Sanity check - DC filtering

[PedalBuilder]

It's kind of curious that not many pedal makers have adopted this way of wiring the circuit. It seems most reverse power supply with a charge pump. I wonder why?

This topic has come up many times over the years. Flipping the power supply in a fuzz face works well some of the time, but it's unreliable and prone to oscillation. As RG Keen noted:

It is indeed possible. However, this method sometimes unpredictably produces oscillation that cannot be cured by methods other than un-converting the power supplies. It works often enough to keep people suggesting it, but every so often it doesn't. It's kind of a self-built trap for beginners.

 

[MBFX]

A self-built trap for beginners.... lmfao. OK, I will power my breadboard backwards and also buy a 9V and build it the right way.

 

[Locrian99]

This topic has come up many times over the years. Flipping the power supply in a fuzz face works well some of the time, but it's unreliable and prone to oscillation. As RG Keen noted:

I have to wonder if the oscillation people experience is from using a 1044cpa vs scpa I believe the 7662/7660 also has a version that produces the oscillation therefore shouldn’t be used in audio circuits.

 

[PedalBuilder]

I have to wonder if the oscillation people experience is from using a 1044cpa vs scpa I believe the 7662/7660 also has a version that produces the oscillation therefore shouldn’t be used in audio circuits.

The oscillation that Keen was discussing results from using a flipped power supply rather that using a voltage inverter circuit. But you are right, you need to make sure that you aren't using a low frequency charge pump as a voltage inverter in a positive ground fuzz circuit.

 

[Big Monk]

Fuzz Faces are noisy enough as it is. NEVER make it worse by doing things the hard way.

 

[MBFX]

I got home from the lab a little early tonight, made my wife dinner and sent her to bed with some wine because she isn't feeling well, and I have a little time to play with the circuit I built from the schematic I posted initially. I am going to write out what I am doing as it happens for posterity. What if it works? I don't want to leave all you kind, helpful people without closure! PG-13 language ahead.

First, I verified that everything was connected correctly, and plugged it in to my amp. Loud-ass dirty boost, no change since this morning, cool! Then I measured Q1 and recorded the following initial values:

• Q1-B = -0.288
• Q1-C = -2.783
• Q1-E = -0.225

Hmm... well, my suspicion that | (negative voltage) = [(total voltage) - (positive voltage)] | has proven true. It also appears that the negative voltage at Q1-C is exactly 4X higher than it should be. I decided to try adding a 1K resistor to Q1-E to see what would happen. I recorded these values: Q1-E= -0.424, Q1-B = -0.484, and Q-1C = -2.933.

Shit. That obviously made things worse, so I tried switching out the Q1-C resistor for a 330K because it was on my desk and a conveniently large value to measure which way the voltages swung. I recorded these values: Q1-B = 0.188, Q1-E = -0.177, and Q1-C = -0.496.

330K was a good guess. I needed a little more current, so I messed around until I narrowed it down to a 220K resistor. I now had the values I was looking for: Q1-B = -0.207, Q1-C = -0.702, and Q1-E = -0.185. I can't find anything regarding the necessary emitter voltage on Q1, so I decided to ignore it and see what was happening over in Q2 land after increasing the Q1 resistor by so much. I recorded these initial values:

• Q2-B = -0.702 (of course)
• Q2-C = -8.58
• Q2-E = -0.614

We're closer! Q2 is fucky, though. The collector voltage is way off, and the emitter is off by >20%. I already had a 10K trimpot in place of the 8.2K resistor, and I messed with it. At the full 10K, I got about -5VDC. I'd probably need around 12K to get -4.5VDC, I think. I'm tired and need to go to bed soon, so I left it and I recorded these values:

• Q1-B = -0.210
• Q1-C = -0.726
• Q1-E = -0.186
• Q2-B = -0.726
• Q2-C = -5.12
• Q2-E = -0.604

That looks pretty good! The apartment next to me is unoccupied, so I'm going to try it out...

...YES!!!!!!!!

Oh man, this sounds SO GOOD! Since the booster was still on the board, I tried switching it on too. Instant Tone Bender insanity. The booster is an Electra without the clipping diodes, using a BC238 transistor. It actually still responds really well to the volume knob, despite being much more aggressive.

WHEEEEEEE I MADE A GERMANIUM FUZZ FACE AND IT'S FUZZY NOW!!!!!!!

Looks like I still have a little too much voltage to Q2-E, though. Does anyone have any ideas on how to remedy that? Is that voltage proportional to the voltage at Q2-C, and if I were to adjust the resistance to Q2-C until it read -4.50VDC would I find -0.500VDC at Q2-E?

EDIT: punctuation, a word, some details

 

[Amplified Parts]

Q2E should be Q1C/Q2B minus the diode drop of the base-emitter junction of Q2. Not proportional, but Q2E should always be just a bit below those voltages, ~0.1-0.3V closer to zero in the case of germanium.

Q1E should be grounded, so it should always be 0V in a Fuzz Face unless you've modified the circuit. Make sure that measures 0V or else check the connection there.

 

[MBFX]

Q2E should be Q1C/Q2B minus the diode drop of the base-emitter junction of Q2. Not proportional, but Q2E should always be just a bit below those voltages, ~0.1-0.3V closer to zero in the case of germanium.

Q1E should be grounded, so it should always be 0V in a Fuzz Face unless you've modified the circuit. Make sure that measures 0V or else check the connection there.

The battery in my multimeter was old, and my readings became erratic over the course of a couple days. I noticed it when it was reading a small amount of DC voltage at idle with nothing connected, replaced it, and it worked. Go figure

 

[Amplified Parts]

Nice. Typically, I wouldn't sweat my bias voltages too much as long as it sounded nice, particularly if they're just a bit off, but I still feel like there's something a little odd at play here. 220k for Q1C is pretty massive, and your Q2E voltage before that tweak tells me that a lot (relatively speaking) of current was flowing through the feedback resistor. I expect to see that with the stock 33k resistor when Q1 is very low hfe, which makes me think that (assuming your measurement of 70 / 300uA is correct) you might have the collector and emitter swapped for Q1. Otherwise, I'm not sure what to make of some of those voltages with those transistor measurements.

 

[MBFX]

Nice. Typically, I wouldn't sweat my bias voltages too much as long as it sounded nice, particularly if they're just a bit off, but I still feel like there's something a little odd at play here. 220k for Q1C is pretty massive, and your Q2E voltage before that tweak tells me that a lot (relatively speaking) of current was flowing through the feedback resistor. I expect to see that with the stock 33k resistor when Q1 is very low hfe, which makes me think that (assuming your measurement of 70 / 300uA is correct) you might have the collector and emitter swapped for Q1. Otherwise, I'm not sure what to make of some of those voltages with those transistor measurements.

This was a couple weeks ago - the current iteration uses GT402s in both spots and works fine. Working up the courage to solder it together and drill up my expensive enclosure I got for it.

I assume the AC128 is just faulty. It hisses like mad wherever I put it.