Voltage controlled resistance...

[iampoor]

Hi all

Currently looking to build an LFO driven filter. I like the bootstrapped Twin T filter, an plan on using R.G.'s design. I already have the lfo designed, and ready to go.



Now comes the fun part, the part that I have no experience with. 
How do we get an effective, and affordable voltage controlled resistor?
Ideally I would like to be able to control the resonance and frequency, but I am not sure how to accomplish this.

Below is a short summary of what I (think) I know about the only 2 feasible options. Any input or direction would be welcome, as I am quite ignorant on this subject!

Options: Led/LDR, or optocoupler
I like this option, but from my understanding, matching photocells can be a pain. I dont see many datasheets with particularly consistent performance curves. Maybe an NSL-32 would be a good option?

Fet: Limited cv range, high distortion. How is the consistency? Is it possible to get reasonably repeatable performance?

Digital pots:
Too complex for this application.

Could anyone point me in the right irection? I really want to learn about this subject.

[garcho]

looking to build an LFO driven filter

IMO, the LM13700 OTA is definitely the way to go in terms of easy voltage controlled filters. it does have an application as a voltage controlled resistor itself but why mess with that when you can build a decent VCF out of it anyway?

check out these as maybe some of the easiest, click on them for the link

[idy]

just buy generic cds cells, enough to test for some consistency, and green leds, stick them together in parallel to a moderate size resistors.
Worth a try.

[R.G.]

Currently looking to build an LFO driven filter. I like the bootstrapped Twin T filter, an plan on using R.G.'s design. I already have the lfo designed, and ready to go.

That's a problem. Until you know that the controlled element needs, having an LFO that needs extensive adaptation to make it work the controlled element correctly may be very difficult, perhaps as complicated as the LFO itself.

The current-controlled devices, like the OTAs, make state variable controlled filters easier, but the need driven from a current source, not a voltage source, and the voltage at the control node is about one or two diode drops. Makes keeping a variable current harder. There are ways around this, too, but you're quickly getting to a complex circuit.

How do we get an effective, and affordable voltage controlled resistor?

I've been chasing that one off and on since at least the mid 1970s. No good answers yet.

Ideally I would like to be able to control the resonance and frequency, but I am not sure how to accomplish this.

\
For best technical performance you control all three resistors in the T in synchronism for frequency, and the amount of signal into that lower opamp for Q.

Options: Led/LDR, or optocoupler
I like this option, but from my understanding, matching photocells can be a pain. I dont see many datasheets with particularly consistent performance curves. Maybe an NSL-32 would be a good option?

You're correct. Photocells vary, and getting the matched is a problem. The best I ever came up with was to use a center tapped photocell with the centertap grounded, and drive the LED with a signal from an opamp that used the resistance of one half as feedback. You can be reasonably certain that the other half of the photocell will track modestly well. But it's complicated.

Fet: Limited cv range, high distortion. How is the consistency? Is it possible to get reasonably repeatable performance?

Correct on the problems. Consistency is poor from FET to FET. Repeatable performance is a problem. No good solutions yet.

Digital pots: Too complex for this application.

Too bad. They're controllable and easy once you get over the uC hump.

The only other technology that shows promise is one of duty-cycled resistors or switched capacitors, both of which run on duty-cycling a fixed component. This can at least be done with an analog PWM modulator. But this itself is more complex that what you're controlling, and keeping switching noise out of the audio path in a small box is difficult.

You've run into one of those problems where it seems like there ought to be a simple, easy, cheap solution if you know the trick; but sadly there isn't.

[garcho]

one small possible bonus is that twin T filters can sound good in a musical way, with asymmetrical values. the Q gets all goofy and it might not give you the swept filter sound you're looking for though, so, blah. but it's worth breadboarding the thing and messing with values and using only one LDR and the other a variable resistor. sometimes you can still get something useable out of it, despite the hamfisting.

Makes keeping a variable current harder. There are ways around this, too, but you're quickly getting to a complex circuit.

good point. but, again, you might get musical results from something as simple as a resistor in series with an LDR, between a pot wired to V_CC and V_EE, and I_ABC. then again, you might not. i love LFOs!!! hard to resist, the LFO sings quite the siren song. it's the perfect addition to almost any audio signal processor, except it adds a bunch of knobs and ticks and hick ups and real estate goes way up, etc.

[amptramp]

You may find a multiplying DAC offers something you can't get otherwise - control of resistance through digital inputs but the voltage swing may be limited.  Check out the last picture in this spec sheet / application note:

http://www.ti.com/lit/ds/symlink/dac0808.pdf

The reference voltage is the input and the digital input is the attenuation.  You can do something similar with the DAC1208 but you have 12 bits to play with.  I am not sure how easy it would be to modify the circuit so you can use a voltage in / current out device as a filter element, but stranger things have happened.  A multiplying DAC has an R-2R ladder network inside that acts as a digital potentiometer but with restrictions on how it can be used.  The LFO drive can be an up/down counter.

[~arph]

Thanks for that app note. Looks like we can use that DAC as a multiplier for a ringmod too (albeit with a stepped control waveform). I will try this. I have some of those DACs laying around. See how it sounds compared to a true sine wave modulated tone.

[iampoor]

looking to build an LFO driven filter

IMO, the LM13700 OTA is definitely the way to go in terms of easy voltage controlled filters. it does have an application as a voltage controlled resistor itself but why mess with that when you can build a decent VCF out of it anyway?

check out these as maybe some of the easiest, click on them for the link

Good idea! Building some voltage controlled resistors out of it would probably be almost as many parts!
The 100k resistor in the top schematic functions as a voltage to current converter, correct? Would this have to be buffered if driven straight from an LFO? In the second schematic, there appears to be a current source that is biased at the base by a CV, correct?

just buy generic cds cells, enough to test for some consistency, and green leds, stick them together in parallel to a moderate size resistors.
Worth a try.

I tried a couple last night, wasnt too happy with the results. I wish it worked altho~! 

Currently looking to build an LFO driven filter. I like the bootstrapped Twin T filter, an plan on using R.G.'s design. I already have the lfo designed, and ready to go.

That's a problem. Until you know that the controlled element needs, having an LFO that needs extensive adaptation to make it work the controlled element correctly may be very difficult, perhaps as complicated as the LFO itself.

The current-controlled devices, like the OTAs, make state variable controlled filters easier, but the need driven from a current source, not a voltage source, and the voltage at the control node is about one or two diode drops. Makes keeping a variable current harder. There are ways around this, too, but you're quickly getting to a complex circuit.

Thankfully its not a special design for this application, its just one that I already have on hand that I like. Modifications wont be an issue, I just know what I would like. Say I need the LFO to an LM13700, dont I have a few options? Would just a simple resistor (like the 100k one above) function as an effective voltage to current converter? If not, couldnt I use a current mirror and drive that with an external voltage? And if that didnt work, couldnt I use a transconductance amplifier to function as a voltage to current converter?
Or is my logic being deceptive? 

I have been pondering this all night, and I could see this becoming an issue. I would like the LFO to modulate the frequency of the filter, not solely control it. I could use a dc coupled mixer with a rail-to-rail opamp to combine the two into a solid control signal, right?

How do we get an effective, and affordable voltage controlled resistor?

I've been chasing that one off and on since at least the mid 1970s. No good answers yet.

The traditional method is to include a trimmer to adjust for gain variations in the transistor on the collector, right? Seems like in many applications FET's create more issues than they solve. I have read pretty extensively about using them to switch signals, and even that is quite a challenge! Lets not even get started on FET based limiters. 

Ideally I would like to be able to control the resonance and frequency, but I am not sure how to accomplish this.

\
For best technical performance you control all three resistors in the T in synchronism for frequency, and the amount of signal into that lower opamp for Q.

What general range of frequencies (ie 1:10, 1:20 1:30 etc)would I be able to get if I controlled all 3 resistors versus just a pair or 1?

Options: Led/LDR, or optocoupler
I like this option, but from my understanding, matching photocells can be a pain. I dont see many datasheets with particularly consistent performance curves. Maybe an NSL-32 would be a good option?

You're correct. Photocells vary, and getting the matched is a problem. The best I ever came up with was to use a center tapped photocell with the centertap grounded, and drive the LED with a signal from an opamp that used the resistance of one half as feedback. You can be reasonably certain that the other half of the photocell will track modestly well. But it's complicated.

Very cool idea! Do you have a schematic? I want to make sure I understand how the feedback is working. Otherwise Ill draw up how I think this works, and then ask for a critique!

Fet: Limited cv range, high distortion. How is the consistency? Is it possible to get reasonably repeatable performance?

Correct on the problems. Consistency is poor from FET to FET. Repeatable performance is a problem. No good solutions yet.

Digital pots: Too complex for this application.

Too bad. They're controllable and easy once you get over the uC hump.

The only other technology that shows promise is one of duty-cycled resistors or switched capacitors, both of which run on duty-cycling a fixed component. This can at least be done with an analog PWM modulator. But this itself is more complex that what you're controlling, and keeping switching noise out of the audio path in a small box is difficult.

You've run into one of those problems where it seems like there ought to be a simple, easy, cheap solution if you know the trick; but sadly there isn't.

Im trying to refrain from digital pots since I have alot of circuity to fit, and I havent finished my C programming class yet. Plus I figured I should know how to do it both ways! In the future, I plan on using digital pots, and seeing what weird and wonderful filters can be created!

The PWM idea is interesting too. I think the super old PYE recording studio compressors used PWM to modulate the attack/release constants across some capacitors.

Well, if there is one thing Im glad to know, its that all the research an googling Ive done over the past few days hasnt turned up empty!

one small possible bonus is that twin T filters can sound good in a musical way, with asymmetrical values. the Q gets all goofy and it might not give you the swept filter sound you're looking for though, so, blah. but it's worth breadboarding the thing and messing with values and using only one LDR and the other a variable resistor. sometimes you can still get something useable out of it, despite the hamfisting.

Makes keeping a variable current harder. There are ways around this, too, but you're quickly getting to a complex circuit.

good point. but, again, you might get musical results from something as simple as a resistor in series with an LDR, between a pot wired to V_CC and V_EE, and I_ABC. then again, you might not. i love LFOs!!! hard to resist, the LFO sings quite the siren song. it's the perfect addition to almost any audio signal processor, except it adds a bunch of knobs and ticks and hick ups and real estate goes way up, etc.

Good idea. I will have a play with this. I dont think thats the solution for this particular effect, but I love playing with filters and seeing what "organic" results can come from it. If I used digital pots it would be easy to recall particular settings. Might be a good application.

LFO's are addictive. SO many creative possibilities. I always find myself brainstorming new ideas with LFO's. I think there is still quite abit of new musical ground that can be covered with them. Especially in the guitar world. We are still years behind those eurorack synthesizer guys. 

You may find a multiplying DAC offers something you can't get otherwise - control of resistance through digital inputs but the voltage swing may be limited.  Check out the last picture in this spec sheet / application note:

http://www.ti.com/lit/ds/symlink/dac0808.pdf

The reference voltage is the input and the digital input is the attenuation.  You can do something similar with the DAC1208 but you have 12 bits to play with.  I am not sure how easy it would be to modify the circuit so you can use a voltage in / current out device as a filter element, but stranger things have happened.  A multiplying DAC has an R-2R ladder network inside that acts as a digital potentiometer but with restrictions on how it can be used.  The LFO drive can be an up/down counter.

Very cool. I will be studying the datasheet, and thinking of some applications. Not sure if this would be the right fit for this application, but I could see this being useful! I wonder if a free running 8 bit counter could provide "randomized" data to it. That might be interesting in a glitchy way!

[garcho]

We are still years behind those eurorack synthesizer guys.

big time!

have you checked out Electric Druid's TAPLFO? It's a great little PWM PIC style LFO with the basic wave forms and ability to distort them. also with tap-tempo, not something i care so much for, but it makes it quite simple to add external sync, and choose how it's sync'ed (the tempo-taps, or 5V pulses, can control the LFO tempo at 1:1, 1:3, etc.). great for playing with hi-tech bandmates.
also, Aron and Andrew have the taptation PIC for sale in the store, i haven't used it yet. it looks geared towards PT2399 projects but it looks like it does everything the taplfo does, too.

[amptramp]

One DAC application I have seen is a DAC with the reference input being the signal input and the output (since it is already a current) going to the inverting input of an op amp which has a capacitor from output to inverting input.  This is a lowpass filter with the digital inputs selecting the cutoff frequency.

[iampoor]

We are still years behind those eurorack synthesizer guys.

big time!

have you checked out Electric Druid's TAPLFO? It's a great little PWM PIC style LFO with the basic wave forms and ability to distort them. also with tap-tempo, not something i care so much for, but it makes it quite simple to add external sync, and choose how it's sync'ed (the tempo-taps, or 5V pulses, can control the LFO tempo at 1:1, 1:3, etc.). great for playing with hi-tech bandmates.
also, Aron and Andrew have the taptation PIC for sale in the store, i haven't used it yet. it looks geared towards PT2399 projects but it looks like it does everything the taplfo does, too.

Thats actually the LFO Im planning on using! I love love love the TapLFO!

I think the taptation controller is targeted at delays like you mentioned. Havent had a chance to use that chip yet.

One DAC application I have seen is a DAC with the reference input being the signal input and the output (since it is already a current) going to the inverting input of an op amp which has a capacitor from output to inverting input.  This is a lowpass filter with the digital inputs selecting the cutoff frequency.

Love it. I actually have another application where that would be quite a handy feature. DO you know of any resources or papers for learning more about general audio applications for DAC's?

[amptramp]

Your wish is my command:

http://www.ti.com/lit/an/snaa097/snaa097.pdf

http://cds.linear.com/docs/en/application-note/an87f.pdf

http://www.analog.com/media/en/technical-documentation/application-notes/AN-1094.pdf

http://users.ece.gatech.edu/lanterma/sdiy/datasheets/opamp/AN-299.pdf

You might like the digitally programmable bandpass filter in the last app note.

[iampoor]

Your wish is my command:

http://www.ti.com/lit/an/snaa097/snaa097.pdf

http://cds.linear.com/docs/en/application-note/an87f.pdf

http://www.analog.com/media/en/technical-documentation/application-notes/AN-1094.pdf

http://users.ece.gatech.edu/lanterma/sdiy/datasheets/opamp/AN-299.pdf

You might like the digitally programmable bandpass filter in the last app note.

Thank you very much. Will be reading through these. On first glance, I am rather fond of the bandpass filter. Might be good for a programmable crossover idea Ive had!

[Blitz Krieg]

maybe worth looking at:

http://www.diystompboxes.com/smfforum/index.php?topic=107197.0

http://www.diystompboxes.com/smfforum/index.php?topic=88582.0

[anotherjim]

Korg Monotron and their much earlier Korg35 filters used bjt's in the T, but their normal form (MS20) is OTA.

Also see the Monotron schematic, but that's set up for 5V.