MXR Phase 90 Models

[soupbone]

I have a MXR '74 Script Logo Phase 90,and a Script Logo reissue,with power jack and led.I noticed between the 2,The '74's Effect was so more pronounced than the other Script Logo.(I believe it's the same circuit as the evh phase 90,minus the script switch.) I want to have another Phase 90 to have as a backup for my '74 reissue.Would I be better off getting a Block Logo Phase 90,and doing the R28 Mod?

[Mark Hammer]

"so more pronounced" = what?

The schematics I've seen posted show two major differences: the amount of feedback, and the width of sweep, both of which are fixed.  A wider sweep will seem more pronounced, as will greater resonance.

So, when the resistor between the LFO and the JFETs is 3M3, rather than 3M9, you'll get a wider sweep.  When the feedback resistor is 22k instead of 24k, you'll get a more resonant phasing.

Neither of us has any idea whether the 2N5952s produced way back when have the same drain-source resistance range as those produced more recently (or by another company).  "More pronounced" can also arise from where in the spectrum the notches move around.

[soupbone]

"so more pronounced" = what?

The schematics I've seen posted show two major differences: the amount of feedback, and the width of sweep, both of which are fixed.  A wider sweep will seem more pronounced, as will greater resonance.

So, when the resistor between the LFO and the JFETs is 3M3, rather than 3M9, you'll get a wider sweep.  When the feedback resistor is 22k instead of 24k, you'll get a more resonant phasing.

Neither of us has any idea whether the 2N5952s produced way back when have the same drain-source resistance range as those produced more recently (or by another company).  "More pronounced" can also arise from where in the spectrum the notches move around.

pronounced,meaning the effect is much stronger.

[armdnrdy]

There is also the issue of MXR's JFET matching. Are all Phase 90s from a certain era created equal? I doubt it.

It is highly unlikely that the JFETs were matched to the same exact parameters. It wouldn't be cost effective to do so.

This is why some effects have that "magic" tone that guitar players covet.

It would probably be in your best interest to build a P-90 and experiment with the components Mark mentioned as well as socketing and trying different sets of JFETs. 

[Mark Hammer]

I'm curious.  Do service manuals or guides indicate setting the bias trimpot for some specific value/target, or are they merely adjusted for a desirable sound?

I ask because, even if they are lacquered into place, the range of viable trimpot settings in a standard P90 build is wider than you'd think; enough that you could install it as a panel-mount control and use a big chunk of its rotation to achieve different tonal colours, some throatier and more gargley than others.

[armdnrdy]

I ask because, even if they are lacquered into place, the range of viable trimpot settings in a standard P90 build is wider than you'd think; enough that you could install it as a panel-mount control and use a big chunk of its rotation to achieve different tonal colours, some throatier and more gargley than others.

Maybe incorporate a pot with fixed resistors to keep the bias voltage centered in the "usable" area.

[Fender3D]

I'm curious.  Do service manuals or guides indicate setting the bias trimpot for some specific value/target, or are they merely adjusted for a desirable sound?
....

Internal service manual surely does, it would be too much worker-depending otherwise.
If you stop LFO (ie by disconnecting the pot), you got a static point where you can measure, phase or level reduction at a given frequency. It should be enough to write down a procedure.

[armdnrdy]

If you stop LFO (ie by disconnecting the pot), you got a static point where you can measure, phase or level reduction at a given frequency. It should be enough to write down a procedure.

But...depending on the Vgs off and Rds between different sets...this wouldn't give more than a ballpark figure.

[Fender3D]

^^
oh well, it's the same as looking for a desirable sound...

You might get different modulation width, but if you measure phase or level reduction at the allpass filter corner frequency (given by the FET resistance), you shouldn't be too far away

And don't forget that MXR used to buy truckloads of FETs...

[Mark Hammer]

If you stop LFO (ie by disconnecting the pot), you got a static point where you can measure, phase or level reduction at a given frequency. It should be enough to write down a procedure.

But...depending on the Vgs off and Rds between different sets...this wouldn't give more than a ballpark figure.

That's a fair point, and one of the reasons why I asked what I did.  Simply aiming for a given bias voltage would not assure the same sound when the Vgs varies between sets.

So is there some sort of whacked-out lissajous figure, or whatnot, on a scope that would identify the "stock" sound, irrespective of JFET set, and the assembler at the factory just tweaks until they hit that...or should we expect unit to unit variation even within a given issue?  I suppose it is always possible that the manufacturer could have some benchmark pedal, an amp, and an A/B switch, and the assembler merely adjusts until B sounds just like A to their ears.

My whole thrust here is that what Mr. Bone is hearing may not be anything inherent to the changes in design, but rather stems from differences in fine-tuning.  Like I say, move the trimpot a little this way or that, add or subtract 5% to/from the 1M resistor that sets the bias current, and you can end up in different sonic territory, even though you have the same ostensive "design".

[Fender3D]

Simply aiming for a given bias voltage would not assure the same sound when the Vgs varies between sets.

So is there some sort of whacked-out lissajous figure, or whatnot, on a scope that would identify the "stock" sound, irrespective of JFET set, and the assembler at the factory just tweaks until they hit that...or should we expect unit to unit variation even within a given issue? I suppose it is always possible that the manufacturer could have some benchmark pedal, an amp, and an A/B switch, and the assembler merely adjusts until B sounds just like A to their ears.

I wouldn't work in such a noisy workplace... 

Obviously you can't just measure bias voltage, that's why I suggested to check phase or level reduction from a static point at a given input signal frequency.

You might expect variations in width from FETs, variations in bias voltage from res/caps tolerances...

But by applying a (static) voltage to FETs, you can always set the trimmer to reach the desired allpass corner frequency

[Mark Hammer]

But by applying a (static) voltage to FETs, you can always set the trimmer to reach the desired allpass corner frequency

Which would be measurable how?  Via a scope or via a DMM?

[Fender3D]

Well I'm lucky enough I have my old Neutrik A1 audio analyzer... 

If your ideal phaser sweeps from, let's say, 300Hz to 2KHz, and you have an LFO sweeping from 1V to 7V, you can feed 1V to FETs, feed 300Hz at pedal input, then set the trimmer for a 180° (or 360° or more depending on filters number) phase change with a scope in xy setup looking at lissajous figure; otherwise you may check for the lowest level reduction again with scope or whatever can measure dB or AC voltage.
Of course scope will be more reliable

[zombiwoof]

The  '74 Custom Shop RI pedal is the exact same circuit as the vintage models, whereas the "Script Logo" RI is just the stock modern Phase 90 circuit with the R28 mod, the circuit is not the same as the original MXR pedal.  IMO the '74 RI is a better Phase 90, however it doesn't have the LED and true bypass modern features.  I suggest getting the Whirlwind Rochester Series Orange Box, which is also the original circuit, with the added features of LED and true bypass switching.  It is handwired like the '74, and designed by a guy from the original MXR company.  IMO it is the best handwired script Phase 90 pedal around, with the original sound and components, and the modern features.  I love mine.

Al