It was a good read. I suppose we should be testing our caps that we suspect have failed before we re-cap.

At my work (nuclear plant), we replace all electrolytic capacitors on a 10-year preventative maintenance cycle to align with EPRI standards. (Electric Power Research Institute) This is for non-safety related / non-critical systems. Safety system's and critical system's electrolytic capacitors get replaced more often depending on duty cycle and environment. EPRI has tons of whitepapers and studies on capacitor degradation, along with preventative maintenance templates on hundreds of other component types and studies of their failure mechanisms. Unfortunately, the majority of the whitepapers and preventative maintenance templates aren't free for download unless you're a member.

Anyhow, you can't always judge an electrolytic capacitor by a bulge on top. By that time, it is far too late. Long before that it is usually out of factory specifications and can cause instability in the board it is installed on.

uscleo wrote:

...I suppose we should be testing our caps that we suspect have failed before we re-cap.

I don't think that you can test a capacitor without removing it from the circuit board. And if you're already removing the capacitor from the board, then you might as well finish the job and re-cap it.

Personally, I'm not very enthusiastic about re-capping. It didn't fix the motherboard that I tried it on. Maybe other people have had better luck.

STX wrote:

I don't think that you can test a capacitor without removing it from the circuit board.

You can test electrolytics in-circuit, proper ESR meters will let you do just that. The idea is the ESR should be the lowest value (miliohms to single ohms) of any impendance in a circuit so other components will not affect the reading much, if at all. There are a few cases where it's not possible to get a correct reading, like an electrolytic cap bypassed with a ceramic one for better frequency response for example. Still, it works most of the time.

I don't do "recaps" unless I suspect power delivery issues, and usually I'd check the ripple level with a scope before attempting that. There is no golden rule though, it depends on what device it is, what state it's in, what age is it, etc. On the other hand, if I find even a single failed capacitor in a PSU I'd usually replace all of them with low-ESR 105C ones.

Deunan wrote:

STX wrote:

I don't think that you can test a capacitor without removing it from the circuit board.

You can test electrolytics in-circuit, proper ESR meters will let you do just that. The idea is the ESR should be the lowest value (miliohms to single ohms) of any impendance in a circuit so other components will not affect the reading much, if at all. There are a few cases where it's not possible to get a correct reading, like an electrolytic cap bypassed with a ceramic one for better frequency response for example. Still, it works most of the time.

I don't do "recaps" unless I suspect power delivery issues, and usually I'd check the ripple level with a scope before attempting that. There is no golden rule though, it depends on what device it is, what state it's in, what age is it, etc. On the other hand, if I find even a single failed capacitor in a PSU I'd usually replace all of them with low-ESR 105C ones.

I recap based on "age", but mainly appearance.

STX wrote:

uscleo wrote:

...I suppose we should be testing our caps that we suspect have failed before we re-cap.

I don't think that you can test a capacitor without removing it from the circuit board. And if you're already removing the capacitor from the board, then you might as well finish the job and re-cap it.

Personally, I'm not very enthusiastic about re-capping. It didn't fix the motherboard that I tried it on. Maybe other people have had better luck.

Cases and cases, i have 2 boards that recapping made it work, i have 3 boards where the problem were different one, for example the last socket 7 board that i recapped(it had bloated caps anyway) didn't solved the issue, after investigation i discovered that it was a faulty transistor, now i'm gathering replacements part to replace the affected parts.

I thought it was interesting that a number of people who seem to have more experience than I do (with electronic *work*) have much less experience with specific failures (tantalum cap explosions, failed caps in monitors/motherboards/car ECUs) - I've a number of each of these under my belt.

Maybe I just have bad luck. 'Swhy I don't buy lottery tickets.

Deunan wrote:

You can test electrolytics in-circuit, proper ESR meters will let you do just that....

Thanks! I didn't know about ESR meters.