More Harmonic Percolators

[PedalBuilder]

After tracing the Latent Lemon Hurts, an interesting contemporary take on the venerable Interfax Harmonic Percolator, I laid out a couple of PCBs for a similar Harmonic Percolator build. I couldn't make up my mind as to whether I wanted to build a 1590BB2 pedal like the Hurts or whether I wanted a more compact 125B version. After hemming and hawing back and forth, I ended up getting both fabricated. I've really enjoyed the look of some Robert's recent PCBs with their curvy traces, so I leaned into that aesthetic for these builds.

I deviated somewhat from the Latent Lemon circuit in both builds. The actual circuit draws about 55µA, excluding the LED, so I was able to increase R101 from 47Ω to 4.7kΩ while only dropping the voltage by 0.28v. Theoretically it should reduce any noise coming in via the power rail. I added a 1kΩ limiting resistor to the Harmonics control so that it didn't mute the pedal when turned all the way down. I also replaced the 5M trim pot that the Hurts uses for the Q2 collector-base resistor with a 2.7M resistor and a 2M trim pot because the 2M is the largest 3296 that Tayda carries.

The two builds were largely identical, with a couple differences:

• Transistors: Both builds used Texas Instruments 182505 germanium PNP transistors for Q1. The 1590BB2 build used one with a slightly lower gain (hFE 44, leakage 7µA) than the 125B build (hFE 57, leakage 13µA). That small difference was enough to make a very noticeable difference in sound and dynamic behavior. Despite having otherwise identical biasing (and identical silicon transistors for Q2), the 1590BB2 build is fuzzier and harmonic-laden, and never fully cleans up with the guitar's volume knob. The 125B build is more polite, for lack of a better term, and while it certainly fuzzes, it always sounds a little more polished and cleans up nicely as you roll back the volume knob. Both have a nice and very prominent suboctave undertone.
• Diodes: The 1590BB2 used 1N914 for the silicon clipping diodes; the 125B used 1N456. The 1N456 looks cooler, but both sound identical to my ears.

Both are fun builds and sound different enough to justify keeping both of them around.

Here's the schematic that I used for my builds; gut shots are below.



1590BB2


125B

 

[jimilee]

That’s really a fascinating schematic. I’m very curious to hear it. It looks very cool. Are the internal trimmers for calibration?

 

[PedalBuilder]

That’s really a fascinating schematic. I’m very curious to hear it. It looks very cool. Are the internal trimmers for calibration?

Yep. One of the internal trimmers (Octave) controls the resistance between the collector and base of Q2, which sets the intensity (and location on the fretboard) of the suboctave undertone. It's very much a set-to-taste control. I had good results from starting with the total Q2 collector-base resistance at 3.9M and adjusting it until I was satisfied with the sound.

The other internal trimmer is part of the Deep switch circuit. One pole of the Deep switch adds a parallel input capacitor, which lets more low frequencies into the circuit. Unfortunately, this causes unwanted oscillation. To counteract the oscillation, the other pole of the Deep switch inserts the other internal trimmer in parallel with the Q1 collector-base resistor. That decreases the Q1 collector-base resistance, lowering the gain of that gain stage, and stopping the oscillation. To set it up, start with the trimmer set to maximum resistance, and then turn it down until the oscillation stops.

One other thing to note—C5 in my schematic, the interstage coupling cap, is often listed as 1µF or 2.2µF in the various Harmonic Percolator schematics that are floating around the internet. If you use large caps like that for C5, you will not get any suboctave undertones. The key to the drawing out the suboctave is using a 100nF cap for C5 and a very large Q2 collector-base resistor (3.9M is a good starting point).