maestro mps2 ..... where to start?

[pinkjimiphoton]

hey fam

so years ago, picked up a maestro mps2 phaser to repair one of these days.

i just found it under my bench, after kinda wondering wtf had happened to it.

so i pulled it out, its been opened up forever... battery was even still good in it <i detached the snap when i stashed it>

its dead as it was then, so i guess a long nap didn't let it heal

so i've replaced the electros. voltages seem to be decent. i have to go thru and figure out which dang ic is which, cuz the schematic doesn't label them and i'm not super good at tracing stuff with my wonky ass eyes.... speed knob has no effect, can't seem to find any sign of oscillations as i'd expected to...
and dealing with the bane of my existence, fets. i @#$%ing HATE fets. they ALWAYS bust my stones.

this beast in particular is known for being a ball breaking sack of nixons, soooooo i'm thrice screwed.





so my big questions....

where the hell do i start? what should i look for? passes signal just fine, so i assume the buffers and stuff work.

if i mess with the trimmers, i can hear marginal differences in the thru-put tone, but, heck, i can't tell if its off or on ffs. i'll have to beep it later to be sure, i guess.

i did also try cleaning the pot.

wondering if it could be the ground of the input jack?  i just had a shin ei wah surf hurricane vol siren pedal arrive doa, and it was the sleeve connection to the input jack that laid it all low.

i'll gladly post voltages, but first i gotta know which ic's are which, etc. oyyyyyy.

makes me widdle bwain hurt.

any guidance for my dumb ass is gratefully accepted. yes, i did search for info. i found a plethora. but still can't seem to figure out how to start.

thanks peeps.

[pinkjimiphoton]

updated the schem slightly and numbered the chips

[Mark Hammer]

I'd say start by verifying that the LFO is O-ing.  So set the rate control for something that looks like it ought to be a medium rate, and measure the AC voltage at pin 6 of IC5.

If that checks out, I'm not sure whether the R1 trimmer results in a greater or lesser amplitude at the junction of those two 390K resistors.  But if you know you have a working LFO, set R1 for the highest amplitude, and then back it off just a bit.  That still won't necessarily make the FETs sweep, so now you want to play with R2 to find the right bias.  Remember that the LFO is going to sum with the bias voltage, which is why I suggested NOT aiming for the highest possible LFO amplitude.

If the unit behaves anything like the Phase 90, then R1 and R2 could actually be panel-mounted controls to set the sweep width and offset or "initial" notch frequency.

We'll leave verifying phasing for later, but for now let's verify that each op-amp - IC1 thru IC4 is passing signal.  Feed something to the circuit input and measure AC voltage at pin 6 for each of those op-amps, in succession.  They're ostensibly unity gain, so you should get something very close to the same AC voltage at pin 6 for all 4 op-amps.  You say that the unit "passes signal", but we don't really know if that's wet, dry, or both.  So temporarily lift one end of that 2k2 resistor by the output to  eliminate the dry signal.  This way, everything you hear will be wet-only, making it easier to make adjustments.

[pinkjimiphoton]

gonna try all that in a minute.
lfo is lfo-ing, so that's good

here's voltages:
8.86 b+

ic1
1 -0.011
2 3.57
3 3.79
4 -.047
5 0
6 6.37
7 6.52
8 0

ic2
1 -0.004
2 3.81
3 3.41
4 -.041
5 0
6 1.61
7 5.92
8 0

ic3
1 -0.005
2 3.80
3 3.72
4 .039
5 0
6 5.32
7 6.52
8 0

ic4
1 .002
2 4.3
3 3.7
4 0
5 0
6 1.8
7 6.4
8 0

ic5
1 0
2 3.7
3 3.7
4 0
5 0
6 modulating
7 6.4
8 0

ic6

1 0
2 3.7
3 modulating
4 0
5 0
6 modulating
7 6.4
8 0

q1-q4

d 3.5 modulating
g 3.8
s 3.8

q5
c 6.53
b 1.34
e 1.32

q6
c 8.73
b 7.20
e 6.53

q7
c 6.52
b 4.39
e 3.80

pot changes modulation speed

switch functioning

z1
7.2 stripe

[pinkjimiphoton]

lifting the 2.2k on the output kills the signal, so tho the osc is working, its not passing audio thru the wet path.
now i gotta find the psychotic break in the circuit lol

thanks mark

more shortly gotta find my audio probe and a signal gen

[Rob Strand]

q5
c 6.53
b 1.34
e 1.32

Assuming your Q5 and the schematic Q5 are the same part, the Q5 voltages look wrong.  That's why the voltage is off at all the opamp outputs.

Maybe easier just to replace Q5 and see what happens to the voltages.

[pinkjimiphoton]

good call, rob. betting its a mistake on my part. seems to me if q5 were bad, it wouldn't be able to pass audio for the dry OR wet signal tho, right?
can hear the emitter follower they used for q5, it seems to drop the output just slightly from unity so i assume that q is alright.

my gut says the most likely culprit to fail would be them tropical fish coupling caps. in my limited experience with them, they often seem to go open internally eventually.

soon as i find my audio probe and a 9v batt to power my sig gen, i'll walk thru the circuit path and try and find where it ends.

i will check the b connection to q5 again. sometimes funk that builds up over the decades can make reading wonky.

thanks rob

stay tuned

[Rob Strand]

good call, rob. betting its a mistake on my part. seems to me if q5 were bad, it wouldn't be able to pass audio for the dry OR wet signal tho, right?
can hear the emitter follower they used for q5, it seems to drop the output just slightly from unity so i assume that q is alright.

Well, the transistor is kind of biasing so it might still pass the dry signal.  However, the bias voltage on the emitter is totally off and that's screwing up the phaser part.    I suspect Q5 is low gain.   That could be due to damage or a natural death.  If it has been replaced in the past it's possible the replaced part has  e and c reversed, that will also make it low gain - maybe worth checking the pin-out and PCB.



Actually the b and e voltages aren't even consistent with low gain.  Maybe b and e are shorted or got shorted when you took the measurement.

[pinkjimiphoton]

 

or.... i didn't get a good connection.
i'll check it out tomorrow, or later tonite if i can't sleep like usual.
its the original transistor, so could be dead i guess. shorted b to e?

i gotta look up the tis97, see if i have something close.

[pinkjimiphoton]

replaced q5 with a bc549. voltage is better there, but still no audio thru the phaser stages.

i am assuming every 741 should be showing the same voltages, pin 5 on ic 2 and 4 is way off from the other two however. gonna replace the caps first, as they seem the most likely place to mess up. goin' in... stay tuned!

[pinkjimiphoton]

got it ALMOST  passing signal... can hear the modulation, faintly and distorted if i strum the guitar hard. speed knob working, gonna reconnect the 2.2k mixing resistor. its making trem sounds, but very distorted.

just messed with it some more, but all i get so far is a very distorted trem... half voltage seems real low too, about 1/2 v shy

man this sucker is annoying 

[Rob Strand]

replaced q5 with a bc549. voltage is better there, but still no audio thru the phaser stages.

i am assuming every 741 should be showing the same voltages, pin 5 on ic 2 and 4 is way off from the other two however. gonna replace the caps first, as they seem the most likely place to mess up. goin' in... stay tuned!

just messed with it some more, but all i get so far is a very distorted trem... half voltage seems real low too, about 1/2 v shy

man this sucker is annoying 

It's a crap design because the DC voltage on the emitter of Q5 needs to be close to the voltage on the JFET sources (ie. the 4V rail).   There's nothing forcing that to be true so there's a high probability of failure!

Can I suggest lifting the end of the 100k that goes to the emitter of Q5 and inserting a 100nF cap in series with it.  The 100k is the one between the emitter of Q5 and pin 2 IC1.  That should fix the DC matching problem.   If it doesn't work the problem is elsewhere but leave the 100nF in until you find the problem since it will remove a lot of freakin headaches along the way.   When you do that the DC voltage on all the opamp pins should be the same as the DC voltage on the rail that goes to all the JFET sources (ie. 4V).

[Fender3D]

Hmm, voltage near power supply on IC1 and 3 output pins, is also suspicious...

[pinkjimiphoton]

i will try your suggestion when i get back to it later, rob.

i agree federico, that voltage looks weird. everything about this sucker is weird

[Eb7+9]

Can I suggest lifting the end of the 100k that goes to the emitter of Q5 and inserting a 100nF cap in series with it.  The 100k is the one between the emitter of Q5 and pin 2 IC1.  That should fix the DC matching problem.   

Rob, don't go launching poor ol' pInK down a nasty rabbit hole now ...
by doing what you suggest the signal path will have no DC reference to GND
if/when the FET's open up completely ... nasty noises will ensue

the most direct (and safest) way to fix this is by fudging the 4v reference circuit
ie., by replacing the 22k/47k biasing the base of Q7 by a 50k or 100k trimpot

set the existing BIAS trimpot half-way first, then set new trimpot to taste

---

incidentally, this is a Maetro sub-unity gain design that sucks
if you convert to direct bypass ...

just so ya know

[Rob Strand]

Rob, don't go launching poor ol' pInK down a nasty rabbit hole now ...
by doing what you suggest the signal path will have no DC reference to GND
if/when the FET's open up completely ... nasty noises will ensue

the most direct (and safest) way to fix this is by fudging the 4v reference circuit
ie., by replacing the 22k/47k biasing the base of Q7 by a 50k or 100k trimpot

What I'm suggesting prevents the DC from Q5 stuffing up the opamp biasing.
It doesn't cause any more problems.   What you are suggesting won't solve that issue
but it does solve yet another issue - I'm all for it, it can only help.

These old pedals often have design issues.   They certainly don't help when
you are trying to debug problems.

If the bias on the opamps can be made right then it's possible to trace the signal through
the opamp chain and see where it stops, then take it from there.

[Rob Strand]

The 2k2 in the dry path looks wrong to me.  My guess is the intended value should be perhaps 22k.  Otherwise when the phasor is kicked in it has virtually no effect.

At the moment you have a bigger problem in the phase shift path, since after following Mark's good suggestion of lifting the 2k2 there was no audio. 

[pinkjimiphoton]

Can I suggest lifting the end of the 100k that goes to the emitter of Q5 and inserting a 100nF cap in series with it.  The 100k is the one between the emitter of Q5 and pin 2 IC1.  That should fix the DC matching problem.   

Rob, don't go launching poor ol' pInK down a nasty rabbit hole now ...
by doing what you suggest the signal path will have no DC reference to GND
if/when the FET's open up completely ... nasty noises will ensue

the most direct (and safest) way to fix this is by fudging the 4v reference circuit
ie., by replacing the 22k/47k biasing the base of Q7 by a 50k or 100k trimpot

set the existing BIAS trimpot half-way first, then set new trimpot to taste

---

incidentally, this is a Maetro sub-unity gain design that sucks
if you convert to direct bypass ...

just so ya know

thanks brother
yeah, i know, tb means no bueno
but i love these shitty old phasers.

right now, its acting like a dirty tremolo. i found/ordered more jfets if i need 'em, but next trick will be likely lifting resistor legs to measure... replacing the caps in the actual phaser stages made an improvement..
but i think the big issue is the freekin voltage is just plain too low in the circuit. it won't pass signal without the bias full up, and the osc most of the way up too.

i will try your suggestions when i get back down the dungeon!

rob... no worries... but so we're on the same page, we're now passing signal and have oscillation... but it sounds like a distorted trem.

so 741's will be coming out, sockets going in, and may end up converting to ne5532 cuz i dunno if i have enough 741's unless i use old metal can ones i usually save for more mojo builds.

more later... thanks brother

stay tuned

[Rob Strand]

As for the schematic:

There's a pic of the PCB here,
https://tagboardeffects.blogspot.com/2014/02/maestro-mps-2-mini-phase.html

I think the obscured 2k2 at the top right is for the power supply.

Just under that near the green wire and above the switch is a 43k and a 22k. 
I think that 22k is incorrectly shown as 2k2 on the schematic.
Tracing the PCB starting from the output jack wire should confirm it.

So for the Mark's 2k2 lift test you want to lift that 22k.

[Rob Strand]

rob... no worries... but so we're on the same page, we're now passing signal and have oscillation... but it sounds like a distorted trem.

OK cool.

Not sure if the distortion is from the JFETs or clipping or a fault.   

There's no linearization (RC networks on the gate) on JFETs so the distortion depends on the JFETs.   

However, it could be just outright clipping.   The 4V bias point isn't bad but it's non-optimal.  Then the last opamp  stage (IC4) has a gain of 2 set by the 2x22ks.    If you decreased the 43k after the phaser switch, say to 22k, then that will let you dial back the gain of the last opamp.   To dial back the gain of the last stage decrease the value of the 22k on pin 6 of IC4.   With the 43k replaced with 22k, you might even be able to just short out the 22k on pin 6 of IC4.

Before you bother with such a mod you would want to make sure the distortion isn't caused by a faulty opamp (or the JFETs).  I guess you will find that out pretty soon  .