bad cap? (slightly high ESR reading, slightly low capacitance reading)

[kq702]

Hi I had a circuit board that had a problem with the monitor not turning on. Sometimes it would turn on, other times it would stay black and would not be detected by the computer. The problem got worse and worse, eventually the monitor did not turn on at all when connected.

I recently purchased an ESR meter and was testing the caps on this board, and I came across two posable caps that may be bad, (though no physical damage shows).

Below is a chart I made with all the info I gathered. Note all caps are rated at 6.3 volts, 1000 uV.




Cap 11 has a 10% lower reading then what its rated at, so could that be the cause of the LCD problem? Cap 13 also has a low reading, allthough not quite as bad as cap 11 so would that one be considered bad as well?

Note- I am only doing this as a test to see what a bad cap reading looks like. Unfortionatley I can not test this board even if I replaced the cap because this board was labeled as "scrap" several months ago and has since picked up a bit of damage on the CPU pins where the cpu attaches. But I am just using this as a test as I said, because when I was working on this several months ago I thought the problem was due to a bad capacitor but was not sure. Now that I have an ESR meter I figured I would try to see if I can locate the bad cap(s).

One other question I had is why the caps had a higher ESR reading when they were removed from the board? Seems a bit odd.

[mariushm]

Re: bad cap? (slightly high ESR reading, slightly low capacitance reading)

Capacitors 11 and 13 are a bit low in capacity.

Even though the capacitors are rated +/- 20% in general, they really shouldn't be more than +/- 5-10% out.

The ESR is however still low enough that the board should work with those capacitors.

Just because you can measure in circuit, it doesn't mean you'll get accurate results. ESR meters advertise they ability to measure in circuit because they use a voltage low enough not to enable other components in the circuit.. so it's more of a safety thing, it's not really saying it's "accurate measurement in circuit".

Some LCR/ESR meters use 0.5-1v pulses, which can conduct on shottky or even small signal diodes, so the voltage can go further on the board and potentially mess something up.

The ESR measurement is also not accurate when you have two or more capacitors in parallel - the esr decreases just like the resistance of two or more resistors decreases when in parallel:

1 / Rtotal = 1/R1 + 1/R2 + ... + 1/Rn

That's why often more capacitors are used than required - for example the VRM circuitry near the processor doesn't always need 6-8 capacitors there, but the manufacturer uses so many instead of 1-2 big ones to get lower resistance overall.

[kq702]

Re: bad cap? (slightly high ESR reading, slightly low capacitance reading)

thank you for the info.

I guess the part I did not understand is how when you look at the "in circuit ESR values", we can clearley see that cap 11 is unusually high compared to the others, followed by cap 13. But if you look at the ESR value for out of circuit (where the reading SHOULD be more accurate) they all show the same reading of 0.10. So in other words it seems the reading is more accurate when the chips are in circuit. Or could that be a coincidence maybe?

[mariushm]

Re: bad cap? (slightly high ESR reading, slightly low capacitance reading)

It can be lots of things, even something as simple as flux remains on the solder or oxides making poor contact on the leads.
Even heating up the solder when you extract the capacitor could heat up the internals of the capacitor (as the leads heat up as well) enough to get you a bit lower esr for a short time.

I wouldn't know what to tell you.

[PeteS in CA]

Re: bad cap? (slightly high ESR reading, slightly low capacitance reading)

As mariushm noted, 'lytics are typically +/-20% tolerance. So none of those capacitances are out of spec. However - as he also commented - good manufacturers like Nichicon or United Chemicon usually deliver product that is much tighter than the ratings. I looked up 1000uF, 6.3V in Nichicon's low impedance PW series, and the max. impedance at 20*C is .090 ohms (I'd expect real parts to measure in the .070-.085 ohms range). So maybe the caps you have aren't quite as good as PWs, maybe your meter's accuracy isn't engineering lab grade, but I don't think .10 ohms impedance is high enough to cause a problem.

You have to use in-circuit readings intelligently. If you are measuring a 10K resistor but your reading is 6K or 7K, the resistor is probably OK; if it's a 10 or 100 ohm and your reading is several K-Ohms, the resistor is probably bad; if it's a cap with high resistances across it (resistors or semiconductors other than schottky rectifiers), a low resistance reading means the cap (or some other component) is probably shorted. Measuring capacitance or capacitor impedance in-circuit and getting useful readings is pretty doubtful, I think.