[R-390] Postage stamp mica caps and AGG problems

[Tisha Hayes]

I have been seeing about a 25% capacitor-fail rate on NOS paper/wax when
tested with a Secore LC75. Most of them will easily pass testing with a
DVM-type tester that only puts a few (tens) of volts across the cap and
doesn't do anything to measure ESR. While mica caps fail at a lower rate
that has been observed as well.

Micas are weird; near rated voltage tests don't always show the problem,
capacitance values look somewhat good, ESR.. didn't show anything; but in
an RF test at a few hundred KHz they are no longer resonant when compared
to control samples.

I try to test at least at 75% of rated voltage (up to 600 volts on 1KV
caps) and failures sometimes happen with spectacular, fireworks type
results. A few times the cap would wander around in value, up and down,
seemingly in a random way as internal areas of the cap are changing
characteristics.

I don't even mess with reforming any more; other than for curiosity's sake;
I might be able to get a cap to not be as leaky but the recovery is
temporary at best and the next time the cap goes on an excursion it becomes
a hard failure that takes out resistors, or with electrolytics in the power
supply, a choke or transformer.

*Ms. Tisha Hayes, AA4HA*

*Sr. Engineer, 4RF USA.*