[DEBUG] MXR Phase 90 LFO doesn't oscillate. Help?

[drea]

1.What does it do, not do, and sound like?
Treble boost/bass cut, slight level increase. No phasing.

2.Name of the circuit
Mxr phase 90

3.Source of the circuit (URL of schematic or project)
http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=2381&g2_serialNumber=2

4.Any modifications to the circuit? Y or N
No

5.Any parts substitutions? If yes, list them.
C3, C7, C8, C9, C10 - 47nF
C4 - 22uF
Q5 - 2N3906
D1 - 1N4148

The 2N5952s are matched to between -2.054V and -2.071V Vgs(off)

6.Positive ground to negative ground conversion? Y or N
No

7. Battery
No battery, using a bench supply.

Voltages.

All taken with the bias voltage trimmed to 4.6V.

IC1
P1 = 5.03
P2 = 5.03
P3 = 4.80
P4 = 0
P5 = 5.62
P6 = Slowly climbs to 4.5 (~3mV/sec). Once at 4.5 slowly falls when probed (~20mV/sec)
P7 = 7.75 stable (I think this should be oscillating?)
P8 = 9.08

IC2
P1 = 5.03
P2 = 5.03
P3 = 5.03
P4 = 0
P5 = 5.03
P6 = 5.03
P7 = 5.03
P8 = 9.08

IC3
P1 = 5.32
P2 = 5.19
P3 = 5.21
P4 = 0
P5 = 5.22
P6 = 5.19
P7 = 5.07
P8 = 9.08

Q1
D = 5.03
S = 5.03
G = 4.19

Q2
D = 5.03
S = 5.03
G = 4.21

Q3
D = 5.21
S = 5.03
G = 4.23

Q4
D = 5.21
S = 5.03
G = 4.23

Q5
E = 5.02
B = 4.44
C = 3.18

D1
A = 0
K = 9.08

Z1
A = 0
K = 5.02

[Luke51411]

Have you double checked for track cuts, bad solder joints, jumpers in the right place? Does the trimmer have any effect? Are transistors oriented correctly, did you socket them? These are the usual culprits for me.

[drea]

Cuts, jumpers, solder joints:
I've re-checked all these and they look fine. I've also continuity checked for shorts across cuts and between tracks and didn't find any.

Transistors:
Yup, all socketed (ICs too) and oriented correctly. The 2N3096 I've used has the same pinout as the 2N4125, unless Google's lying to me. I don't think that would affect the oscillator though?

Trimmer:
Not a lot of sanity here.

With the bias voltage set between 0V and 1.25V, all the non-power pins of IC 2 and 3 are at 5.02V.
IC1 is unchanged except for pin6 which is hard to measure but seems to be doing the "slowly climb to near vBias then slowly fall when probed" thing.

As vBias increases to 2.9V, IC3 changes to:

P1 = 6.55
P2 = 5.88
P3 = 5.80
P4 = 0
P5 = 5.80
P6 = 5.80
P7 = 5.06
P8 = 9.08

IC2 doesn't change. IC1 pin6 goes to 2.9V and falls while I'm probing. The drains of Q3 and Q4 are at 5.80V. The gates of all the fets are at 2.70V.

With vBias at 4V I have:

IC1 pin6 goes to 3.9V and falls.
IC2 still unchanged.
IC3 voltages have gone down again.

IC3
P1 = 5.50
P2 = 5.27
P3 = 5.30
P4 = 0
P5 = 5.30
P6 = 5.30
P7 = 5.11
P8 = 9.08

Q3 and Q4 drains at 5.29V. All fet gates at 3.71V.

With vBias at 5.03V:

IC pin6 goes to 4.6V and falls.
IC2 still unchanged.

IC3
P1 = 5.29
P2 = 5.17
P3 = 5.19
P4 = 0
P5 = 5.19
P6 = 5.16
P7 = 5.06
P8 = 9.08

Q3 and Q4 drains at 5.19V. All fet gates at 4.52V.

And I'm officially out of my depth now

I'm mostly a software guy and although I have done several digital hardware projects (arduino robot etc), this is only my second pedal (first one was a tube screamer and works great). So I really appreciate any help 

[Luke51411]

Q3 and Q4 drains at 5.19V. All fet gates at 4.52V.

I pretty much am too  You tried to set the trimmer by ear correct? In my experience with p90 type circuits the section that produces audible phase shift can be quite small. Do you have a multi turn or single turn trimmer? Do you have any other matched sets of jfets? If I understand correctly, they can be matched in one spec and be off on other specs which can effect the phase shift though that closely matched should at least give you something. I'm sure one of the heavyweights will weigh in soon.

[drea]

You tried to set the trimmer by ear correct?

Yup, the original set of voltages came from setting the bias by ear. I think that the bass cut at the output suggests that I am getting some phase shift, and the resulting notch in the bass frequency response, but because the oscillator isn't working there's no swishiness^tm

Do you have a multi turn or single turn trimmer?

It's a bourns multi turn.

Do you have any other matched sets of jfets?

I do, but they're not as well matched as the set I'm using (~37mV range vs 16mV for the set I'm using)

[Luke51411]

I meant to quote I'm out of my depth now part haha. My post doesn't make much sense as is. It couldn't hurt to try the next best matched set though it doesn't sound like that's the problem. Have you checked the part values in the lfo section?

[armdnrdy]

Since there is no schematic that matches the vero layout linked.....and I don't feel like reverse engineering the layout....
and I'm not sure which op amp and pin would be the oscillator output....

Check lug 2 of the rate pot for oscillation. Also, check the JFET gates again for oscillation.

If there is no oscillation double check the component placement...if it's all good...then hunker down and look for a solder bridge or a trace that isn't cut all of the way.

I know that you stated that you went through your board already....and didn't find anything but....
If I had a dollar for every time that statement was made and then the poster found a teeny solder bridge or a tiny little piece of copper trace left uncut....I would be filthy rich! 

[Luke51411]

I know that you stated that you went through your board already....and didn't find anything but....
If I had a dollar for every time that statement was made and then the poster found a teeny solder bridge or a tiny little piece of copper trace left uncut....I would be filthy rich! 

If I had a dollar for every time I said that to myself... I seem to have trouble with jumpers finding their way to the wrong row.

[drea]

If I had a dollar for every time that statement was made and then the poster found a teeny solder bridge or a tiny little piece of copper trace left uncut

Well here's another dollar then.

There was a springy sliver of copper from the cut between Q3 and Q4. It wasn't shorting when I had the board upside down to check continuity, but putting it right side up on the desk to measure voltages caused it to bridge the cut

But it's not all good news. Fixing the short has got rid of the weird voltages on IC3, Q3 and Q4, but I still dont have any oscillation. I've re-checked the component placement and confirmed continuity on all the jumpers.

Am I right in thinking that it should be oscillating whatever the bias voltage is trimmed to?

BTW, this schematic matches the layout I've built from http://www.tonepad.com/getFile.asp?id=42

And thanks for the nudge back to first principles.

[jimilee]

Have you checked the pin outs? You can use a couple of transistors and the most common ones have a reversed pin out if I'm not mistaken.

[drea]

Have you checked the pin outs?

Yup. The jfets I'm using are the 2N5952s specified in the source circuit. I've replaced the 2N4125 PNP with a 2N3906, which has the same pin out. I don't think any of the transistors are involved in the oscillator though.

[drea]

Breakthrough!!! We have oscillation!

I was reading through the build reports on the Tonepad site and found one of a guy who added a switch in the pot wire to pause the oscillation. Hmm, he's done deliberately what I've done by accident.

So I desoldered the pot, replaced it with a couple of 150k resistors and suddenly pin6 on IC1 is bouncing up and down

The pot's reading open circuit, that's about the last thing I thought would be wrong.

I don't have a spare 500k pot of course, but as soon as I get another one we'll be rocking.

Thanks for the help everyone.

[Luke51411]

Excellent! Glad you got it solved!

[armdnrdy]

that's about the last thing I thought would be wrong.

I concur!

The pot's reading open circuit,

Be careful spending too much time soldering the off board wires to the pot lugs. The pot can be damaged.......as you may have just learned. 

[drea]

So I got a new pot, carefully (and quickly, thanks armdnrdy) soldered it up and it works!

I'm not sure I've quite got the bias voltage dialled in yet, but that creamy, shimmery phasing's there. Just got to get it boxed up now

[armdnrdy]

Glad you got it up and running!

Don't feel bad...I've fried a few pots in my time as well! 

[drea]

So I've been fiddling some more and there's a couple of things I need to sort out.

Thing the first:

The trimmer's a 20(?) turn and I find there's phasing over about 6 or 7 of those. At one end of that range the sound is quite bassy but gets more sparkly as I get towards the other end. Might that be because not all the FETs are turning on/off? Is there an objective way to set the bias voltage or do I just have to fiddle until I'm happy?

Other thing:

All the speed adjustment is in the last 5% of the pot's travel.

I've used a 500k linear but I think I might need a log/audio instead?

Oscillator speed is inversely propotional to resistance? Because it's governed by the current into the cap and higher resistance means lower current?

[armdnrdy]

a 20 turn may be a bit hard to dial in. With a regular single turn, there is a small area of "good phasing"

Set it by ear what sounds good to you.

The rate pot should be a 500KC (reverse log)

[drea]

Set it by ear what sounds good to you

Fair enough, just trying to placate my latent OCD

The rate pot should be a 500KC (reverse log)

Ah, ok. I guess that means I don't fully understand how the oscillator works. More thinking needed.

Thanks.

[duck_arse]

some of us are known to measure the range of the trimpot while phasing, then string-up a panel mount pot to match, cause we can't decide which phase sounds bester^tm.