Constant Amplitude Phase Shift EQ design

[TomWaterman]

Thanks for chiming in Ed...

Do you not get any severe curve interaction on the Odd?

In the simulator (these things aren't particularly real! LOL) I get pretty strong curve interaction even with all inductors.

The curves become heavily asymmetrical.

I suppose your bands don't overlap much... I'd expect to see it when the LF is set to 150 and the LMF @ 300Hz?

Thanks Tom

 

[peter purpose]

My eq uses the caps circuit for the mid frqs. V.nice.

 

[TomWaterman]

Peter yours is SWEET!

Did you sell any yet?

How are the curves on that one? I guess as you say it sounds very nice, you have no issues! LOL

:guinness:
-T

 

[TomWaterman]

Also Peter, you have a switched Q... do you find you lose overall gain when Q is at its widest?

Cheers T

 

[TomWaterman]

I have been messing with more bands and have got a three band in the simulator with a 50Hz inductor LF, a CAPS MF and a 10kHz inductor HF.

First thing I noticed is the interaction between the bands, the Dove article shows individual band bypass which can also be seen on the Opamp Labs EQ of similar nature.

I notice the Buzz MPE1.1 has individual band bypass switches, but the Purple, Peters EQ and the Buzz 500 EQ do NOT.

Have you guys - not found an issue with bands interacting in a manner that didn't always sound good?

This picture shows the 50Hz inductor LF peak affected by a CAPS MF whose gain pot is set at midway but the resonant circuit is not bypassed. Look at the effect on the 50Hz curve as the freq pot is varied on the CAPS circuit:

Also, with the bands set with gain at midway (no cut or boost), the CAPS freq pot has an affect on the overall frequency response.

This is the effect of the inductive LF band on the MF CAPS band (with the LF gain pot set midway) - loss in MF boost as freq gets nearer the LF band:

Bizzarely the 10k HF band produces some funny interaction - the curves are OK when the MF freq is high, but when low, the shape change is drastic (again this is with the MF gain pot at midway):

These are the lowest and highest freqs on the CAPS MF boost/cut, with curve interaction from the LF and HF (again both set with gain midway):

This is the change to the LF with the CAPS MF bypassed and unbypassed as shown by Dove:

This is the CAPS MF sweep with the LF unbypassed by gain set midway:

This is the CAPS MF sweep with the LF bypassed:

Hmmm interesting.
-Tom

 

[jeth]

I have been looking at the CAPS article and some others trying to gather thoughts for a simple 4 way eq section.

I will be using standard opamp gyrators with a swinging input amp similar to the CAPS set up, or in fact closer to the eq described in Forssells article.

Currently thinking about bypass switches and was checking the way they are implemented in the caps circuit.. and, I'm not entirely sure what's going on. The input of the gyrator/caps section is switched to ground, leaving the pot wiper floating. On all but the lf gyrator section the input is shunted via a 10K resistor.
So..why shunt the input to ground? The capacitors in the filters have a discharge route to ground via the pots. It would seem to make sense for the pot wiper to shunt to ground via a resistor so that with all bands bypassed the eq amplifier would still have a reference??? (Or perhaps it's assumed that you would never use an eq with all bands bypassed?

I'd be grateful if anyone can help me understand this arrangement, the 10K R value and its position so I can work out if this is a viable method of bypass for my circuit. It appeals as it avoids a double pole switch directly in the audio path to bypass the whole eq amp and will be easy to wire between pots and switch.

 

[TomWaterman]

From memory, the resonant section is switched away from the wiper to ground via a 10k as you say... but can also be left floating as in the API553.

Bypass is essential otherwise the filters really interact and change shape, even if they are set to midway on the gain pot.

In terms of the circuit, you'll really want a 10k resistor at the positive opamp input to ground bypassed with a small pF cap, equivalent to your feedback network to get the H.F stable.

10k and 120pF wouldn't be a bad starting point.

-T

 

[jeth]

Ok, Thanks Tom.. I thought that adding a ground reference resistor at the input of the swinging input amp might affect boost/cut symmetry? Or did you mean at the input of a seperate buffer amp?


Would I be right to assume that the 10K resistor via which the gyrator inputs are shunted to ground would be chosen to match the feedback value R?

 

[barclaycon]

What software are you using to plot those curves ?

Regards.