1000V capacitors

[ppppenguin]

I've found a few of these NOS capacitors. They read zero leakage at 1100V, I haven't tried a destruction test up to 5kV.

I know the RS "Mixed dielectric" types should normally be treated as faulty and the white ones are from RS. Though the MDC caps were usually grey in my memory. There were also some similar blue and white caps (Dubilier?) that are as bad as the grey ones.

But these white RS ones and yellow TCC "Duomold" seem OK.

Any thoughts as to whether they are safe to use? If they were only 600V I wouldn't bother to keep them as I've got a supply of modern 630V parts but I don't have any other 1000V parts at this value.

[Diabolical Artificer]

I've a few of those too Jeffrey and the white and blue two tone Dubilier's, if they test ok, they should be fine, i test all NOS caps anyway.

Andy.

[ppppenguin]

Even if they test OK now, I'd still be wary about using them in high stress positions such as TV line output stages.

In ordinary valve radios the only 1000V cap was usually the tone corrector, which could be subject to high peak voltages. I believe that a modern 630V is OK in this position, though can't give good evidence for this. Also the tone corrector is often connected from the OP valve anode to 0V. If it fails then this fries the OP transformer. Better to connect across the transformer primary where failure won't cause damage. Remember that to a decent approximation, HT and 0V are conencted together as far as AC is concerned.

[boater sam]

I have lots of the white ones in various values and find that they are OK despite not knowing how old they are. 
The blue and white ones I have found to be very variable. Strangely they seem not to increase in measured value like most damp paper capacitors but show a reduction in withstanding voltage.
Sam.

[ppppenguin]

Is that a genuine increase in value or an artefact of measuring capacitance by timing how long it takes to charge? The "Peak" and Chinese testers all use this method which will show a leaky cap as having increased in value.

[boater sam]

Yes, its just an erroneous reading on the Chinese Magic Box due to leakage. Its handy though, as a certainty that the cap is no good. Doesn't seem to do it with electrolytics, but then you get a leakage percentage and an ESR reading.

[ppppenguin]

For high voltage caps my quick and dirty test is a 500V Megger. If any cap from 350V up shows any leakage then it's duff. My main bench HT PSU, a Solartron SRS153, does 0-600V on a good day with the wind behind it. For higher voltage I've got an old Unilab 5kV PSU. That with a DMM on uA range and a 1M limiter resistor does the job nicely. The internal resistance of the supply is high enough that the 1M resistor isn't really needed.

[Mike Watterson]

Is that a genuine increase in value or an artefact of measuring capacitance by timing how long it takes to charge? The "Peak" and Chinese testers all use this method which will show a leaky cap as having increased in value.

Um, even simply measuring voltage drop at RF gives an apparent increase in capacitance with a leaky cap.
I use the electronics of a flash on a "disposable" camera with the 300uF cap replaced with 1uF, 2 x 1M in series as one would have voltage rating exceeded. Remove xenon tube.
The "earthy" end of the cap connects via a neon with 0.1uF. A solid "on" is a fail. A good cap flashes once and neon goes out. You can use it as a 100s of M Ohm meter by timing the flashes.
Measure PIV of diodes or transistor Vce if you bypass neon and use a 20 MOhm or higher DMM (or a VVM), non-destructive due to low current.
I've a 2nd higher voltage one from a standalone flashgun that has a uA meter in the case. It uses 4 x AA. The modern ones made from guts of "one time" film cameras use a single AA. An old one uses a PP3 and sprung on switch.

Higher value low leakage caps are rather dangerous when charged to 600V.

[ppppenguin]

Higher value low leakage caps are rather dangerous when charged to 600V.

What's more, due to stresses in the dielectic they can sometimes recharge themselves. Serious capacitors (big metal can 4uF+, 2000V+ for example) are best kept stored with a shorting link across the terminals. I've been bitten by a CRT that recharged itself and was lucky not to drop it. In that case the shock wasn't serious, it's the surprise and involuntary reaction.