4 pole Multimode VCF Rev 2 (og3)

I've always liked the CEM3320 filter chip, which consists of 4 OTA stages, an exponential generator, and a VCA to control resonance feedback. You can get the different fourth order filter responses (HP, LP, BP) depending on how you interconnect the OTA stages. I hit on the idea of using analog switches to change the stage interconnections, and therefore build a circuit which could switch between the multiple filter responses. The CEM3320 Multimode Filter is my first design using this idea.

I wanted to build another circuit that uses NO CEM CHIPS, and which has all three filter responses available. This circuit uses easily obtainable LM13700 OTA's for the transconductance stages, and MAX413 analog switches.

• Block Diagram. Shows how to connect the transcaonductance stages to get the different filter responses.
• Schematics
• Bill of Materials

Results

The lowpass and highpass modes work excellent. I can get a low distortion self oscillating sine wave from the lowpass, and a somewhat distorted since out of the highpass.

Rev 1

"The bandpass mode however doesn't quite respond to increased resonance as the CEM version does. I can't get self oscillation out of it, and it doesn't seem that the response gets sharper as you increase resonance, but the amplitude gets smaller. It may be that the BP configuration, which is a cascaded 2 pole highpass then 2 pole lowpass, needs to have the stages matched better, as they would be on the CEM chip."

Rev 2

Grant Richter (of Wiard) pointed out that I needed to invert the resonance feedback signal to get it to work right. I just stuck in an opamp amp inverter, and that definitely fixed the problem with the resonance response. However, it exposed another problem which is inherent in this type of bandpass - the gain goes up as the feedback does, so that means at high resonance you go right into clipping.

So I went through a lot of gyrations, but came up with this scheme: I made the resonance feedback voltage controlled, and also put a VCA on the input. The resonance and input VCA's have their control voltages inverted; when the the resonsance increases, the input VCA gain decreases thereby decreasing the input signal. The result is that you get the sharpening of the band with increasing resonance, but the amplitude stays fairly constant.