Faulty capacitor, DMM/LCR measurement

[Ohmy]

Hi all,

I recently fixed something for a friend where the problem turned out to be a faulty electrolytic cap. I got the problem narrowed down to a specific
circuit but could not determine the exact fault by in-circuit measurements. I had it narrowed down to 3 possible components. I ended up "fixing" it by just throwing new components at it and it worked when I changed the cap. I initially checked if the cap was shorted, and it wasn't, it wasn't measuring a particularly low resistance either and it had none of the typical physical signs of being worn out.

The cap is 220uF. I tested the cap out of circuit with first an LCR meter and then a DMM. Pictures are below. (Sorry for the crap quality. I smashed the lens on my phone)
LCR measurements:


Nothing about these LCR readings scream dead electrolytic capacitor to me? Capacitance on point ✔ ESR OK ✔ DCR a little low but ✔
At this point I don't see why changing this cap fixed anything, so, I decided to also check the capacitor with a DMM.



DMM shows very low resistance and a super wonky capacitance measurement, so, now the DMM has told me the cap is faulty.
LCR works by analyzing an impedance and, I think, the DMM may be analyzing a time constant or something to get the values. So then I'm wondering,
which one is lying? What is going on here? Am I missing something? Is the LCR meter faulty? Anyone? 

[Phil1977]

Electrolytic caps often show their failure only when a certain voltage is exceeded. Or to put it in other words, they sometimes fail by lowering their max. voltage more and more.

Can you just connect the cap to a lab power supply and slowly increase the voltage? If you see a constant current flowing without exceeding the printed on voltage then this is the defect.

[Ohmy]

I did as you suggested and it is definitely, definitively, dead. With a constant current of 5.9mA at 5.3V putting it's DCR at ~900Ω



Just to be clear, are you saying these meters may show "correct" values because their test signal levels are, relatively, low? compared to whatever the maximum voltage of the cap is at the time of measurement.

[Kleinstein]

The LCR meter and DMM do the test with different voltage and frequency. With a good LCR meter the conditions should be noted and often one may also have some choices, e.g. in the frequency.
How the DMM reacts to leaky capacitor is often not well defined - an LCR meter may have a more suitable model, though with capacitors this is usually an RC series model and thus ESR.

For the the test with the supply it looks like there is excessive leakage, need for refomation. After some time, that capacitor may actually come back to life. I would still not reuse it.

[Phil1977]

Just to be clear, are you saying these meters may show "correct" values because their test signal levels are, relatively, low? compared to whatever the maximum voltage of the cap is at the time of measurement.

Yes, exactly. As Kleinstein wrote, some LCR meters have the possibility to superimpose a DC voltage during the test - that may help if the voltage reaches high enough.

But frankly speaking the Lab power supply test is the easiest. There are lots of more advanced tests possible, but if a cap has the specified capacity, a reasonable ESR and low leakage then it´s usually a good cap.

With some working experience a curve tracer is a great alternative.

[Ohmy]

Interesting. I don't often repair old equipment, repairs are not what I do, where failure modes like this show themselves. I will take this one as a lesson learned and remember that for the future. 

@Kleinstein the cap is definitely not going back in and in fact I'm going to change the rest of them as well. There aren't that many of them and since this one failed then the rest may be in a poor state as well.

[Martin72]

The pictures with the Fluke clearly show that the capacitor is defective.
The Fluke can only measure DC resistance and the resistance of a capacitor should be in the multi-digit M-Ohm range after a more or less short time after the meter has charged the capacitor.