Im wondering which resistors (around opamp or LM13700) to change in order to get zero resonance when resonance potentiometer is fully closed (ccw). It now shows a little spike as if resonance is dailed in a tiny bit. Any advice highly appreciated.
How have you set R47 ? This looks like it’s a trimmer to adjust the resonance level? Usually you would be focused on making sure the self oscillation is not too extreme. Otherwise you could try reducing R52.
R48 and R52 have values so close to 50k that together with R47 they must be meant to divide the 24 V range in 3 parts. -12V to -6V, -6V to 6V and 6V to 12V. The pot R47 will allow to set a voltage between -6V and 6V. But 49.9 k is a resistor that is hard to come by and if you use 51k or 47k you will achieve almost the same result. And those resistors are more easily available. My 2 cents.
Especially since pots typically have tolerance ±20%, which easily nullifies any difference between 49.9k and 47k.
Thanks for confirming my initial suspicions that those resistor values are weirdly spec’d.
Hi all, thanks all for your replies. Ive changed R52 for a 50k trimmer and no matter how i set it, the reso peak is still there. Let me first upload some pictures ive taken, and to completion also sheet 1 of the MFOS filter. There is no explicit explanation regarding trimming the amount of resonance but. I see comments that the BP is feedback into opamp that feeds the resonance. I’m still unsure which resistors to change. I can ofcourse take a wild guess. 150k resistors over pin 1 (output pin) and the inverting input?? Well. Hope this helps you to help me trimming down this peak of reso. Again thanks in advance.
3 Likes
Hi Jos. Thanks for your reply. Ive tried replacing them both but indeed with the same results. Uhm, where to look next in order to get zero resonance when the pot is fully ccw? Thanks in advance for your advice
Try changing R54 to a 43K or 47k resistor, that should force IABC of U6-A to a lower value and further reduce the feedback/resonance when R47 is at minimum.
EDIT: My reasoning, assuming XRES is left unconnected (more on that later).
With R54=54k, the voltage at the non-inverting input of U5-B is 30/130x-12V=-2.8V
When R47 is at minimum, the inverting input of U5-A is at -6V and its output is (0 - -6)x120/100 = +7.2V
The output of U5-B is then (-2.8 - 7.2)x100/100= -10V
I’m not completely sure how to compute IABC, but I would estimate it at approximately 60µA which is low but maybe not enough to completely kill the resonance.
Changing R54 to 43k, would set the non-inverting input of U5-B to 43/143x-12V=-3.6 and then the output of U5-B would theoretically go to -10.8V, reducing IABC a bit.
With R54 at 47k, U5-output would theoreticaly go to -11V further reducing IABC.
I say theoretically because even with the original value of R54=30K, the output of U5-B is already pretty close to its limit, as it typically doesn’t get less than 1.5V away from its power rails.
So changing R54 might get its output closer to -12V, but it’ll never get there and there will always be a non-zero IABC and thus some resonance.
If the filter was designed by ear rather than with an oscilloscope, it is possible that what the author meant by “no resonance” was actually what you are seeing on your scope.
Can you actually hear that level of resonance? (It does look strong enough to be heard.)
Another point to notice is that the filter is probably not meant to have “no resonance” when the XRES input is not connected, it is only when XRES sees -5V that the filter is meant to provide no resonance. The voltage at XRES is summed with the voltage provided by R47. So R47 is probably more meant as an adjustment to the maximum resonance rather than the minimum. To barely keep the filter from oscillating at maximum resonance.
If you really want zero resonance, you might have to replace U5 by a Rail-toRail opamp.
3 Likes
Hi Antoine. Wow! What a great explanation and thanks for the effort. I really appreciate it. As soon as I get home, I’ll try I’ll changing R54 and let you know the results.
I have another VCF (Moritz Klein) based on two vactrols and when I turn down the resonance knob, the resonance is completely off as it should be imho. I like to get a clean square wave if the resonance is turned down, but since the MFOS filter is another design with another components which behaves differently, I can imagine that this could happen. It is not a very big problem, but its just my autistic mind that says, if no resonance is dailed in, no resonance should be visible or audible 
anyways, again many thanks and have a nice day
1 Like
R48 and R52 are part of a voltage divider, so the little that I know tells me that changing them alone should make no real difference. I think that @antoine.pasde2 might be on to something with the rail to rail op amp suggestion. Assuming that you only need the op amp to swing down to the negative rail, you might be able to use the lowly and cheap LM358. Might worth the try as this is a 10c common part.
1 Like
I have to mention here your feat of fitting this VCF on a 14x34 piece of stripboard, including the standoffs. That’s a hell of an elegant layout!
2 Likes
Hi K. Ostas. Thanks. Yeah i try to make it as tiny as possible leaving just enough space to make repairs if nessesary. Panel is 20cm long and the width depends on how many pots and jacks.