Monday, September 16, 2013

messing with the MFOS variable state filter -- Sagittronics -- Charlie Slick

I will start this post with a modular confession.  I have yet to utilize, on any filter, the ability for it to act as a sign wave oscillator.  I mean, like in a performance, I mostly like to use filters as filters and oscillators as oscillators.  That being said, I've been questing for a nice sounding, versatile, simple filter and I don't care so much about 1v/oct tracking.

One of the really amazing things about the MFOS website is that Ray sometimes has the same circuit in 3 different stages of complexity.  As a person trying to learn as much as I can about circuits, this is an unmatched resource!  I like to look at the early versions of circuits and compare them to the later more complex circuits, to try and understand why things were changed.

Earlier this year, maybe as earlier as February, I built the 12db state variable filter from Ray's "Oldies but goodies" section, on a breadboard.  I liked the filter but the filter had some problems.  The first problem was that the filter didn't self-oscillate. The second problem is a bit harder to explain.  It seemed that when the CV input voltage and Cutoff voltage combine to more than about half a volt, the filter would self-oscillate at a full 30v spread.  I likened these problems to the fact that I was using +/-15v rather than +/-12 and could be solved by some small resistor tweaking.  I took video of my oscilloscope to demonstrate the problem.

it happens at about 10 seconds in, then it happens a bunch more times.

I was unable to solve the problem then, so I decided to shelve the project for a later date.

After 8 or so months, I figured I'd learned enough to attempt to correct this problem.  I assembled the circuit on a breadboard and ran into the same problems.  No self-oscillation and whatever you call what is happening in this video.

I toiled over it for a whole day, trying to set a limit to incoming CV but it never truly felt like I was attacking the problem at the source.  Finally, right as I was about to give up for the day, I decided to change R27 and R28 to 100k and the problem in the above video went away.  I'm not sure how this change affected the topography of the filter but I did notice that the bandpass output got quieter which I can't really explain why.

At this time, I decided to try and implement the darlington transistor pair inside the LM13700 that Ray mentions trying to integrate on the "oldies but goodies" page.  I guess I wanted to simplify the circuit even more and I was in a general electronics exploration mood.  He presents a circuit in the MAKE: Analog Synthesizers book that unitizes the transistors in this way, so I had a nice guide to work from.

I was able to to do it, but not without some trouble.  I now had the opposite problem, when the incoming CV and Cutoff voltage was so low that it turned of the Darlington transistors, causing the outcoming voltage to swing all the way to -15v. I solved this with the 2m2 resistors connected to ground, they are intended to keep the transistors on.  I'm sure this affects the 1v/oct tracking but again, I can't even get the filter to self oscillate.

In the end, I'm not even really sure what I have here.  I ended up with a similar circuit to the one in the MAKE: Analog Synthesizers book but with a Hi-Pass Filter and slightly more precise CV input.

I'm not happy enough with the circuit to build a module, and I'm so burned out on it, I can't tell if it even sounds good anymore.  I guess I will shelve it again for a while and let the ideas stew.  I did benefit from toying with the filter in my attempts to tame/simplify the design.  Comment with any suggestions.  I guess I'm moving on the the YuSynth Steiner filter next, or perhaps this MS20 design I found.

No comments:

Post a Comment