MOStache Overdrive [documentation]
It's a fairly straight forward amp-in-a-box type of deal, but here I'm using P-Mosfets (hence the name) instead of Jfets because they are much more consistent device-to-device thus easier to bias and, in my experience, a bit quieter than NMOS. While I was inverting a 9V supply for P-type devices I thought it would be fun to double the voltage too, and add a capacitance multiplier to keep supply noise down. I designed it on the board with a BC516 PNP Darlington, but I made by own 'integrated' Sziklai pair for the final assembly. Also, I managed to fit the entire power section on a sub board where the battery usually sits (I do not like batteries), leaving more space for the main board and also providing a bit of isolation. Note for the vero layout, I used an ICL7660S instead of an LT1054 (since I was running low, and something about using a CMOS chip seemed thematically appropriate)
I was roughly aiming to replicate a Bassman/JTM45 type of tone, and however close it got (or didn't) I'm very happy with how it sounds.
The first toggle switch changes the gain of the first Common-Source stage, somewhat akin to swapping out a 12AX7/ECC83 for a 12AT7/ECC81 or 12AU7/ECC82. The second toggle controls the bright cap, which I have placed parallel to a fixed resistor into a pot, to get a more useful amount of treble bleed across the whole gain sweep.
Next is another CS gain stage feeding the tone control. I designed this to work like a simplified FMV stack roughly as I like to set them; Bass around 3, Mids around 7 and Treble set to taste depending on the guitar/pedals I am using. Simulated responses at 0%, 50% and 100% below:
The output of the tone control then hits the final stage, a Mosfet Long-Tailed Pair with a drain-to-drain diode network, to simulate phase inverter and (some) power valve clipping. The resistor values were chosen with a bit of trial and error to get both drains to within about 100 mV of each other at idle. The diodes and snubber cap were added after I had done the layout for everything else and run out of room on the main board. I've used a whole bunch of BAT41's and a 1k series resistor to get the clipping very soft.
This is followed by the volume control and an output buffer/2nd order active lowpass to help smooth out the higher treble range. The 560pF capacitor was chosen to account for the roughly 120pF of gate-drain capacitance (bringing it to ~680pF in total) and the whole thing has a cutoff at about 6-7kHz with a small amount of interaction from the volume control. The volume has a stopper since I only had 10k pots with the right taper, didn't want to load down the previous stage too much, and wasn't particularly worried about output level.
Just for fun here are my veroboard layouts. I'm actually quite happy I managed to squeeze the whole thing into a 1590B. The lonesome sockets are voltage testing points. I recommend setting the idle drain voltage of Q1 (and thus Q2) to about 2/3 of the supply voltage, and the LTP drain voltages about halfway between the supply and the source voltages.
Yes. It's on the schem and the layout but I probably could have made it clearer.
I don't have a fantastic recording setup at the moment but I can try to get something
recorded if there is interest.
also thank you, that is almost exactly what I designed this for, to be the last drive in my chain before mod/time effects into a clean amp, and to play nice with other boosts, drives or even fuzzes before it.