Phase 45 clone troubleshooting

[Kipper4]

Well done mate. The 2n5458 seems to have the same pinout as the one in the ggg layout.
Now you know how to source the pinout and identify it form the data sheet.
From the same sheet you can compare the on and off characteristics. Etc.
Your on your way.

[cnspedalbuilder]

Thanks @Kipper4, good to know I am making progress 

To continue on the path: Do the on and off characteristics correspond to the voltages you would record when pedal turned on or off? Or are they absolute limits so if your voltage is higher then you know something is wrong?

[robthequiet]

To continue on the path: Do the on and off characteristics correspond to the voltages you would record when pedal turned on or off? Or are they absolute limits so if your voltage is higher then you know something is wrong?

Pretty much true. Another thing I would check in general is current.

I am definitely not trying to send you on a wild goose chase with this, but what I would be doing is "walking the circuit" against the schematic, or layout. In other words checking with power off and the meter set for resistance (on "beep"  mode if available) to check for continuity from pin to pin between components. Tricky, of course, because some connections are common to multiple parts. But if you found something open where it should show a join it might be a clue.

I'm going back to look at the voltages and pics you got earlier to see what else they might tell.

[duck_arse]

that diagram of the TO-92 package is the way/the one to draw on all your diagrams whenever you use a TO-92. of course, it it was a TO-18 package, you'd draw that instead.

[cnspedalbuilder]

Quick questions before I continue:
-@rtq dumb question, but are you saying to run a continuity check or to test resitance b/w points. My multimeter has different settings for those two options, and continuity tells me "short" vs "open" and resistance tells me resistance in ohms. [as an aside, I've run a lot of continuity tests in a semi-systematic fashion. I can't say that I've checked all connections, but a lot]
-I have a transistor tester coming in the mail today. Is it worth desoldering the transistors to test them out? I suppose I can also use the probes to check them in the PCB.

[robthequiet]

Oh, what I do is set my meter on "beep" just to get a quick answer as to whether something is open or not. Hopefully the resistance across a short is zero so it's kind of moot . Testing components for resistance in circuit is something I usually avoid, because everything is connected to something else. But what I'd be looking for is something that is not actually connected even though it seems to be OK visually. As far as pulling the FETs out, in this case I guess it wouldn't hurt given the mystery of the unit. Might give us a chance to microscopically examine what's going on with the circuit board in their vicinity. Good chance to install sockets in the process.

edit: PS Sockets might be useful if you think you might swap the FETs out for fresh ones, but as Mark Hammer said on another thread, it's more secure in the long run to solder them in, your call, of course.

[Kipper4]

Please repost your voltages. I can't remember if they where all ok or not.

Also does the circuit have a bypass switching arrangement already in place?
This feedback things bugging me. I'm sure it's not bugging me as much as the OP.

Does the circuit work without the switching?
Have you tried testing the circuit before you made the switching arrangement.

If there's an issue with the switch. It might possibly be causing the "feedback"

Sorry to ask. I know I'm a PITA.

[Renegadrian]

I dismantled my previous p45 (which worked perfectly), and used almost all the components (ics and trannies, resistors and caps) to gave birth to a new one using this layout above, which is the one you find online but made it smaller to save 1 row and 1 column, and include univibe mod at the same time.

obviously, it doesn't work. signal pass, but no phasing...

mandatory voltages.
WW 8.86v
upper IC
1 3.79
2 3.83
3 3.82
4 0
5 3.82
6 3.83
7 3.86
8 8.86

lower IC
1 3.86
2 3.85
3 2.60
4 0
5 3.27 <-> 5.43
6 3.66 <-> 5.63
7 1.37 <-> 8.21
8 8.86

2n5458
3.82
3.82
1.58 <-> 1.75

[cnspedalbuilder]

Please repost your voltages. I can't remember if they where all ok or not.

Also does the circuit have a bypass switching arrangement already in place?
This feedback things bugging me. I'm sure it's not bugging me as much as the OP.

Does the circuit work without the switching?
Have you tried testing the circuit before you made the switching arrangement.

If there's an issue with the switch. It might possibly be causing the "feedback"

Sorry to ask. I know I'm a PITA.

1. OK I'll check voltages again. I just got a cheap transistor tester, so I'll also try to figure out a way to test them w/o removing them from the PCB.
2. Yes the feedback is bugging me even more. But thanks it's really cool that you and other forum experts have been sticking through this.
3. Does the circuit have a bypass switching arrangement already in place? I'm not 100% sure what you mean, but the footswitch is installed and it is wired for true bypass. The sound is perfect when bypass is on, but the feedback will go through both inputs and outputs when pedal is powered on. The gain is so high that you can hear it even with volume at or slightly above 0.
4. Right now I'm still early on the learning curve so, I did not test the circuit before installing the footswitch. Is there a way to easily test the circuit w/o switch that does not involve desoldering all wiring to the switch? I really cleaned up the wiring and I think I might cry if I have to undo all that work 
5. I have been wondering if the 3PDT switch might possibly be causing the "feedback". Can you think of how a switch problem might do this? Is there a continuity test I can run to test for this possibility?
6. thanks Kipper, no you're not a PITA

[cnspedalbuilder]

Follow up:
So I got a cheap transistor tester today (http://www.ebay.com/itm/321963237498?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT) and tested the JFETs. I was nervous about desoldering them b/c I didn't want to damage them during testing, so I used some wires for leads to hook around the legs of the transistor on the PCB. As a result, these readings might be off. My tester said that Q1 showed 2.3A and 1.1V, and Q2 showed 3.2A and 1.0V. I don't know what the voltage and current correspond to on the datasheet  and again I'm unsure of the readings. But I am guessing that these matched JFETs should show approximately equivalent numbers??

Voltages:
Q1 (D S G)= 3.92;   4.02;  1.4-1.7 [voltages on G seem to oscillate]
Q2 (D S G)= 4.01;   3.99;  1.3-1.7
IC1: 4.15   4.12   3.83   .6   3.9   4.06   3.87   9.11
IC2: 3.8-3.97   3.93-3.95   3.88-3.89   .7   0-5   0-5   0-8.5  9.13 

[Kipper4]

Your voltages seem to check out. And if the gates wiggling the lfo is working .great.
I'd unsolder the in out wires to board end and put temporary ones in and test it without the switch

[cnspedalbuilder]

Your voltages seem to check out. And if the gates wiggling the lfo is working great.

@Kipper, I just updated the voltages to include IC2 (see above). In the build instructions it says voltages for pin 4 of both ICs should be zero, but in all my tests, it has been pretty solid in the .6-.7 range. Is that weird or normal [I dk how to interpret these voltages]?

I'd unsolder the in out wires to board end and put temporary ones in and test it without the switch

I think I get it, but let me make sure. Right now wires go from "I" and "O" pads to switch. So you are suggesting that I remove those wires and remove connections b/w jacks and switch, so I can put a temporary wire linking jacks directly to I and O. Right? It makes me shudder to think of desoldering more connections  , but I can do it in the name of science!

[cnspedalbuilder]

Another?: I feel silly having bought a transistor tester and not knowing what the results mean. Is it unusual that the current and voltage readings from the tester do not match across the JFETs? For that matter, when I see results about current, "HFE", and voltage, how do I know whether results are bad or good? Sorry for the lame questions, but I will learn!

[robthequiet]

The transistor tester I believe should be used on components out of circuit. Any other components attached to the legs would present other avenues for current to flow and skew the results. Also note the diff between testing FETs and BJTs, as I am fond of saying things that go without saying.

The sound is perfect when bypass is on, but the feedback will go through both inputs and outputs when pedal is powered on

<---  This does raise a question about the switch, so I'm with Kipper4 on testing with switch out of circuit. In true bypass, the signal from the guitar completely disconnects from and goes around the box. So the bypass part works fine, it's when we stomp on the switch that the fun starts. I'm not big on adding solder joints, but you could just snip the wires at their mid-point and alligator-clip connections for testing the switch, then resolder and shrink-wrap them back together later. Saves the switch and board from desoldering trauma.

As far as pin 4 on each chip being at .6 or .7, that would coincide with the voltage drop across a diode junction, but could just be a coincidence. I would expect zero volts at ground, too. One way to look at that is to put the meter leads carefully on pin 4 and pin 8 of each chip, then try the same but put your leads on pin 4 and the ground side of the battery terminal. If the ground on the chip is not at circuit ground, that would be something to ponder. According to GGG, pin 3 on IC2 should "vary with speed control" -- does it vary or just stay locked on ~4V? A 0.2V swing may be normal, but imagine what happens if IC2 pins 2 and 3 were shorted?

Also,

If it's any consolation, my main axe is a strat that I've been tinkering with for 20 years. I'm reheating solder that came from the factory in 1991. Just have to heat sink the parts when you apply the iron.

[robthequiet]

Sorry for the double post!!

[cnspedalbuilder]

As far as pin 4 on each chip being at .6 or .7, that would coincide with the voltage drop across a diode junction, but could just be a coincidence. I would expect zero volts at ground, too. One way to look at that is to put the meter leads carefully on pin 4 and pin 8 of each chip, then try the same but put your leads on pin 4 and the ground side of the battery terminal. If the ground on the chip is not at circuit ground, that would be something to ponder. According to GGG, pin 3 on IC2 should "vary with speed control" -- does it vary or just stay locked on ~4V? A 0.2V swing may be normal, but imagine what happens if IC2 pins 2 and 3 were shorted?

These are really good suggestions. Will follow up.

If it's any consolation, my main axe is a strat that I've been tinkering with for 20 years. I'm reheating solder that came from the factory in 1991. Just have to heat sink the parts when you apply the iron.

Sorry for yet another ?, but how do I heat sink the parts?

[Kipper4]

I know your reluctant to remove the I and Out wires.
It's just a hunch, based on your feedback comment.
Usually I test all circuits work before doing the bypass switching. Just one less thing to debugg. If I know the circuit works and I attach the switch and it no longer works, I'm all over that switch.
Next I box it up, it now doesn't work. Geez just an ordinary day at the bench. Start looking for shorts and stuff. Anyway you get the picture and if not here's a little saying we have around these parts.
Rock it before you box it.

Heat sink. Clamp the needle nose pliers, haemostat , tweezers what ever you have, to the leg of the transistor/componant your about to solder.
This will absorb some of the heat from the soldering and save the body overheating.
Maybe do A leg then do something else to allow it to cool a bit.
You shouldn't need to be too cautious. Get in with the iron heat up pad and leg, flow the solder and get out, a few seconds should be enough.

Let us know if you decide to test withou the bypass switching. And the results.

Rich

[robthequiet]

To heat sink or sink heat, clamp an alligator clip or forceps or tweezers on the leg, if possible between the component and the iron -- it serves to draw heat out of the component so that the component does not take all the heat from the iron. Right down on the board, it may not be possible, so use your needle-nose pliers on the component side and the iron on the copper side, and pull very gently. Using solder-sucking mesh could help to remove excess solder, but just to lift a leg of a diode or transistor, you only have to heat the join enough that the leg will come out with a very gentle action. I'm sure there's a couple of videos around on solder technique, hopefully something about heat sinks. Good skills to have if you go about salvaging germanium transistors from old radios and TVs.

[cnspedalbuilder]

Got it, thanks to you both!

So, I desoldered the I and O wired from the board. Then I did a quick and dirty solder of wires from the board to a couple of mono 1/4 inch jacks and ran a ground wire between them. Now I'm getting no sound. I certainly can fix the soldering job on the 1/4 inch jacks, but wanted to ask: Is it just better to wire the board directly to the original jacks on the pedal enclosure? I soldered to new jacks because I figured that all the extra connections to the switch, DC jack, etc. could mess up this experiment. But now I'm thinking that I might have been better off just replacing the in and out jack wires with the wires that I've hooked into the PCB. Thoughts?

[Kipper4]

Ok great. That's cleared that up.
Back to the circuit then.
I'll take another look through all through the info tomorrow as it's very late here now.
Ps you did have the fets in when you did the test didn't you?