance from terminal ▲
to terminal B.
Therefore, the actual length of the breakdown path must
be measured along a path sim ilar to path p.
This is supported by the
fact that brush corona was observed at the outside corner of terminal C
extending out a short distance along a path similar to path p.
This te s t is described here to point out that the actual break­
down path may be dependent upon the exact configuration employed. No
single method of determining the actual length of breakdown path will be
accurate for a ll electrode configurations.
The use of the clearance dis­
tance w ill be reasonably accurate in most b arrier design.
If the clear­
ance distance is incorrect in a given case, the resulting design must be
on the conservative side since the clearance distance is the shortest a ir
path from metal to metal.
-43-
EiFECT Oi KLECTfiODE SHAPE
The fourth condition of electrode-harrier configuration em­
ploy# a slig h tly modified electrode shape.
The conclusions drawn to th is
point are based upon electrodes equipped with flat-head screws counter­
sunk level with the top surface of the electrode.
In th is section, round-
head screws were used in place of the former for purposes of comparison.
The resu lts of th is comparison show that allowance must be made for any
p a rts of the electrode which would tend to decrease the clearance dis­
tance.
In the p ractical design of relays, terminal s trip s , switches,
e tc ., such allowance must be made fo r lugs, washers, and screw heads
which li e above the plane of the metal connecting strap.
That is , the
clearance distance should be determined with a ll connections assumed to
be in place.
-
44-
SJPFKCT OI humidity
Throughout the testin g of the various electrode-harrier con­
figurations, sotce factors were studied which are independent of the con­
figuration used.
Some data were obtained to determine the effect of a ir
humidity upon breaJcdown.
There is no significant difference between the
d-c breakdown voltages obtained with saturated a ir or with dry a ir pro­
vided no moisture is condensed on the te s t block.
Results obtained with
condensation of moisture on the electrodes and b arrier were not con­
sis te n t.
However, breakdown voltages are significantly lowered by-
moisture condensation on the b arrier under te s t.
In th is case, the
breakdown voltage rise s with continued sparking because the heat of the
arc rapidly removes the condensed moisture in the region of the te s t.
The resu lts agree favorably with previous work of th is nature.
1.
Berberlch, L .J ., Moses, O.L., S tile s, A .M ., Veinott , 0.6. EFFECT OF
ALTITUDE OS ELECTRIC BRBAKDOWE AJiD FLASEOVER CF AIRCRAFT INSULATION.
AIEE Transactions, Volume 63, 1944, pages 346-63.
2.
Delerno, M.J. POTENTIAL BREAKDOWN OF SMALL GAPS UNDER SIMULATED
HIQH-ALTITUDFi CONDITIONS. AIEB Transactions, Volume 63, 1944,
pages 109-12.
-
45 -
m ’^c$ a t czois
The effect of ozone concentration was studied on a qualitative
ta s le .
A Wastinghouao WL-794 oaone-yreducing lerop was nounted inside the
h e ll ja r close to the b a rrie r under te a t.
The lamp provided a continuous
supply of ozone in the region of the te s t electrodes.
I t would seem that
the presence of ozone in the teat region would promote a higher degree of
ionization in the a ir thereby lowering the voltage required to cause
breakdown.
Data obtained with the lamp seemed to yield smoother curves,
but no significant change of d-c breakdown voltage was noted as a resu lt
of i t s use.
However, even without the lamp, ozone is present in the b ell
Jtur fc.6 soon as the a ir is broken down in te s t.
Therefore, a ll te sts re­
corded here were made in the presence of some ozone, a small blower unit
was mounted inside the b a ll ja r which directed a b la st of a ir to the te st
region.
This effectively kept the a ir in constant motion past the te st
configuration.
Again no significant change of breakdown voltage could be
observed.
An attempt was made to obtain breakdown voltages fo r fresh a ir
by evacuating the a ir and replacing i t with fresh a ir a fte r each data
point.
The data obtained in th is a tte .p t did not p lo t into smooth curve.
The breakdown voltages measured in th is te s t were within 25 percent of
the corresponding voltages obtained with the ozone leap, but individual
points were both high and low. The readings obtained with the d-c te s te r
are questionable because of the te st procedure required here.
Consistent
re su lts were obtained with the Takk te ste r throughout th is study by using
-46-
the method described In the Test Procedure.
However, the te s t with
fresh a ir required that the breakdown voltage be determined at the f i r s t
arc.
Time would not allow rep etitio n of points because the te s t chamber
was evacuated twice between each voltage readings.
Inconsistent resu lts
may be expected with th is equipment In any te s t if the breakdown voltage
must be determined from the f i r s t arc.
I t Is the opinion of th is w riter
that breakdown voltages are lowered by the presence of ozone in the te s t
region, but additional data is required fo r proof of th is statement.
HIQH-TOLTAOS TESTS Of AlBCRAJT EQPIFMMT
E lectrical equipment intended for use at high a ltitu d e must
meet various requirements of an e le c tric a l and mechanical nature. One
of the most important of these is the requirement of adequate clearance
distance to insure against failu re caused by arc-over.
Specifications
for a irc ra ft e le c tric a l equipment usually state the minimum allowable
clearance distance, but do not require an adequate high voltage te s t to
check the clearance distances used,
fo r example, the specification used
for a irc ra ft relays (AH-B-BOb*) specifies a minimum clearance of 1/8 inch
from liv e terminals to ground and a minimum clearance of 1/16 inch between
liv e term inals.
This specification also requires a 1000-volt (rms) sea
level te s t to check the adequacy of the solid insulation.
I t has been demonstrated e a rlie r in th is report that clearance
distance may become c r itic a l in the region of altitu d e from 50,000 feet
to 100,000 fe e t.
The rela y specifications cited above do not insure ade­
quate protection against breakdown fo r a ltitu d e s above 60,000 feet.
The
1000-v o lt (rms) insulation te s t does imply that the solid insulation w ill
not breakdown under 1414 v o lts peak at any altitu d e ,
fo r the specified
minimum clearance of 1/16 inch, however, the surrounding a ir w ill break
down at 52,000 feet i f 1414 volts peak were applied (figure 6 ).
At
80,000 fe e t, a ir breakdown w ill occur a t 870 volts peak (figure 6 ) which
is fa r below the requirements specified fo r the solid insulation.
I.
AMEHICAH STANDARD DBfIHITIONS OF ELECTRICAL TSBUS, AIEE.
20.10.415-420, page 85.
Section
-4 8 -
I t ie recommended here th at specifications fo r electric al com­
ponents, which are intended for use above 50,000 feet, include a highvoltage te s t to determine the adequacy of clearance distance.
Such a
te s t cannot be made a t normal pressure because the insulating properties
of a ir w ill be most c r itic a l a t the maximum design a ltitu d e .
!Rie only
conclusive single te s t which w ill guarantee adequate clearance distance
would be the application of the minimum allowable breakdown voltage at an
a ir density corresponding to the maximum altitu d e of the design.
If no
breakdown occurs in th is te s t, the clearance distance is d efin itely ade­
quate at any altitu d e lower than the simulated te st a ltitu d e .
The voltages recorded here are based on complete breakdown of
a ir .
At high altitu d e , th is breakdown was in the form of a glow dis­
charge.
fo r lower a ltitu d e s, only a v isib le spark was considered to con­
s titu te breakdown. A brush corona discharge was observed in some te sts
at voltages one to ten percent below the breakdown voltage.
This effect
was noted only from sea-level to about 30,000 feet for larger clearance
distances.
In p ractice, corona discharge may be objectionable even thou^i
actual arcing does not occur.
However, th is brush corona effect ie not a
lim iting factor in the design of high-altitude a irc ra ft used at 60,000
feet or higher.
I f equipment w ill not breakdown at a given voltage at
50,000 feet or above, th is voltage is not high enough to cause brush
corona under 30,000 feet.
-4 9 -
COHCLUSIOHS
1.
The measurement of clearance distance between electrodes should be
made along the shortest a ir path from metal to metal.
All lugs,
washers, and fasteners should be in place for th is calculation.
2.
The use of b arriers to increase clearance distance is found to be
most p ra c tic a l. Most effective use of a barrier may be made if te r­
minals are mounted close to the b arrier on each side.
Clearance die-
tance is then determined by the height of the b a rrie r above the
term inals.
A clearance distance measured over a b arrier is ju st as
effective for protection against arc-over as the corresponding planesurface creepago distance.
This statement is true fo r d irect voltage
and 60-cycle alternating voltages and fo r any shape of b arrier.
3.
The beneficial effect of increasing creepage distance fo r protection
against arcing becomes lim ited a t altitu d e s above 50,000 feet.
The
properties of a ir as an insulating material becomes questionable above
th is a ltitu d e .
Results obtained here indicate that solid or liquid
insulation may be required in place of a ir at extreme altitu d es,
4.
I t is recommended that specifications for e le c tric a l equipment in­
tended for use above 50,000 feet, include a te s t to determine the
adequacy of clearance distance.
In th is te a t, the minimum allowable
breakdown voltage should be applied to the configuration under te s t in
an a ir density equivalent to the maximum altitu d e used in the design.
High-voltage te sts performed in normal atmosphere w ill not assure ade­
quate protection from breakdown at high altitu d e .
Breakdown voltage is not a function of the relativ e humidity of a ir
provided that no actual moisture condensation e x ists.
Breakdown
voltages are lowered significantly i f moisture is condensed on the
h arrier and electrodes.
-5 1 -
Appmroix i
Beductlon of Laboratory Data to Standard Atmosphere
A.
Sample reduction using U-tube manometer.
Data:
Temperature: 77°1. or 25°0.
Manometer readings: 14.30 and -13.33 in. Hg.
Barometer reading: 29.82 in . Hg.
D-C breakdown voltage: 3.2 kilovolts
The height, H, of the mercury column represents the difference be­
tween the a ir pressure in the b e ll ja r and atmospheric pressure:
H = 14.20 - (-13.33) = 27.53 in. Hg.
The height, H1 is corrected to O0C by multiplying by the ratio of
the density of mercury at 25°C to the density of mercury a t O0 C.
H (corrected) = (27.53)
f |} = 27.42 in . Hg.
The a ir pressure in the b e ll ja r, P, is obtained by subtracting
H (corrected) from the barometric pressure
P = 29.83 - 27.53 = 2.29 in . Hg.
The density of the a ir is calculated from the formula
Air density - -
3*2® —
^SnS - 0.00598 ponnd,/ca.ft.
The simulated altitu d e may be read from the HACA density vs. a l t i ­
tude curve (Figure 2 )
Simulated altitu d e = 64,000 f t .
3.
Sample reduction using Wallace and Tiernan manometer
This manometer is temperature compensated and was connected so that
the absolute pressure in the b ell ja r was obtained by d irect reading.
Consequently, the a ir density could be calculated from direct tem­
perature and pressure readings. Simulated altitu d e was obtained from
the same curve (Figure 2 ).
-5 2 -
- APPENDIX II
l i s t of Equipment
1.
Hi^ci Voltage Insulation feste r (D-G); The Iskk Corporation.
Model SG; Serial 31.
2.
Greet Voltmeter (Mji) ; General E lectric.
Serial 1287278. Battery operated.
3.
Vaomun Tube Voltmeter (Mg); General Badio. Type 726-A;
Serial 3972; BAG 6699. Beads rme of sine wave or 0.707 of peak
of complex wave. Calibrated with voltage divider to read peak
value d irectly .
4.
Voltmeter (M4 ); Weetinghouee.
5.
Vacuum Tube Voltmeter (Mg); Hewlitt-Packard.
F691; D.S. 4SPB 62127.
6.
Vacuum Tube Voltmeter (Mg); General Radio.
D.S. 4SZB 63090. Battery operated.
7.
Vacuum Pump, Duo-Seal; W.M. Welch Mfg. Co. Serial 5842-0; BAC 35032.
8.
Manometer, Reservoir-Type; Wallace and Tiernan.
Serial WP5151G; BAG 62140.
9.
Potential Transformer (Tji); Weetinghouee. Type PT; 11500/115 volte;
50-60 cps; 400 va; Style 958997-A; Serial 2299858.
10. P otential Transformer (Tg).
11.
Type A3 ; Cat. 5993877G6;
Serial 2016060; U.S. 4SZB 62896.
Model 40QA; Serial
Type 727-A; Serial 379;
Type FAl35;
Same as above expect Serial 2299859.
Induction Regulator; General E lectric. 1.2 kva; Form HK; Serial
7571361. Connected fo r llo /llO volt operation.
-5 3 -
APPSNDIX III
Description of Test Slectrodee and Test Blocks
Test Electrode#:
All electrodes were cut from 1/8 inch brass sheet stock.
Type At
Ground Electrode - I x l inch: sharp corners.
High-Voltage Electrode - i - X J inch; s h a r p corners.
Type Bt
Ground Electrode - ^ x 4 inch; sharp corners.
Higb-Voltege Electrode - £ x I inch; stjarp corners.
Type Gt
Ground Electrode - Same as type B except that two #4 flat-head
braes screws were mounted a t one end of the electrode and two
#4 round-head braes screws were mounted at the opposite end of
the electrode. The electrodes were countersunk fo r the f la thead screws so that the screw heads were level with the top of
the electrode. See Photographs 122031 and 122032.
High-Voltage Electrode - Same as type B except that two #4
flat-head brass screws were mounted in sor.o of the electrodes
and two #4 round-head brass screws were mounted in the others.
®ee Photographs 122031 and 122033 for mounting configurations.
Test Blocks:
The te s t blocks were labelled 3-1, B-2 , 3-3, e tc ., according
to the design of the te s t block. In some cases duplicate te s t blocks
were made to replace defective blocks (b arriers cracked, chipped, or
punctured). Duplicate blocks were distinguished from each other by
using capital le tte rs ; for example, B-2A, B-2B, B-2C, etc.
All te s t blocks were made of I inch black laminated mlcarta
except B -l, which was made from 1/4 inch black laminated micarta. See
Photographs 122031, 122032, and 122033 for sample configurations.
B-I B arriers - None.
Electrodes - T^pe A.
See Photograph 122033.
B-2 B arriers - l/lB inch wide x 1/8; 3/16; 1/4 inch high.
Electrodes - Type C.
B-3 None.
S-4 B arriers - l / l 6 inch wide x 5/16; 3/8; 7/16; l /2 inch high.
Electrodes - Types B and C.
-5 4 -
B-5 B arriers - 1 /1 6 inch wide z 9/16; 5 / O inch h ig h .
SLectrodes - T y p o C.
B-6
B a r r ie r s - 1 /1 6 inch wide x 11/16; 3 /4 inch h igh .
B lectrodea - rJJype C.
B-7
B a rriers - None. P lan o-su rfa ce creepage d ista n c e s o n ly .
E lectro d es - Type C. The ground e le c tr o d e was noVahle so th a t the
plane-surface creepage distance could he adjusted as
desired.
B-S
B a rr ie rs - 1 /1 6 in ch wide x l / 2 in ch high (4 h a r r ie r s ).
B lectrodea - Types B and C.
B-9
B a rr ie rs - 3 /l6 inch wide x 1 /2 inch high; 5/16 in ch wide x 7/16
inch high; 7 /16 inch wide x 3 /8 inch h ig h .
Electrodes - Tlype 0 .
-55 -
'#
,
G E N E R A L ® ELECTRIC
I
C R E S T VOL T M E T ER
nPE -I ESiC A T 5 9 9 3 8 7 7 6 ^ 1 N O -H B .
c
.
u
n
o
.
i
' I
" tu.mioi^
SCHENECTADYCT^G
-
CREST K V
G E N E R A L ® ELECTRIC
‘
-
V ‘
DANCKM
HICH VDLTAtiK
“
TEST
BLOCKS
59-
12-12-51
122031
-60
TEST
BLOCKS
12-12-51
12 2 0 3 2
-6 1
TEST
BLOCK
—
122033
-62-
LITJGBATUfiS CITED ASD CONSULTED
1.
Effect of Altitude on ELeotric Breakdown and Plaahover of A ircraft
Insulation, L .J. Berberlch, G.L. Moses, A.M. S tiles, C.G. Veinott.
AIEE Transactions, Volume 63, 1944, pages 345-63.
2.
P otential Breakdown o f Small Gaps under Simulated High-Altitude
Conditions; M.J. Delerno. AIEB Transactions, Volume 63, 1944,
pages 109-13.
3.
Altitude-Pressure Tables Based on the United States Standard
Atmosphere. National Advisory Committee for Aeronautics, Technical
Report So. 538, 1935.
4.
Standard Altitude Table. SACA Technical Report No. 837 and Super­
sonic Aerodynamics Handbook.
5.
Standard Handbook for E lectrical Engineers, A.S. Knowlton.
Edition, 1941, Section 4-654, page 405.
6.
Breakdown Studies of Compressed Gases, A.H. Howell. AlEB Trans­
actions, Volume 58, 1939, May Section, pages 193-206.
7.
American Standard Definitions of E lectrical Terms, Al EE.
20.10.415-430, page 85.
8.
Army-Navy Aeronautical Specification for Belays; Direct Current.
March 1946.
9.
Corona in A ircraft E lectric Systems as a function of A ltitude,
W.R. Wilson. AIBE Transactions, Volume 63, 1944, pages 189-94.
10. Measurement of Sparkover Voltage a t High A ltitude.
Company, Document No. D-7074 #8.
Seventh
Section
Boeing A ircraft
11. High Voltage Breakdown Tests of Insulating M aterials.
c raft Company, Document No. T-24750 #7.
Boeing Air­
12.
Elektrische Durchbruchfeldstaerke von Gasan, W.O. Schumann. Berlin,
verlag von Julius Springer, 1923.
13.
Minimum Surface Leakage Distance in D-C Power Systems, J .1. Hart,
W.W. Bosenberry, A.T. MoClinton. E lectrical Engineering, October
1951, pages 910-14.
103028
__ _
,in o iB T E S
3 1762 10020808 9
N378
—
^
UhSh_____cop,2
105028
A U T H O R
Uhlrich. P au l__________
High voltage flashover of
insulating b a rrie rs.
t i t l e
D A TE
B O R R O W E R 'S NAME
D U E
0CTT?<j W
Q-&><
/
"
r
i; <' / S n j Y
f?=*=***/
/,
IOi-*
_______ _______
/<i
i
>
M Zt t ^ V
/-
2
7
Q
IL
S
I-
(J h fo h
O
c T
z
103023
M IN T E D
IN
U .8 .A .