4069 State Variable Filter (no VCF)

synthmonger

I'm trying to build a 4069 that is just a state variable filter like in the wasp schematic minus the CA3080's and cv control.

Any help would be much appreciated!

JovianPyx · Posted: Tue Oct 21, 2008 7:09 am Post subject:

I've worked with 4069 linear circuits before - what questions do you have?

I haven't built the Wasp, but I am familiar with it.

Are you trying to un-voltage control it (so that it tunes with only a pot)? IIRC, the Wasp used OTAs to allow voltage control over frequency while the 4069 "gates" are used as soft clipping distortion generating buffers.

You could possibly replace the OTA and cap combination with a pot and cap, but the pot would have to be dual ganged for this to work in such a circuit.

Personally, I wouldn't remove the voltage control part of the circuit. If you buld it as is, you can use a single simple pot to generate the CV. You may decide after building it that it would be fun to be able to modulate the flter electronically. Then you will want/need the CV input.

synthmonger · Posted: Tue Oct 21, 2008 9:35 am Post subject:

JovianPyx · Posted: Tue Oct 21, 2008 10:32 am Post subject:

I wouldn't expect the performance to be identical since an OTA is really a voltage controlled current source. A resistor is a crappy current source and the 4069UB output isn't very "stiff" so it will perform poorly with low values of the timing resistor. Assuming the resistors are variable resistors, they should be dual ganged so they track and they should be fairly accurately matched. The more mismatch, the more deviation of the poles' positions from ideal there will be.

You might be able to improve the performance by multiplying the size of the resistor by a factor N and dividing the capacitor by N. This would not change the frequency range, but it would change the impedance of the system. The inputs of the 4069UB are very high impedance, so this should be possible. Doing this should result in the resistance being the major player instead of the driver's output impedance.

synthmonger · Posted: Tue Oct 21, 2008 12:04 pm Post subject:

I'm using two of my VCORs (voltage controlled oscillator resistor) that are dual ganged together. I tried it with CGS BPF, the A-100 Steiner Vactrol Filter, with great results. I didn't notice any change in sound (other than the added control options of the VCOR). I can tailor the VCORs to be matched and pretty darn stable too. They have a max size of 500k per switch...and I just realized I could run them in series to up the resistance. Duh!

The other reason why I want to use a 4069 as the filter is that it can run off a single supply. I tried a couple standard filters with my VCOR's and they sounded great on a bi-polar supply and kinda iffy and at times unstable on a single supply with voltage referencing...Though I'm probably just being too picky. Mad

Incase you didn't see my mini guitar synth post you can find my VCOR schematic there.

loss1234 · Posted: Tue Oct 21, 2008 12:14 pm Post subject:

wow
this sounds cool

may i ask why you are ditching the resonance control?
it hadnt hit me that your vco's were actually meant to be used in OTHER circuits as VCR's...that is weird!

how do they perform verus vactrols or an fet?

how is the guitar synth going?

synthmonger · Posted: Tue Oct 21, 2008 12:20 pm Post subject:

ohhhhhhhhhhhhh! I think I know why it isn't working properly. I completely forgot to add the necessary caps. I won't be able to test it out until later tonight after work Crying or Very sad

Now it's down to figureing out the cap values.

JovianPyx · Posted: Tue Oct 21, 2008 12:30 pm Post subject:

I didn't realize you wanted to do single supply.

Look at the Wasp schematic, - see the odd biasing at the OTA inputs? That's because it's running from a single supply. Not sure how you would emulate that for your circuit other than creating a virtual gound with one or more 4069UB gates. It's easy enough to do, put 1, 2 or 3 gates in parallel (the more, the stiffer, try one first) and run a 100 K resistor from output back to input. Connect nothing else to the input and your output should be Vdd / 2. Then use this point as a ground for input and output signals instead of the power supply ground rail.

If you need your signal and supply grounds to be the same, you might try using a resistor to run the Vdd / 2 reference voltage into the 4069UB integrators at the CMOS input. A guess at the size of the resistance would be to use the same as the feedback resistor. This will add a DC voltage to the AC input potentially canceling the offset caused by using a single supply.

synthmonger · Posted: Tue Oct 21, 2008 12:37 pm Post subject:

synthmonger · Posted: Tue Oct 21, 2008 12:47 pm Post subject:

JovianPyx · Posted: Tue Oct 21, 2008 2:14 pm Post subject:

synthmonger · Posted: Tue Oct 21, 2008 6:42 pm Post subject:

V/2ing the input helped amplify the signal to the normal level, but for some reason I'm not getting a frequency change even with a dual ganged pot instead of the VCORs.

My prototype is laid out exactly like Fonik's except with no OTA's. I just replaced the input and output of the OTA with the dual ganged VCORs. I biased the gates to the same bias the OTA's were once connected too. This seems like how it should be setup for an SVF right?

JovianPyx · Posted: Tue Oct 21, 2008 9:40 pm Post subject:

You have the right structure. Perhaps a wiring error somewhere? Are you sure the VCORs are doing what they are supposed to? It sounds like the resistance isn't changing if Fc won't move.

synthmonger · Posted: Wed Oct 22, 2008 2:05 am Post subject:

I'm getting a very slight frequency control. I do have something wired wrong though, I need to test it out on a new breadboard the one I'm using it too old and crummy. Ah well I'm too tired now to do that. Tomorrow will be the day! ;D

synthmonger · Posted: Wed Oct 22, 2008 11:29 am Post subject:

Do you think the 51k resistors in this schematic http://modular.fonik.de/pdf/Wasp%20Filter%20clone.pdf
should be removed?

I'm getting some frequency change but it doesn't sound like a wasp. Sounds more like a ladybug or something Very Happy

Hrm I think I have the other two V/2 sections in the wrong place. Right now they're after the resistors on the outputs of pin 2 & 4. Should they be on pins 3 & 5?

JovianPyx · Posted: Wed Oct 22, 2008 12:26 pm Post subject:

This is where I would start, not sure what you actually have - do you have a drawing of what the circuit currently looks like?

Resistors R ?? It may also require that Vref is not Vdd/2, but rather some function of the voltage or current driving the VCOR.

The more I look at this, the more I think it _should_ work without the Vref resistors altogether.

Yeah, the place that needs the Vref bias is only the input to the filter. Everywhere else, it's not needed. I'd try it with no bias anywhere first.

SVF.GIF

Description:

Filesize:

13.98 KB

Viewed:

981 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

synthmonger · Posted: Wed Oct 22, 2008 12:38 pm Post subject:

After my last post that's exactly how I pictured it should look in my head.

I'll give this a shot when I get home and if all goes dandy expect audio samples!

synthmonger · Posted: Thu Oct 23, 2008 12:10 am Post subject:

Aha! I finally got it somewhat working! The resonance isn't working all the way but the frequency sweep is mighty fine sounding! Next post will have samples.

synthmonger · Posted: Thu Oct 23, 2008 11:31 am Post subject:

Oh fishsticks! It looks like I made some sort of smooth sounding filter (does sound neat but no resonance control = boring). I'm stumped. I'm going to rebuild the whole thing using OTA's then swap them out with resistors then swap resistors out with the vcovrs.

This thing is a lil' whacky. I was able to get it to work without the vdd/2 somewhat, self-oscillate, and work with vdd/2. The chips work fine so I can't blame it on that.

JovianPyx · Posted: Thu Oct 23, 2008 1:11 pm Post subject:

Here's what I would do:

First, ignore the diode mess for now, it can be added later. First you need to see if you can control resonance with a VCOR. Not one - two in series seems about right. I would replace all of that stuff from pin 1 of U1 to pin 4 of U1. then Run 2 VCORs in series between pin 1 and pin 4 instead. I'd start there. You should get max resonance when they are at max resistance.

ADDED:
The diodes perform a distortion function. This requires impedance characteristics that are present within this circuit. In order to replace the 47K pot in this circuit, you'd have to leave the rest of it there. It might be possible to remove the 47 k pot and replace it with a 50 K resistor. Then across the 50K resistor, you would put two VCORs in series. The center connection is the pointer on the pot. Make sure the control to the two VCORs is opposite. That might work.

synthmonger · Posted: Fri Oct 24, 2008 6:12 am Post subject:

Been up all night so this will be a quick post. I tried a different wasp-like schematic and everything worked right off the bat. I realized that the resistors I were using were 1k and not 100k like I thought, which I can only assume was the reason for it failing.

Anywhos, I haven't setup the resonance under vcovr, but the frequency control is working mighty fine. I used one of the extra 40106 gates to sync the vcovrs which gave a really neat robotic vocal filter sound -kinda digital but in a really cool sounding way. I'm curious how it the vcovrs will perform with this circuit under exponential control and not linear. I won't be able to test that out under after mid november -on a tight schedule. Shocked

One thing I'm surprised with my 4069 vcf self oscillates. I thought they weren't supposed to?

JovianPyx · Posted: Fri Oct 24, 2008 6:16 am Post subject:

synthmonger · Posted: Sat Oct 25, 2008 3:49 am Post subject:

synthmonger · Posted: Sat Nov 01, 2008 7:55 am Post subject:

Here's a short clip showing the scope of the 4069 vcf under sync control

http://www.youtube.com/watch?v=RFMUEMy4KI8

JovianPyx · Posted: Sat Nov 01, 2008 8:09 am Post subject:

Can you tell us more about the conditions of this setup? I'm unsure as to how synch applies to a filter?? I get synching an oscillator, but not a filter.

Also, can you post your final schematic? (assuming this is the final design)...

Thanks for the demo. Interesting, wierd sounds.