dielectric absorption

[Bernhard]

http://www.national.com/rap/Applicat...570,28,00.html

[Christer]

I think that paper has been mentioned a couple of times before
in the numerous threads on DA. It is a good paper nonetheless,
although it seems to deal mainly with S/H applications and it
seems not immediately obvious to what extent it carries over
to audio (not meaining to claim DA is irrelevant for audio, just
that this paper may be).

NB, this was not meant as a complaint against you posting it.

[Bernhard]

There is an equivalent circuit on that website:




If this is used as a coupling cap between a low impedance output stage and a high impedance input stage, and there is a 1 minute positive single square halfwave, both Cs will charge and after termination of the halfwave, the parasitic C will discharge into the primary C (until both voltages on poles of the model and poles of the parasitic C are equal) and into the input stage.

The discharge into the input stage will be slow because of the high impedance.

Also the low impedance output stage can be seen as a virtal ground, so the whole thing is a charged cap with one side to ground and other side to a voltage divider.

If the input impedance is changed to a low value, the discharge will be faster and the error voltage much smaller.

Right ?

So coupling caps should be sounding better with lower impedance input and DA is not to worry about if the cap is used as PSU bypass cap

[hjelm]

I agree,
This problem should be most critical if you have say a rc filter on the input of a high impedance op amp. Not sure of any direct application for it, DAC filter maybe.

Maybe we should build compensation circuits for this on the inputs of all amplifiers. Could be interesting to try.
I suppose this could also be a reson for not using decoupling capacitors.