Various Ways of Bypassing Phaser Stages?

[mth5044]

I've got a few questions on what happens when 'removing' phaser stages. Using the GGG Phase 45 as an example:

http://www.generalguitargadgets.com/pdf/ggg_p90_sc.pdf

I've had three different thoughts on taking it from 4 modulated stages to 2.
1) Tap the output of IC3/pin6 to the output mixer. Standard enough.
2) Lift the left side of C4 and C5, the side that connects to the output of the previous stage. I believe this pretty much makes the last two stages simple unity gain inverting stages. Besides needing a DPDT instead of a SPDT to this, are there any drawbacks when compared to the first method?
3) Lifting the drains of Q3 and Q4 to stop the modulation of the last two stages, making them a static allpass. The Phase 100 and Drolo's Liquid Mercury both have static allpass filters in them. I somewhat understand what the allpass filter is doing thanks to GEOFEX, the combination of the two making a notch at a certain frequency, but without a modulation source, and the first two stages being modulated, would this notch be noticeable? What is the purpose of these static filters in the phaser designs?

Thank you!

[Digital Larry]

Last summer I was messing around with DSP based phasers and multiple stages and individual modulation control of each stage, and spent quite a few hours in the mile high elevation of Lake Tahoe pondering such things.  Whether a phase shift stage is modulated or not, it will contribute ultimately to 90 degrees of phase shift (as the signal frequency goes higher and higher.

Modulating any given stage by e.g. reducing the FET impedance, makes that 90 degrees happen at lower frequencies.  Still though, you are shifting through all stages.

I wondered, does it matter, if I have 5 identical cascaded stages (or pairs, required to make a single notch), whether I modulate the first one, the last one, or one in the middle?  And I have to say, I don't think it does, because your end result is taken at the end of all of them, so how does "which one" was being modulated matter?  At one point I thought "but that's the one that controls the third notch, so that one will move back and forth, while the others stand still".  This is what thin air can do to your thinking.

I tried (in DSP) a 10 stage phaser, with individual control for each of the 5 notches stage pairs, and it was a total waste of time to do it that way.  At least to the extent I messed with it.  To my ears, the magic happens really close to one extreme of the LFO sweep, and not so much on the other end.

So, if you are shifting through 4 stages, and modulating 2 of them, is that different than modulating all 4, but not quite as much?  In this case, I'd think the answer is likely yes.  To fully understand it you'd probably have to get pretty detailed about what phase shifts happen at what control voltages and which frequencies and all that good stuff.  I wish I had a more solid answer!   I remain enamored of and fascinated by the phase shifter sound and also somewhat baffled by the interaction of the electronics, mathematics, and psychoacoustics.

[mth5044]

Thanks for the thoughts! I moved to Denver recently - I'm thinking the thin is air my problem too (I wish). I was near Sacramento for a stretch and hit up Tahoe a few times. Nice place!

So you're saying that placement of the modulated all pass filters doesn't really matter. In the phase 45 example, you could turn the first set of stages into the allpass or the second, and it would likely sound the same. You're also saying that adding unmodulated allpass filters would be like turning the depth down? I don't quite understand that, but I don't really understand too much about phasers to begin with.

[Mark Hammer]

Phase shift is additive.  So, the phase shift at frequency F in stage 1 is added to the phase-shift at that same frequency in stages 2, 3, 4, etc.  If some stages  are fixed, while others are swept, you're still adding phase shift across stages.

Normally, in a stage where you have a cap linking two stages and a FET or LDR to ground from that cap, more phase-shift is added the higher up you go.  The reverse can be true as well, where the two op-amps are linked by a resistance, and the cap goes to ground.

In both instances, you have a single-pole filter: cap with R to ground is a 6db/oct highpass, and R with C to ground is a 6db/oct lowpass.  Both of these are relatively shallow, having a gentle slope across the frequency spectrum.  So, if you have a couple of fixed stages of the C-with-R-to-gnd type, they add a fixed amount of phase shift, increasing as you go up in frequency.  Where that phase-shift becomes conspicuous will depend on the component values chosen.  If we were to use a .047uf cap, and 22k to ground, we would be increasing phase shift (up to a max of 90 degrees per stage), from a little above 150hz upwards.

The simplest, and least problematic method for having more or less stages is just to build a circuit with as many stages as you want, and simply tap the output of a given stage.  The qualification is that you need to be sure that the feedback signal has a way to return to the desired point.  Of course, by tapping a stage, you do not disconnect it.  So if you have 6 stages, and tap after 2, your feedback signal may be coming from stage 6 back to stage 2.  Whether or not you like the sound is your decision.

[mth5044]

Danke Mark, makes sense!