Help - Klon clone debugging

[Dangaridoo]

Howdy Guys, having some issues with a Klon clone and I'm pretty dang stumped.

Project: Refractor (Klon Centaur clone) from Aion electronics

Project documentation/URL: https://aionelectronics.com/project/refractor-centaur-overdrive/

Schematic/IC Voltage measurements: http://imgur.com/VvWixhk

Modifications to original project: I'm using D9Es for D1 and D2

The situation:
-Tested all Resistors/Caps/Diodes prior to soldering onto the board
-Build and soldering looks alright
-Bypass works perfectly
-Pedal works perfectly for about 5secs after power is applied
-After 5sec or so, with gain knob at max (5 o'clock) signal begins to fizzle out (starts with tail jumping in/out and dying faster than natural - progresses to almost no signal coming through at all)
-Can dial back the gain knob to regain functionality, but over time, less and less range is functional
-After 30sec or so, gain knob cuts signal in all positions except in min (7 o'clock) position

Actions taken:
-Due to the gain pot having some effect on the problem, I've replaced the Gain pot. This made no difference.
-Due to the IC voltages (on the op amp outputs) being pretty crazy I've replaced IC1 and IC3 (I've also swapped IC2 in/out with the old IC1). None of those changes made a difference.
-Due to the fact the problem appears to be affected by time, I:
   -checked all caps to see if they were leaking/deformed (all were fine)
   -checked all polarised caps to ensure that they maintain steady voltage (all held steady voltage)
-I was pretty stumped at this stage so I built an audio probe - found that there is no signal output from IC1 pin 7
-Probed around and replaced C3, C7, C8 based on flawed logic. This made no difference.

IC voltages in text format in case that's preferred:
Vin: 9.25

IC1
1: 4.4
2: 4.4
3: 3.45
4: 0
5: 3.78
6: 4.24
7: 1.33
8: 9.25

IC2
1: 15.64
2: 4.15
3: 4.34
4: -8.86
5: 4.33
6: 4.35
7: -6.87
8: 16.67

IC3
1: 9.25
2: 4.47
3: 0
4: -4.55
5: -8.84
6: 5.16
7: 6.73
8: 9.25

Any hints/tips on where to go/what to do next would be really appreciated!

[duck_arse]

hello Dangaridoo, and welcome.

your IC1 pin 7 shows a problem, it should be about the same as pin 6. you might have a shorting component somewhere in that area. you'll need to test continuity (POWER OFF) in and around that area. can you post some photos of your build, front and back and off board wiring, please?

[EBK]

I'm wondering if C9 is bad.

[Dangaridoo]

hello Dangaridoo, and welcome.

your IC1 pin 7 shows a problem, it should be about the same as pin 6. you might have a shorting component somewhere in that area. you'll need to test continuity (POWER OFF) in and around that area. can you post some photos of your build, front and back and off board wiring, please?

Thanks for the welcome and suggestions!

They're a bit nasty but photos can be seen here: http://imgur.com/a/c5iHD

Some explanation with regards to the offboard wiring: the enclosure is intended to contain both King of Tone and Klon circuits. The king of Tone side works fine. Both circuits are connected to the 9v supply.   Although not pictured (due to using loose wires), the jack input is connected to both boards currently, but only the klon output is connected to the output jack.

With regards to testing for continuity, I did a quick check prior to heading off to work, and found nothing out of the ordinary.
IC1 pin 7 to ground = increases with time (due to C9).
C9 to ground = total resistance equals R13 + diode.
IC1 pin 7 to Vb = total resistance equals R10+R11+R12 when gain pot is at max (5 o'clock).

When I get home I'll do some more checks to confirm that no solder joints around IC1 pin 7 are shorting. (And maybe I'll replace C9 just in case)

Edit:

Alright. Got back from work and jumped straight on this.
-Checked continuity around IC1 pin 7 - it's all fine
-Checked continuity from IC1 pin 7 back to the gain pot - it's all fine
-Checked continuity from IC1 pin7 to the GND via the diodes - it's all fine
-Replaced C9 - made no difference
-Checked diodes - .286v across each diode in each direction (though - because these are wired to the board, I'm pretty sure this doesn't indicate the diodes are busted? happy to replace if it's recommended)
-Found another new TL072CP - threw that in the IC1 socket just to double check that IC1 wasn't busted - unsurprisingly it made no difference.
-Discovered that the audio probe DOES actually pick up the signal on the output of IC1 pin 7 - it's just REALLY QUIET
-And before I could throw the audio probe on C10 it broke

I figure at this stage - I'll replace C10 just for kicks. If that doesn't work, I'll give D2/D3 a shot. After that, I'll be well and truly confused.

Happy to take suggestions in the mean time!!!

Edit 2:
-C10 replaced. No difference.
-C15 replaced, thought I spotted a leak. No difference.
-Lifted D2 to check the diodes with multimeter - they're both fine (.286v in one direction only per diode)
-Reflowed solder on all joints because I'm out of ideas. No difference.

Final voltage measurements on ICs (thought I'd take some measurements with gain pot at min and max):

IC1	min	max
Pin 1	3.83	4.1
Pin 2	3.83	4.09
Pin 3	3.14	3.28
Pin 4	0	0
Pin 5	3.81	2.94
Pin 6	3.82	3.97
Pin 7	3.83	1.31
Pin 8	9.25	9.25

IC2	min	max
Pin 1	15.64	15.55
Pin 2	3.66	3.54
Pin 3	3.78	4.08
Pin 4	-8.85	-8.8
Pin 5	3.79	4.08
Pin 6	4.18	4.14
Pin 7	-7.13	-7.03
Pin 8	16.69	16.6

IC3	min	max
Pin 1	9.25	9.25
Pin 2	4.45	4.45
Pin 3	0	0
Pin 4	-4.57	-4.57
Pin 5	-8.85	-8.85
Pin 6	5.16	5.16
Pin 7	6.66	6.66
Pin 8	9.25	9.25

I'm flat out of ideas on what to do next.

[EBK]

I hope it's only the fact that D1 and D2 are backwards keeping you down.   
Flip them around the right way and report back (fingers crossed).

(I also hope the pad labeled "IN" got soldered to something after you took those pics....)


Update: Hmm...  Looking at the schematic, it shouldn't make any difference if D1 and D2 are backwards (as long as they both are, which is the case here).

[bean]

Also, pin1 and 7 of IC2 are way off. Make sure to check those voltages again after flipping the diodes like EBK said.

For reference here are voltages measured on a different clone
9.42v Supply                     

IC1   TL072
1   4.6
2   4.6
3   3.6
4   0
5   4.58
6   4.6
7   4.6
8   9.17

IC2   TL072
1   4.64
2   4.6
3   4.59
4   -8.9
5   4.59
6   4.6
7   4.55
8   17.7

IC3   ICL7660CPA
1   9.17
2   5.47
3   0
4   -3.59
5   -8.9
6   4.49
7   5.7
8   9.17

[duck_arse]

I don't much like klones, too many funny voltages. but, after doing EBK's stuff, can you measure your VB voltage, at R29//R30? (my documentation doesn't match your board layout, I think aion said he was going to update the docs.) does VB change with gain pot turnings? it shouldn't; it may indicate a bum connection betwixt pot and vb.

can we also have the voltage at R5//R8?

[EBK]

R20, perhaps (IC2A appears to be running open circuit and saturating). What is that in R20's spot anyway?  Doesn't look like a resistor to me.  Is it a wirewound or something?

I'm also concerned about R12 (this is actually where I'm placing most of my bet).  Looks like you have two resistors in series with a very questionable looking solder joint.  A bad solder joint could easily cause the time-based failure you are observing:
Joint heats up and expands slightly, breaking continuity.
After continuity is broken, joint begins to cool (because there is no longer current heating it up) and reconnects.
Sputtering is caused by a loop of heating up, breaking, cooling, reconnecting....
Reducing gain lowers current through R12, giving you a little more time before the inevitable thermal break. Over time, R12's surroundings are heating up (especially that op amp being switched in and out of saturation), and you reach that thermal break quicker (less and less range being available). With gain at 0, you finally reach an equilibrium where the current through R12 doesn't heat it up enough to break the connection.

[EBK]

Stephen, it looks from the latest posted measurements like VB is varying between 3.8 and 4.1V, which seems a tad low, but I think that is explained by a higher than expected current draw around IC1B (which is also why it is varying slightly with gain pot adjustment).

[Dangaridoo]

Thanks for the replies 'erryone!

(I also hope the pad labeled "IN" got soldered to something after you took those pics....)

Not soldered - but it certainly is being connected up prior to each test (one less wire to break!)

Update: Hmm...  Looking at the schematic, it shouldn't make any difference if D1 and D2 are backwards (as long as they both are, which is the case here).

That's my understanding. My diode test (on circuit board and then after lifting a leg from D2) confirm the theory. However, even though it doesn't matter, there's some fun trivia here: due to these being D9Es I believe that the anode is marked with the band (rather than cathode) - apparently that was soviet practice at the time?

I haven't changed out/re-orientated the diodes for the above reasons, but am happy to if ya'll still think it's worth a shot.

For reference here are voltages measured on a different clone                  

Cheers for the info! I've just observed that the voltage across the circuit slowly drops with time - that might be why my second set of readings are so much worse than my first, I had the board running for a while...

I don't much like klones, too many funny voltages. but, after doing EBK's stuff, can you measure your VB voltage, at R29//R30? (my documentation doesn't match your board layout, I think aion said he was going to update the docs.) does VB change with gain pot turnings? it shouldn't; it may indicate a bum connection betwixt pot and vb.

can we also have the voltage at R5//R8?

Of course:
R29//R30 approximately: 4.2v (gain min) 4.3v (gain max)
R5//R8 approximately: 4.2v (gain min) 4.3v (gain max)

I say approximately because the voltage appeared to be slowly dropping...  I figured this might be due to C18 leaking to ground slowly? So I replaced it, but that didn't make a difference.

R20, perhaps (IC2A appears to be running open circuit and saturating). What is that in R20's spot anyway?  Doesn't look like a resistor to me.  Is it a wirewound or something?

I'm also concerned about R12 (this is actually where I'm placing most of my bet).  Looks like you have two resistors in series with a very questionable looking solder joint.  A bad solder joint could easily cause the time-based failure you are observing:
Joint heats up and expands slightly, breaking continuity.
After continuity is broken, joint begins to cool (because there is no longer current heating it up) and reconnects.
Sputtering is caused by a loop of heating up, breaking, cooling, reconnecting....
Reducing gain lowers current through R12, giving you a little more time before the inevitable thermal break. Over time, R12's surroundings are heating up (especially that op amp being switched in and out of saturation), and you reach that thermal break quicker (less and less range being available). With gain at 0, you finally reach an equilibrium where the current through R12 doesn't heat it up enough to break the connection.

R20 is a 0.1% tolerance resistor - I brought it for kicks when ordering my parts. It's simply a metal film resistor with an epoxy coating for environmental protection. The resistance measures in tolerance of the expected 392k.

ARGH - I forgot to talk about R12!!! I'm really sorry about that. You're correct, it was two resistors in series - in the picture it was half covered in heat-shrink. I've replaced it with a 390k resistor to be safe - it didn't make any difference unfortunately

[EBK]

However, even though it doesn't matter, there's some fun trivia here: due to these being D9Es I believe that the anode is marked with the band (rather than cathode) - apparently that was soviet practice at the time?

Strange, but very interesting. 

I say approximately because the voltage appeared to be slowly dropping...  I figured this might be due to C18 leaking to ground slowly?

Also strange.  This voltage quirk bugs me.  I was so hoping to be right on R12 (and not just for my ego  ).  At least I only placed most of my wager on that.   We'll have to keep looking.

[EBK]

Edit: Just modifying this message to put Replace IC2 first. 


Could you verify the color bands on R11 and R17 for me?
R17 should be Red, Purple, Black, Red, (Brown), but I can't make out the first two due to that shadow.  Likewise, R11's first two bands are obscured (should be Brown, Green, Black, Red, (Brown)).

I've been visually double checking your parts and connections.  I'm unable to verify the markings on those two resistors and all the caps that aren't marked on top.  Can't read the markings on IC2 or IC3 either.

Additionally, can we have a peek at the gain pot wiring under that tape?

[duck_arse]

are there any straws left for the grabbing? the Vb should not change at all, any pot setting, certainly not at the R29//R30 point.

can you lift the "GAINA" pot wiper from the Vb, and the "GAINB" pot pins 1 and 3? you should get signal good thu IC1 at least, if that's where the problem is.

[digi2t]

However, even though it doesn't matter, there's some fun trivia here: due to these being D9Es I believe that the anode is marked with the band (rather than cathode) - apparently that was soviet practice at the time?

Strange, but very interesting. 

http://www.diystompboxes.com/smfforum/index.php?topic=79791.0

[Groovenut]

Could you do me a favor and measure your bias voltage with the gain at max after it stops passing signal? My theory is that the circuit is pulling the bias to ground with the gain pot at max, for some yet undetermined reason (bad cap in the bias supply maybe).

Pin 1 on IC2 is way too high, should be at the bias voltage (4.5 or so) and pin 7 is way too low, should be at bias voltage.

I recommend replacing IC2 with a new chip, not the old IC1 and see if there's a difference.

BTW, here are my confirmed good voltages. They echo Bean's

Measurements taken with a regulated bench supply set at 9 volts using a Fluke model 87V meter with fresh calibration

IC1 - TL072

1 = 4.49
2 = 4.49
3 = 3.54 stable, 3.95 initial
4 = 0
5 = 4.49
6 = 4.49
7 = 4.49
8 = 9


IC2 - TL072

1 = 4.55
2 = 4.49
3 = 4.49
4 = -8.59
5 = 4.49
6 = 4.49
7 = 4.44
8 = 16.23

IC3 - ICL7660SCPA or TC1044SCPA

1 = 9
2 = 4.59
3 = 0
4 = -4.22
5 = -8.59
6 = 4.09
7 = 5.59
8 = 9

[EBK]

are there any straws left for the grabbing? the Vb should not change at all, any pot setting, certainly not at the R29//R30 point.

can you lift the "GAINA" pot wiper from the Vb, and the "GAINB" pot pins 1 and 3? you should get signal good thu IC1 at least, if that's where the problem is.

There are a few different things going wrong. 
1) Too much current is flowing from that VB divider.  I'm wanting to blame the gain pot (a short hidden by that tape, perhaps).
2) IC2A and IC2B are both apparently running open loop and saturating.  But, why??  Looks like we've only tried replacing IC1 so far.  Maybe replace IC2 (seconding Lawrence's suggestion).
3) That IC1B output voltage (pin 7)....  Might come down to that gain pot current.


I'm wondering if my cracked solder joint theory was correct, but on the wrong part.  Maybe the cracked solder joint is on the gain pot under the tape....

[EBK]

One more theory.  If IC2 is bad, perhaps it is the thing sucking a lot of current from the VB node, which, through the gain pot connections, is messing with IC1. 

[Groovenut]

My guess is an internal short in IC2

[EBK]

My guess is an internal short in IC2

+1

[digi2t]

Are the IC's socketed? I once came across an internal short in the socket itself (poor molding?). Drove me insane before I found it. It was one of these types of sockets;