I have in several posts made comment on these. It is well established that "Waxed Paper" types are now a type rarely good. The reference that I have is from an "Honor" (built for Lafayette USA) "CR Analyser" As noted often; any Screen decoupling cap below 50M is considered failed & any coupling cap below 200M is also a fail. My parameter is if it leaks its a dud.

It might be a surprise to some that Waxed Paper types are not the only caps in these positions. There was a tendency in some equipment to use "Oil filled" types that look like an Aluminium Electrolytic cap. I have found these in "Leader" signal generators for one. These I consider no better than Waxed Paper types, as I have found many failed, & two exploded, where they were on the mains.

So I think it pertinent to remind those attempting to fix to beware. There is currently a Trio 9R-59DS on my bench. There were several (one to go) oil filled caps in it. I tested one & it failed, so I tested a second one, after which the lot will gradually end up on the bench: None have made it to 200M.

These are tested at as close to working, marked working or surge voltage as is possible using mainly an insulation tester.

Yes Marc it's the paper. Wax or oil, doesn't matter, paper is hygroscopic and the water will always find a way in, given enough time.

Recently I found some paper caps in my Kriesler 59-03 - that's 1976 production. They were leaky sure enough. All were ERO branded, plastic encapsulated and in high voltage high current positions in the hor. OP.

Plastic encapsulated can be a trap, there can be a tendency to think that they are modern substrate like Polyester, Mylar or some form of Polystyrene when they are not. Even some of the caramel polyesters, cracked there case like the "Mud" types & went short.

The only brand specific ones I note are "Micamold" anything, and those little axials about 3/8" diameter & 1/2" long often found in sixties car radios & such: Cracking, shorting & burning is popular with them.

Here's what I said on another forum about a recent experience with Mica caps in a TV chassis recently:

You know how you never replace mica caps because everyone knows they never fail?

Well, I just had three go leaky in the same chassis!

In all cases they were the smaller phenolic encased "Simplex MS" type.(Australian) from the mid 50s.

In all cases they had over 100V across them and were coupling into a high impedance circuit.

I think here, after over 60 years, we are looking at the onset of silver migration, since mica caps use silver as a conductive element.

Silver migration takes many years to develop, moves VERY slowly but is hastened by high voltages and moisture. Its existence was largely unknown when these caps were made.

Here is a link:

http://www.kurtzersa.com/electronics-production-equipment/solderlexicon/begriff/silver-migration.html

Old micas are still OK in tuned circuits, connected across a coil for example. But now I guess we should regard a mica with some voltage across it as automatically suspect.

http://www.kurtzersa.com/electronics-pr ... ation.html

You might want to fix that link...

Yes, the first faulty (leaky) mica I came across was over 20 years ago. The last three radios I have restored required the replacement of almost every mica, as well as all the other caps of course. Leakage isn't the only problem though. High ESR is prevalent. They measure good for value on a capacitance tester but have no null on a bridge. I've recently had to change mica capacitors in I.F. transformers, because they all had developed high and constantly changing ESR. The same radio also used polystyrene capacitors, but these had developed an intermittent open circuit condition and needed replacement. It seems that the silver mica caps, such as those marked PT and MS, develop both conditions, others develop the high ESR (or open) condition. I regard all micas with suspicion now. I have also come across faulty fixed padder capacitors, causing wrong frequency range and/or intermittent oscillation.

Cheers.

Intermittent OC polystyrene caps have been with us for many years.

If you have an oscillator circuit that relies on the cap to start and you fit a polystyrene cap, it WILL be unreliable as many TV manufacturers discovered in the '60s. You can't replace them with polyester either, because polyester caps have a temp coefficient. You can use NPO ceramics for small values but once you get into the tens of nF range there is a problem.

Kriesler had some special "extended foil" polystyrenes made to fix this one.

I have been known to use four 1nF dipped micas in parallel to replace a suspect 3.9nF polystyrene cap.

With modern SMT caps you can specify "C0G" and get a cap with close to zero temperature coefficient. More expensive though.

Link fixed. Here it is courtesy of my browser history...

http://www.kurtzersa.com/electronics-production-equipment/solderlexicon/begriff/silver-migration.html

Some sort of copy and paste issue.....

My CYA comment with Mica to the Americans is that here they "rarely fail" They have some horror ones that consistently have Silver migration issues, particularly in IF transformers.

In practice here, they are rarely found failed: That clearly means that they can fail. There is one AWA Black one I have never had an electrical fail. And even the later open types are extremely rare fails in the older sets past 1935. I think it a lightning strike but I did have an Astor RK where every cap & resistor had to be replaced: Not the valves.

There was a thing on "Tin Whiskers" on the NASA site, which covers The forming of "Whiskers" that short from Tin, Silver, Zinc & others. I have seen those whiskers from lead free solder. It was known in the forties that 3% of Lead killed Tins migration.

Sirwin comment re voltage: Agree in that a high DC voltage can encourage Silver migration albeit, whiskers seem to be independent of environmental influences. The interesting one is the IF: The primary caps have high DC voltage on them, but rarely across them, so that is possibly a capacitive effect like happens with the blackening of light bulbs.

Additional to first post: C18, 0.22μF 160V. Leakage test 100V gets to 94M flashes over & starts again to recharge.

Silver migration can happen in IF transformers where both caps in the LC circuits were printed on the same sheet of mica. There is usually about 100VDC between the two caps, and the silver ions are all too happy to migrate over the mica sheet surface between the caps.

and some repair procedures

The tendency here was to have slug tuned with fixed caps mostly encapsulated & the top trimmers were separate trimmers, I have rarely had issues with either, from anything other than physical damage. That being in several hundred radios.

G'Day Folks.
Is it possible to blow away the tin whiskers from silver migration as you do with germanium transistors.
I had a couple of transistors that were shorted in the IF stages of an 1960's national panasonic radio.

I charged up a cap to around 500 volts and soldered the EBC leads together and discharged the cap through the shied lead and the EBC soldered leads.

The short went away, as measured on the meter and the radio was back in working condition.
just wonder if it would work on the mica caps

Cheers
Peter.

Is it possible to blow away the tin whiskers from silver migration

We use a pair of 9 volt batteries to burn away zinc whiskers (dendrites) inside Hammond organ vibrator scanners. Not a long term permanent solution but it sure beats disassembling them for a cleaning.