FW: SSM2040 Secret unveiled !

[Haible_Juergen#Tel2743]

 ----------
From: Haible_Juergen#Tel2743
To: Joachim Verghese
Subject: Re: SSM2040 Secret unveiled !
Date: Tuesday, 2. April 1996 15:59

> So you do get *symmetrical* clipping, after all. Or?

Yes, I think so.

Even in a stand-alone 2040, the clipping would mostly be symmetrical,
because the 2nd stage cuts what is left from the 1st stage and so on.
The only exception is the one I had reported: Large signals at rather
high frequencies.

At high Resonance settings, most of the power circulates inside the filter, 
so
*there* it may apply to the P5 as well. But then again, high resonance
only leads to much overdrive in the neighbourhood of the resonance peak
frequency, and not way above it ...

So really, I am still learning ...


BTW, found another interesting thing about the OB-8 filter. While the
architecture is really the same as in the OB-Xa, there is a *slight*
difference in component values.
The 3320 data sheet suggests a 91k/100k/200R feedback network
(instead of the usual 100k/100k/200R for standard OTAs), because of
some imperfections in gm stage output impedance.
If you use the 100k/100k/200R nevertheless, you get less than unity
gain, but all you have to do is adjusting the resonance feedback gain and
vca gain. *But* the OB-8 mixes stages with 91k and 100k - this means that
in 4-pole mode the poles are a little detuned to each other !!!
(Just from first glance, so don't take it for truth!)

Regards,

JH.