FML – 4 operator FM percussion drone module

While doing research for the oscillator I stumbled –  seems to be a common theme — upon these IC’s. The ICL8038 promises to be a stable oscillator with very few external components. It’s an obsolete product, but I could get a few of them via Ali-express cheap enough to experiment with. Since it offers a sine output I immediately thought about doing some FM type percussion. That way I wouldn’t need 1/v per octave and pitch fluctuations wouldn’t be that much of a problem either. I also had a V2164 quad VCA chip around, so a 4 operators (aka oscillators) seemed a perfect fit. Enough to provide some interesting FM timbres at least.

FML prototype build

While planning on how to go about it I decided each oscillator should have it’s own decay setting, pitch control and while I’m at it, decay to pitch. Naturally a volume control and a switch to route either oscillator to the main output or to modulate another oscillator. That way one should be able to come up with all possible routing combinations.

Spending a few evening with them on a breadboard made the dream of cheap simple and stable oscillators crumble fast enough. They proved to be simple enough to set up and get going, but getting a wide enough range out of them to cover the whole audio spectrum is a whole other thing. And stable.. well not really, but that could be due to the prototype soldering. Apparently somehow they just ‘give up’ when you move them out of their comfort zone, which seems to be defined by the timing capacitor. You probably can coax a better range out

of them then I did if you actually know what you’re doing, but the best I got was from around 10Hz to around 5kHz. Which, although not ideal, should do the trick.

I had no clue how it would sound in the end, so I started of with 2 voices and see whether the concept worked. It became apparent that having a VCA on the voices modulating the source wasn’t always desirable as you’ll always end up with a ‘sine’ tail on the sound. So I added a way to bypass it by pulling the CV input on the V2164 to ground. Simple and effective, probably not a good engineering practice though.

First real test

First tests with 4 voices has proven to be very hopeful. Still a bit rough, but nothing that couldn’t be easily addressed. A surprising range of sounds could be coaxed out of the thing. Even on it’s own it was capable of generating wobbling rhythmic patterns just being provided with a steady pulse. And the VCA bypass can turn it in a rather unique sounding drone machine. The oscillators ‘cutting’ out on the edges of their range added some unintended chaos and unpredictability to the mix and I’m quite torn between trying to address the issue or just leave it be. Using something like a V3340 as an oscillator would be far better and more reliable across the frequency spectrum, but also a lot more expensive. Maybe in the future.

The limited frequency range really did bother me a bit though, because it ment the frequency content of the output was also limited. Therefore I added a switch per oscillator to change the source waveform, since the IC’s also provided, next to a sine, a triangle and square wave. This mitigated the issue a bit since both these waveforms have more harmonics then the sine, which — in an ideal world – has none.

What I hadn’t thought of when I came up with this idea was the sheer amount of pots and switches needed. 15 pots, 11 switches is quite a lot and therefore need a lot of room on the front panel. So it’s either going to be a huge eurorack module or I’ll house it separately.

Sound examples

FYI There is some interference from my soldering iron in some of the recording and recording has been done, less then ideally, on a very very old macbook which still services me gratefully as an audio scope/analyser/recorder/player/… . Also, it’s only the sine wave, the other waveforms were not yet done.


But enough of the bantering, some schematics. As always, these schematics are provided ‘as is’. You should expect some errors to be present as I drew it up from notes I’ve taken while building. I might have forgotten something.




  • 4 oscillators
  • Sine, triangle, square waves on each oscillator
  • external modulation inputs per oscillator (not 1/V oct)
  • independent decay and VCA bypass
  • decay to pitch modulation
  • A lof of different routing possibilities (19, I’ve counted them, yes, sometimes I do stupid things like that)


03/04/2018: Added some pics of the first build and updated the schematic to fix an error in the decay to pitch pot (thanks to user Smrl on Muffwiggler for pointing that out)