1920.] Birks and Ferguson.—High-tension Insulators. 183 
gives the results of tests made on one batch of insulators. The megger 
and 50-cycle pressure tests are tabulated side by side. Several insulators 
were first tested on the 125-kilovolt high-frequency oscillator, and then 
on 50 cycles ; but the oscillator tests proved misleading, and this test was 
omitted on insulators which had been in service. Good results were, 
however, obtained with the oscillator as described later. In all, eighty-five 
insulators were removed from the lines and tested. Of these sixty-six 
were taken off at random and nineteen were suspected from observations 
made at night. The whole of these insulators were subjected to the 
50-cycle test as described above, and Table 2 shows the summarized results 
of the tests. Of the sixty-six taken off at random, five insulators, or 
7*6 per cent., failed at or below 100,000 volts from head to pin, and of these 
one broke down at 42,000 volts, although it had just come off the fine 
where it withstood 66,000 volts. A further five of the random insulators 
broke down in three shells at or below 50,000 volts. These ten insulators, 
representing about 15 per cent, of those taken off the line at random, were 
apparently on the verge of failure, and their removal probably obviated 
several line-breakdowns. Insulators which failed in two shells constituted 
16-7 per cent, of the sixty-six random insulators. Such insulators are not 
very reliable, but could probably be utilized for further service until 
opportunity for replacement was favourable. Of the balance of the sixty- 
six, twenty-two, or 33*3 per cent., failed in one shell, and twenty-three, 
or 35 per cent., withstood the 50-cycle tests on all shells. 
From the results of these tests it would appear that about 15-7 per cent, 
of the insulators installed on the line are in a dangerous condition and 
constitute a menace to continuous service. It is probable that conditions 
may be somewhat better than the tests represent, because of the fact that 
the greater percentage of the insulators tested were removed from those 
portions of the lines most subject to mists and fogs, and it is possible that 
absorption, and therefore deterioration of the insulator, is more pronounced 
in these localities. A few insulators selected for the purpose because they 
meggered low, most shells under 50 megohms, were placed in a drying- 
oven after being megger-tested and dried out at a temperature of 200° F. 
for about two weeks. It was observed that the resistance dropped appreci¬ 
ably after heating for a few hours,* and after the first few days the resist¬ 
ance gradually rose until in about a week the megger showed infinity. 
The few specimens tested in this way withstood the 50-cycle test without 
failure, demonstrating fairly definitely that moisture formed the conducting 
path in those insulators which failed under the 50-cycle test. The cement 
joints of one or two of these dried-out insulators were painted with water¬ 
proofing cement varnish, and the insulators allowed to remain outside in 
the weather for about three months. On meggering it was found that 
the resistance was still infinity, possibly the insulating layer of varnish 
having something to do with resistance. However, after water was poured 
on the joints and allowed to remain for an hour the megger tests showed 
that while some of the shells still gave a high resistance, others were down 
almost as low as before drying, moisture having penetrated by way of fine 
cracks in the varnish and cement. Possibly those which had absorbed 
moisture to such an extent after drying were either slightly cracked or 
punctured with high-frequency punctures, but when dry would survive 
the milder 50-cycle test imposed on them. It is very improbable that 
* A. Bang, Trans. Am. Inst. Elec. Eng., p. 1243, 1915. 
