Prince of Tone - A particular problem

[LaloFP]

Hi people!

Im having an issue with a Prince of tone (in a PCB I used other times).

It works beautifully in all settings except when I activate the hard clipping, and in particular with the diode that is pointing to VREF.

The symptom: With Gain in MAX, if I hit the guitar hard, the volume drops, the sound dies with fuzziness (like a fuzz face low biased) and returns back with a fade in. If I keep playing it works Ok, but if I stop, wait two seconds, and hit the strings hard again, the issue repeats. Replacing the IC doesnt solve the problem.

Clue1: If I take out the hard clipping diode or if I cut the soft clipping part pointing to that same direction, the problem disappears.

Clue2: In the "safe settings" I get 4.5v in PIN3 of the IC. With the "bad settings" I get 7.5v! (it doesng happends in another POT that I already have). So it make sense the IC is clipping in the input. But I cant trace the root cause.

Marked in green:

[baloubass]

I had this problem in the past and as i remember, it was a bad +9v connection. Check your solder on the 9v jack and the pad on the pcb.
Remove the solder and check if the pad are not damaged.
Then you can check your power capacitor pad and trace.

[LaloFP]

I added a battery clip directly to the Pads in the PCB and it doesn't solve the problem.

Another clue: I replaced both power 100uF capacitors to new ones and the problem persist BUT the time it takes to the sound to return now is longer and the volume drop is deeper.

[antonis]

Try to to replace (connect) D5/D6 VB with a "real" (DC+AC) Ground..

[LaloFP]

Antonis, I tested that and worked, but the original POT design is with the VREF and my old builds never had that issue.

I want to know whats happening here so I learn more 

[antonis]

It's happening exactly what is happening to many circuits using AC ground for an "absolute" ground..

VB is formed via a voltage divider from PS to GND.. PS to GND andVB to GND are regulated with electro caps..
These caps are there NOT to form an AC short to ground - although they manage it well in most of cases..

In fact, their issue is more"current" than "voltage"..
(just for fun, try to make R9 10k or greater.. or connect C4 lower leg to VB..)

[LaloFP]

It's happening exactly what is happening to many circuits using AC ground for an "absolute" ground..

VB is formed via a voltage divider from PS to GND.. PS to GND andVB to GND are regulated with electro caps..
These caps are there NOT to form an AC short to ground - although they manage it well in most of cases..

In fact, their issue is more"current" than "voltage"..
(just for fun, try to make R9 10k or greater.. or connect C4 lower leg to VB..)

Hey, the C4 change to VB worked...

But, I dont understand whats happening yet. I understand the VB and voltage divider thing (just in a general way...), but cant see why It works most of the time and now it doesnt (even changing those two 100uF caps for new ones). Is it due to components tolerances and having a more gainy unit because of that, that the current is larger than the caps can manage?

I would love to fix the issue knowing what is happening and without changing the design. (It feels like a workaround for me)

[antonis]

I dont understand whats happening yet.
I would love to fix the issue knowing what is happening and without changing the design.

That said, it should be more ethical to ask the original designer for it, shouldn't it..?? 

P.S.1
Don't get me wrong.. I'm in a spamming mood...

P.S.2
Doesn't the fact that, changing the design results in working circuit, makes you suspect that maybe it could be just a practical design issue..??

[LaloFP]

I retract myself...

I changed C4 to point to GND instead of VB again and the problem is still there...  (I dont know If I tested it right before. I wasnt using all the diodes in the clipping sections)

When I hit the strings really hard the volume drops to zero and after a second fade in again. If I play continuously it works. The problem appears after some seconds of silence

Any other ideas? Since its a fade in effect, I suspect it has something to do with caps, but I already replaced the two 100u ones

[LaloFP]

Okey, I think I found the problem.

I replaced every component (no kidding) and the problem still remained. Until I checked the soft clipping section and found that one of the rows of clippings has one of the small MA856  looking to the other side... Corrected that and worked perfectly...

Im kind of sad and happy, you know... hahaha. But worse would have been to go rebuild the pedal without knowing what happened

Thanks guys!

[antonis]

Until I checked the soft clipping section and found that one of the rows of clippings has one of the small MA856  looking to the other side... Corrected that and worked perfectly...

Can't figure out the way of asymmetrical clipping could result into your isuue, but if it works OK now, let it be..

[LaloFP]

Yeah, I didn't suspect that, and because of that I looked there last.

The effect worked then but all the changes made the board look ugly, so I built it again and it worked at first.

So... just to test the previous diagnostic, I replicated the diode error (to left open one of the soft clippings lines) and, in the first try (hard string hit) the error came back (the sound muted and then faded in again) but this time really quick, like 1 second total. But after that, I couldn't replicate the error.

So I think there is something there. Maybe the kind of diodes that the POT uses let pass a lot of current and that makes it hard to the VB capacitor and put it in a not so safe workload.

Thanks guys!!! Happy with my refurbished POT haha 

[antonis]

Well done..!! 

As you might understood, "open" half-wave NFB loop (due to reverse diode placement) resulted into op-amp working at its open-loop gain mode for the half of complete signal duration..
(not a good thing for anything else than a halfwave oscilator..)