Poll: electrolytic capacitors

Tantalums are surprising that way. The inductance is package limited. The leaded part inductance is all in the leads. The surfmount parts are a couple of nH.

The ESR of a 10 or 22 uF surface mount tantalum is just right to stabilize a lot of linear circuits, like LM1117s or from the output of (some) opamps to ground.

Once I figured them out, maybe 30 years ago, I've had no more tantalum failures. But I could say that for lots of different parts, not just tantalum caps.

One trick for a 317-type regulator is to use two tantalums, one at the output and one from the ADJ pin to ground. The second one reduces output noise and also reduces the dV/dT of the output, so you don't have to derate the output cap.

My big problem with tantalum is that they can survive a surprising long time when reversed biased. (Some one stuffs them in the board backwards. I guess this is mostly a problem with thru hole and humans.) I had one fail after years of intermittent use. (35 V tant on a 5 V supply.)

George H.

There was a reasonably useful discussion of this about 5 years back in EDN (believe it or not). I have a copy at

Cheers

Phil Hobbs

Sorta. Aluminum capacitors are dog food parts. ;-)

On a sunny day (Mon, 9 Feb 2015 13:44:33 -0800 (PST)) it happened George Herold wrote in :

Maybe because for some reason unknown to me the tants have a '_' sign where the + is supposed to be, and normal electrolytic caps a '_' sign where the - is supposed to be?

Yes I once fell for it. Nothing to do with through hole though.

I had one fail after years of intermittent use. (35 V tant on a

I have replaced tons of electrolytics in TV sets. Not shorted, but usually high impedance or high inductance (foil came lose). So the failure mode is less spectacular. The result is the same: It does not work.