recurring problem with geofex simple parametric eq

[blackieNYC]

Need advice.  Built as shown, using the given cap values of .15u and .022u, and also with the pair .022u and .01u, this EQ oscillates (at the selected center freq) on power-up unless you are in the lowest resistance/highest frequency of the 1meg center frequency pot.  If there is too much resistance between the 51k resistor and Vbias (the 1M B-pot is in the last 10%, so somewhere around 100k) the oscillation will occur on power up.  Occasionally it may oscillate after having been powered on for a while, the the immediate power-on oscillation is easy to recreate.
I thought maybe the gain (low to mid frequency hump) or the narrow Q I desire were the cause - but I can back off on both substantially and the same thing occurs.  If I were to fire up the pedalboard with the frequency maxed, I could avoid the problem.  But with this EQ, my 2nd, I had planned to find the sounds I want, and then substitute fixed values.
This is currently on the breadboard, which one might suspect is part of the problem, but on the parametric (same circuit, different freq) I have already boxed up, the oscillation will somewhat quietly bleed everywhere if the pedal is bypassed, if I power up with the freq pot set below its maximum.
http://www.geofex.com

            EDIT: see fix below!

[R.G.]

How clean is your power supply and ground setup?

Power bypassed with electro and ceramics near the devices?

[blackieNYC]

A boss pca supply, 100u, no puffs. I'll stick a couple in there.

[R.G.]

OK, which circuit exactly are you using? Need the schematic as you built it.

[blackieNYC]

The circuit is exactly as your drawing, but a single band. The spare half of my TL072 is properly terminated. No subs. When I experienced the problem, I started using the cap values listed in the text - those are the ones in my post.

Edit: for anyone playing along, this is under RG's "FX Projects"

[R.G.]

I'm at least temporarily stumped. It's been a while (the article is from over 16 years ago  ) but I have used those same circuits before and since. That's why I was pointing you at power supply and grounding issues. Opamps can get fussy about grounding and feedback both when run at unity gain and when simulating resonant circuits, both of which are happening here.

Here's one you might try. I'm still thinking it's an opamp quirk. But the gyrator opamps are running as voltage followers. You can sub in NPN emitter followers for the opamps in the gyrators if you arrange to bias them. That circuit has been used in a lot of Boss and Ibanez pedals as a resonator. The Boss SM7 uses a batch as I remember. Or change out the JFET input opamps for a bipolar input.

If not one of those, it's get-out-the-scope time. 

[blackieNYC]

Looks like the BJT op amp works!
The 5532 has yet to oscillate.  Looks good.  Went back to the TL072 and it goes into oscillation pretty consistently.

[R.G.]

Looks like the BJT op amp works!
The 5532 has yet to oscillate.  Looks good.  Went back to the TL072 and it goes into oscillation pretty consistently.

Yeah, I figured it had to be some opamp quirk. The 5532 has that low input impedance thing, but it's probably OK for this. The very high input impedance of the TL0xx series and also the CMOS opamps make even a few pF of stray PCB capacitance able to start oscillation. Mostly you get away with it, but sometimes it bits you.