VCF Chips?

[Michael Krusch]

Does anyone knows of a voltage controled filter chip for building a variable lowcut filter (12dB/O or more)? Audio quality is not an issue.

 

[Michael Krusch]

Thanks! I thought of something more cheap. It's not for a synthesizer, but for a compressor sidechain.

 

[peterc]

What about a switched capacitor filter? I know that these have been used for DAC filters before.

http://www.national.com/pf/MF/MF10.html

Regards
Peter

 

[bcarso]

Following Peter, I would use a switched-cap filter with a VCO for setting the filter frequnecy (if voltage control is what you really want).

Linear Tech used to have some nice parts---probably still in production. The cutoff freq was typically 1/100 of the clock, iirc.

 

[Michael Krusch]

I think, switched-cap filter is a little bit to much. I thought of a little chip (1$) doing this, like the ones used in stereo's for high and low boost.
But googling for vcf is hard because of that synthesizers around.

 

[bcarso]

Well---what is your desired freq range? I don't know that you will find a single chip, but a state-variable lp topology can be made with one dual OTA* like an NE5517 and a few (well, several) more parts. Basically each OTA with a C load on the current output becomes an integrator, and you need two integrators and a summer for the filter. Noise and distortion will not be exemplary but it will be pretty cheap at least.

Actually I think there is a way to do it with just the OTA, taking advantage of the ability to make a positive-gain integrator as well as the more typical negative-gain one.

Is the control ultimately a voltage (or current)?


*OTA = operational transconductance amplifier, for those not familiar with the term: differential voltage in, bipolar single-ended current out, with the diff voltage-to-current conversion gain or so-called transconductance set by a control current. Typical commercial chips also have darlington emitter followers on board that can be used as unity-gain buffers.

 

[Michael Krusch]

Control is just a pot, but for multichannel application you need to control more sidechains with one pot.

Here I found a schematic of that filter (Figure 9 on page 11) http://moon.feld.cvut.cz/hw/philips/acrobat/5017.pdf

 

[bcarso]

Yeah that's the baby---I had forgotten that was part of the app notes.

There is a way to use the linearizing diodes and a couple more parts including a couple of voltage inverters, to make the distortion much lower, but I don't know what your needs are there either.

 

[adamasd]

There is a way to use the linearizing diodes and a couple more parts including a couple of voltage inverters, to make the distortion much lower, but I don't know what your needs are there either.

If it is the one i am thinking of it just uses the other OTA and Buffer in the IC. It does do a good job, but from teh audio standpoint I thought it sounded better running without that or the linearizeing diodes. Though I have not played with this setup in quite awhile, and that was aimed towards modular synth, so the distortion was not an issue. I have the schematic for it, and I think an article that is pretty in depth on what is going on with the circuit, let me know if you want it and I will dig it out.

adam

 

[bcarso]

This circuit already needs both OTA's of course.

I've developed an approach with the linearizing diodes that works a lot better* than the one recommended in the datasheets for these parts. However, it does require differential drive, so that adds parts. If National et al. had managed to bring out the collectors of the darlingtons independently then it would be a snap, but that would have made it an 18 pin package.

The approach also makes the gain a lot more stable and less dependent on the particular sample.


*Better meaning less colored---but that color may be desirable in some cases so NB.

 

[adamasd]

I forgot that he was going for 12dB rolloff, but adding parts is adding parts I guess. The linearizing setup in the data sheet is kinda useless, I am not sure how much of an improvement it is supposed to be, but it really is not very much. It always seemed like trying to get 1 more dB out of an amp so you can get to that nice round 50dB. I never have tried the data sheets approach to it in anything but VCAs though, and they seem to only use it when the 5517 is used as a voltage controled resistor, maybe it has a more noticable effect in that application

adam

 

[Michael Krusch]

Is it me, or does the high pass filter look the same as the low pass filter?

 

[bcarso]

Hah. That's funny. I guess what they neglected to show for the single-pole examples is that with the HP they want you to inject the signal at the bottom of the cap, and ground it when it is the LP. (Referring here to fig. 8 and fig. 9)

The classic three or four op amp Kerwin Huelsman SV filter has three or four outputs, low pass, bandpass, highpass, and notch. So those will look the same and it justy depends on what output you use. The abbreviated versions with positive and negative gain integrators have two outputs, BP and LP. I'm not sure if you can get a BP and HP out of two integrators or not.

 

[StephenGiles]

Maybe you need CA3094 OTAs, unfortunately now discontinued but there are a few still about. Electroharmonix did a lot with them.
Stephen

 

[maneco]

i think a multimode filter around an lm13600 will do the trick...www.smallbearelec.com sells an equivalent for something like .90
In synth sites you'll find many filters based around this chip... :thumb:

 

[bcarso]

And 13600/700 = NE5517 btw