looking for a resonant shelf EQ schematic

Fenris said:

Designing something from scratch is beyond me.

Click to expand...

You don´t need to, practically everything has been done and published on the net already.
There is a difference between a synth sytle voltage controlled filter, like the steiner vcf, which is mostly used in
music ´generation´, and a filter like the K&H or Neumann w495, etc with the purpose of recording and mixing. You´re probably aware, but just in case...

For the latter kind the only SallenKey implementation that I know of is the drawmer ds201, which i mentioned above.
These are used for tailoring the sidechain bandwith, and are the opposite of subtle tone controls!
I think all your descriptions scream ´state variable´! Here´s an easy one:
http://www.diystompboxes.com/smfforum/index.php?topic=45772.0

Does the resonance control in the Craig Anderton EQ actually affect the HPF and LPF? It looks like a standard state variable circuit. I found a PDF of the book but it's not explicit.

Yes.

Fenris said:

Does the resonance control in the Craig Anderton EQ actually affect the HPF and LPF?

Click to expand...

Yes it does. When the resonance is high, all three modes (LP, BP, HP) exhibit a sharp boost, which becomes the dominant character. At extreme settings all three modes will sound almost undistinguishable.

Sorry to be a pest, but I still can't get it to work. Here's the schematic of my EQ section:
http://fenrisaudio.com/monoeq.html

The high band is on the right. I cut the connection from U6b pin 7 to R38, while keeping the connection from U6b pin 7 to R63. Then I connected U1a pin 1 to R38. All I get is oscillation. Does the boost/cut circuit need to be hooked up differently?

Fenris said:

Sorry to be a pest, but I still can't get it to work. Here's the schematic of my EQ section:
http://fenrisaudio.com/monoeq.html

The high band is on the right. I cut the connection from U6b pin 7 to R38, while keeping the connection from U6b pin 7 to R63. Then I connected U1a pin 1 to R38. All I get is oscillation. Does the boost/cut circuit need to be hooked up differently?

Click to expand...

Look at the resulting arrangement. The output of U1a goes to the inverting input of U1b (via R38 & CR13). The resulting inverted output (pin 7 of U1b) goes back to the inverting input of U1a (via R65). Perfect example of POSITIVE feedback. You need to either add an inverting stage or reinject the signal to the non-inverting input of U1a.
Please note that, after applying this fix, additional phase-shift at high frequencies, due to propagation delays in the opamps, will produce positive feedback, hence oscillations, if you don't reduce adequately the BW.

It worked! Thank you so much! I got the Z-shaped curve I was looking for, with a bump at the top of the slope and a dip at the bottom. I can probably eliminate the bump and keep the dip by combining the BP and HP outputs.

I used the unused section of U15 for the inverting amp. I should also add a DC blocking cap. How do I limit the bandwidth? Increase the value of feedback cap on U1b? Add a feedback cap to U1a?

Fenris said:

It worked! Thank you so much! I got the Z-shaped curve I was looking for, with a bump at the top of the slope and a dip at the bottom. I can probably eliminate the bump and keep the dip by combining the BP and HP outputs.

Click to expand...

The bump and dip are related to the slope and amount of boost/cut. You can't get rid of them without seriously altering the slope. Indeed I find the dip very useful on the LF section because you can add some beef without mud. You must play with the Q control. [/quote]
How do I limit the bandwidth? Increase the value of feedback cap on U1b? Add a feedback cap to U1a?
[/quote] Don't fix a problem that doesn't exist. If the oscillations are gone for good, leave it alone.

Boost is working fine, but I'm getting what appears to be DC in the output when I cut. The sound is distorted and the channel fader crackles like it's getting DC. I put a 33 uF electrolytic cap in between the HP output and the boost/cut section, but it's still there.

Fenris said:

Boost is working fine, but I'm getting what appears to be DC in the output when I cut. The sound is distorted and the channel fader crackles like it's getting DC. I put a 33 uF electrolytic cap in between the HP output and the boost/cut section, but it's still there.

Click to expand...

That may very well be an oscillation. Try adding small caps across the FB resistors.

Here's a video that demonstrates the Undertone Audio EQ.

http://vimeo.com/26815302

Starting at 9:00 it explains how they used a potentiometer instead of a switch to fade between a bell and a shelf. You use the "Shape" pot in conjunction with the "Q" pot to get a variety of shelf shapes. With "Shape" at 50%, increasing "Q" creates a large bump at the top of the slope and a smaller dip at the bottom of the slope. With "Shape" at 100%, it's reversed, you get a large dip and the bottom of the slope and a small bump at the top of the slope. I was able to get the 1st type of shelf, but not the 2cd type, which would be a lot more useful. It has a silky sound that I really like. Any ideas on how they accomplished this?

John Roberts mentioned combining the BP output with the HP or LP output. I tried it, but it prevents the SVF from functioning properly.

I finally found an explanation of how SVF's work, courtesy of Handbook for Sound Engineers.

At high frequency, the low-pass sections attenuate the signal so that the feedback signal is small, leaving the unaffected signal at the high-pass output. As the input frequency approaches the center frequency, the levels at both the bandpass and low-pass outputs begin to increase. This leads first to an increase in positive feedback from the bandpass section giving a damping dependent overshoot. When the input frequency is below the center frequency, the net phase shift of both low-pass sections is 180 degrees, leading to negative feedback and an attenuation of the high-pass output. The cutoff frequency of the filter in Fig. 23-18 can be changed as in the preceding circuits by varying R 1  and R 2 or  C1  and  C2  while keeping other values identical. The damping factor is varied by changing the band-pass feedback gain, controlled by the ratio of R3 and R4.

Click to expand...

The Undertones sound good and look very flexible!
I bet it´s (at least the parametric section) done with state variable filters also.
And the shape control fades between states/outputs aka mixing different amounts of lp/bp or hp/bp outputs.
The amount of bp mix (post the feedback tap) needs to be gain x2.
The q would be bp feedback damping...

What was your issue with combining lp/hp with bp?
(I think it´s also needed to get a gentle -6dB slope from a two pole filter...?)

A short, but comprehensible explanation is here about output combinations and bp gain, etc:
http://www.analog.com/static/imported-files/tutorials/MT-223.pdf