Help analyzing TC Electronics Phaser Controls...

[JKowalski]

I've been trying to figure out how the TC electronics phaser accomplishes it's "function" knob control.

This is the only schematic I can find for the phaser...

And here is a demonstration video showing the control in use.


Now, the schematic that I linked above is not very well drawn - it looks like a total mess, basically. Half of the phase stage outputs don't even exist in the drawing for example, its like a jigsaw puzzle of circuit snippets.

I've been going mad trying to make sense of it, there are like 5 feedback loops (if the schematic can be believed...) The lower right corner op amp appears to be the main loop, mixing the raw input and the after all-pass signal and then moseying over to the function pot (which is a volume control for this signal) that outputs to a mixer (which has four inputs!) right before the all-pass stages...

I tried simming a basic model of what I took as a block diagram of the circuit in falstad's filter app, and it gave me some strange results that I couldn't really intuitively connect to the sound change so I figured I (misled by the schematic) screwed up somewhere along THAT line.



Basically... Can anyone make heads or tails of how this control works or how one might go about creating a control that accomplishes a similar function? Before I had the schematic I assumed it mostly dealt with flipping peak/notch on the frequency response (that can be done by switching between input->output mixing or output->input feedback) but after messing with that idea and getting nowhere close, I turned to digging for a schematic. Frustratingly, it hasn't really helped except to tell me that my first notion does not look like the right method!

I am going for that ability to move the phasing into the really hollow sound and up to the piercing sound. (2:30 in the video) I haven't seen a control like this on any other phaser but if you know of something, please mention it!

[StephenGiles]

I don't think it can do much more than a Badstone.

[JKowalski]

I don't think it can do much more than a Badstone.

I'm not really sure what you mean. What control are you referring to that is similar?

[Taylor]

That schem makes my eyes bleed so I can't analyze it. But, I think a continuous notch/peak control could be done by using the "switch hitter" amplifier that I've recently been turned onto by this forum - place this amp in the feedback loop. Its control pans between inverting and non-inverting, so at one end you get maximum negative feedback and at the other, max positive feedback.

http://www.diystompboxes.com/smfforum/index.php?topic=87327.msg734038#msg734038

[JKowalski]

That schem makes my eyes bleed so I can't analyze it. But, I think a continuous notch/peak control could be done by using the "switch hitter" amplifier that I've recently been turned onto by this forum - place this amp in the feedback loop. Its control pans between inverting and non-inverting, so at one end you get maximum negative feedback and at the other, max positive feedback.

http://www.diystompboxes.com/smfforum/index.php?topic=87327.msg734038#msg734038

Yeah, that's the basic idea that I started out with and didn't get far with... Even though it does create what you would think a notch/peak control would do (i.e. pans between sweeping notches and sweeping peaks in a frequency plot) I could never really get it sounding like the TC phaser - it tended to color the sound differently but didn't give you that intense change from really hollow to really sharp. Checking out the terrible schematic I have, it looks like something else is going on...

I think I'll go back to that inv/noninv concept for a little while to refresh my memory as to its possibilities. (My phaser has been on a breadboard so long that I can barely remember all the stuff I've done with it!)

It's frustrating when you try to design a circuit based not around a definable goal but rather a sound... Experimentation, experimentation!

Thanks for the input

[slacker]

Looks like what it's doing is taking a mix of what ever the signals at D1 and B1 are then depending on the position of the control feds some of it back into the non inverting input of  U6. U6 looks like a differential amplifier, it's got signals going into both its inputs, so my guess is depending on the position of the pot something either gets cancelled or boosted, maybe

[JKowalski]

Looks like what it's doing is taking a mix of what ever the signals at D1 and B1 are then depending on the position of the control feds some of it back into the non inverting input of  U6. U6 looks like a differential amplifier, it's got signals going into both its inputs, so my guess is depending on the position of the pot something either gets cancelled or boosted, maybe

D1 is actually an output from that point at the function control to the 4-input op amp... the signals being mixed into the function pot can be found at the leftmost lowest op amp stage - mixing the output of the all pass filters and the input of the clean signal...

C1 is what I don't really get. the signal comes from the first op amp stage in the all pass path and goes to both inputs of the 4 input op amp... the C1 input seems to be a VCF because of the photocell? So maybe there is a filter wobble thrown in with the phase shifting?

[slacker]

Oh yeah, D1 is an input to the opamp, I misread it. That makes even more sense then because with the function pot on max, you've got the same signal going into both inputs, although maybe not at the same level so they will cancel out.

[JKowalski]

Oh yeah, D1 is an input to the opamp, I misread it. That makes even more sense then because with the function pot on max, you've got the same signal going into both inputs, although maybe not at the same level so they will cancel out.

It makes sense, but it seems like a kind of rube-goldberg way to do a simple volume control... AHA but maybe it lets it flip polarity?

I got to check it out again, haha


EDIT:

YES, it does flip polarity as you turn the knob, so Taylor, you were right in your suggestion! Its a "switch hitter", just buried in a badly drawn schematic and a bundle of other components.

Now I just got to figure out why it didn't sound great the first time round....

Thanks a lot guys.

[JKowalski]

Well, for some reason the switch hitter isn't really "hitting" the right sound in the ways I have tried it. On a frequency plot, it mainly serves to "tilt" the strength of peaks and notches from side to side (one extreme the lower frequency peaks are higher, the other extreme it looks like you flip the entire frequency plot horizontally, almost perfectly. It's a definite difference but not in the way I am looking for - strange, I was so excited that I could hit the sound after my epiphany yesterday....

I went back to sim some of my original work, the mixing between input+allpasses->output and output->input using a dual ganged pot and it definitely gives you what you would describe as a peak notch control, i.e. in one extreme there are perfectly formed peaks and the other extreme there are perfectly formed notches...

Going back to the breadboard tonight to screw around some more, argh! Hopefully something interesting and unique will come out of all this.

[slacker]

I don't think it is just inverting the polarity.
What I think the part of U6 with D1 as one of the inputs does is actually boost signals that are out of phase on both its inputs and cut signals that are in phase on both its inputs. This is different than what normally happens in a phaser, usually where the wet and dry signals are mixed you don't get any boost, where the signals are the same phase the output is the same amplitude as the inputs and signals with the opposite phase are cut. If you invert the polarity you just get notches in a different place.

[JKowalski]

I don't think it is just inverting the polarity.
What I think the part of U6 with D1 as one of the inputs does is actually boost signals that are out of phase on both its inputs and cut signals that are in phase on both its inputs. This is different than what normally happens in a phaser, usually where the wet and dry signals are mixed you don't get any boost, where the signals are the same phase the output is the same amplitude as the inputs and signals with the opposite phase are cut. If you invert the polarity you just get notches in a different place.

Well, even though it may be boosting its still not going to do what you'd intuitively think it should do - it would still be basically a mirrored frequency plot, left & right, no matter the gain. The peaks and notches will always be the same, just moved around when the polarity flips like you mentioned - the boosting just adds peaks into the mix as well.

So maybe we are looking in the wrong spot? Maybe this control has some interplay with the actual all-pass network values or some other section that gives it its abilities?

In any case I am heading off right now to hook up the mentioned IN-OUT/OUT-IN mixer (which sims to an ideal peak-notch mixer, but I really don't know if an ideal peak-notch mixer IS that sound)...

[JKowalski]

RESULTS:

Not what I wanted. Flipping the response from peaks to notches just sort of mitigates the phasing, which makes sense looking at the frequency plot - even though it's a totally different shape (peaks vs notches) it has the same GENERAL shape, just kind of the other way around.

Now my new theory is that the flipping left to right is what causes the sound change... maybe the function knob actually focuses the depth of phasing either in the lower spectrum or up in the higher spectrum... The 'hollow' sound is just deeper bass phasing and the "sharper" sound is just the deeper treble phasing... Would that make sense, listening to the sound clip?


This thread is kind of my thought process, if you didn't notice... I'm just hoping that someone might see something I don't  

[slacker]

It could be doing something like, I had a phase 45 with variable bias and you could get similar sounds to the notch setting with that, adjusting the bias just made it sweep lower frequencies. The peak setting sounds almost like a high pass filter to me.

I simulated a bit of the circuit including U1 and the bit of U6 that "B" comes out of and the "switch hitter" and as well whatever it does to the phase, you get a big change in frequency response with different settings of the pot. There seems to be a lot of treble boost at one end and sort of a mid boost at the other, so I don't know if that's got anything to do with what it does.