How to revive old capacitors?

[Conrad Hoffman]

Being frugal, ok, cheap, I want to use whatever good caps I have, and avoid the bad ones. The answer is two-fold. First, they have to support their rated voltage with minimal leakage current. This number can go down for weeks after installation and regular operation, but it has to be reasonable after a few hours of reforming. Actually, I'd be very suspicious of any cap that doesn't have its leakage under control in a few minutes. I always use a large resistor, not relying on power supply or the Variac to limit current. Next, the capacitor has to measure correctly on a bridge, and have a reasonable dissipation factor. Everybody should buy or build a bridge that measures both C and D! C without D can deceive you because series C rises if D rises. Thus, a bad cap can measure fine on a meter that only measures C. If the cap meets spec on both DC and AC, chances are it will have a decent service life, and sound just as good as any other cap with similar measurements.

 

[sklimek]

Everybody should buy or build a bridge that measures both C and D! C without D can deceive you because series C rises if D rises. Thus, a bad cap can measure fine on a meter that only measures C. If the cap meets spec on both DC and AC, chances are it will have a decent service life, and sound just as good as any other cap with similar measurements.

Is it a Wheatstone Bridge, scroll down -

http://www.tpub.com/neets/book21/88.htm

do you have a link that shows how one may make this type of bridge for measuring??

Thanks

 

[Conrad Hoffman]

There are a bunch of way to set up a bridge for caps. I think the old GR 716C used the Schering configuration. If you have giant variable caps, that's a nice way to go, but most don't. I'd do something very simple and just balance the unknown cap and it's dissipation, with a known cap and series resistor in the other leg. Remember that a bridge doesn't have to be anything more than some parts tacked together on the bench. Measure the final settings with a DVM and use the bridge equations. Easiest in Excel. Of course a dedicated box with numbers on the panel is always preferred ;-) I'll search around and see if I can find or post a reasonably easy to build circuit.

 

[Conrad Hoffman]

No doubt everything is on the 'net, they just didn't leave it where I could find it. Here's a simple to build bridge, copied more or less from the series C part of the GR 1608A bridge. The original had unique pots and some compensation parts so it could get to 20kHz. This one will have reduced accuracy at higher frequencies, but should work fine to 1kHz. PS caps are usually checked at 60 or 120 Hz anyway. Note are on the schematic, but since the formulas are there, you can use pretty much any parts you want. The original used a 0.15uF mica and polystyrene reference cap, but the final range was limited to 1100uF. I'd use several values so you can check large value caps, maybe 0.15, 1.5, and 15. Or whatever value series is handy, .1, 1, and 10, for example. Use film! RN was two pots in series for better resolution. You can use a single, a multi-turn, or two pots in series. If you use a single pot with a big dial, you can mark it directly in capacitance. The original values were odd, so the dial stopped at exactly 1100, but I've rounded up to standard parts. Any filament transformer should work for the input. A 1:1 tranny would be best with a sig gen. If you don't have a scope, this works fine with sensitive headphones- that's how they balanced the very old bridges before sensitive meters were common. If I've made errors, no doubt somebody will point them out.

Cap Bridge

 

[sklimek]

I think the old GR 716C used the Schering configuration.

Yep, it was the Scherling Bridge. I found the 716C GR service manual as a free download on the BAMA site if anyone is interested:

http://bama.edebris.com/manuals/gr/716c

The 9mb file works, not sure about the other PDF...

 

[Conrad Hoffman]

The 716C is desirable because it can measure extremely low dissipation factors. It can easily tell the difference between Mylar, polypropylene and Teflon. In fact, it can resolve the difference between air and Teflon. Using differential techniques, it can measure less than a picofarad. It can also work up to a megahertz with somewhat reduced accuracy. Avoid the special HF version, as it tends to cost more and is much more limited in range. On the downside, the maximum value you can measure on a 716C is 1.1uF, unless you add external components. Even then, it's not practical for large value electrolytic caps. It's also not something one could easily build as a general purpose instrument for any good accuracy, due to the specialized components and non-linear scale on the DF dial. IMO, if you like to fool with RIAA components and coupling caps, the 716C is a prize. OTOH, for electrolytics, the modified Wheatstone is more useful. This is drifting kinda OT, and I've yet to see much interest in traditional bridges, so anybody can feel free to email me and I'll rattle on about bridges for hours...

edit- DL the manuals for the 1608 and/or the 1650 from the source above for more info. The 1603 is also quite easy to build, but quite complicated to understand- but very good for characterizing speakers and transducers.

 

[then_dude]

hello people, i was about to ask the same question.

i have some old, quite old at least 30 years elna electorlytic. they are both coming from an tube amp, and although JJ manufactures cheap caps, i thought of reusing them.

they are : 100uf/250v
40-40-40/350V

i have taken the 350v cap, and felt that the weight was not evenly distributed along the cap, so i put it up side down, probably the electrolytic paste. anyway, i hooked up a 2k ohm and a 22vdc which had lying around
i measured the voltage acros the resistor and was below 90mV immediatly, i don't know but i think the cap should be quite alright. so should i trust it ? would it be as good as a descent new cap in the power supply ? or should i not take the chance of blowing it ?

what do you think. i will try to get 160vdc on it first, and then 280vdc on it. that's all i have.
what do you think people ?

antoher question: the only thing a cap can become by aging is dried out ? can you insert new oil in a cap ? or some other fluid ?

thanx a lot
greetings

 

[then_dude]

after 20 minutes the the voltage over the resistor is 2mv and the 3*40uf are parrallel to this resistor. so i think the cap is fine, at least at 22vdc. my multimeter says nicely 39.82uf. so up for more

 

[Conrad Hoffman]

That's a good sign, but you really have to get the voltage up to the rated limit. Be careful! A lot of old caps will be just fine until you start raising the voltage, then they short, or the leakage never comes down to a reasonable level. No, you shouldn't try to add anything to an old cap, and I wouldn't pay any attention to the weight distribution. That's just determined by where they put the "guts".

 

[then_dude]

thanx a lot,

at this moment i haven't got any source getting it higher then 22vdc, which is quite ridiculous,

i have no adjustable power supply, so i will have to connect the cap straight away on the transfo, getting 160vdc over it. after that i can switch to 300.

i know with a variac you can adjust, so maybe gotto look for a scheme (which is also secure).

thanx