Capacitor ESR, how much is too much?

[Randall]

I am looking to purchase or build from a kit an ESR meter. I am a little confused by some of the info and discussions I read concerning what the readings actually mean in terms of using them to help determine the health of a cap. I even see on a eevblog discussion about different brands of caps being taken into account. I'm not sure I need to get that geeked out over this, I'd just like to know how to apply the measurements in a practical and timely manner. I find some charts, but am not sure quite how to apply them.

[Enzo]

I will totaly make up some numbers:

What if you have two caps, and they both measure ESR of 5 ohms. Look at the data sheet for each. One has an average spec of 4 ohms and the other has an average spec of 6 ohms. One is higher than normal, the other lower. SO if I wondered if the ESR was climbing on a part, I'd need to know what to expect in the first place.


Just my opinion here: Think about all you have ever done in this area - amp electronics - so far without an ESR meter. Now specifically, what are some instances when you really needed one, and how would it change things? It is easy to get a new tool that makes new measurements you never had before and now you wind up wondering if every cap has a bad ESR reading.

[olddawg]

Back in the day I repaired a lot of camcorders. They had a lot of miniature smd electrolytic caps that caused all manor of problems in the frequency range that they were applied. I had 3 different ESR meters that could be used in circuit. All would give different readings. I got tired of callbacks eventually from caps I missed and would shotgun dozens of caps. Eventually the manufactures actually provided "fix it" kits with all of the caps. It was a royal pain in the ass. For audio I have usually found it less time consuming just to replace a suspected cap. In newer equipment with switching power supplies I would shotgun the 5 or 6 caps in a heartbeat. That's just my experience. It's like ESR meters were a passing fad that generated callbacks. But your mileage will vary. I had one good Sencor cap checker. I wouldn't let anyone else in the shop use the damn thing. It tested caps at operating voltages and worked very well. But it was difficult to set up and the other techs were always blowing up caps with it and causing a lot of mayhem. It was a nice expensive toy for a while but by the time you removed the cap, set up the checker, then performed your test... you could have just replaced the damn cap anyway. AND.. I learned a long time ago. That with a $200 repair most customers would rather see a bag with a few parts than one $.50 capacitor which they would then give you grief over!

[Mick Bailey]

I use my ESR meter all the time and I find it to be a really useful tool. I don't get hung up on minor differences in readings between caps, or look at charts. Just whether the reading is high for the particular size of cap. I'm mainly looking at electrolytics and in (say) an early Boss digital delay can very quickly identify bad caps without even powering it up. Some of the Peavey DFX boards have about 40 SMD electrolytics and a single bad one with high ESR can cause all kinds of problems. It takes just a few minutes to check every single one on a board to find out where the problem lies.

After 50 years of sanding internal curves by hand, I've just bought myself an oscillating bobbin sander. What a difference - wish I'd bought one years ago. The same with my ESR meter; when I started using it it made certain jobs so much quicker that I wondered why I'd ignored them for so long. Horses for courses, as they say. Different people work in different ways and it's whatever works and makes sense to you.

[nickb]

The key is you are looking for a significant difference from the norm. Practically speaking, oftentimes an amp will have a number of identical or similar capacitors. When you run down them with your meter and one 2x the others, you know it's going bad.

[mozz]

Some of us not only repair but we build, modify, and design. We like to know all the specs we can get. If you download Duncans PSD power supply designer, it has a editable resistance for the power supply caps, change that and see the results, never too late to learn. Just have a 68 dripedge princeton on the bench, filter caps are working fine. But the owner wanted me to check it over and replace what i though was needed. Filter caps were a little high and a little low (16mf-25mf) (20,20,20,20). The new can cap is a lot tighter spec and was lower esr. I trust the new cap, not the old one even though the amp was working fine, it does idle 3 watts lower now on 120v. There was leakage, i can't test at rated volts so i use value, esr and dissipation on my test equipment to help me decide.

[Enzo]

To expand on olddawg's point, I figured out years ago that parts are cheaper than labor. At a dollar a minute or so, how many minutes do I want to spend testing a 30 cent cap I suspect?

There are many tools that some never use, and others wouldn't do without. In my day the scope component checker was all the rage. The Huntron Tracker was a commercial model of it, but you could make one with a few simple parts. Scope across a part, and you get a curve on the screen. Many techs swore by them, I never found them very useful. Even a staple like the oscilloscope, some guys reach for it first in almost all cases, other guys tend to use it as a last resort. Look up "scope component tester".

I think cap testers that can apply real world voltages to caps are very useful. Like old heathkit and Eico units.

[olddawg]

I use my ESR meter all the time and I find it to be a really useful tool. I don't get hung up on minor differences in readings between caps, or look at charts. Just whether the reading is high for the particular size of cap. I'm mainly looking at electrolytics and in (say) an early Boss digital delay can very quickly identify bad caps without even powering it up. Some of the Peavey DFX boards have about 40 SMD electrolytics and a single bad one with high ESR can cause all kinds of problems. It takes just a few minutes to check every single one on a board to find out where the problem lies.

That was my attitude with the camcorders at first.... but over time I found that if there were one or two smd electrolytics of similar values with a bad ESR... there would probably be more within 6 months.

[Mick Bailey]

Taking the Peavey DFX board as an example; would you shotgun it anyhow and replace all 40, whether any were bad or not? how would you know if there was a cap problem if you didn't check?

Here's what I do; it takes just a few minutes to run down the rows of caps. If none are high I know the fault isn't a cap problem (though it often is). It takes a lot longer to replace every cap on one of those boards than it does to check them. If one is bad I know the others will follow and replace the lot. The same with Fender Deville/Hotrod/etc amps - it takes 5 seconds to check every PSU electrolytic. They may not show any signs of leakage or causing trouble, but I know if only one shows higher than usual ESR I can bet those caps have a limited life. Why would I not spend 5 seconds?

I look at ESR like reading resistors - I don't look at whether a resistor is 99K or 102K if it's supposed to be 100K Ohm, I don't even look if its close - just whether its open or (near) shorted. Kind of like logic states. So a cap to me reads good or bad, high or low, depending how you look at it. I'm not advocating the use of ESR meters, just saying how I do it. The same with a ring tester - I personally wouldn't be without one.

[olddawg]

Taking the Peavey DFX board as an example; would you shotgun it anyhow and replace all 40, whether any were bad or not? how would you know if there was a cap problem if you didn't check?

Here's what I do; it takes just a few minutes to run down the rows of caps. If none are high I know the fault isn't a cap problem (though it often is). It takes a lot longer to replace every cap on one of those boards than it does to check them. If one is bad I know the others will follow and replace the lot. The same with Fender Deville/Hotrod/etc amps - it takes 5 seconds to check every PSU electrolytic. They may not show any signs of leakage or causing trouble, but I know if only one shows higher than usual ESR I can bet those caps have a limited life. Why would I not spend 5 seconds?

I look at ESR like reading resistors - I don't look at whether a resistor is 99K or 102K if it's supposed to be 100K Ohm, I don't even look if its close - just whether its open or (near) shorted. Kind of like logic states. So a cap to me reads good or bad, high or low, depending how you look at it. I'm not advocating the use of ESR meters, just saying how I do it. The same with a ring tester - I personally wouldn't be without one.

I’ve been out of game for over 10 years. But just before I got off the bench full time it seemed that practically everything made had dozens of 22uf and 47uf @ 10 and 16v smd electolytic caps the size if a corn kernel on double sided boards. They would dry out in droves. You could measure them with an ESR meter and usually identify and replace a few obviously bad ones. Trouble is if you sent it out like that, you risked weird intermittent problems as the other ones aged and had a similar failure mode. Now maybe things have changed. Maybe the components are more reliable and you have the luxury of replacing a single bad cap with a catastrophic failure. But I (and it seemed everyone else) had to replace buckets of the damn things daily.

[g1]

I wonder if this was around the time of the "capacitor plague" (1999-2007)?
https://en.wikipedia.org/wiki/Capacitor_plague

[The Dude]

I'm in Mick's camp on this one. I do use my ESR meter quite a bit. It's a quick way to suss out a bad cap or caps. You can quickly buzz down a row of caps and find inconsistencies. IMO, you aren't so much looking for some number in a book as just something inconsistent or way off the mark. In other words, I won't worry about a cap that's ESR reads 4 ohms on the meter when it's supposed to read 2 ohms. I will worry about it if I see 100 ohms. And, it's taken only seconds to find that cap.

[Enzo]

As long as you also look at the circuit context. The leaky one may have a parallel resistance or some such.

It is also what you are used to doing. I agree it is more about looking for odd results than it is looking for precise results.

I used to work on logic boards a lot, and I might see rows of parallel outputs, like to a switching matrix. (or inputs) And they would go to some octal chip and I'd use an ohm meter to go down the row and measure resistance to ground from the IC pins. It didn't matter what the reading was, but if 7 pins measured 10k and one measured 450 ohms, that was a clue. That was easier than figuring out which pin was which strobe line.

[olddawg]

I wonder if this was around the time of the "capacitor plague" (1999-2007)?
https://en.wikipedia.org/wiki/Capacitor_plague

Makes sense to me. I left the bench in 2003 and was servicing all of the major brands.

[g1]

I think the capacitor plague was responsible for a lot of the younger techs now wanting to do total recaps on everything as routine maintenance. We never used to assume that a 10 or 15 year old unit needed all new caps, now it seems to be the accepted idea.