I really liked the sound of the very popular darkglass pedals, like the vintage and the b3k, but looking at the schematic it felt wasteful of components and with some weird values maybe dictated by smd. So i embarked on a journey to make my own workalike with the goal of keeping parts count and board size down, while trying to keep the same sound and staging. Doing so, the circuit went down from using a quad opamp, a dual opamp, an hex inverter and two jfets to just a dual opamp, an hex inverter and a bjt as active components. Here's the vintage inspired one:
Let's go through it: the mu amp stage has been replaced by an inverter with roughly the same gain, frequency response and headroom. This has been checked in simulation with the inverter being an op amp with an open loop gain of 30, which matches the documented value at this voltage. Here it's critical to consider this stage together with the loading of the hall notch filter after it, as considering them separately leads to very different results and it was an error I made in the earlier versions, leading to an harsher sound. The following two stages and bridged T are basically the same, just using E6 values without compromises. The two op amp sallen key filters have been replaced by inverter multiple feedback filters with the same response, and the output buffer is a bjt to round the op amp count to two.
I'm proud of having used all six inverter sections, altough the current draw reflects this being at 50-60mA. The first gain stage uses three of them in parallel to try and lower the noise of this critical stage. Speaking of which...
I really like this pedal and how it sounds, but the noise is higher than I'd like. Not too bothersome but noticeable. This is definitely due to the first inverter gain stage I've concluded by testing and measuring, and it's no news inverters are noisy. V1.3 improved on this with the different frequency response and lower gain in this stage. The LC filter on the inverter was another attempt at keeping the noise down, altough I don't know how effective.
Here's a vero for it. It's not pretty beecause it went through many revisions, and the back jumpers are adventurous, but it works if you want to use it. There's no inductor because i tacked it on the back later.
I've also drawn a B3K version of it since it's just some small changes.
Finally, here's an attempt at further streamlining the circuit by getting rid of the inverters altogether. The dry blend buffer is in parallel so it doesn't add noise to the effected path, the input stage is the highest gain it can do cleanly, the hall notch has become a T notch with a very similar response, and the opamp and cmos clipping stages have been rolled into one with roughly the same gain that uses the principle of diode ladders to get very soft clipping. No big changes in the second half. I've tested and it's promising, altough the clipping is different and I don't know how much quieter it is, It should be though.
demo of V1
demo of V1.3
thanks again to aotmr for their support.
- i played around with cmos many years ago - even built a complete guitar amplifier with one hex inverter 4069 chip. it worked well but was just too noisy, no matter what voltage I ran it off.
yeah that's my experience too, fun to work with but noisy no matter what.