Capacitors as diagnostic tools

[glebert]

I've thought about this a few times and don't know that I have seen anyone say anything for or against the idea. The question is if you have a situation where you suspect a capacitor is bad, can you tack another capacitor in parallel to see if it fixes the problem (or at least improves it). For example maybe don't have the right size or don't want to teardown/rework/reassemble the amp just to find out it that wasn't the problem after all. I am working on a Peavey Triumph where some of the radial lead caps are mounted between the mezzanine style PCBs, so they have to be under a certain height and I don't have any caps that will fit in that space, but I could tack caps to the back of the board (which faces up) to test a theory.

Similarly I have considered using caps to create AC short/ DC block to ground on small signal circuits to troubleshoot noise problems.

So is this something that people do? Is it a good idea, bad idea, or just weird idea?

Greg

[Chuck H]

Most often the failure mode in power supply electrolytics is such that tacking a good cap across a bad cap can restore performance and this actually has been discussed here a lot. Though you're not likely to find any thread titles that address this for a forum search.

This common failure mode for electrolytics may not always be the case. In the event that too low of a parallel resistance has developed in the failing capacitor then the tacked in 'good' cap would still have this same parallel resistance across it. In many applications, like power supply and decoupling circuits, a parallel resistance defeats circuit function regardless of the capacitance across it. And, you didn't ask but the common failure mode for film caps would be to develop such a parallel resistance, aka: "leaky cap". Same as above for the electrolytics. Paralleling a good cap can't do anything to correct or detect this.

[glebert]

This common failure mode for electrolytics may not always be the case. In the event that too much parallel resistance has developed in the failing capacitor then the tacked in 'good' cap would still have this same parallel resistance across it. In many applications, like power supply and decoupling circuits, a parallel resistance defeats circuit function regardless of the capacitance across it. And, you didn't ask but the common failure mode for film caps would be to develop such a parallel resistance, aka: "leaky cap". Same as above for the electrolytics. Paralleling a good cap can't do anything to correct or detect this.

You kind of lost me on this one. Using a model like this for the parasitics of the capacitor, an increase in the parallel resistance makes the capacitor more ideal. Did you mean an increase in parallel conductance (aka decrease in parallel resistance)? I would agree that putting a second cap in parallel does reduce the parallel resistance.

[Helmholtz]

I can't think of a case where temporarily paralleling a good capacitor with similar capacitance would hurt (one exception could be the reservoir cap after a tube rectifier where too much capacitance can damage the tube).

If problems are caused by low capacitance or high ESR (Rs) like excessive hum/ripple, oscillation or motorboating (= low frequency oscillation) the parallel cap should show improvement.

If amp problems are associated with excessive leakage (reduced Rp) the parallel cap can't help.

Using caps for AC coupling, DC decoupling and AC shorting of inputs is a standard procedure.

[Chuck H]

Did you mean an increase in parallel conductance (aka decrease in parallel resistance)?

Yes, that is what I meant. I was using a hypothetical "ideal" cap as model for the discussion. In which case what I meant by "too much of a parallel resistance" is too much conductivity. I edited my post above and I hope it makes more sense.

EDIT: American is a pretty clumsy language in some ways. A statement like "The man looked at the girl with his glasses on." is common prose. Though it actually means the girl was wearing his glasses. Properly written, that sentence should be "With his glasses on, the man looked at the girl."

[Enzo]

I can't think of a case where temporarily paralleling a good capacitor with similar capacitance would hurt... the parallel cap should show improvement.

I can, a shorted filter cap. (yes, I note you essentially said that a line or two later) Hell a shorted cap anywhere. And by "shorted" I include very leaky caps. if a coupling cap between stages is leaking DC, tacking a parallel cap on won't stop the DC. And a shorted filter cap, or a 450v cap that leaks over say 50v, won't be temporarily cured by a tacked on parallel. Parallel a weak cap in a filter, and sure that will show improvement and reveal the weak cap.

It is like anything else, it is a useful technique, but context is everything. You need a sense of when it will help and when it won't.

[Helmholtz]

I can, a shorted filter cap. (yes, I note you essentially said that a line or two later) Hell a shorted cap anywhere. And by "shorted" I include very leaky caps. if a coupling cap between stages is leaking DC, tacking a parallel cap on won't stop the DC. And a shorted filter cap, or a 450v cap that leaks over say 50v, won't be temporarily cured by a tacked on parallel. Parallel a weak cap in a filter, and sure that will show improvement and reveal the weak cap.

I considered a shorted filter cap before I posted my statement above. And I repeat that a parallel cap won't make things worse (maybe my English isn't clear enough).

[Jazz P Bass]

A multi meter to me is the way to go.

DC coupling cap? check for Vdc leakage.
Filter capacitor? check for abnormal amounts of Vac ripple.

[Chuck H]

A multi meter to me is the way to go.

DC coupling cap? check for Vdc leakage.
Filter capacitor? check for abnormal amounts of Vac ripple.

This is what I've always done. I read here about the parallel cap thing sometimes and I always wondered if I just didn't "get it". Like I didn't know enough to see the merits or something. Twice I've opened up amps to repair them and found a new cap paralleled to one of the old ones. I figured it was a "repair" and not just a mod to add more uf's, but wondered what sort of tech just leaves the old turd cap in there. Again I thought I just didn't get it and didn't know enough to realize whatever I thought was wrong about it didn't matter.

I don't think I've ever tested a filter cap I didn't already know was bad. And I've never suspected a bad filter cap only to test it and find out it was fine. In a guitar amp they pretty much let you know they're bad on their own. And most often if I'm replacing one, I'm replacing them all. So if there's one or two good ones in there, how much longer could they have if the others were bad and how much sense does it make to just replace the bad ones!?! But it gets done all the time. You open up an amp and find some original filters, but one or two newer ones.

EDIT: Just now considering that caps will often test fine on a meter and still be bad at higher voltage. I suppose a parallel cap could reveal this but it hasn't come up for me yet.

[SoulFetish]

Specifically, using a capacitor to "probe" suspected failing filter caps (ie, connecting them in parallel) is a great way of determining which node is the problem. If one of the filter caps is laying down, then this technique will almost always provide some audible improvement, or visual improvement when monitoring a scope. When going through and probing each filter node, if there is no improvement, then its often not the caps themselves. (often, but not always; and ignoring broken leads/connections).
I admit, I don't love doing this because it's a brute force kind of technique and makes me jumpy charging and discharging electrolytics this way. It can be tight quarters in some really old amps and it takes a steady hand.
But its a fast and effective.
If I determine a cap to be faulty or failing, then I usually remove and replace the cap in that stage.

[SoulFetish]

And most often if I'm replacing one, I'm replacing them all. So if there's one or two good ones in there, how much longer could they have if the others were bad and how much sense does it make to just replace the bad ones!?! But it gets done all the time. You open up an amp and find some original filters, but one or two newer ones.

Sure, if it were my amp, I would replace them all as well.. and I'm sure I'd end up rebuilding the most of the amp when it's all said and done. But, this is not always practical or warranted in a repair business. This can add significant cost in parts and labor to a customers bill. Particularly if it ends up being an added to a larger job. Often, a customers budget determines that only absolutely essential repairs are performed.

[g1]

I think the key is knowing the limitations of the testing and having a suspected failure mode.
If you think a filter cap is going 'open' and no longer filtering, the paralleling method works great.
But if you are looking for a leaky or shorted cap and unaware that paralleling will not work, you might conclude a leaky cap is good, because there is no change adding a good cap in parallel.