DOD FX20 Stereo Phaser - need output stage explained

[MrStab]

Hi guys,

I was wondering if anyone could explain the reason behind the differential (?) amp in the 2nd output of the DOD-FX20 Stereo Phaser (U10B). Schematic here: http://www.freeinfosociety.com/media/images/1905.gif. Is this a method of minimising potential cancellation in the stereo field? If so, why does it work?

I understand that the other output (U10A) is an inverting summing amp because the input buffer is inverted, so i guess that's moot. It's what the differential amp does relative to the other output that matters.

I did a bunch of vague reading on this kind of stereo output ages ago and i'm only now realising i mis-read a lot of it. So currently the phaser schematic i'm working on has plain inverted & non-inverted outputs which, when i actually thought about it, seemed like a bad idea. Before that, i had the output of the last phase stage inverted one extra time and mixed with the non-inverted signal (again, caused by misinterpretation). Just for curiosity's sake, would either of those work out reasonably? my mind says no. I've seen such a raw configuration in other stereo effects on Google Images, but they weren't phasers.

So anyways, an overview of why the FX20's output stage is the way it is would be really appreciated.

Cheers!

[Mark Hammer]

I have one of those.  One output sums the phase-shifted and dry signal, and the other provides their difference.  The Regen control works nicely for one of those outputs, and does diddley squat for the other.

[MrStab]

the other provides their difference.

I'm kinda embarrassed to admit this, but i think it's the root of my problem here: mathematically i more or less understand explanations of what differential amps do to the numbers, but i just can't picture what it means in the real world whenever i read about them amplifying "the difference". i just don't get it. the only example i kinda understand is recombining balanced signals, but that doesn't really give me a good all-round insight as those signals are usually exact opposites. i've read at least 3 or 4 different explanations, but they just seem like number machines to me. i think i'm a visual learner. or aural learner? er...

[Mark Hammer]

Well just imagine you have two sets of graphic EQs, one with every other slider set to boost, and one with every other slider set to to cut.  Now move the center frequencies around.

[MrStab]

i'm still confused, to be honest. for theory's sake, say you have a unity-gain differential amp with 2V vocals going into the non-inverting input, and 1V guitar going into the inverting input. 2-1=1, if i'm not mistaken (that took me hours). would you hear the mixed signal at 1V, but in the same proportion as the sources? so the output is basically dragged down by the inverting input voltage?

[MrStab]

assuming i have the above right, the purpose is to change the output voltage on the 2nd channel in accordance with the sweep, so the notches are never at the same level as the 1st. so on its own, that channel might have a sort of pulsing, semi-tremolo sound, though it's probably kept really subtle to the point it's not noticed. that would prevent problems caused by a live or studio engineer mixing em at the same levels (unless they compressed it or something). umm... right?

i'm jumping the gun a bit here because i'm not sure if i understand it right yet, but i have another question: if the dry signal is in-phase, could the differential output stay the same, or would that opamp's inputs need reversed?

[PRR]

> ...vocals going into the non-inverting input, and ...guitar going into the inverting input.

Two different signals. Phase hardly matters.

Take guitar. Call it A.

Pass a copy of A through an INVerting stage. Call that B.

Mix A and B equally. They cancel. (+1)+(-1)= 0

(Does not happen when you mix voice and guitar, they are never equal or even similar.)

To mix A and B without cancellation:

1) Invert one of them and add. (-1*(+1))+(-1) = 2

2) Subtract them. (+1)-(-1) = 2

> DOD-FX20 Stereo Phaser

I am not going to trace-through that mess tonight. We seem to have an original signal (A), and a processed signal (B). Somehow the B path has an odd number of INVersions. Simple additive mix would cancel. An inverter could be added but that's over 19 cents. A subtraction gets the right result and can be done with about the same parts as a good adder.

[MrStab]

it's all simple stuff, but i'm borderline numerically illiterate so forgive any misunderstandings:

so invert + combine = cancelled, then subtract = uncancelled? if so, i think i get it.

when you mix voice and guitar, they are never equal or even similar

[MrStab]

been racking my brain a bit and it seems to me both the dry and the wet signal are inverted, passing through an odd number of stages in total, as stated. but the wet signal probably passes through an even number of stages in the 3x LM13600 version mentioned on the linked schematic - in which there's no mention of a change to the output stage. this makes me think that if the input buffer were made to be non-inverting, the OUT1 channel output mixer could also be non-inverting and the OUT2 channel could be the exact same as in the schematic, because the inversion in that stage only existed in the first place to reverse that of the input and the differential amp would still apply the same effect in output without phase cancellation issues.

tl;dr: if the input stage and OUT1 were non-inverting, would OUT2 be the same as in the schematic or would it need changed?

i think i'm uncertain about the differential amp's effect on phase, but i've read that inversion isn't really an applicable concept with that type of amp.

[MrStab]

I am not going to trace-through that mess tonight.

Duly noted Paul, but...

We seem to have an original signal (A), and a processed signal (B). Somehow the B path has an odd number of INVersions. Simple additive mix would cancel. An inverter could be added but that's over 19 cents. A subtraction gets the right result and can be done with about the same parts as a good adder.

this confused me - wasn't sure if you were talking about the circuit or what could be done to improve the circuit. the reason for my confusion is because as far as i can tell, both the dry and wet signal are inverted until the mixer, and i took your post to imply only the B path was.
as far as i can tell, OUT2 remains inverted post-mixer. gah, i'm clueless!

[PRR]

Are you fixing one? Building one?

If so, you can observe what it does.

If this is only theoretical-- some times an inversion puzzle is such a brain-strain that it is easier to just build it 95% and work out the last bit (to invert again or not invert?) by hit-or-miss.

[MrStab]

Apologies for the delay, got a bit sidetracked. I'm building one. i basically have board space for stereo - it's non-essential, but i figure it could come in handy and relatively-few additional parts would be needed. plus the board contains a buffered-only bypass system for potential stereo use.

I'll pick up another breadboard at some point (ran outta room) and see what i can come up with. i was hoping to have a better understanding of the theory before either building a stereo headphone amp or carting a half-stack and a combo around to test stuff, but it seems this one's a bit on the obscure side. i'll try to analyse it in my head a bit more when my 1 byte brain cache clears up.

[amz-fx]

The inverted input is passed to both output opamps and re-inverted so the output has the correct polarity. Good so far.

The processed signal is phase delayed by the four LM13600 stages. As in a standard phaser, the phase shifted signal is added with the same polarity to the input signal (at U10A) and two notches are produced in the output signal.

When the processed signal is added in opposite phase to the other opamp (U10B) then the non-shifted signal is cancelled and the phase shifted section produces a pair of peaks instead of notches.

So, you get pseudo-stereo with two sweeping notch filters coming out of the Out1 jack, and two sweeping bandpass signals coming out of the Out2 jack.

If you were to combine the two main outputs, the signals will all cancel (or nearly so).

regards, Jack

[MrStab]

Thanks for the explanation, Jack, and sorry for the delayed response.

So U10A is inverting purely to compensate for the input stage inversion, but U10B is configured to deliberately ensure that the wet and dry signal are out of phase with one another? do i have that right?

if so, maybe my original plan to create a 2nd channel by inverting the wet signal and regular summing with the dry signal would work? just more parts. i think this is what Paul was alluding to. i'm still a bit uneasy on why the inversion had to happen in the first place, is it an unrelated design choice, maybe based on the fact that inverting amplifiers are more stable (afaik)?

cheers!

[Dreverb]

Hello, any idea about how to mod this pedal to sound more univibe??
Thanks anyway