[sdiy] IR3109 (OTA-based) filter question

[jh.]

> Now that's where OB confused me. It uses switch as you need, according
> to the theory to dich that pole, so OB opted for much smaller cap
> making it much farther. I don't now how ladder would react to such
> thing, maybe you could eliminate cap entirely,

That's what I would expect: small cap needed at CEM for stability
of local feedback system - omit it in ladder version (2044) entirely.

>  but yes, there's also
> the issue of the switch that would add/disconnect larger cap to maybe
> smaller one or none (like 4066 or 4052). Note that 4th integrating cap
> in OB is much smaller due to the CEM's lack of internal buffering
> between stages (thanks to JH for clarifying that), and 2044 in poly6
> does also have one cap smaller that others (thou IR chip in JX3P is
> standard buffered so it uses all the same caps).

I have made some Spice simulations with unbuffered filters, and whatever
the topology, I always got bleedthru "backwards" over the resonance
OTA if I didn't make the 4th stage different. Maybe that's why all
such filter chips have it, too.

JH.