Building an MS-20 filter almost identical to the LMNC one using the provided schematic. testing it by playing music from my phone into headphones. I don’t own a scope currently.
the output is muddy and distorted, and at lower volumes only lets low frequencies thru with a low crackle. the cutoff technically works but not for the distortion.
the resonance screams, way below the expected threshold, is very very loud, but does not self-oscillate with no audio which is strange. it chirped once while doing a sweep but that’s all.
differences between mine and schematic:
the cutoff pot is a 10k pot due to lack of 4.7k or 5k.
using 6 1n4148 instead of green LEDs. tried yellow first. resonance was worse with yellow
using 2n3906 instead of bc558. tried first with bc557. sounds better with 3906.
other things to note:
shoddy serial 1.8k resistors.
tried to learn some schematic to breadboard practices but gleaned from most people that it’s more of a vibe than exact science
let me know if I can provide anything else, thanks!
Try and connect an amplifier via a capacitor and a cable to various stages of the breadboard build and listen whether the signal you hear corresponds to what you expect. Use a 10muF capacitor or something similar and connect the - to the amp and probe the circuit with the + side (as long as the DC voltage you expect at that point is positive otherwise reverse the capacitor). Instead of an elco it is more convenient to use a non-polar capacitor if you have one of 10muF or so.
The audio from all opamp outputs you should be able to interpret depending on the functon of the subcircuit the opamp is in.
Yes it does, lacking a Greek character mu on my keyboard or the knowledge to somehow imput it. You may also encounter uF as micro Farad.
The output of the LM is a current output so there may not be a noticeble voltage change there (unless you add a resistor of say 1M Ohm or so to GND to the ‘probe’) but the opamp following it converts that current into a voltage, so that’s output you could probe.
I finally got around to testing this a little bit. Apart from the fact that 50% of the issue was pots not connecting in the breadboard, i’m getting some interesting results. I made a few videos in a Google Drive folder marked with each stage to supplement my description here in case it’s helpful. At this point, I may just try again but wanted to follow up since this was an interesting experiment, and your advice was very helpful.
Output A - Resonance: It’s the very mid-heavy resonance (does not go high-pitched as expected), with very very faint sound of the oscillator (square out of an 0-coast). I suspect that the resonance and my lack of a 4.7k pot for cutoff are the largest overall issues. Using a 10k and 2.2k interchangeably.
Output B - LPF: With resonance at 0, this is a pretty nice sounding low pass filter. With the resonance all the way up, the sound cuts out.
Output C - Main output: It sounds like a very choked filter, resonance or not. Almost in bandpass territory, though I know that’s not what it’s doing (my kingdom for a scope!). Basically the same sound as when the main output (jack on the other side of the output capacitor).
Output D - CV output: Strange behavior, but I assume that’s because the CV section doesn’t handle audio(?)
Posting this just in case anyone says “ahh, that sound is a telltale sign that your op-amp is backwards!” or something revelatory like that.
I think you can add an Attenuator pot on the input
(i do this on all my MS20 to fix that problem, and it’s also very usefull to decrease like you want the signal in, to increase the Resonance sound)
Great idea, thank you. The first night of making it, I was using music through the aux cable on my phone and found that the square wave input I was using today sounded far better.
This remedy worked for me. I had the same problem so I searched the forum and found this advice from @Dud in several threads. I put a 10k pot in the circuit to attenuate the incoming signal and it cleared the problem right out.
A bit late saying this but if you have a 10k input resistor I think a 100k pot would behave somewhat wonkily. Aaron’s 10k pot should be a better fit.
Then again I’m kind of surprised the circuit uses 10k; that’s a much lower input impedance than the typical 100k. And it seems to me 220R is too large to be used with 10k. As discussed here you want the input voltage to the LM13700 to be about 60 mVpp. (They’re designing a VCA but the same applies for a VCF.) Here with a typical 10 Vpp input it’s 10 * (220/(10k+220)) = 215 mV. No wonder it distorts! In the ED article they use 100k (10 times as large) with 620R (only 3 times as large): 10 * (620/(100k+620)) = 62 mV.
Then again again this supposedly is the actual MS-20 filter schematic, and it uses 10k and 220R too. Though I don’t know if Korg used a 10 Vpp signal; it might have been smaller.
Barton’s vaguely similar Sallen-Key Filter uses 100k with an even lower 470R, giving an OTA input around 47 mVpp.
