Combined Q&D VCF, Phuncgnosis and Dr. Quack to make the perfect auto wah

[BluffChill]

Sorry for the misleading title, I didn't have enough space to say I was 'trying' to make the perfect auto wah!!

So I've been messing around for a while trying to make a really good auto-wah, and have borrowed elements from a few sources to make it:

- Dr. Quack (based on Dr. Q)
- Phuncgnosis (Escobedo)
- Phuncgnosis (Deadastronaut mods)
- Q&D VCF (Escobedo)

I originally tried the phuncgnosis and it was ok but a little too resonant, and sounded kind of...blowy. So after a couple of iterations I came up with this.



It sounds pretty decent, but occasionally a little fuzzy, like the clipped signal from the envelope is bleeding in.

You'll notice as well that I've added a resonance control - some resistance where C4 (680pF) meets the output of Op Amp B (I'm using a TL072, 4558 worked pretty good too). This is one of the features I borrowed from the Q&D VCF but a 10k doesn't seem to give much variation - a 50K was better.

I also found the 'release' pot added by deadastronaut didn't do very much so left that out and replaced it with a 470k fixed resistor. I also omitted the much maligned Up/Down control featured on the phuncgnosis - I'll be damned if I could get that to work in either Escobedo or Deadastronaut's ways.

I also omitted the 330k resistor to ground after the envelope (D2 and R6 on my schematic) because it didn't seem to do anything.

Just interested to hear people's thoughts on this. Will post a demo soon!

[garcho]

are you looking for the perfect *one-chip* autowah?

[BluffChill]

are you looking for the perfect *one-chip* autowah?

Fair point - yes I am! Obviously there are a lot of more advanced models which trump this, like the Mu-Tron etc.

[bool]

Could you use a BJT instead of the diode for the control element? That would decouple the envelope from the audio signal.

[BluffChill]

Interesting - do you mean D2? How would I orient it?

[Mark Hammer]

I'm going to recommend using a photocell/vactrol for the control element.  Either that, or use a VERY good fullwave rectifier circuit, or aim for as quick a decay as you can tolerate.

The problem is that envelope ripple is very audible in such circuits during the decay phase.  If the ripple is reduced, then you don't hear it.  If the filter has completed its sweep before the ripple becomes objectionable, you don't hear it.  Photocells have a built-in smoothing characteristic that smudges out whatever ripple remains, making for a nicer-sounding sweep.

My experience is that downward sweep requires more than merely inverting the envelope.  Because of where the harmonic content occurs in the lifespan of a plucked string, the time constants need to be different, and the start and stop points of the filter also need to be a little different.  That can certainly be done, but requires a more complex circuit than many of us - myself included - are aiming for.  So, you were wise to simply sidestep it.

Different decay times can be achieved by varying the resistance of the component you have labelled as R12.  But as you noted, one tends not to hear big differences by installing a pot.  This can be addressed by using something like a 3-position on-off-on toggle that selects between sufficiently different resistance values, where decay-time differences are more obvious.  To do that, start with 1M as a "default" value of R12, and use the toggle to add resistances in parallel to achieve values closer to 22k and 150k, which achieve approximately equal jumps in decay time (150/22 = 6.8, 1M/150k = 6.7).

[garcho]

^ Speaking to that point, some old synths and many VST type filter plug-ins that have envelope followers have something called "inertia". Basically, that's referring to how much ripple you hear. If you hear none, then it no longer behaves like an envelope follower. If you hear too much, it becomes audible and behaves more like a glitchy ring mod. Disco duck is in the middle somewhere. If you happen to have a plug in with inertia control for the envelope, try messing with it to get an idea of what you like. It really depends on what you think you'll play on the guitar while using it. Different styles, sensibilities, aesthetic and all that might guide you to picking a decay/ripple compromise that you'll be more satisfied with. I think that's usually a big part of why people often get disappointed with autowah - it's either too much or not enough. And a big reason why people end up using OTAs and NE571 and all that jazz noise.

[anotherjim]

Here's an example of BJT filter control. It's from a Farfisa Bravo organ, but is single supply albeit 15v. The unseen control to the Q12 is a keydown envelope trigger. As such, it won't suffer from control envelope ripple.
Note the placement of Q12 emitter and collector. This is deliberate.
I keep meaning to breadboard this. It might work on 9v with little or no changes.

[Kipper4]

Nice. I like it Jim.
Bridged T iirc.

Might be worth playing with c65 47pf.
Keydown envelope..... might have to google the organ.
Cheers

[anotherjim]

Keydown just means it can be worked by a keyboard contact and that can trigger a generated envelope not much different than a synth AD envelope. Even if you used an envelope follower circuit running off the notes as you would a guitar effect, the dependable on-off nature of organ notes would make it much more controlled and predictable.

[garcho]

bridged t yes, but the transistor makes it *special*

[Kipper4]

Ok I'll ask.
How intresting.
Presuming the transistor dumps current ?

Am I over thinking.

[anotherjim]

The transistor is a voltage controlled resistor. There's a resistor across it (R137) to range the sweep. The swapped e & c of the transistor is an old trick to get better control response. To be fair, there isn't much that Google will find you about this application. Searching for "Transistor as a voltage controlled resistor" invariably throws you a JFET, even if you add BJT in the terms.
It certainly isn't a Farfisa speciality. An older version uses a BJT amplifier stage instead of the op-amp.