Variable filter methods

Thread Starter

Veracohr

Joined Jan 3, 2011
772
I'm working on designing an audio line mixer (well, it's been on the shelf for a few months but I'd like to pick it up again) and I'd like to put in variable 4-pole high pass filters that would be controllable from a single knob to two filters (a stereo signal).

After I learned a bit about MOSFETs in a class last term, I tried simulating a filter with a MOSFET as a voltage controlled resistor, but I couldn't get it to work at a low enough frequency due to the limited resistance range. Since this is for audio applications, I'd like a filter that can go at least down to 30Hz, possibly as low as 20Hz. Upper limit isn't as critical, probably somewhere around 500Hz. I thought if I replaced the resistors in an active filter design with MOSFETs, and controlled those with a single source, it would accomplish it.

What other options would I have for varying two 4-pole filters with one control, without a lot of deviation in frequency between the two? I could go with an OTA-based filter, but that requires one OTA for each stage. Plus, a lot of the OTA filter info I've seen is kind of confusing. Any other common methods that could maybe use fewer components?
 

Audioguru

Joined Dec 20, 2007
11,248
A Mosfet of one part number has a wide range of spec's so matching two or four to make an adjustable frequency filter is almost impossible.
All IC tone controls filters have matched transistors. Most are single pole, a few are two poles but none are four poles.
 

Thread Starter

Veracohr

Joined Jan 3, 2011
772
So no suggestions? I'd already ruled out the MOSFET idea.

The only other idea is cascading 4 stages like this one from the NE5517 datasheet, with Vc of all 4 fed from one source:


 

Thread Starter

Veracohr

Joined Jan 3, 2011
772
I was under the impression switched-capacitor filters weren't generally used in audio applications because of noise.
 

Audioguru

Joined Dec 20, 2007
11,248
My very low distortion oscillator uses a switched capacitor filter IC and the noise level is not measureable. The clock is 100 times the cutoff frequency so a simple single RC removes its already low level.
 
Top