PedalBuilder
Well-known member
Bottom line up front—this thing is seriously awesome. For those not in the know, it's a modified fuzz face that uses an op amp servo circuit to keep the bias stabilized at a wide range of temperatures.
So having said the most important part, on to the build. I went slightly premium on this one—Wima, Nichicon, and silver mica caps and mil-spec 1T320V transistors. I have a sizable stash of low leakage, mid-to-high gain Soviet transistors with BCE pinouts and thought that they would be good fits for this build, so I laid out the board with in-line BCE footprints rather than the usual TO-5 footprint. The I/O board uses P-channel MOSFETs, a few resistors, and a couple zener diodes to provide reverse polarity and over-voltage protection. The relay bypass board at the bottom has tap for latching, hold for momentary bypass/engage functionality.
There are a few hidden components. Tayda does not stock a 2.5mm thick 6.8nF Wima cap, so I went with a 2.2nF Wima capacitor with a 4.7nF gray box cap in parallel on the opposite side of the board. There is also a 100nF ceramic bypass capacitor on the back of the board under the op amp. It's best practice to place the bypass capacitor as close to the V+ pin as possible, and in this case the best location was directly under the chip.
I made a couple minor changes to the schematic. First, I used a nominally 390pF silver mica that measured as 440pF for C7. The CA3130 data sheet says that a 47pF capacitor is usually sufficient here, and it worked on the breadboard, so I felt fine making the change. Second, I replaced the 100pF Miller capacitor in C10 with a 53p silver mica. The 1T320V transistors have a warmer sound than the other transistors (GT308V, P416B, 2N2043) that I tried, so lowering the value of C10 helped prevent the tone from getting too dark and wooly. I could have used ceramic for both of the caps, but silver mica looked cooler and fit the vibe.
I tested a few other changes on the breadboard, but they did not make the cut. I tested increasing the Q1 collector resistor from 10kΩ to 33kΩ to match the Fuzz Face, but the bias stability suffered and the transient response/picking sensitivity was underwhelming.
Let's see, what else? The 1T320V's that I used for Q1 and Q2 have hFEs of 82 and 114, respectively, and less than 20µA leakage at 70ºF. I tried other gain buckets and all of them sounded good. That said, Q1 should not have an hFE above ~100, or the bias stability gets wonky.
Per Chuck's recommendation, I set the trim pot so that with the external Bias pot at 12:00, Q2's collector voltage is at 1/2 VCC, i.e. 4.5v with a 9v power supply. With the bias control turned fully counter clockwise, the tone gets spitty, synthy, and gated; when turned fully clockwise, the tone is smooth and compressed but not quite gated.
I tested the pedal at temperatures as high as 120ºF and as low as 50ºF. I didn't bother testing at temperatures above or below that because who wants to play in those conditions? Not me. Regardless of temperature, Q2's collector voltage quickly stabilized at the same voltage, ±0.005v. Pretty impressive!
One quirk of this pedal is that due to the auto-bias feature it needs to be powered on for ~15 seconds before the bias current starts to flow and the servo stabilizes it.
Obligatory gutshot:
I went with Tayda's metallic candy red enclosure, which I had UV printed by Tayda with a double pass on the white layer and matte gloss finish. I had Tayda do the drilling, too. The outie DC jack has flattened sides, so I used the drill shape option for its hole to hold it more securely. The resulting hole was a tiny bit too small for the jack, but I was able to expand it to fit the jack using a small curved file.
So having said the most important part, on to the build. I went slightly premium on this one—Wima, Nichicon, and silver mica caps and mil-spec 1T320V transistors. I have a sizable stash of low leakage, mid-to-high gain Soviet transistors with BCE pinouts and thought that they would be good fits for this build, so I laid out the board with in-line BCE footprints rather than the usual TO-5 footprint. The I/O board uses P-channel MOSFETs, a few resistors, and a couple zener diodes to provide reverse polarity and over-voltage protection. The relay bypass board at the bottom has tap for latching, hold for momentary bypass/engage functionality.
There are a few hidden components. Tayda does not stock a 2.5mm thick 6.8nF Wima cap, so I went with a 2.2nF Wima capacitor with a 4.7nF gray box cap in parallel on the opposite side of the board. There is also a 100nF ceramic bypass capacitor on the back of the board under the op amp. It's best practice to place the bypass capacitor as close to the V+ pin as possible, and in this case the best location was directly under the chip.
I made a couple minor changes to the schematic. First, I used a nominally 390pF silver mica that measured as 440pF for C7. The CA3130 data sheet says that a 47pF capacitor is usually sufficient here, and it worked on the breadboard, so I felt fine making the change. Second, I replaced the 100pF Miller capacitor in C10 with a 53p silver mica. The 1T320V transistors have a warmer sound than the other transistors (GT308V, P416B, 2N2043) that I tried, so lowering the value of C10 helped prevent the tone from getting too dark and wooly. I could have used ceramic for both of the caps, but silver mica looked cooler and fit the vibe.
I tested a few other changes on the breadboard, but they did not make the cut. I tested increasing the Q1 collector resistor from 10kΩ to 33kΩ to match the Fuzz Face, but the bias stability suffered and the transient response/picking sensitivity was underwhelming.
Let's see, what else? The 1T320V's that I used for Q1 and Q2 have hFEs of 82 and 114, respectively, and less than 20µA leakage at 70ºF. I tried other gain buckets and all of them sounded good. That said, Q1 should not have an hFE above ~100, or the bias stability gets wonky.
Per Chuck's recommendation, I set the trim pot so that with the external Bias pot at 12:00, Q2's collector voltage is at 1/2 VCC, i.e. 4.5v with a 9v power supply. With the bias control turned fully counter clockwise, the tone gets spitty, synthy, and gated; when turned fully clockwise, the tone is smooth and compressed but not quite gated.
I tested the pedal at temperatures as high as 120ºF and as low as 50ºF. I didn't bother testing at temperatures above or below that because who wants to play in those conditions? Not me. Regardless of temperature, Q2's collector voltage quickly stabilized at the same voltage, ±0.005v. Pretty impressive!
One quirk of this pedal is that due to the auto-bias feature it needs to be powered on for ~15 seconds before the bias current starts to flow and the servo stabilizes it.
Obligatory gutshot:
I went with Tayda's metallic candy red enclosure, which I had UV printed by Tayda with a double pass on the white layer and matte gloss finish. I had Tayda do the drilling, too. The outie DC jack has flattened sides, so I used the drill shape option for its hole to hold it more securely. The resulting hole was a tiny bit too small for the jack, but I was able to expand it to fit the jack using a small curved file.
