Voltage to resistance?

[Taylor]

I'm working on some pitch-to-voltage stuff (yes, I know, very tough to do well, filtering, fundamental extraction, compression, etc.) and was wondering how I might then convert my voltage to a scalable resistance, to control parameters that use that as a control.

Would an optocoupler setup be the simplest way to do this?

[R.G.]

It would sure be nice if there was something that did that. 

Actually, there are things that do that, but the trouble is that they are complicated, non-repeatable, nonlinear, and have small usable ranges.

JFETs are actually kind of voltage controlled resistors. But getting them to be repeatable is as bad as making JFET amplifiers repeatable.

OTAs can be used to make a voltage controlled resistor, but they're complicated, needing two OTAs and supporting circuitry to do it right.

About the best performance is complicated: voltage controlled current source pumping current into the LED of an LED/ center tapped LDR. One half of the LDR is used as a resistance to ground, and feedback from this half is used to linearize the curves of voltage to current to light to resistance. The other half of the LDR is used as the controlled resistance.

I think I first started looking for a good, simple voltage controlled resistor in about 1971. Still looking.

[brett]

Hi
MOSFETs can be useful, especially if combined with some negative feedback to help linearise the response.  Unlike JFETs, which are all over the place (-1 to -6V for common SS types), many MOSFETs go low resistance at 2.5 to 3V. ie. a much narrower range (and positive voltage, which may simplify the circuit)
cheers

[solderman]

I'm working on some pitch-to-voltage stuff (yes, I know, very tough to do well, filtering, fundamental extraction, compression, etc.) and was wondering how I might then convert my voltage to a scalable resistance, to control parameters that use that as a control.

Would an optocoupler setup be the simplest way to do this?

Hi
I (think) I had the same sort of problem and asked about it here a while ago. My need was a way to control the resistance in one component from an array of many. I tried bread boarding LDR/LED combos and JFETS but nothing worked. (The LED/LDR worked but was to unprecice. The resistance was all over the place and too slow)  until R:G so kindly pointed out that I could thing out of the box and use real live pot:s in stead of "components wanebees" Unfortunately you have to scale your self and that might be out of the question for your need I guess.

The setup is to use a transistor as a switch and switch in the pot /controllable resistance) where and when it's needed. I used a 4017 to control "when" and connected to control the potential between signal and GDN. I guess it can be used to control "whatewer" and be triggered by some circuit that sense the guitar signal etc.

For what its worth here is the thread

http://www.diystompboxes.com/smfforum/index.php?topic=77564.msg636949#msg636949

[edvard]

I agree with most of the opinions here: too slow, too tricky, range too small, etc.
I'd say don't give up hope though, because it has been done, therefore it can be done.

Look in your supply catalogs for "Voltage Controlled Resistor". They do exist, and they serve a purpose.
Also, look into the world of digital potentiometers (Hey! Put away the spikes and garlic, it's ok...) I've heard they are getting better every year.

[e45tg4t3]

look over at  Squarewaveparade.com, there´ll you find a bunch of voltage to resistance circuits...

Benny

[WhenBoredomPeaks]

I had the same problem:

I tried the LDR, LED combo, it kinda works for me because i can control it with an Arduino board. (you can write code for the useful voltage ranges and waveforms, etc.) But it is not so reliable ad very unpractical. (it is just a breadboarded version so i have to cover it with boxes, it takes up a lot of space, etc.)

My other choice looked very good at first glance but i failed to make it work. I never gonna give up altough.

I tried this: Digital pots and Arduino. I found it's failure paranormal because it is not too complex, but i never made it work.
I tried it with a Arduino Diecimila and an AD5204 then with an AD8403 digital pot.

[valdiorn]

The world needs better Voltage controlled resistors! There are very few possibilities out there that are actually any good. Why won't someone design a chip like that!!?
I'm doing a 3-pole Low pass and high pass filter circuit for my amp with controllable cutoff. I rolled my own vactrols from a diode and Light-dependent resistor. Works pretty good, has a natural logarithmic scale (linear current to diode give log scale for resistance). I used an op-amp and trimpots to adjust the range of voltage the diode gets. It seems to work pretty well.

[R.G.]

Ooops! Sorry, I forgot to mention the complex, but good solution - pulse width modulated resistors, either using a switched resistor or a switched capacitor.

If you put a fast switch in series with a resistor and then open and close the switch with a PWM signal, the apparent resistance goes up as the PWM duty cycle goes down. If the switch is on 100% of the time, the resistance appears to whatever its actual value is. If the PWM duty cycle is 50%, then the resistor is conducting half the time, off half the time, and only half the charge goes through. The resistor appears to be twice as big.

Likewise, if you have a cap to ground with a switch on one side and a switch on the other side, opening and closing the switches alternately lets in a cap-ful of charge when you close the input switch; closing the output switch and opening the input switch lets the cap-full of charge out the output side. In this case the apparent resistance through the switches is linearly proportional to the frequency at which you operate the switches. This switched-capacitor variable resistance is what almost all clock-driven active filters use.

The disadvantage to both of these is that you have to have a clocking signal made somehow, and that is complicated; also, they are sampled-data systems inherently, and therefore subject to the Nyquist criterion that you get aliasing unless your input signal contains no frequencies over 1/2 the clock rate.

[Taylor]

Wow, I didn't realize how deep a topic this is. I couldn't find any purpose built voltage controlled resistors at Mouser or Digikey or with a Google search - is this an actual part that exists? Seems like the sort of thing where somebody makes a chip that has 4 of them in a 14-pin DIP. Guess not.

I'll first try some optocoupler ideas, and see if that works for me.

The application is this: I wanted to play around with having pitch control of, say, tremolo speed, or delay time in a PT2399.

[R.G.]

Yeah, we all get to there. After playing with Light Dependent Resistors, the idea for a Voltage Controlled Resistor hits us pretty quick.

There *are* components sold as voltage controlled resistors. See

http://www.specteclogistics.com/docs/70598/70598.pdf

for some discussion of VCR FETs. There are a few JFETs which have the drain feedback resistors added to them and are sold as VCRs. They have notably not taken the world by storm.

[edvard]

...
There are a few JFETs which have the drain feedback resistors added to them and are sold as VCRs. They have notably not taken the world by storm.

You noticed that too?


Here's a little write-up I found on the topic:
http://freespace.virgin.net/ljmayes.mal/comp/vcr.htm

Since you were originally doing frequency to voltage conversion, have you searched for that?
National makes a part for it: http://www.national.com/mpf/LM/LM2917.html

Oooh, it interfaces with magnetic pickups 

[Taylor]

...
There are a few JFETs which have the drain feedback resistors added to them and are sold as VCRs. They have notably not taken the world by storm.

You noticed that too?


Here's a little write-up I found on the topic:
http://freespace.virgin.net/ljmayes.mal/comp/vcr.htm

Since you were originally doing frequency to voltage conversion, have you searched for that?
National makes a part for it: http://www.national.com/mpf/LM/LM2917.html

Oooh, it interfaces with magnetic pickups 

Hmm, very interesting. I am using Harry Bissell's P2V:

http://www.synthdiy.com/show/?id=4127 (PDF)

Which has quite a few more components. I think the basic idea is the same as this tachometer chip, but working with audio presents a lot of issues that I suspect don't exist in measuring a motor. I will look into this LM2917, though, that's pretty cool.

[Taylor]

Yeah, we all get to there. After playing with Light Dependent Resistors, the idea for a Voltage Controlled Resistor hits us pretty quick.

There *are* components sold as voltage controlled resistors. See

http://www.specteclogistics.com/docs/70598/70598.pdf

for some discussion of VCR FETs. There are a few JFETs which have the drain feedback resistors added to them and are sold as VCRs. They have notably not taken the world by storm.

Well, I certainly didn't think I was the first to ponder this, I just didn't expect that there wouldn't be any really good answer. It seems like something that people would need enough that a good solution would have been devised in 1940 or so... I'm thinking PWM looks good, but then I suppose I have to deal with removing the oscillator click/whine.

Arr, why isn't there a 40XX Quad VCR for 40 cents apiece? Has technological progress been halted entirely?

[puretube]

Cc. reply#8:
Simple [PWM]
[PWM]VCR

t.b.c. ...

[brett]

Hi again
thanks for the effective explanations, RG.

re: opto-coupler
I've used these at fairly high speeds (100kHz, 10us) with great success.  IIRC my load was on the collector, which allowed slightly faster switching (maybe 2 x faster).  Another way to greatly speed up the switching rate (several x faster) is to use a base-emitter resistor of about 220 kohms (I got that off an appn note somewhere on the net).  It probably lowers the current transfer ratio, but that shouldn't be a problem in this application.  I use the old workhorse of optocouplers, the 4N25.

cheers

[cloudscapes]

I use a led/ldr combo, ho's minimum resistance is as close to zero as possible. then use a resistor in paralel with the ldr for the upper limit.

I'm sure there's a better way of doing it