[Collins] dipped mica caps do fail

[Dr. Gerald N. Johnson]

The IF tubes in the S-line receiver are run at fairly high plate 
currents to get enough gain and so they often fail with grid current 
from grid emission as particles of cathode coating have migrated to the 
control grid. They are best tested with a grid emission tube tester, 
which some vintage TV service shops felt was their best money producer 
and made for the happiest customers because it detected tubes that were 
destined to upset the TV operation before they did and so service 
recalls were way fewer.

The difference between grid current induced S-meter drift and this 
partly self curing capacitor leakage is that grid current doesn't show 
immediately, it takes a few minutes for the heat radiated from the 
cathode to heat the grid coating to emission temperature.

Another way to detect a leaking coupling capacitor (most critical for 
any receiver's audio output stage) is to pull the suspect tube and check 
for a positive grid socket pin voltage with power applied using a VTVM 
to have high DC input resistance.

Another super critical capacitor in the S-line is the 8 uf filtering the 
bias supply to the output tube. The output already runs hot, if that 
capacitor gets leaky it makes the output tube run hotter and sometimes 
takes the audio output transformer with it. A couple years ago it cost 
more than $125 to get that output transformer rewound.

Also critical to the receiver bias is the bias rectifier. Orignally it 
was an encased selenium rectifier that goes bad with age lowering the 
bias supply voltage. It should be replaced with a good silicon diode on 
a three terminal terminal strip (or two insulated terminals, probably 
center ground on three terminals). 1N4004 or higher voltage is fine. The 
common 1N4007 is super fine and should live a very long life.

In 75A4s the postage stamp mica between the second mixer plate and the 
mechanical filter is known to go leaky and that is not healthy for the 
mechanical filter.

In every tube receiver I've known of with an S-meter the receiver used a 
AVC controlled IF stage as a DC voltmeter displaying the AVC voltage, 
but calibrated in signal strength. Getting two functions from that tube.

There are no absolutely perfect capacitors that last forever, but dipped 
silver mica are generally way better than paper and postage stamp mica 
capacitors of earlier radios. Both micas use natural mica for the 
insulation and reliability only results from good inspection before 
silvering or stacking with foil. Natural mica can have inclusions of 
partly conducting material. Often it adds color to the mica so is 
detectable by a visual inspection of the raw material.

Stacked and rolled foil plus insulation capacitors suffer from slight 
air spaces between the foil and the dielectric. With DC applied, the 
voltage distribution between the air and the dielectric is dependent on 
the dielectric constants and the thicknesses. In most capacitor 
dielectrics, the dielectric constant is high so the voltage distribution 
is low in the dielectric and high in the air. In the extreme case where 
the air and dielectric are the same thickness, the voltage split is the 
ratio of the two dielectric constants, so if the insulation has a 
dielectric constant of 10, 10 parts of the DC are applied to the air and 
1 part to the dielectric though the dielectric can often stand more than 
100 times the voltage per mil of air. So the air arcs, and in time 
converts the dielectric at one terminal of that arc to carbon causing a 
very leaky capacitor. Oiled paper capacitors hope to fill all the air 
gaps with oil but were leaky when new because the paper wasn't perfectly 
clean. When the oil leaks out they go bad faster.

Silver mica capacitors and metalized plastic like Orange Drops solve 
that chance of an air gap and the localized dielectric leakage defect. 
First the silvering or metalization prevents air gaps, and second the 
metal film is so thin that leakage current can often burn the metal back 
from a dielectric defect. Self healing they say.

I'd say that 100 pf capacitor down to 55 pf has tried to self heal at 
the cost of metal plate area.

Also in mica capacitors, getting rid of the air gaps and the foil has 
made them more stabile for capacitance which is a sure benefit. Some of 
the modern ceramics are very good and some are worse than horrible for 
both temperature and voltage stability. Some will lose 75 or 85% if 
warmed up or with rated voltage applied. And some are piezo electric.

73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association.

On 6/21/2014 2:23 PM, Al Parker wrote:
> Hi folks,
> We don't expect dipped micas to fail, but some of us have seen it happen
> occasionally. Here's another example:
> In the 75S-1 which I've posted about recently, the S-meter has pegged
> downscale on warmup, coming back to zero pretty consistently after a few
> minutes. After hunting a bit, checking both IF tubes pin voltages at
> turn-on and after a few minutes, it appeared that the 2nd IF had a
> slightly positive voltage on the grid, and the tube was drawing too much
> current. A good pointer at the grid coupling cap from the previous
> stage. That cap was 100 pf brown dipped mica, like we regularly use to
> replace the red/pink rectangular micas, etc.
> I put the bad cap on my bench megger, it started showing leakage as soon
> as I brought up the voltage, and showed about 20 megs at 400vdc. But it
> was erratic, and settled down (up) to higher values fairly quickly with
> steady voltage of 4-500. I think they're rated at 400vdc. it also
> measures about 55pf vs 100.
> 73,
>
> Al, W8UT
> www.boatanchors.org
> www.hammarlund.info
>
> "There is nothing -- absolutely nothing -- half so much
> worth doing as simply messing about in boats"
> Ratty, to Mole
> ______________________________________________________________
>