MXR Phase 90 Modifications
By Douglas Kovach
March 28th 2015
How It all Began.
First let me state for the record that though I have often heard of and read about the MXR Phase 90, the brand name was not on my list of coveted vintage pedals to own. As for phase shifters I am a died in the wool Electro Harmonix Small Stone fan with its watery swishing swoosh sounds. I have owned a couple other makes and models of phase shifters over the years that simply did not impress me. The MXR brand was further tainted in my mind by the purchase of an MXR Dynacomp compressor pedal a few years back. After my Electro Harmonix Soul Preacher I owned as a much younger being and the Dan Armstrong Orange Squeezer I had cloned a few times, I simply was not impressed with this compressor. It’s optical based though so maybe thats why. But the Phase 90 kept beckoning me and with a user list that starts off with Jimmy Page of Led Zeppelin and Eddie Van Halen of Van Halen it wasn’t long before I actually found one at a decent price that I was willing to pay. So, a couple years back I purchased an MXR Phase 90 in a pawnshop for the very reasonable price of $60.00. I brought it home and hooked it up and toyed with it for a while before deciding that it sounded okay but had a rather distorted overdriven sound that I didn’t much care for and could not get rid of by tweaking its only knob labeled “Rate”. This was far from what I heard when I listened to Van Halen and Led Zeppelin! I used it on and off for a while before just relegating it to my box of pedals I might use someday. I could hear its potential by the way of the sweep of the LFO and I knew it could sound like I had envisioned while listening to these bands. So one day as I was perusing the interwebs looking for some schematics of this pedal, I found there were several versions. I knew about the differences between the original Script Logo “Holy Grail” version and the later Block Logo version but was pretty unaware of its history otherwise and of all of the various versions and flavors of this pedal. So I took the time to educate myself and opened up my pedal and disassembled it enough to try to identify a circuit version. In tracking down the age and version of my pedal I found it to be a REV E design with a date of 4-27-2001 both of which were printed on the PCB. After some research I found that this is a re-issue pedal. I also found that my pedal had some oddities which were not listed in some schematics and yet other schematics had some of the oddities but not all of them and I found other schematics that had the other oddities but not the first ones. Therefore I think my pedal is a transitional model block logo re-issue version made in or around 2003. For instance mine used a 2N4126 PNP as the mixer transistor and its Miller capacitor was a 680 PF while most schematics called for a 2N4125 transistor and no Miller cap or a much larger value. so I started looking around with this info in mind.
I quickly found that there were lots of versions of this pedal as it changed over time and especially after the Dunlop company took over the MXR brand. I guess it has been in pretty much constant production for almost 40 years! In addition to all of these bewildering versions there are several clone kit schematics available as well as a few modified and suggested modification schematics. Talk about convolution, invigilation, and obfuscation! I don’t give up easily however and with dogged determination I searched and settled on a version that appeared to be stock and had the same sort of component set as mine did with a few small value differences but ironically was dated 1993.
The Various Possible Mods.
I then began gathering up information about various modifications and then reading up on the perceived results of each of these particular mods. Of course I found a variety of opinions about each modification along with of course, the purists who state “A few modifications will not make your Block Logo pedal into a Script Logo pedal” and that the whole circuit would need to be built from scratch in order to obtain original Script Logo Phase 90 realism. While true as this statement may be, a few mods can certainly clean up some of the harsh overdriven sound of the pedal which was my entire goal as well as to add a couple simple mods to extract added features from this pedal.
Eureka, I’ve Found It!
I found a small collection of modifications printed on what appeared to be a mostly REV E schematic along with a REV E board with all the parts numbered and overlayed on the board which made it extremely simple. There were a few mods I wanted to do however that were not entirely clear that were not included in this mod list and so it took a bit of experimentation to puzzle out. The two images I used as schematic and board overlay which included almost all of the mods to give the pedal clarity and depth were found with the name “Nero F. Rox” posted in red on both so thanks out to Nero F. Rox whomever you might be. The other schematic I used came from RG Keens site GeoFex and was titled “Phase 90, 180 Plus” and can be found here: http://www.geofex.com/FX_images/p180plus.gif so thanks out to RG Keen for his schematic idea which was the mix knob to allow access to the vibrato. This schematic is the one I had to do some experimentation with to find the correct resistors. The part the caused all the confusion is that there are 6 or 7 150K resistors that all come together to form junctions with a few other parts and only 2 of these need to be modified. Now, after figuring it out, I can assuredly tell you that on “Nero F. Rox” schematic and board they are labeled as resistors number R8 and R29. The description of that mod is posted below.
Hmm, I Must Ponder These A While.
An additional couple of mods I want to do eventually is to make the LED flash in synch with the LFO and add in a small circuit that is switchable to give the LFO more of a sine wave sweep as opposed to its current triangular sweep. This would benefit the Vibrato to sound much more pronounced in my opinion and should be simple enough using an oppositely polarized pair of diode clippers and some voltage divider resistors to round off the peaks of the triangle waveform. It would have a swoopier sound which would excite the vibrato a bit more than the smooth up and down-ness of the current triangle waveform of the LFO. I also want to examine the “LERA” LFO modification at some later time. This mod is supposed to emulate the speeding up and slowing down action of a Leslie rotating speaker when switched from fast to slow and vice versa, specifically not the Doppler effect that Leslie systems are noted for, but simply the speeding up and slowing down effect of the speakers rotations. I also want to ponder a modification that will let me switch between a Phase 45 and a Phase 90 also developed by RG Keen by shunting a couple of the phase shifting stages.
Potential Upgrades Maybe.
The opamps used on my board are TL062 and TL064 which are okay for the purpose but certainly are not high fidelity chips. The TL082 and TL084 would probably be better chips to use or perhaps the NTE equivalents. Considering the compactness of this board I will not attempt to replace these chips but if ever I build one from scratch I will certainly be using hi fidelity parts. In considering the original version of these pedals, they were built using 741 chips which are also very much not high fidelity chips but indeed they do have that vintage and desirable sound. They aso used cheap ceramic disk capacitors and non-descript resistors. Vintage tone using cheap parts. This is starting to sound like a recipe for a vintage Fender tweed amplifier. Cheap parts equal great tone! Audiophiles and tone-hounds hate me now, I am quite certain. Either way its luck of the draw or a happy accident.
I am considering purchasing an additional Phase 90 kit board to have a stereo version of the pedal. I also have an Electro Harmonix Russian built Small Stone phase shifter pedal that I have been experimenting with. Thoughts of a stereo or multiple Phase Shifter synthesizer with a twin unsynched multiple selectable waveform LFOs including perhaps a sample and hold feature with a slew or lag option and possibly an envelope generator or envelope follower seems to be quite doable and a pretty interesting concept. If I pursue this I will document and release the circuit as well as samples and perhaps a video. I have always liked the swirling odd and random sounds of 2 unsynchronized phase shifter pedals running at the same time. This collection of ideas will need time to simmer and stew in my brain before they come to fruition.
Is It Really Worth It?
I have decided that the common “Univibe mods” presented for both the Phase 90 and the Small Stone are not worth the effort in my opinion. I get better satisfaction and more variety from the added depth and mix controls personally. My Phase 90 after the modification process has a clearer, throatier, vowel-ish sound to it that I am really enjoying a lot. Wobbling the wobble per the Univibe cap selection is not going to make me sound like Jimi Hendrix anymore than flipping my guitar over and restringing it will.
Before we begin I recommend that you have a very strong background in electronics, reading schematics and layouts, pedal disassembly and reassembly, and know how to wield a soldering iron well enough to know that you will need a fine pointed tip to do these mods. If you don’t know how to take something apart and put it back together again without having leftover parts then just stop right now and take the pedal, the schematics, and this text to a knowledgeable technician and pay them to do this for you.
The Mods, Lets Begin.
Anyway lets start with the ubiquitous R28 removal. I just straight up removed it and labeled the part number and value and put it in a bag. Folks have mentioned putting this resistor on a switch or a knob and having tried adding and removing it while playing I have decided that I didn’t need nor want it in my pedal at all. The only logical reason I can see for keeping it would be in having more variety and a larger palette of sounds. If I did keep it I would put it on a knob to vary its strength as well as to completely remove it from the circuit.
Then I also removed R32, C11 and C12 and tagged and bagged all of these as well. R32 and C11 formed a filter to ground and perhaps also to set a bias point for the buffer it was attached to. C12 is a Miller cap between the base and the collector of the 2N4126 PNP transistor. Its purpose is to act as a low pass filter and to eliminate any oscillations that may occur. After removing it I hear no oscillations at all and a larger bandwidth passing through to the output.
As to R31 I simply just jumpered it with stripped wire and left it on the board thus bypassing it. This resistor sets the gain of the buffer and simply jumpering it we have effectively set the gain of the stage to unity or 1. This will help to eliminate some of the distortion overdrive sound the pedal had originally.
I then replaced C2 and C4 with 0.047uf or code 473 caps rated at 50 volts. The original 0.1uf caps were tagged and bagged. This prevents a small amount of lower frequencies from entering the pedals stages allowing a clearer voicing of the pedal. You can think of this mod as raising the lower part of a “tone window” a bit. I searched through my parts stash and all I could come up with were a pair of matching cheap ceramic disk caps. They work and sound great so no need for a more expensive option here.
R20 was originally a 750K resistor which I removed, tagged and bagged and replaced it with a 12mm 1M linear pot (its all I had but it fits surprisingly) with wiper 1 connecting to pin 2 of R27 and pins 2 and 3 connected together and then soldered to the gates of all the JFETS via R20s other hole. This is now my depth knob. I originally tried a 9mm 1M Audio taper pot but found that all of the sweepable range was bunched up on one end of the wiper (doggone logarithms anyhow!). I then switched to the 12mm 1M Linear taper which solved this problem.
R30 and R34 were both jumpered with bare wire and left on the board. Since we are removing R27 and replacing it with a larger trimmer we will not need these.
R27 was removed and a 470K trimmer was put in its place since I didn’t have a 250K trimmer. It doesn’t seem to effect it in a bad way so I adjusted it and left it and it works fine. This allows a broader range of biasing for sweep range and the waveform offset and duty cycle of the LFO stage.
I jumpered R16 on the LFO leaving it in place. This helped to increase the range of the Rate knob. This leaves a tiny deadspot at the very top of the fast rotational range. The original value is 4.7K so perhaps a 470 ohm or 100 ohm resistor or a 1K trimmer might fix this problem. For now I am ignoring it until I get a trimpot to replace the resistor with.
After much experimentation and circuit tracing I then removed R8 and R29, tagged and bagged them and since I didn’t have any 47K resistors laying around (as suggested by RG Keens Phase 90, 180 Plus schematic) I replaced them with 68K resistors of 1/8 watt and left the ends that would have connected to the base of the 2N4126 sticking up in the air. I soldered wires on each of these and then trimmed and heatshrunk them all the way down to the board to prevent shorts. I connected each of these wires to the outer pins of a 9mm 250K linear pot (I could not find a linear 200K 9mm pot as RGs schematic suggested) and the center pin was connected to the base of the 2N4126. I had originally tried using 27K resistors (Ohms law failure!) but there was a lot of nasty sounding distortion at either end of the pot rotation. I will eventually replace the 68K resistors with 47Ks since I feel this would open up the range of variance between the pots stops. 56K resistors would probably work just as well if not better. I am sort of leery after hearing that nasty distortion of the 27K’s if 47K will be a high enough value.
I then measured and drilled holes on the pedal case on either side of the speed pot. I made sure the wires were long enough to facilitate putting the new pots in and tightening them down to ensure no twisting of the pots or wires to cause short circuits while the board was outside the case. I really dislike drilling aluminum without a drill press and a way to prevent the bit from taking a different route than I planned. One of the holes came out a little uneven even though I used a punch to set the point of the hole centered correctly. It still works though. It just looks a little lopsided. As long as the pot mounted in that hole is not shorting something out then all is well. Purely aesthetic accident.
Then I replaced the board and put all the washers back and tightened up all the nuts.
I then covered the back of the board with a piece of cellophane packing tape to prevent any of the jumper wires from shorting against the back of the case. It is highly unlikely that any shorts would occur due to spacing and mostly concave shape of the backplate, I just did not want to take any chances. I will probably cut a piece of gasket paper to fit the entire back of the board to be certain.
I screwed the case shut. 4 screws… Easy Peasey!
I labeled the new “Depth” knob and the new “Mix” knob with my handy Dymo labeler. I had to trim the labels down a bit as there was a lot of blank white paper around the wording which would have been unappealing to the eye on this orange pedal case.
I hooked it up and now I also have a Depth knob and a Mix knob as well as the original Rate knob. The overall tone is much clearer and does not have that nasty midrange overdrive sound that it previously had. I have tried it both in front of my amp as well as in the effects loop and it sounds great in both places. The Depth knob goes from clean to full warble and the mix knob goes from full phase to full vibrato. The Rate knob now goes much slower and much faster but I noticed at the extreme end in a very small part of the fastest setting the LFO simply stops. Whether it is just so fast that it can’t be heard is unlikely. It is more likely that it has crossed some bias or trigger voltage point which halts the sweep of the LFO. A trimpot or resistor of a low value in place of the currently bypassed R16 would probably correct this. R16s original value was 4.7K so perhaps a 470 ohm or 100 ohm resistor or a 1k trimpot would be enough to fix this artifact. On the other hand it now goes so slow that instead of cycles per minute it does minutes per cycle! This is definitely a very subtle and usable tone. I can play an entire song before it goes through a complete cycle of phase! I had to actually record it and time it to discern if it was actually phasing and indeed it is.
I tuned the 470K trimpot by ear since I don’t own a scope. Lacking any instructions on tuning it and ignoring dire warnings of “NEVER MESS WITH THIS TRIMPOT OR YOUR PEDAL WILL DIE!” I did some experiments. Since I completely replaced it, it needed to be adjusted anyway. What I found to be the most accurate way to tune it was turn the LFO speed “Rate” knob up as far as I could (almost into ring modulator territory) before it faded out and then adjusted the trimmer until it gave the strongest warbling effect and did my best turning it back and forth several times to center it on this warbling strength much like one would tune a guitar. Then I reassembled the pedal and it all seems to work great. The LFO sounds even on both the upswing and the downswing with no lopsidedness of the duty cycle noted at either fast or slow speeds. It sounds good with no hum or nasty fuzzy distortions present and I would definitely use it on stage or in a studio now.
In summary, even though the purists are correct, that to obtain the true Phase 90 Script Logo sound you need to just build it from scratch or climb the Himalayas searching for an original one and spend a lot of money to just purchase it, the Phase 90 Block Logo versions can be made to sound very pleasant and have hidden features as do all phase shifter effects, that are just waiting for some imagination to break them free. After a few days with my “new” pedal I am totally digging the sound. It is very throaty sounding and I can dial in lots of variations of the phasing sound. My fellow musician friends are also loving the new found sound of this pedal. It definitely has earned its place on my pedal board! In my opinion these re-issue Block Logo Phase 90 versions are cheaply acquired and are pretty easy to modify into something that can sound beautiful and has more features than the “Holy Grail” expensive one trick pony versions.
Just Do It.
If you own one of these later Block Logo pedals prior to the Surface Mounted parts versions (and NOT INCLUDING THE EVH VERSION!!!) and know your way around a schematic and the innards of a pedal, I strongly urge you to consider performing these operations. Don’t listen to the naysayers or harbingers of bad opinion. Just do it. You will be very pleased with the improvement in sound and the added features you can obtain. I turned my $60.00 pedal from ho-hum into hell-yeah in just a couple of hours and with $6.00 worth of added parts. This is definitely a win! These pedals as new are around $99.00 USD so a good score at a cheap price makes it even sweeter when you realize its true potential!
Do not attempt these mods on an original Script Version of this pedal. They are quite unnecessary and would simply devalue the pedal as well as ruin its tone. Additionally I would also avoid modifying the ’74 Hand Wired Custom Shop Vintage version as well as some of the untouched earlier versions of the Block Logo in order to keep its value and to prevent ruining its tone. I would avoid modifying the Surface Mounted PCBs as well as the Eddie Van Halen Signature versions simply because of the difficulty level in dealing with these ultra tiny SMT parts which are GLUED and then SOLDERED to the boards during assembly. You would need a PCB holder, a magnifying glass, tweezers and forceps, an X-acto knife, and super ninja leet de-soldering skillz in addition to all of the other regular tools required to perform these mods. You can’t forget the semiconductor rule of 3 seconds at 300 degrees Fahrenheit while soldering can ruin a part. Even if you manage to unsolder an SMT part without ruining it, it is still glued to the board. Bad voodoo and no mojo dudes and dudettes!
Just A Word…
My pedal is not true bypassed. Most likely your Block Logo Phase 90 pedal is not either unless it was previously owned by a solder iron wielding ninja with super leet skillz. The footswitch is soldered to the PCB and would require a lot of work as well as cutting an accurate square hole in your PCB and attaching large gangly wires to very tiny circuit traces about a tenth of a millimeter in width and then painted with that clear red or green goop to seal the board. I would not attempt this even with my years of experience unless I was going to put it into a completely new box with the added room to do it. This pedal is very tightly packed and I was not even sure if I could add a set of 12mm and 9mm potentiometers. As others before me have stated, “Do not ask me how to add true bypass to this pedal” If you are not bright enough and experienced enough to do this and figure it out on your own then find someone who is and pay them to do it for you. Truly you might be better off financially if you simply purchase a kit board from General Guitar Gadgets or Tonepad and build your own as you see fit. You will certainly enjoy the experience and perhaps learn a few things. The pride and reward of building your own pedal and with the features you want will put a lump in your throat and a smile on your face when you finally get it working and plug it in and play with it. You will be bragging to your friends for years about it.
If you want to delve deeply into the theory of why and how the Phase 90 works along with all the math calculating power supply and signal filtering frequencies and a good math based overall primer on how phase shifting actually works in general then I highly recommend two pages.
The first is here and gives a thorough physics and mathematical analysis of the Phase 90:
And secondly head over to RG Keens GeoFex “Technology Of” site located here where he gives a good overall description of phasing and other modulation schemes in general theory:
If history is more your thing then this page describes how phase shifting and flanging are created and were discovered and invented:
In my vast research into this project I wish to thank Nero F. Rox formerly of Jeanhatesjohnmusic.com where, who, and whenever you are, RG Keens http://www.geofex.com/article_folders/phasers/phase.html site at geofex.com, General Guitar Gadgets at http://www.generalguitargadgets.com/ , Tonepad at http://www.tonepad.com/ , http://www.electrosmash.com/ , http://stinkfoot.se/ , http://www.performing-musician.com/…/oct07/ar…/retrospec.htm , and all the guys and gals over at http://www.diystompboxes.com/smfforum/index.php… for the theories, debates, and potential modification ideas.
Here is the demonstration video of the pedal after the mods: