I got this PCB in a trade with soggybag over at Super Freq. This is his implementation of Tim Escobedo's "PWM". It includes the optional LFO. It's a square wave synth with modulated pulse width. In a track it could be mistaken for an old analog synth.



I tried a spray-can finish for the first time in a long time. Historically, I've had a hard time getting the paint to adhere. I'm optimistic that the few tricks I've learned over the years will help. Aesthetically, I wanted it to look like a Lisa Frank fever dream about fuzz.



You could fit all the components in a 1590A. The extra room is for all the attitude.



Of course, I can't leave well enough alone, so I made some mods.

• I swapped lugs 1 and 3 of the Rate control so that clockwise was faster.
• I jumpered the LFO switch and opted to put a big-ass 10mm bypass LED in its place. Setting Depth to zero still kills the modulation.
• I changed C2 to 6.8nF and added a 47k resistor between lug 3 of the PWM (labeled "Width" on my build) and ground. Without the minimum resistance, it's possible to set the pulse width so narrow that the signal cuts out entirely.
• I changed C2 to 10uF and R2 to 4.7k. This widens the range of the LFO.

All and all this is a fun circuit and unlike anything else I've built!

Nice build! I've tinkered with a number of Tim's circuits, but never got to this one. Time to give it a go!

I wonder if you could add envelope control by disconnecting pin 3 of the depth pot and connecting it to the output (pin 5) of the LM386?

Quote from: moosapotamus on March 19, 2025, 11:33:34 AMNice build! I've tinkered with a number of Tim's circuits, but never got to this one. Time to give it a go!

I wonder if you could add envelope control by disconnecting pin 3 of the depth pot and connecting it to the output (pin 5) of the LM386?

Thanks!

I'm tinkering with envelope control on the breadboard right now. I haven't made much progress yet. I was planning on making a deluxe version of my own comparator fuzz which uses a 4558, but depending on how the experiments go, it may end up looking more like the PWM.

Initial experiments suggest that the 40106 needs a large input signal to overcome the hysteresis/input threshold. This could make it less susceptible to string noise. You could go really simple and hang a diode off of pin 5 and connect an LDR in place of the PWM knob, but I don't think that would give you much dynamic control.

Of course, if I ever get something working, there will be a build report!

Sounds promising.

Quote from: Aleph Null on March 19, 2025, 12:12:19 PMInitial experiments suggest that the 40106 needs a large input signal to overcome the hysteresis/input threshold. This could make it less susceptible to string noise. You could go really simple and hang a diode off of pin 5 and connect an LDR in place of the PWM knob, but I don't think that would give you much dynamic control.

What if the LDR went in parallel with the PWM knob, instead of in it's place? The 386 can definitely drive an LED. I guess you're making me think of something along the lines of how it's done in the Clari(not).
https://www.diystompboxes.com/smfforum/index.php?topic=74294.0

At any rate, you certainly seem to be more knowledgeable than myself. So, please forgive my ignorance. Back to my first thought of switching from the LFO to the 386 output for an envelope response, well... I guess I don't really understand how that Schmitt trigger LFO works. The output of a Schmitt trigger is basically a square wave, yes/no? Then it gets smoothed out by R1/C3? If so, other than being a periodic function, how is the LFO signal that reaches pin3 of the 40106 different from the output of the 386?

Sorry if I'm being an ignorant pest. I do have my own breadboard. So, feel free to just tell me to go use it! Thx! 

Quote from: moosapotamus on March 19, 2025, 04:06:04 PMWhat if the LDR went in parallel with the PWM knob, instead of in it's place? The 386 can definitely drive an LED. I guess you're making me think of something along the lines of how it's done in the Clari(not).
https://www.diystompboxes.com/smfforum/index.php?topic=74294.0

The 386 will definitely drive an LED! I think you should be able to place the LDR either in series or parallel to the PWM pot. I'm sure they would interact differently, but I haven't thought about it enough to have any insight into what the difference would be. (I'll probably just try both when the time comes.)

Quote from: moosapotamus on March 19, 2025, 04:06:04 PMAt any rate, you certainly seem to be more knowledgeable than myself. So, please forgive my ignorance. Back to my first thought of switching from the LFO to the 386 output for an envelope response, well... I guess I don't really understand how that Schmitt trigger LFO works. The output of a Schmitt trigger is basically a square wave, yes/no? Then it gets smoothed out by R1/C3? If so, other than being a periodic function, how is the LFO signal that reaches pin3 of the 40106 different from the output of the 386?

There's also a lot that I don't know! I believe the output of the LFO (What lug 3 of the Depth knob sees) is closer to a triangle wave and R1 and C3 are smoothing out the action of the 2N3904, which, in turn is acting as a variable resistor. The resistance of the 2N3904 and the PWM knob form a low-pass filter with C4. That low pass filter (via a mechanism I don't understand) is setting the pulse width of the square wave.

My first strategy for envelope control is to forego the LED and just use a BJT as a variable resistor. I'll probably have to partially rectify the output signal—something like what's happening in an Orange Squeezer, but with a BJT instead of a FET.

Time to get breadboarding!

Quote from: Aleph Null on March 20, 2025, 11:45:41 AMI believe the output of the LFO (What lug 3 of the Depth knob sees) is closer to a triangle wave and R1 and C3 are smoothing out the action of the 2N3904, which, in turn is acting as a variable resistor. The resistance of the 2N3904 and the PWM knob form a low-pass filter with C4. That low pass filter (via a mechanism I don't understand) is setting the pulse width of the square wave.

Thanks for that! Sounds good to me! My simple thought was that the 2N3904 is providing some variable voltage, from the LFO. And, if the output of the 386 could be coaxed into that same range, it would generate an envelope response that could be switched between or maybe combined with the LFO. I don't know what that range needs to be, so I guess I just need to try it on the breadboard, too, and see what works, or doesn't. Thx!