Re: MS20 filter questions

Started by snk, June 28, 2023, 12:54:59 PM

Previous topic - Next topic

snk

#20
Hi,
I changed the resistors and capacitors (470pF / 51K), and the leds as well (one green, one blue, both 5mm, instead of 3mm red).

I must have done something wrong, as now the wet sound is quieter than the dry sound, and rather weird... It also looks weird on an analyzer, with the bottom half of the wave shaven (like an half wave rectifier, which makes me think I should look into the leds wiring)...
Let's go back investigating!



garcho

#21
I can't tell you why Escobedo chose to use two LEDs instead of another resistor, but the LEDs are there to make a reference voltage. If changing them changes the sound, it's because of the DC the AC signal is "riding on" and how the internal circuitry of the OTA responds to that. It's not the "sound" of the diodes. You could use two resistors if you wanted, or use a trim and dial it in until it sounds right to your ear.

Quoteboth 5mm, instead of 3mm

LEDs have different forward voltages, different current requirements, but 3mm vs 5mm really doesn't matter.

A lot of the distortion/overdrive/fuzz pedal threads here are people throwing random diodes in and discovering magical "mods". This isn't one of those circuits. The whole magic of OTA filters is the IC and the control current for IABC.
  • SUPPORTER
"...and weird on top!"

ElectricDruid

Quote from: garcho on June 30, 2023, 12:25:13 PM
A lot of the distortion/overdrive/fuzz pedal threads here are people throwing random diodes in and discovering magical "mods". This isn't one of those circuits. The whole magic of OTA filters is the IC and the control current for IABC.

+1 agree.

The Tim Stinchcombe paper discusses the difference between the 'Early' and 'Late' versions of the MS20 filter. One significant difference is where they place the clipping diodes. The Late version with the OTAs (like the filter we're discussing here) has them in the feedback loop, which is a fairly popular trick. You see it a lot on state variable filters for synth use, and the SH101 filter uses it too, with the Roland IR3109 quad OTA chip. I love that SH101 filter, so fat and fruity. That's why it's a shame that this stripped-down 9V version has taken them out.

The other big contributing factor in the sound of an OTA filter is the input level. Given that OTAs have a differential pair of transistors as the input stage, they have a very nice soft-clipping characteristic as you turn up the input level. The distortion doesn't all appear at once, but increases gradually. A big part of the "sound" of most OTA filters is a few percent of distortion that you don't really hear as distortion, especially after it's been through a filter. Gives them character and a bit of grit.