[sdiy] discrete SSM2018?

[jhaible at t-online.de]

Czech Martin schrieb:
> Well, I've listened a lot to the 2218 chip, and I 
> couldn't 
> hear anything unpleasant. I mean there was attenuation,
> or not, as desired, and that's all, no artefacts,
> almost ideal. Since the device could take 20Vpp
> I never got it into distortion, so I never experimented
> with that. Certainly distortion would be no good
> application because of sharp clipping anyway.
> I just wanted a near to ideal vca for audio
> and dc voltages at the same time.

I cannot decide that - it may as well be a myth.
Did't want to give it that much importance.

More interesting: What artefacts will be produced with
not ideally matched transistors (diskrete solution) ?

Is there an advantage for the either the 
[unity gain, variable gain] method or the [a, 1-a] method,
or is it basically the same ?

JH.