Capacitor Response Time? Blue Meter?

[bigbeark]

I have recapped several boards that were unresponsive and ran hotter than I thought they should. All the old caps were high-quality Rubycon MBZ and Sanyo WG.

When I tested them in circuit with the ESR meter, their values were all low.
When I remove them from the board and test ESR and Capacitance, all are within spec.

But when I recap with the same high quaility NEW caps, the boards run cool and without annoying lagging.

My theory is that as caps age, they become less responsive, IE they take longer to reach the deisred ESR and capacitance. I have seen this for sure on caps that fail ESR tests, the ESR SLOWLY falls point by point, wheareas the new caps seem to fall almost instantaneously to the low reading.

Do those of you who use the Blue meter observe this phenomenon of fast versus slow response in your meter readings?

I invariably end up recapping the entire board, good cap and bad cap brands all get tossed.

It would seem to me that the boards with the good caps never really die, they just sort of fade away..

Whereas the Jackcons, JPCons etc fail early in the VRM section. I brought an ECS board back to life by replacing only the VRM caps and the board is REALLY quick, I suspect because there were very few hours on the board when the VRM caps failed.

Your comments welcomed.

[severach]

Re: Capacitor Response Time? Blue Meter?

I had a 3.3v buck circuit supplying about 2 amps through a FDS6982S SOIC-8 dual N channel MOSFET. With the original dented Panasonic FJ 1500uf 6.3v ESR=0.05 the FDS6982S would burn itself out within a minute. The buck inductor would also run hot. With a new Nichicon HM 1500uf 6.3v ESR=0.01 the circuit runs forever cool as cucumber.

Perhaps the dent damaged something that doesn't show on a capacitor/ESR meter. Perhaps cutting ESR by 5 greatly increased the current capacity of the circuit.