I’m started a new “small” project where I will have a dual SVF filter. Almost as the dual SVF for Shruthi. with the exception that it will be able to handle 2 different inputs in serial or parallel. As an example: Filter 1 is fed with osc1+osc2 and is set to HP, Filter 2 is fed with osc3+output from filter 1 and set to BP
I started with looking at the MI filters, but for a prototype (there is more to the project) I don’t want to base it on V2164, so i went for a 13700-solution.
I googled a bit and found: http://www.electronics-tutorials.ws/filter/state-variable-filter.html. I modified it to have OTA-C integrators and replaced the 1K input-resistor to 10K to get 0db DC gain. The problem I have now is to: a) get some decent resonance out of it (that is 0-> self oscillating) b) design the VC resonance circuitry.
With 1K inputresistor I get a nice resonance “bump”, but on the other hand the filter have a ~5,6dB DC gain, with 10k I get 0dB DC Gain and no resonance
So… Where do I start?
I’m pretty sure there are already many schematics out there for LM13700-based SVFs.
One trick to get SVF to self-oscillate is to have a tiny bit of negative feedback from the BP output to the input (This compensates for the small phase shifts introduced by the op-amps).
Also: keep in mind that the TL07x can phase reverse. Your circuit will crash if the input level is too high and/or if the resonance gets out of hand. It’s safer to use the op-amp after the OTAs as inverting integrators (with the capacitor in the feedback loop) and rearrange the circuit to take into account the sign change.
Can this sort of stuff be modeled in Spice or similar? as a precursor to using a breadboard.
@pichenettes. Actually i had the op-amps as inverting with the cap in the feedbackloop. I never got it working at the time. Then I looked at MFOS SVF (http://www.musicfromouterspace.com/index.php?MAINTAB=SYNTHDIY&PROJARG=STATEVARVCFFEB2006/STATEVARVCFFEB2006.html&VPW=1670&VPH=797)
and went for a “stripped down” version. And combined it with the electronics-tutorials.ws solution for the first op-amp (instead of having a inverting or non inverting amp there as usually seen).
I’ll do some fiddling around to see if I can get it to work this time.
@6581punk Yepp. The schema in my first post is from TINA-spice
I feel its easier to check if something works as intended and to change component-values in spice. And if I do a stupid mistake I get an error instead of a puff of magic smoke
Now I actually got it to work with the caps in the feedbackloop (no Idea why it didn’t work last time)
I reconfigured the first Op-amp to inverting instead of that “summer adder” (as the tutorial called it)
The problem is that with a 10 Hz sinewave with 1V p-p, LP and BP outputs becomes square-waves with ~10V p-p in simulation.
HP is a inverted Sine at (as expected?) But with p-p 1V and spikes correlating to the BP/LP squarewave edges.
With the “summer adder” there is a nice 1V p-p sine on the LP output as expected
Obviously something isn’t right. The ideal would be to breadbord it and check with a scope… if I had one here at home (its on the to-buy list)
OK. Problems in the post above solved. Now I just have to figure out a decent solution for VC resonance.
The MFOS SVF seems to be a bit “overcomplicated” with extra opamps and voltagedividers all over the place.
At the moment I’m experimenting with some kind of VCA solution for the BP-feedback. All form of bright ideas is welcome
Q1 is the BP-output and Q2 goes to the input of the filter