Compensating caps and filters

Magnus Danielson cfmd at
Sun Jul 2 14:12:48 CEST 2000

From: "jhaible" <jhaible at>
Subject: Compensating caps and filters
Date: Sun, 2 Jul 2000 02:01:47 +0200


(I keep misspelling his name since I do not know how to get the german u in

> >To make the filter oscillate you need positive feedback!!
> >IMO this doesn't count as the purpose isn't to improve a gain stage.
> >And quite tricky to compensate. *grin*
> You're partly right - the overall negative feedback becomes positive
> with the additional phase shift from the caps. What I forgot to mention is
> that each stage has a NFB loop to make the integrator a RC-type filter.

Yeah, it is all this gain-margin and phase-margin stuff. That is usefull
keywords if you want to dig into op-amp stability.

> The interesting thing is that a typical OTA filter unit, complete with
> gm cell, integration cap and voltage follower, plus NFB around that all,
> is quite similar to the ordinary opamp circuit. In fact, you could use
> an ordinary opamp, with its internal compensation cap, as a OTA-type
> VCF, if you could control the Iabc current of the internal differential
> amplifier. For fixed cutoff frequencies, this is widely known as
> "capacitorless filter". But it's going to be even better. In some old
> Farfisa schemos, I found a capacitorless VCF. They used a quad opamp
> with programmable current sources, the LM346. They actually built a
> one-chip capacitorless 4-pole LPF. The *topology* is remarkable similar
> to the SSM2040 and similar OTA filters. The difference is that instead of
> external 1nF caps it works with the internal (5pF ?) compensation caps,
> and a much lower Iabc to match. What performance would one expect
> from something like that ??
> I was curious, and so I breadboarded it. It works remarkably well for
> cutoff frequencies of a couple of hundred Hz or higher, but below that
> (nA range ?) stability becomes a problem. No wonder, as the external
> current not only sets the internal Iabc current, but also *every* other
> bias current of the opamp.
> Nevertheless, the results were astounding. I guess in the Farfisa
> Synthorchestra
> it's only used in the upper range (?) to control the brilliance of some
> sounds.
> But it's complete with resonance and all. When I'm doing the next batch of
> scans for my homepage, I'll probably include this circuit. Not recommended
> to build, but interesting for being the "missing link" between opamps
> and OTA filters. In a way, most opamps with NFB are OTA filters @ fixed
> cutoff frequency.

Thank you Juergen for this information. I knew that many op-amps have builtin
compensation caps but I didn't know that they have been used for
"capacitor-less" filters in products. On the topic of being interested in this
topology I raise a hand.

I also didn't know you where sitting on Farfisa schematics.


More information about the Synth-diy mailing list