Noob Question: Possible to mod Phase 90 for stereo out?

[Jrockpsycho]

Is it possible to modify Phase 90 output mixer to be stereo?

[Phoenix]

OP was edited after I replied, I'll post a correction shortly.

[Phoenix]

[Mark Hammer]

What sort of "stereo" are you thinking of?  There are several different formats that all call themselves such.

• phaser effect on one out, clean signal on the other
• only phase shift on one, only clean on the other
• clean plus phase-shift on one, clean minus phase-shift on the other

[Phoenix]

What sort of "stereo" are you thinking of?  There are several different formats that all call themselves such.

• phaser effect on one out, clean signal on the other
• only phase shift on one, only clean on the other
• clean plus phase-shift on one, clean minus phase-shift on the other

Quite true. Without a more specific details, any number of answers could qualify.

[Jrockpsycho]

Hi everyone! Thank you for replying! I am trying to mimic the output of the eventide instant phaser, a "pseudo stereo" which they describe as such:
"The left and right outputs were also out-of-phase with respect to one another".
"each output comes from a different analog phase shift filter and, as a result, the outputs are decorrelated"

So here is the full schematic, as I'm guessing from that statement each of the outputs would have to connect at a different point...




Could this work?

[idy]

The simplest thing is to try dry out on one side, phased on the other, seperating the two inputs to the mixer. See if you like. Otherwise you will have to have some more phase stages... Daughter board...? the Evantide had tons of stages right? Like 6 plus... I think the ones that "do that" have more than 4 stages. You could try "tapping" between stage 2 and 4, maybe that's the next easy thing to try....

[Jrockpsycho]

It had 8 stages, yes. I will be using 8 stages as well. Here:

[Mark Hammer]

I will offer the standard advice I offer to anyone aspiring to go beyond 4 stages.

• There is a risk of accumulating noise - hiss in particular - with ever more stages.  Stick a small value cap in the feedback loop of the last phase-shift stage.  So in your case, one would stick a 1000pf cap in parallel with R18.  This is especially recommended if one plans to include any feedback/emphasis.
• There is a risk of oscillation with adding more stages.  Each stage is supposed to be unity gain, but even with 1% resistors, it is possible for a stage to have a teensy amount of gain.  With more stages, each "teensy" gets multiplied by another teensy and another, such that even a small amount of feedback from the last stage results in oscillation, as that "teensy" gets reamplified and reamplified.  There are two partial solutions.  First limit the bass content in the feedback path.  Second, use a trimmer to bring the signal back from the brink of breaking into oscillation.
• Your drawing indicates you are using TL061 single op-amps.  If you are not planning on running the pedal off batteries, abandon those for TL072 dual op-amps.  MXR used them because everything ran off batteries or an onboard transformer in the '70s.  Look for pics of the early Phase 90s and you will not find a jack for external power.  Why are TL072s a better choice?  First, singles require more layout space and a bigger board than duals.  Second, TL06x chips are roughly 3x as noisy as their TL07x counterparts.  Their redeeming feature is that they do not draw much current, which is great for LFOs and such.  But for audio purposes, a TL072 is a much better choice.

Dual outs complicates matters for bypass switching, since one may have to switch not only the input and a status LED, but TWO outputs.  That latter requirement will depend on the kind of stereo you envisage.  Many phasers that provide dual outs will use electronic switching, rather than a simple stompswitch.  As Boss discovered in the late 70s, many modulation effects work by combining a dry and wet signal.  If you simply block the wet signal but leave the dry as is, you "cancel" the effect.  So they would use (and continue to use) a single FET to either block or unblock the wet signal from the mixing stage where dry and wet would normally be combined.

How would that work for a "stereo" phaser?  One way is to feed the phase-shift signal to two mixing stages.  One takes the phase-shift and combines it with dry in-phase.  The other takes the same dry, but inverts the phase-shift to yield a different product.  Blocking that single phase-shift signal from reaching the two mixing stages cancels the effect...in stereo.  I have several phasers that do this.  Increasing resonance does nice things for one of those combinations, but nothing for the other.  As well, running those two outputs simultaneously cancels out the phase-shift effect "in air" for a wide arc of the listening area.

Probably a better approach is to run two sets of 4 stages in parallel, feeding them off a common input, but each set having an independent LFO.  If using the P90 circuit, you can easily obtain dual-ganged 500k pots to simultaneously adjust the speed of both LFOs.  The dual LFOs will not be in sync with each other, leading to a richer-sounding effect.  Moreover, you won't get the "cancelling out in air".  You'd need to use electroinc switching to lift the wet signal on both subcircuits at once.  You could use a stompswitch, but it would need to be 4PDT which are obtainable, but big and expensive.

So, as noted earlier, which kind of stereo is a pivotal question to be asked and answered.  There ARE suitable solutions for every type, but you need to have a clearer sense of what will be required for accomplishing the switching by considering the stereo format first.

[Jrockpsycho]

Ok so after a ton of searching, I found this project online with eight phase stages and pseudo stereo, This is exactly what I want to do, although their phase stages seem somewhat more complex.
I see that output 1 is happening before the phase stages, and output 2 is coming out of stage 8, which was an idea someone mentioned earlier in this thread. So I'm just hoping someone could show me how to integrate output 1 in the p90 schematic.
Thank you all again for the help! 

[slacker]

Output 1 on that schematic isn't before the phase stages, it just looks like it is the way the schematic is drawn.If you look at the switch marked "S2" that is taking the output of the phase stages and mixing it with the input signal so output 1 is similar the normal output on the Phase 90. It's not exactly the same because that design is using feedback which the Phase 90 doesn't have.
Output 2 is the output from the 4th phase stage. So for your second output you could copy the existing Phase 90 output mixer but connect it to the output of the 4th phase stage.

[Mark Hammer]

The late great Ray Wilson designed a few things for guitar, but primarily for synth.  This phaser sort of straddles domains.  It is not really stereo, though.  Ray himself describes it as "pseudo-stereo"

Again, not to belabour the point, but what sort of stereo do you want/need?  What is it you want the stereo to accomplish for you?   Do you simply want a "bigger" soundstage?  Are you aiming for some sort of spatial swirl?

[Jrockpsycho]

Hey mark, I hope to mostly use it on vocals, and use the stereo for widening effect. Going for the Beatles kind of wide phasing you hear on their records (ie vocals on across the universe or Lennon's gimme some truth). I know ADT was used but just trying to get in the ballpark of the phase sound this created. Todd Rundgren used the eventide to similar effect in his productions. Anyway I am hoping this explains what I am going for a little better, and I appreciate your help! Thank you

[Mark Hammer]

Yeah, after listening to Lennon's vocals on that tune, I doubt that any sort of phase-shift effect is going to get you in the ballpark.  It could interesting, but it won't be that sound.  Now, a fairly short fixed delay (e.g. 10-15msec), going to two lightly-swept phasers might do it, but the "spaciousness" is decidedly coming from use of delay.

You have to consider that voice doesn't have all that much bandwidth beyond a wee chunk of the mids.  Phasers don't create many audible notches and peaks (8 stages will only give you 4 of them), and they tend to be rather spread out.  So it would be a bit of a stretch to expect any comb-filtering of voice to be audible.  In contrast, delay-based comb-filtering, as found in rotating speakers, flangers, chorus and doublers, yields many more notches and peaks, which are also closer together, such that even if they only sweep over a narrow range would still be easily heard with voice.

It is a delicious sound, to be sure, but it's gonna take more than Ray's or anyone else's phaser to do it.

That's my take on it.  I'm certainly open to dissenting opinions and contrary evidence.

[Jrockpsycho]

Interesting, Yeah I kind of figured that ADT or some form of tape delay would be necessary to achieve that exact sound. I do have the Waves ADT plugin and it sounds great, but would still like to build this project as I find it interesting and educational.

Funnily enough, this is what Eventide had to say about their 8-Stage Instant Phaser "Eventide's Instant Phaser unit employed a series of eight carefully tuned, FET-based, analog all-pass filters that delivered sweeping comb-filtration effects that sounded similar to — yet distinct from — tape flanging"

[Mark Hammer]

Phasing CAN be made to sound as dramatic as flanging.  Former member Mike Irwin demoed a 24-stage phaser for me (may still have the soundfile somewhere) that he has made, and you'd swear it was a flanger.  But he demoed it with white noise as the audio signal, such that there was something for every one of those 12 notches to comb-filter.  I've noted here many times over the years that the "classic" flanger sounds that motivated us to buy one were generally used in post-production on a multi-source mixdown, covering a big chunk of the audible spectrum.  As such, all those notches were more audible than using it on a single guitar.

The dramaticness of any sort of comb-filtering, whether time-based or phase-shift-based, will depend on aspects of the circuit design, but every bit as much on the signal to be processed.