Phase 90 MODDING

[Moejax]

Allright, here it is: I've been unable to find a single thread entirely dedicated to modding phase 90. Im talking overkill modding: get the drill, half-dozen 3 way switches, the drawer full of assorted caps and resistors, to be able to tweak and to squeeze out every little bit of versatility out of this nice little phaser! Allright fellow modmen, hit it.

[tele_guitarist]

Most of what you are wanting to know is out there already, it just isn't all in one thread. For starters, experiment with capacitor values everywhere you see the .047mf caps. Raise one up to .22mf, and another down to .01mf and you will hear big differences in this alone.
Experiment with the infamous r28...stick it on a pot to control depth. There are a ton more things to be done to it, but that'll get you started.

[Joe Hart]

What do changing those caps do?
-Joe Hart

[Mark Hammer]

There are a couple of things you can do with just about any phaser, whether a P90 or other.

One is to lift the straight signal from the mixing stage.  When phase-shifted and straight are combined, you get notches.  When you have phase-shifted alone, you get vibrato (pitch wobble).  Looking at the schematic for the Pez 90 from Tonepad ( http://www.tonepad.com/getFile.asp?id=42 ), you will see a pair of 150k resistors on the far right, lift either end of the one that comes straight from the input op-amp and you'll get vibrato.

A second, that Tele_guitarist noted, is to vary the amount of regeneration.  Many phasers have a variable regen control, but the P90 did not.  The amount of regeneration, and subsequent emphasis of the peaks and notches, is governed by the value of the famed resistor R28.  In the Tonepad schematic this is the 22k resistor in red that straddles the 4th and 2nd phase-shift stages.  The lower the value, the more signal is fed back earlier in the signal path, and the greater the emphasis of the notches and peaks.  It is extremely unwise to have a resistor value LESS than the value of the resistor found in either the feedback loop or input of the phase-shift stages, since this will introduce oscillation.  However, a wide range of values greater than that can produce useful variations in tone.  T'wer I, swap the 22k fixed resistor for a 12k or 15k fixed resistor and a 50k pot wired in series as a variable resistor.  This will give a feedback resistance of 12k (or 15k) to roughly 60k (those 50k pots are really about 47k), which should provide a nice range of near-piercing resonance to imperceptible regeneration/resonance.  Probably a good idea to stick a cap in series with that path as well.  The cap would go between the output of the last phase-shift stage and the pot/resistor.  To restrict feedback at lower frequencies, a good value might be .047uf, though you can consider moving up to .1uf or even down to .033uf.

The cap changes suggested by Tele_guitarist redistribute the phase shift, such that the maxima and minima occur elsewhere.  In the case of the Uni-vibe, the cap values are selected to distribute the phase shift more broadly rather than focussing it.  This is what provides a more watery, squishy, chewy sound than any obvious notches and peaks.

Finally, although the P90 is not really set up to provide an initial phase-shift control (i.e., the point where the phase-shift "starts"), the 250k trimpot can sort of fulfill that function since it adjusts the bias for the FETs.  I chassis-mounted a pot to replace the functionally-equivalent trimpot in my Rocktek Phaser, and can get a vastly broader range of sounds from it now.  Just note that the "usable" portion of that pot's rotation may only be as little as 1/3, but within those limitations, you may be able to achieve phase-shift sounds that range from throaty to glassy.

[Khas Evets]

Mark-

I was on your site and noticed that the file for your article on 'Phaser to Phasefilter mod for Small Stone' is missing. Does that mod apply to a P90 as well? Will this be posted soon?

Back on topic. R.G.'s Phase 180 Plus covers much of what Mark mentioned. It also covers adding additional stages. You could also reduce the number of stages for a Phase 45 sound. I'm planning to build a 2/4/6 stage P90 one of these days.

Another idea I had but never tried is to invert the phase of the feedback line to produce negative feedback. Has anyone tried this?

[Mark Hammer]

I was on your site and noticed that the file for your article on 'Phaser to Phasefilter mod for Small Stone' is missing. Does that mod apply to a P90 as well? Will this be posted soon?

1) It's one of those things I have to "get around to".  Now that you've reminded me, I'll see if I can tend to it soon since it is definitely worth doing.

2) No, this mod only applies to those phase-shifter designs using OTAs for allpass filter stages.  This includes the Small Stone, the later Ross phaser using the dual LM13600s (not the earlier ones using FETs), the DOD FX20, and anything with a similar type of design.  In the case of these designs, changing where the cap is tied to changes the actual type of filter mode (from allpass to lowpass).  In the case of LDR or FET-based designs, changing the cap connections will do nothing of any particular use.

[RickL]

Another idea I had but never tried is to invert the phase of the feedback line to produce negative feedback. Has anyone tried this?

I did it with a Ross Phaser. I had to fiddle with the mixing resistor to get a good phasing sound when inverted (lower I think) but aside from that it worked fine. I used Craig Anderton's phaser as a model for the mods I did to the Ross.

I don't know if it will work with the P90 but it's a simple enough mod that it's worth a try.

[Joe Hart]

Very interesting about those cap changes, Mr. Hammer. I can never thank you enough for your great info and insights!!

Another mod is to change the output resistor from a 150K to a 130K. The 150K is a little too high and creates a slight volume boost. 130K should give you unity gain.

And I think I remember hearing about taking out caps 11 and 12. This is supposed to mellow out the sound a little bit (which is what people are generally trying to do when they want it to sound more like the "script" P90's).

I have also found that playing around with the trimpot can give some different sounds.

And don't forget doing a true bypass mod!

-Joe Hart

[Fret Wire]

Another mod is to change the output resistor from a 150K to a 130K. The 150K is a little too high and creates a slight volume boost. 130K should give you unity gain.

If you got that information off Vanhansen's site, remember that's for the Dunlop reissues. That was a cut and paste from one of my old posts. 130k is ballpark, because it depends on whether you do the R28 mod, or both the R28 and C11 and C12 mods (remove). The older block and script models aren't so bad, level wise. Removing C11 and C12 really mellows the phasing past script, it's more what I call the "Phase 70 mod". It does work great with acoustics or clean tones.

The trimpot merely bias's the jfets, it's either optimum or not. But you're right, it does affect the sound.

IMHO, the best mod for the Phase 90 is to replace the 150k output resistor with a pot. Phasers and flangers can be sensitive about what pedals are before and after them, especially OD pedals. A volume pot allows you to get unity, boost, or less than unity. I think it's probably one of the more useful mods. Other than a volume pot, match the jfets, balance the wet/dry mix, and bias it correctly. That's all the Phase 90 needs, IMHO. The P-90 has a character of it's own, why mod it to sound like other phasers.

It's gets to the point where instead of playing through your effects, you end up playing with your effects. Frank Zappa said it best, "shut up and play yer guitar". Replacing every resistor with a pot, and every cap with a switch is ok for home or the studio, not practical to use live.

[Mark Hammer]

The trimpot merely bias's the jfets, it's either optimum or not. But you're right, it does affect the sound.

...The P-90 has a character of it's own, why mod it to sound like other phasers.

It's gets to the point where instead of playing through your effects, you end up playing with your effects. Frank Zappa said it best, "shut up and play yer guitar". Replacing every resistor with a pot, and every cap with a switch is ok for home or the studio, not practical to use live.

In total agreement with all of that.  

Chances are that altering the trimpot affects the nature of the sweep as well as the starting point of the sweep.  At a certain point in its rotation you may well have an interesting starting point in the sweep but at the cost of a clumsy "turnaround" in the sweep cycle.  Tthat can be a musical tool as well.  Of course, whether it's YOUR sort of tool is another thing.

Ideally, one *would* want to make re-instatement of the "classic" P-90 sound easily feasible.  In that spirit, one might want things like a regen-defeat switch that switches from a variable regen control to a fixed value (22k or whatever, depending on your level of script-o-philia).   I imagine for some folks, a 3-position switch (shades of the recovery/attack mod on the Ross Compressor!) with high, medium (classic, actually) and low regen may be more than sufficient.  In such a case, perhaps resistances of 15k, 22k, and 68k may be suitable targets. (If you plan on doing this by switching parallel resistors with a 3-way SPDT toggle, your "default" resistor is 68k for the middle switch position, with a 33k parallel resistor on one side, and a 20k on the other side.  This will get you 68k, 15.5k, and 22.2k.)  Similarly, as interesting as a bias-pot might be, being able to switch to a P-90 preset is a plus.

But, to return to the topic of other possible mods.....

If one has an interest in using disparate cap values to "re-distribute" phase shift across the spectrum, consider use of series caps to accomplish that.  So, those caps that you wish to change would be placed in series with another cap.  When in series, their combined capacitance would be less than the value of the lesser cap.  You can change the effective cap value by means of a toggle that shunts one, neither or the other cap.  Shunting of series caps prevents popping when switching.

Finally, note that the FETs are in parallel to 22k resistors.  In effect, what is being varied is their combined parallel effective resistance.  The product of that effective resistance and the cap value is what determines where the amount of phase shift reaches its maximum.  (Each stage can produce up to 90 degrees of phase shift, but frequencies below the zone determined by the RC combination receive LESS than that.)  While the resistors and FETs are chosen to complement each other and produce an ideal sweep, changing the value of the parallel resistors can alter the nature or range of sweep.  

Note that when two resistors are in parallel, the larger the value of the *fixed* R, the greater the impact of the varied resistance (the FETs in this case) on their parallel resistance.  What that means for us is that for those stages where you monkey around with the value of the 22k resistors, making them a bit bigger will produce more "movement" in the sweep, and making them smaller will produce less movement,  My guess is you probably don't want to vary them too much, since there will be a range outside of which the notches will either be unheard or else annoyingly low.  Perhaps a good move might be to stick a dual-gang 10k pot and 18k fixed resistors in place of 22k resistors for two stages.  That would permit two stages to be staggered over a reasonable range.  The change from 22k to 18k might not seem like much but its a 22% change in each direction and that is likely enough to be a meaningful and audible difference in the context of this pedal.

I suppose I should append this by saying that moving your notches around and altering resonance can have impacts on perceived loudness.  In which case, the variable output level mod that Joe suggests is a smart idea.