How to tell if a cap is bad if it's on a board already?

[Bret608]

Hi all,

On a recent vero build I did, I ended up tracking down a bad cap when it initially wasn't working. Luckily, it was easy because I could see some residue around one of the leads! This is what I get for using caps of unknown age that I find where I work.

Now, I've got another build over in Tech Help that could very well end up being a bad cap too if the problem is in fact on the board. Is there a straightforward way to use a DMM to tell if a cap is fried? Or would this be a job for an audio probe?

Thanks in advance! I know this knowledge will only help me going forward...

Cheers,

Bret

[culturejam]

I *think* you could measure resistance across the leads to tell if it was really bad. You would get a readable number if the cap is bad.

[Bret608]

Ah, because you would be reading the ESR, right? Let me know if any particular DMM range works best. I may try it out tonight.

[culturejam]

Ah, because you would be reading the ESR, right? Let me know if any particular DMM range works best. I may try it out tonight.

You want the lowest range setting possible. A good cap probably is below your meter's ability to measure, unless you have a really nice meter. If you get a solid number (no drifting of value, etc), it's probably a bad cap. ESR should be close to zero on a good cap.

On the other hand, it's possible that the cap can fail the other way (as a short, rather than being open). There's really no substitute for pulling the cap and measuring its capacitance with a cap meter.

[culturejam]

I just did a little googling on this. I'm not really sure you can get an accurate idea of what is going on in a cap with a standard multimeter and with the cap in the circuit.

[Bret608]

No problem! Thanks for looking into it.

Interestingly though, I did try this out on the board in question this morning. I'll update my Tech Help thread to that effect. But basically, I set my DMM for the 200 ohm range and read resistance between the two pads for every cap on the board. None of them read anything except one, a silver mica that was reading a very stable 123 ohms resistance.