Full resolution (JPEG) - On this page / på denna sida - 1958, H. 5 - Flashover Characteristics of Non-uniform Spark Gaps with Standard 1×50 Impulses, by Jörgen Fryxell
<< prev. page << föreg. sida << >> nästa sida >> next page >>
Below is the raw OCR text
from the above scanned image.
Do you see an error? Proofread the page now!
Här nedan syns maskintolkade texten från faksimilbilden ovan.
Ser du något fel? Korrekturläs sidan nu!
This page has never been proofread. / Denna sida har aldrig korrekturlästs.
Fig. 1 a—f. The flashover distribution of vertical rod-rod gap with 500 impulses (1 X 50) per voltage level.
Study of the vertical rod-rod gap
Tests
The main part of the investigation was made on a
vertical rod-rod gap with square 15 X 15 mm
electrodes spaced 50—600 mm apart. The lower rod of
the gap was 1 m long and connected directly to an
earthed sheet covering the whole floor of the testing
plant. The tests were made with both positive and
negative polarity of accurately adjusted standard
1 X 50 shape. The front condenser of the impulse
circuit, connected in parallel with the test object,
was 800 pF.
The test voltage was measured by a sphere-gap of
500 mm diameter using the relevant IEC-spliere-gap
tables. This was used to determine the voltage ratio
of the circuit, i.e. the relation between the voltage
on the test object and the flashover voltage of the
irradiated lowest gap of the self-tripped impulse
generator. The constancy of this voltage ratio was
checked from time to time and at different voltage
levels. A further check of the voltage measurements
was made by an independent method, namely
measurements by a cathode-ray oscillograph and a
resistive voltage divider. All checked values showed an
agreement within ± 2 %. This indicates that both
the absolute voltage levels and the constancy of the
voltage during a number of applications with the
same setting of the generator fell within this ± 2 %
limit.
During the test, the investigated rod-gap was
adjusted to one of the following spacings: 50, 100, 150, 200,
250, 300, 400, 500, and 600 mm. At each spacing a
number of test series were made where in each
test series 10 impulses were applied at each of a
number of different voltage levels spaced 3—5 %
apart. The lowest and the highest levels were chosen
so that each test series had at least two voltage levels
without flashovers and at least two voltage levels
where all 10 impulses gave flashovers. This test was
repeated at nine times in the range from 0 % to
100 % flashovers. The whole test was further
repeated 4 times, i.e. the test involved a total of 500
impulses per voltage level and gap spacing.
The mentioned test procedure is identical with that
used during more conventional impulse tests with the
exception that the total number of voltage
applications per voltage level was higher.
The results of the tests were corrected for humidity
in accordance with ASA-standards. The factor k used
in the denominator of the correction formula varied
between 0.95 and 1.00. No external irradiation was
used on the rod-gap but the testing plant was exposed
to daylight of varying intensity via roof-windows,
and to light from incandescent lamps.
Results
The result of the test was treated by plotting on
probability paper, a paper where the x-axis is
linea-rily graded while the y-axis is graded in such a way
that a Gaussian distributed probability curve will
become a straight line. For each voltage level the
probability pv of flashovers was calculated as the
relative number of flashover in all the 500 shots.
Such curves are presented in fig. 1 a—f valid for
100, 300, and 500 mm spacing positive polarity and
200, 400, and 600 mm for negative polarity.
As can be seen most of the probability curves show
the conventional straight line (Gaussian distribution)
with a standard deviation between 3—4 %. For some
spacings, however, the curve has a pronounced
S-shape. This is most obvious for 300 mm gap
spacing positive polarity and for 400 mm negative
polarity.
The problem is now to determine the 50 %
flash-over probability as a function of spacing. It is of
course desirable to obtain results which may be
compared with those from other laboratories. The
results should also be dealt with in such a way that
sufficiently correct information could, if possible,
be obtained from much shorter test series. The
observations should also be made in such a way that
the spacing between the voltage levels used in a
test does not influence appreciably the results.
In all these cases the Gaussian curve provides the
possibility of just taking two points on the curve
preferably on each side of the 50 % level and
drawing a straight line between them. This construction
gives a fairly correct estimation of the 50 % level
1 66 ELTEKN I K 1958
<< prev. page << föreg. sida << >> nästa sida >> next page >>
