Original MXR Phase 90 doesn't work

[Elijah-Baley]

Thanks!
I'll solder a socket IC.
Should I expect an IC fryer?
Can I try a new IC?

I have some TL072, but I prefer keep them for other projects of mine. In case I can buy later a new IC, probably a TL062 like the original IC.

[Mark Hammer]

The TL062 is a low-current chip.  It is a good choice when there is a need for low current consumption (e.g., to make batteries last longer), or for an LFO that might otherwise risk creating spikes on an audio line sharing power with it.  But is has worse noise specs than a TL072 (roughly 3x noisier).  So if it is the one that has the LFO, use a TL062.  If it is a chip used for the phase-shift stages, use a TL072.  Not many of us use batteries anymore.

[Elijah-Baley]

Thanks, Mark!
I know TL062 and TL072.
I thought to replace this faulty suspected IC with a TL062 just because that's it. I think my friend don't use batteries, anyway.
As the schematic I posted in the previous page this TL062 has one half used like input buffer stage, and one half used, indeed, as LFO. I checked it while I probed some days ago.
All the phase-shift stage is in the TL064.

According how much you said I should buy a TL062.

Just I don't know what is exactly the risk when you tell: «for an LFO that might otherwise risk creating spikes on an audio line sharing power with it.»

I built on verobaord a Phase 90 (I don't have it, now) using three TL072. It was a verified layout by TagBoard Guitar FX. But I had necessarily to do the Script mod, because the Modern Mode it was too distorted, even with single coil, and the trimmer was really hard to adjust unless get the phasing not so nice and well balanced. I don't know if it make some sense.

[jim-analog]

Greetings!

  Elijah, I wouldn't be concerned about TL0 62 vs TLO 72 for testing purposes; use whatever you have to figure out what's wrong and then replace final version as you see fit.

  Question, what does the V Ref at the Zener read now? Do you have 5V1 at any of the input pins to the ICs? You should.............if not you need to track that down. Do this before installing any new ICs. Let us know what you find..........Good luck!

Regards, Jim

[Elijah-Baley]

[...]
Look at the schematic I posted, while...



...I checked the connections of the upper and below diodes and the 10k resistor, just there.

10k resistor
Left side to the positive of the 15uF
Right side to the 9v

Upper Diode
Left side, positive to the Ground
Right side, negative to the 9v

Below diode
Left side, positive to the Ground
Right side, negative to the positive of the 15uF

And, of course, the negative of the 15uF tantalum cap is on the ground.

[...]

This should be the real configuration of the diodes on the original PCB of the MXR Phase 90 Rev D:



I measured these:

0.11v ------[10k]------ 9.00v
0.00v ---[+1N914-]--- 9.00v
0.00v ----[+zener-]--- 0.11v

0.00v ---[-tantalum cap+]--- 0.11v

I expected to get 5.1v, looking at the schematic, instead to get 0.11v.

The TL062 it's still missing.

[jim-analog]

Greetings,

If the figures you just posted are with no ICs installed, try to replace the Zener Diode and see if that get the Vref back. The circuit won't work without that rail.

Good luck!

Regards, Jim

[Elijah-Baley]

I desoldered the zener diode. And I desoldered even the other diode, the one I told in the first post it had a small fracture. Indeed, it was half broken. I used a 1N4733 (5.1v) and a 1N4148.
The place for the TL062 is still empty.

I measured the voltage, but are not so different.
Where before the voltages are around 0.11v now are 0.20v

0.20v ------[10k]------ 9.00v
0.00v ---[+1N914-]--- 9.00v
0.00v ----[+zener-]--- 0.20v

0.00v ---[-tantalum cap+]--- 0.20v

Is this normal?

[Slowpoke101]

I suggest taking the 15uF tantalum out and testing it. It is possible that it has developed an internal short - which isn't unusual for tantalum capacitors.
If it is faulty it can be replaced by a 10uF electrolytic.
The zener was probably cracked during the manufacturing of the pedal but replace it anyway.

[Elijah-Baley]

Actually, the upper diode was cracked, and it was the one across 9v and the ground. I already replaced, it is practically crumbled during the extraction!

I could desolder the tantalum, too. In case I have even a 15uF electrolytic. But I really hope this intervention will solve, I don't know why the issue appear so tricky. I think I should get 5.1v n the negative side of the diode and on the positive side of the 15uF cap, I shouldn't?

After I desoldered the tantalum cap, how I can test it? Should I just to see if the two lead are connected each other? Now, they are not, on the PCB.

[DIY Bass]

I would expect that if the cap is shorted, and you remove it, then you should get 5.1V at Vref.  The cap is there for filtering.  If the voltage goes to normal with the cap gone then replace the cap.

[Elijah-Baley]

I removed the tantalum cap, and still nothing change. I got the same voltage!
The two leads of the tantalum are not shorted.

Anyway, I notice something really strange, just when I power up the board. Sometime my multimeter prove a connection between ground and 9v, but not everytime. Sometime my multimeter behaves weird, the display turn off when I try to see if 9v and ground are connected or I try to measure the 10k resistor. I can measure it just when I take off the power. I even try to insert in the DC jack a snap battery with jack, but when I plug the battery the board do all this.

[Slowpoke101]

I removed the tantalum cap, and still nothing change. I got the same voltage!
The two leads of the tantalum are not shorted.

Oh well, that is a bit disappointing. At least the tantalum is not causing the problem. The Vref rail may be being pulled down by something else or is not being supplied with voltage. Check that one end of the 10K resistor is connected to 9V then check the voltage on the other end ( I think that you have indicated that you are reading 0.11V or 0.20V ). Then remove power and try to test the resistance of the 10K resistor (in circuit it will read anywhere between 3K to 8K - best to take it out of circuit to test ). If it is 10K and is OK check the resistance across the 5V1 zener - the measured resistance should be greater than 1K. If it is less than 1K then the zener make be faulty or the 47nF capacitor may be faulty.

Your multimeter problem is odd. It may be developing a fault or its battery is going low. Replace the battery anyway and see if it helps.

[Elijah-Baley]

Probably my multimeter is not at 100% ok, but it still work.
The side to the right of the 10k resistor is definitely connected to the 9v. The left side measure 0.20v (it was 0.11v with the original zener diode).
I can measure the 10k resistor just without power up the circuit, and I get 9.89k, considering it's a carbon film probably that is the real value, I guess you expectet other measurement, instead I  I can measure it with none influece from the circuit. But it seems I can't measure it with power up.
The resistenza of the below diode, the zener, is just 232R. It's a new one and the only one I have, now.

The 47nF cap is a 100nF on this board, the +Vbias side of this cap measures 0.07v, and the +Vref side measures that 0.20v.

[Slowpoke101]

232R is actually what I was expecting if the 10K resistor was OK. Ohms Law is quite useful but it does not really provide the answer to the pedal's problem. The 5V1 zener is most likely OK but if you do lift one of its leads from the circuit the zener will probably test OK as a diode - but it is worth doing just to be certain. If it is OK then something is dragging Vref down. If the zener has failed then the question is - what killed it?

Assuming that the zener is OK you will now have to try and find what is dragging the Vref rail dowm. Vref goes off to the trimpot and to a 470K (or 1M depending upon which revision board you have ) resistor which connects to the input buffer OP amp. The Vref rail also goes to all four JFETS and to a single transistor making up the output buffer. It is unlikely that any of those devices could pull the Vref rail down that far but a loose piece of solder or wire could - have a very close look at the board and you may find it.

The trimpot (250K ) should be fine but still check it out just in case. I have only ever had two trimpots that failed in a manner that could cause the problem - one had contamination that tracked across the ceramic substrate between pins 1 and 3. The other failure was caused by the wiper contact snapping and moving back towards pin 1 and 3 shorting them.

I may have misunderstood what you have written regarding resistance tests with you multimeter but do not try to measure resistance whilst the circuit is powered - very odd reading will be observed on the multimeter.

[Elijah-Baley]

I'm assuming that the zener is ok without further tests. For the moment.

https://www.electrosmash.com/images/tech/phase-90/mxr-phase-90-script-logo-schematic-parts.png

It's a bit hard to measure R20 (see the schematic), it's over 560k, and slowly rise up. I'm measuring even the 470k in the input buffer it looks different than R20 in the power supply, it's a bit difficult to measure, too, but it's over 900k.

The trimpot on the board is just one, in the power supply, though far away from the diodes. It should be 250k, but this black trimpot is labeled "25K 723K". Pin 1 and 3 measure 10k, as well 1 and 2, as well 2 and 3.

[willienillie]

The Electrosmash schematic you're referring to is for an older "script" Phase 90, although with the wrong opamps for some reason.  The unit you're repairing is a modern variation, somewhat different, but I can't say how.  Maybe you can find the correct schematic for the modern Phase 90.

[bluebunny]

Anyway, I notice something really strange, just when I power up the board. Sometime my multimeter prove a connection between ground and 9v, but not everytime.

You're trying to check continuity with power applied to the circuit?     Power should be off...

[Elijah-Baley]

The Electrosmash schematic you're referring to is for an older "script" Phase 90, although with the wrong opamps for some reason.  The unit you're repairing is a modern variation, somewhat different, but I can't say how.  Maybe you can find the correct schematic for the modern Phase 90.

I didn't find the modern schematic, but it should be very similar. I found some little difference, but the lost 5.1v voltage is really a mistery.

Anyway, I notice something really strange, just when I power up the board. Sometime my multimeter prove a connection between ground and 9v, but not everytime.

You're trying to check continuity with power applied to the circuit?     Power should be off...

Yes, I did it, sometime! I didn't know it was wrong.

Now I feel a little bit sad about this issue. I thought I could fix it.

[Slowpoke101]

Now I feel a little bit sad about this issue. I thought I could fix it.

Don't feel too sad. I think that you should be able to repair the pedal.

Having now reviewed the previous posts you mentioned that the TL062 was getting warm. This does suggest that the pedal has been "spiked" with either reversed power polarity or a low voltage AC supply. The 1N4148 diode near the DC connector may have been damaged by such an event and not a mechanical injury as I first thought.

Anyway, the next step is to replace the TL062 (if you haven't already done so ) and the TL064. I suspect that the TL064 is unwell due to a power mishap. Replace both ICs and then test the Vref voltage - hopefully it will be correct.

The JFETs should be OK but be prepared for them having been damaged. How they connect to the circuit does not really allow much current to flow through them if there is a power mishap so they "should" be OK.

[Elijah-Baley]

Thanks.
The TL062 place is stille empty. I have some TL072, but honestly I prefer save them, I mean, I don't want to risk to fry them, because I need them for other projects and I don't want let my friends to spend too much. He will reimburse the costs, but I don't want ask him too much money to replace this parts and to back the pedal still not working. Of course we are speaking about few money, but this annoy me.

(I really hope that the 2N5952 are ok. I have a quartet for my Phase 90, but I don't want to buy another kit because it will cost too much (kit + shipping delivery). About the same costs if I buy a bunch of JFET to match them.)