Potentially Broken Fuzz, v3

Details

I made a version of this a while ago. I was trying to build a filthier fuzz and misinterpreted the data sheet of an LM386. The error didn’t necessarily sound “good,” but there are companies out there that have sold a factory’s worth of fuzzes that can make a lot of terrible noises, and I took it as an “Isaac Newton Watching an Apple Fall” moment, representative of my latent genius in circuit design and emblematic of the Dunning-Kruger ethos. I did not go on to invent the calculus of guitar pedals, however, and my understanding of electrical circuits has remained impeded by topics as basic as “impedance,” which I still think of exclusively as “inscrutable resistance.”

That said, I have learned a few things. I can do simple layouts on vero now, for example, and can routinely wire up a potentiometer without setting it on fire—although during the development process of this pedal, I did manage to burn myself on an LM386 that had gone China Syndrome and was in the process of melting through my breadboard. I decided to apply what I do know to the Potentially Broken Fuzz concept and see if I could get it to still be audibly broken but less electronically broken, finding a compromise between fun wackiness and annoying instability.

My first instinct was to replace the JFET buffer with a TL072 and use the extra op-amp to maybe run an active EQ, but I got talked out of that. The buffer wasn’t really doing much in this context, which I confirmed via experimentation, and the active EQ was a lot of parts for questionable benefit, so there is no buffer (from neither a JFET nor an op-amp) in the Tokyo Drift version. The old first order low-pass tone control was replaced with some sort of “half tuned bridged tee” notch filter borrowed from dylan159’s Dyl-ei. It was suggested that I check the impedance (i.e., inscrutable resistance) in relation to the LM386, and was also told that it was just the reciprocal of 2π times the square root of R1, R2, C1, and C2, so I was like, “Okay.” I don’t know. I tried to model it in LTSpice and everything, but it’s over my head. It seems to work as-is, and will do a wild amount of adjustment.

I also simplified the starve controls, tuned the values in the FREQ section, and got it all to fit in a 125B enclosure, which is painted with the same “80s Kitchen Appliance Almond” base color as the previous Mk II version.

Sound

It’s delightfully broken. I wonder if I tamed the STARVE knob too much and the FREQ doesn’t seem to do as much as it did when it was on the breadboard, but it does have a more useful range than the previous version (where “useful” in this context means non-self-oscillating). The TIMBRE control, which is the notch filter thing that replaced the old tone control, is bonkers, and has a huge impact on the sound.

I really did burn my finger on an LM386 that was shuffling off this mortal coil, and created a disturbingly square blister on the index finger of my fretting hand, so the following samples were difficult to record. The first section is a baritone guitar that progresses through clean (a SansAmp GT2), then the standard distortion on single notes, the standard distortion on power chords, then the FREQ is engaged and it gets messy. After the long tail with the classic 386 grit, the guitar switches to the Coronacaster into a mostly clean Fender Deluxe Reverb. In this section, the onboard tone knob on the Coronacaster is actually all the way down, and the high frequency content is mostly coming from the FREQ feedback, which is…odd. The final part is the same settings, with some delay from a DD-5.

---