electrolytic caps testing

[Randall]

I have an Acus (made in Italy) acoustic guitar amp in for a sort of low end howling sound with nothing plugged in. It sounds like a low end feedback. All else functions on the amp. It's an SMD torroidal affair, and I usually consider this a board replacement, but I thought I'd have a look at the filter caps since one measured odd in circuit compared to the others.

Pulled the caps (two 4700uF/63v, and two 4700uF 35v), and measured. Three looked good pretty quickly, and one at first glance looked bad, but after a while of leaving the probes on it, the Fluke went from no indication to "Disc" (I think that is disconnect), to finally inching it's way up to a reasonable value close to 4700uF. Not really getting what was happening, I hooked one up to my DC supply to check for leaks. What happens next confuses me.

<+24v> to <+capacitor-> to <+ volt meter lead-> to <PS ground>

This measured 24v, regardless of polarity of the capacitor on all four caps. I've always tested coupling caps this way with good results, why is this time different?

Then I tried measuring current by:

<+24v> to <+capacitor-> to <400 ohm resistor> to <+ammeter lead-> to PS ground.

This time I measured ~42mA after a slow climb of about 40 seconds, with one cap taking maybe twice as long to get there.

Next test was reversing cap polarity

<+24v> to <-capacitor+> to <400 ohm resistor> to <+ammeter lead-> to PS ground.

This time each cap counted down to less than 0.10 mA, again with one taking longer than the others.

So, what am I seeing? Do I have a bad or suspect cap? And how come they are passing DC?

[The Dude]

I'm not clear on some of your tests. I can add (maybe) some things. 1) You shouldn't be putting reverse DC voltage on an electrolytic cap, if I understand what you did correctly. 2) With the negative side of a cap not having a ground path and putting DC on the positive side, I would expect there to be residual DC on the negative side.

I've always found it easier to look at the cap in circuit and see if it's doing the job it's supposed to. In other words, if it's a filter cap, was there ripple on the cap that shouldn't be? Was the DC voltage there what it was supposed to be. Or, if it's a coupling cap, is it coupling and not leaking?

P.S. It's made in Italy. What do you expect?

[Randall]

I did not look at it in circuit. I've never heard of the brand, and I generally do not service SMD switchmode (I think) equipment. My question here is mostly about how to test these caps out of circuit. If it's not the caps, I'm out.

[The Dude]

You can check for leakage (kind of) by measuring with diode check. You should see the cap charge to open. Keep in mind that the diode check doesn't apply the voltage that the circuit does so it's not a foolproof method. If you don't have a cap tester, it's difficult to test them for capacitance and ESR. Substitution is probably best. That is why it's easier to see how it functions in the circuit by measuring voltages.

And, BTW, I was joking about the Italian thing.

[Randall]

"And, BTW, I was joking about the Italian thing."

Good, because my name ends in a vowel.

[Enzo]

Passing DC maybe is charging current.

So you seem to have a 24v bench supply and caps off the board. SO use the supply to put 24v into the cap. It should charge up to 24v very quickly. Now disconnect the power supply, and touch your volt meter to the leads. Is it storing the voltage? I'd expect it to slowly discharge through the mete. A bad cap might not take a charge or discharge fast. Try it with a good cap first to see how it works.

[Randall]

All four caps held their charge during this test, but again with one just a little faster than the others.

[The Dude]

Somewhat speculative, but if one of those 4700UF caps, was starting to dry out and was maybe only 2000UF, it would probably discharge faster.

[Mick Bailey]

Often when caps dry out the ESR increases, in effect providing a longer RC charging time-constant. The nominal capacitance value usually reads fine on a DMM and the DC leakage is also usually not affected. The result is that a cap in-circuit may never charge up to full voltage. The ESR can get to a point that the cap is effectively replaced with a high-value resistor. In decoupling circuits this means motorboating/feedback, in audio stage coupling it results in diminished signal, in power supplies higher ripple and poor smoothing.

[diydidi]

https://www.youtube.com/watch?v=KVkUoljBz80&t=21s
What you guys think about this? Maybe hooked up to a variac and test caps close to their operating voltages.

[Mick Bailey]

The guy states that it isn't really suitable for electrolytics, though he does show at the end of the clip testing a low-value, though a high-voltage cap. Low voltage caps wouldn't work because the neon would never conduct (below about 90v)

[cjenrick]

as long as your in there, why not spend $3.75 per cap and sleep easy?

[Mick Bailey]

as long as your in there, why not spend $3.75 per cap and sleep easy?

Because if the original caps were OK you've just wasted $15. If only one was bad you've wasted $11.25. Maybe if there was a part or two costing pennies I'd replace 'blind' if I'd narrowed the fault down to a specific area, but that would be rare.

[darkfenriz]

Wait a sec... so you connected 24VDC to electrolytic polar caps in either direction?

I've seen caps explode when treated this way... confused

[MarkusBass]

I have an Acus (made in Italy) acoustic guitar amp in for a sort of low end howling sound with nothing plugged in. It sounds like a low end feedback. All else functions on the amp. It's an SMD torroidal affair, and I usually consider this a board replacement, but I thought I'd have a look at the filter caps since one measured odd in circuit compared to the others.

In such a case the easiest test is to plug in a guitar into the RETURN jack in the EFFECTS LOOP and listen whether the problem still exists. Or, disconnect the preamp from the power amp. Have you tried doing so?

Next test was reversing cap polarity
<+24v> to <-capacitor+> to <400 ohm resistor> to <+ammeter lead-> to PS ground.

This is quite risky. Haven't you heard that it is not allowed? By doing this you destroy the capacitor . Now you really need to replace it .
EDIT: here is some more info on the problem: https://www.eevblog.com/forum/beginn...the-wrong-way/
or here (scroll it to 25th second):

You can get in touch with the factory here: Contacts