There’s been some Rat discussion over on the Mail Day Thread, thought maybe it should have its own topic. I’ve been thinking about doing a Rat module but it won’t happen right away, so I figured I’d write up a little of what I’ve thought of in case others want to chime in or use it. I haven’t done any breadboarding so this is all theoretical or based on a simulation.
Base design: I’ve been working off this schematic:
Levels: I’d want to make it synth level compatible by reducing the level on input and boosting it on output, using more or less the Ken Stone design.
One thing puzzling me about that is the input is divided down using a 22k:1k divider. Why not more like 100k:4.7k for a ~100k input impedance? Are there guitar pedals with low enough impedance that you need the smaller resistors? If so, that doesn’t apply here, so the larger resistors I think would be preferable.
For the output level, I did a simulation and it looks like the level after the JFET is about 1 Vpp. If that’s right, then Stone’s ~20x boost is too large. His 470R should become 1k for a ~10x boost.
Stone uses one bypass capacitor between the two rails, I’d change that to two, from each rail to ground.
I don’t like output attenuators, so I’d drop that pot.
Power: We can use a 9 V regulator (L78L09, T92), presumably in its datasheet configuration with 330 nF and 10 nF capacitors, and get rid of everything left of R11 (protection components and bypass caps, not needed here). I’m not sure how necessary C13 is (or what it’s for) so I’d leave it in. In my simulation I tried changing the 9 V to 12 V and the outputs seemed to change very little, so maybe that’s a simplification that can be made.
Redundancies: With the input voltage divider, I see no need to keep R1. Nor do we need Stone’s 1M and 10 uF after the Rat output.
Op amp: Rat mythology notwithstanding, there’s very little difference if you substitute an OP07 ($0.49 at Tayda) for the LM308 (unobtainium and widely faked). (See this video.) Of course that’s up to the builder, the circuit design isn’t affected.
Diodes: People like to experiment with different diodes and maybe it’s worth throwing in a switch to select between different diode options. I’ve been thinking of a DPDT to select either 1N4148 or LEDs. The LEDs could be on the front panel but they don’t have to be; it’d complicate things to have them light up while not affecting the distortion, doing it the simple way means they would not light up if the 1N4148s are selected. Of course the output level would be maybe 4x higher with LEDs than with 1N4148s, meaning the 10x output boost would be too large. So you’d have to either provide a way to change the gain of the output stage (no, not just an attenuator after it!), or put a voltage divider after the LEDs that gets switched in and out along with them. I prefer the latter. With an on-off-on switch you could also select no diodes, but there again you’d have to reduce the level compared to the 1N4148 path, and that gets more complicated to do other than by just manually reducing the output stage gain. Or you could go nuts with a rotary switch, selecting 1N4148, various color LEDs, or no diodes, each with suitable voltage divider.
Panel size: I’d favor 25 mm, with a perpendicular PCB using horizontal board mount pots, Molex to connect the switches and jacks. Of course such a board could be mounted behind a 50 mm panel too, if that’s your preference.
