Understanding odd diode clipping in a fuzz pedal

[jordan s]

I'm building a Harmonic Percolator style fuzz pedal with a pcb from Fuzz Dog. I have built about 20 of these pedals before on strip board with great success but I'm getting some weird results with this batch. What I'm noticing is that when I turn the output volume all the way up, the diode clipping is gone completely. When I bring the output down a little, the clipping is back. Other than this, everything functions as expected. There is a switch to switch between germanium or silicon diodes or to remove the clipping diodes all together.I tried manually just placing a pair of diodes between the output and ground without the switch and I get the same result: no clipping! I can simply place a 4.7k resistor on the output and that seems to work but a resistor on the output is not part of any schematic available for this fuzz so it seems like I've done something wrong but I can't figure out what. The output level is quite a bit more than the bypass level, seemingly more than my first batch of 20.

I have made some minor part value changes on the Fuzz Dog pcb and added one cap to match the Steve Albini schematic released by madbean pedals (both linked below.)
http://pedalparts.co.uk/docs/HarmonicPercolator-v3.pdf
https://www.madbeanpedals.com/projects/_folders/1590A/pdf/PepperSpray2018.pdf

 

[RuudNL]

Do you load the output the same way as you did before?
I can imagine that if the output is terminated with a low(er) impedance, the output level would drop in such a way that the diodes can't conduct. Especially because you get clipping at a lower output setting.

 

[jordan s]

The schematic is the same so it shouldn't be any different. The output doesn't really have any pulldown resistor or anything to deliberately set impedance.

 

[jordan s]

What's even more odd is that it displays this behavior on some amps, on others it acts perfectly normal.

 

[PRR]

> schematic is the same so it shouldn't be any different.
....it displays this behavior on some amps, on others it acts perfectly normal.

The "Whole Schematic" must include the source driving it and the load loading it!!

The 91k at R3 suggests it has a very high output impedance. If pushing a low impedance load the gain will be less. Maybe not enough for your guitar times gain to come up to the clip-diodes' threshold.

 

[RuudNL]

What's even more odd is that it displays this behavior on some amps, on others it acts perfectly normal.

That is an indication that the load impedance matters!

 

[jordan s]

Well, I grabbed one of the pedals I built years ago on strip board and found that it has the same problem. I just never noticed it because it depends on the input impedance of the amp and I guess it wasn't a problem on the amps I tested on previously.

 

[PRR]

It actually should "do OK" to below 100k load.

"Any guitar amp" should be over 100k input (the exceptions are mostly old and notorious).

So what is this 'new' amp which loads so much?

 

[jordan s]

The amps I use at home are a "little gem" LM386 type amp,  an Orange Crush 35RT SS combo and a "firefly" tube 1.5w tube amp. The issue occurs on the Orange and firefly but not the little gem. My guess is that is has to do with the JFET buffer in the gem. Adding a similar buffer to the output of the pedal seems to solve the issue for the other amps. I just wonder why so many people have built this pedal but no one has mentioned this issue.