schematic for a 24db or steeper Bandpass filter????

[Martin Czech]

These emails contain no information of private character.
I think that it may be usefull for the list.
So I send a cc to the list as well!


:::How could I change the design to make it better for higher Q?

As long as you stay with a multi feedback structure, you can
only improove by op amp with large gain bandwidth product and slew
rate.


:::Any suggestions on the best BPF for this?

Basically you have 5 choices:

1. MFB, your circuit: Good passive component tolerance sensitivity
   not so good for high Q (>20 or so, depends on frequency)
2. Sallen & Key: less components, but critical tolerance sensitivity
   The Sennheiser Vc something uses this. Low gain op amp
   thus no problems with higher Q and high frequency
3. State Var: good component tolerance sensitivity, but more components
4. Delyanis: Sallen & Key with additional positive resistive feedback.
   Can have very high Q (I will not show the circuit because E.N.)
5. Latice and other structures: less literature, but advantageous

See: 

Lancaster: filter cookbook (beginners reading) 
Chen: The circuits and filters handbook
Williams: Active Filter Design
and of course E.N.

I think that only one section per channel is not enough.
The Sennheiser had 3, others even 4. Because you need good
far away damping, and this does not depend on Q but on the number
of stages. At the end they all roll off with -6dB per stage.

I'd suggest to use 3 or 4 moderate Q (5..10) stages in series, not staggered.
Maybe you include some switch to select the nuber of active stages.
An overall feedback per channel can give two poles with very high Q.
Very high Q gives lot of ringing, this is not desireable, since it
makes "reverbating" effects, can be good for effects, though.
Some kind of clipping in the front end should kepp all other
internal nodes from clipping (soft characteristic, compression).
Overall feedback around the whole filter bank should be tryed.

m.c.