A simple Phase 90 mod

[Mark Hammer]

Modded a friend's Script P90 reissue.  We were hoping to cancel the feedback but I simply could not identify the relevant resistor unless I spend a few hours deciphering the board.  However, I was able to easily identify the 1meg (105) resistor coming off the trimmer that sets the bias.

The bias trimmer on the P90 and 45 nudges the JFETs into the range where they will sweep when current from the LFO is fed to them.  Normally, the trimmer is set to provide a fixed drain-source resistance for the JFETS and a resulting "offset" that positions the sweep at a given spot in the spectrum.  The LFO makes it sweep from that starting point.

On my own P90 clones, I replaced the 1M resistor that limits the current from the bias circuit with 470k/510k, and added a 500k pot in series, such that the current-limiting resistance can be varied from around 500k to around 1M.  The effect is to move the range of the sweep up and down, without affecting how wide the sweep is.  Bear in mind that the current from the LFO is summed with that coming from the bias subcircuit.

There wasn't much room in the P90 enclosure for a pot, but there WAS room for a SPDT on-off-on toggle, to be installed horizontally in the manner shown.  I bent one of the JFETs over to make room, and carefully measured where to pop the hole (which needs to be close enough to the top surface of the enclosure to fit over the trimpot, and nudged over enough to not have the corner screw post block full insertion of the switch).  The switch and added components fit in perfectly, above the existing components on the board.

I soldered a 1/8W 1M resistor to one outside switch lug, and a 510k to the other, yielding 500k and 338k when combined in parallel with the existing 1meg surface-mount resistor.  The free ends of these resistors were soldered together and a thin-ish wire went from their joined ends and from the center/common log of the toggle to the two ends of the 1M onboard resistor.

Set the bias trimmer for the highest usable sweep.  It will sound "swirly" and more suitable for acoustic guitar and single coils.  When the bias trimmer is adjusted right, the three toggle settings give a high (middle position), medium, and low range of sweep.  Three audibly different and all usable personalities.  The "low" range is the familiar gurgly sound that a fresh-out-of-the-box P90 provides.

If you find your single knob P90 a little limiting, this will add some life to it.  And if you can identify the surface-mount resistor that handles the feedback, post here and perhaps we can add another mod.

I realize there is LOTS possible to expand the vocabulary of the basic P90 circuit.  On my own builds, I always include variable resonance/feedback, sweep width, and offset/initial, in addition to a vibrato option.  The trick seems to be figuring out how to implement it on more recent surface-mount reissues, and do so within the confines of a 1590B-sized box, where space limits the possibilities.

[fowl]

I'm editing my posts in this thread as they were a derail and just muddied the water.

But it seems that the resistor in question was normally 470K in the Phase 45, at least for several years, though most schematics I see online show a 1M there.

[Rob Strand]

Funny you bring this up.  I'm currently building a Phase 45, and I'm going to copy an early (?) version I've seen pics and a video of.  It has a TL044 but I'm going to use a 14-pin socket, to take LM324 or TL074 etc.  This particular version differs from the more common P45 in that it doesn't have a filter cap on +9v (mine will), and three resistors are different.  The 7.5K is 10K, the 20K is 22K, and the 1M that connects to the trimmer wiper is 470k.

FYI, there's quite a few versions of the Phase 45 and Phase 90.   I don't mean the more modern re-issues either.  The early days had variations.  I've got most of it documented but it's archived away.   The thing to notice is the marking on the bottom of the PCB.  Your PCB says 45a  which is an early one.   You will see 45b etc for the changes.    IIRC the phase 90 might have gotten up to a rev 'd'.

[fowl]

edit

[Rob Strand]

Not mine, just stuff I've found online.  I've seen the above setup with the 16-pin TL044, the more common resistors and added filter cap with the TL044, and then the two TL072s.  So three versions there.  The resistors go from carbon comp to carbon film.  I don't know where the logo change fits in.

Ah, OK.   Unfortunately my memory isn't so good on the phasers and MXR stuff anymore since it involves small detailed changes here and there.   I think the logo changed within 1977.   Before that there were a few changes with the pot part numbers on some models - perhaps early 1976 (Can't remember exactly but I may have posted something on the forum as some point).  After that was LEDs and DC jacks.   The Revs on the PCB are perhaps the best indicator.  Sometimes the pot dates and IC dates sometimes - watchout for part swaps (deceptive and/or repairs).  I think some of the early ones have very specific dates on the pots, like they only bought a few large batches.

[Elijah-Baley]

Hi Mark.
I'm not sure I have identified that 1M resistor you talking about for the mod.
Is it R20?

[fowl]

edit

[mac]

IIRC I changed the 3M9 from the LFO to bias + gates to 3M3 in my 15 years old clone. Can't remember why... 

mac

[Mark Hammer]

IIRC I changed the 3M9 from the LFO to bias + gates to 3M3 in my 15 years old clone. Can't remember why... 

mac

Lowering the value of that resistor yields a wider sweep.  I replace the 3M9 with a 4M7, and use parallel resistors and a 3-way toggle to drop the resistance value down to 3M9 and lower.  If I'm not mistaken, one of the issues uses 3M3 and a slightly lower value for the feedback resistance, optimizing it for slower more dramatic sweeps.

Hi Mark.
I'm not sure I have identified that 1M resistor you talking about for the mod.
Is it R20?

Yes.  It is R20 on that drawing.  I didn't make note of the labelling on the commercial reissue pedal, but it is the only 1meg in the vicinity of the trimpot.

I credit RG Keen for drawing my attention to the use of varying that resistance.  The specific values came from experimentation.  In making pedals for myself, with thru-hole components, I can easily use a 510k fixed resistor and 500k pots in series.  The commercial pedal uses surface-mount parts, which impede such changes, so my choice was to reduce the value of that "R20" via parallel resistances, soldered to the contacts of the 1meg resistor (labelled 105 on the resistor itself).

Normally, in the commercial pedal, the trimmer is factory-adjusted to produce the most "gurgly" (low) sweep.  The zener diode drops the battery voltage, and the trimmer adjusts it down from there, with the 1M resistor limiting the current going to the JFET gates.  There is actually a range of viable trimmer settings that can all produce sweep, a little higher and lower in the spectrum.  MXR aims for a given voltage at the trimpot wiper to yield consistent tone across pedals, but if you have an MXR unit, you can play around with the trimmer to get different offsets an sonic "personalities".

The lower the drain-source resistance, the higher up in the spectrum the maximum phase-shift-per-stage is.  A stock P90 will use a 47nf cap and a 22k resistor (in parallel with the JFET) in each stage.  Approach the 47nf/22k like a one-pole highpass filter.  When the FET+22k is at 22k, maximum phase-shift begins around 154hz in that stage (content below that point has less phase-shift).  If the combined parallel resistance drops to 10k, that point of maximum phase-shift now moves up to just under 340hz.  So, the more we can "drive" the drain-source resistance, via current applied to the gate, the higher up in the spectrum notches start to be produced.