replace pot with electrically controlled "device"

[Frances Rhodes]

hey y'all

i've been searching on line for a while now and i can't seem to find the answer
i would like to replace a pot in a circuit and replace it with... "something" that would be controlled by a voltage, but i don't want to have to replace all the circuit.
i guess replacing a pot with 2 "inverted" LDRs or 2 jFETs in series wouldn't be very simple task, but would it be possible to manage something with FETs or LDRs that would be stable enough to be used in a transistor booster like the LPB1, the SHO or jack orman's mosfet booster?

regards

[robthequiet]

I'd be searching for "vca" -> voltage controlled amplifier -- are you going for something like a tremolo or rather something you adjust on the fly with a knob or footpedal?

[Frances Rhodes]

yeah, i already have a few VCA circuits but i what i want is to mod a booster circuit that i have already built (an amz mosfet booster and an LPB1) so i could control the volume or gain with either a pot or anything that could provide a voltage.
i could replace the gain pot with a jfet but i would need to find one with the right resistance value, and i have the feeling that it will be really hard to achieve...
if i have no other choice i'll build a simple VCA, but i'd like to find a way to mod my already existing circuit

[GibsonGM]

yeah, i already have a few VCA circuits but i what i want is to mod a booster circuit that i have already built (an amz mosfet booster and an LPB1) so i could control the volume or gain with either a pot or anything that could provide a voltage.
i could replace the gain pot with a jfet but i would need to find one with the right resistance value, and i have the feeling that it will be really hard to achieve...
if i have no other choice i'll build a simple VCA, but i'd like to find a way to mod my already existing circuit

This is often done using an LDR (or two...).  An LED is what you control using a foot pedal (rocker), and its brightness controls the LDR(s).  Optical wahs work this way...no scratchy pot...

[Transmogrifox]

The problem with LDR's is the control range is not linear, and you can't have a "10k pot". 

Maybe more complicated than what you want to do, but this is an interesting idea.
https://www.maximintegrated.com/en/app-notes/index.mvp/id/4051

This concept would be most useful as a little SMT board with DIP-compatible header pins, or maybe a vertical SIP format that you just stick into a perfboard or project of choice.

I was also thinking to increase the resolution to use basically a brute-force DA converter so you can get the resolution of the incremental changes up to 12 bits or so, although 8 bits would probably be enough for something that you "set and forget"....but something you want to continuously control like a phaser or tremolo then you likely would want more than 12 bits.

Anyway the "brute force" approach is to make an array of, say, 12 analog CMOS switches in parallel and each in series with a resistor scaled according to a rmin*2^x graduation.  It would still fit onto a small PCB using quad CMOS switch packages and a microcontroller. Of course it takes 2 of these to make a pot.

I haven't spent much time looking at digital pots to see if maybe they can do better resolution than 8 bits....maybe that is a lead.

[amptramp]

Use a multiplying DAC where the digital inputs are your selected tap points and the reference input is the analog input.  See Figure 9 here:

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

You can set up an up/down counter in CMOS or a PIC to use as the control.  You could also use values from a memory but using an A/D converter either standalone or ina PIC to generate the digital multipliers is a reasonable solution.

[Transmogrifox]

Use a multiplying DAC where the digital inputs are your selected tap points and the reference input is the analog input.  See Figure 9 here:

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

This IC used as suggested will be a digitally controlled VCA with an irritating DC offset.  It will not emulate a resistance element.

For this application an OTA will likely be less problematic and a simpler application.  If willing to go digital than I think a digital pot will be better suited to the intended purpose.

If not willing to go digital then an OTA or LDR is probably your best bet. 

For many applications the nonlinear voltage-resistance relationship is beneficial (like an optical wah) because it behaves more like a log taper pot. 

For applications where this isn't desirable, or you need more exact control over the actual resistance you can add circuit complexity to overcome the nonlinear voltage-to-resistance characteristic on the LDR.   You can servo the system with an op-amp.  For this you get two reasonably matched LDR's.  You put one in the feedback of an op amp so the output voltage is a direct linear relationship between LED and LDR.  You put 2 LED's in series.  One illuminates the reference LDR and the other illuminates 1/2 of your pot. 

Then you reference all of this back to your CV so that you get a linear change in servo op-amp voltage in relationship to the CV.  This results in a LED current characteristic that drives the LDR along a linear curve and then you can control its resistance in a relatively exact linear relationship to the CV.

With 2 of these circuits where the relative CV is driven differentially then you have a voltage-controlled pot.

This system being a feedback control loop will need some work to ensure it is stable.  Probably you can reduce loop gains enough to stabilize it and still have suitable performance.

[R.G.]

This is a longstanding issue. I was looking for a voltage controlled resistance with low distortion, isolation, accuracy, and repeatability in the 1970s, and periodically ever since. Sadly, there isn't one of those, at least not at the low level of complexity you're describing.

There are voltage controlled resistors in the form of JFETs with variable gate voltage. They suffer from limited voltage and current in the "resistor" and distortion. There are servo'ed LDRs in the form of a center tapped LDR with one half providing the feedback to get the other half to behave in a controlled manner, but they are very difficult to hook up without grounding the center tap. The circuits to float the center tap are as complicated as other solutions.

There are three semi-solutions, both requiring selling a piece of your analog soul to the digital gods. Switched capacitors make a real variable "resistance" for switching frequencies meeting the Nyquist criterion for all signal frequencies. Switched/PWM resistors can do a similar thing for making resistors larger than the resistor being PWM'd and also meeting the Nyquist criterion. Finally, you can use digital pot ICs, and drive the digital pot with a microcontroller that has built-in A-D and converts the (presumably analog) control voltage to a digital drive for the digital pot. I've done all of these at least once.

The uC and digipot is the most conceptually simple thing in this list. You could also use a uC to read an analog control voltage and output a PWM or frequency to control the switched capacitor and resistor, probably more cost-effectively than doing analog PWM.

[Frances Rhodes]

thanks for all the answers!

ok, i guessed it wouldn't be easy, but i think in my case, i'm just looking for a "punk" way of doing it.
i was thinking of replacing the emitter resistor in a simple booster circuit by an FET for instance, and make a cheap VC gain control.
since the resistance of an FET goes from around 100 ohms to several M ohms, that way i could make the gain of the booster vary from a maximum value set by the ratio of Rfet and Rcollector. and i would just need to make a simple LFO to drive the FET.
would that work or is it impossible to achieve?

[wavley]

You need Neil Young's Whizzer.

[Frances Rhodes]

You need Neil Young's Whizzer.

i have no idea what this is, but i sure want one!!

[R.G.]

i was thinking of replacing the emitter resistor in a simple booster circuit by an FET for instance, and make a cheap VC gain control.
since the resistance of an FET goes from around 100 ohms to several M ohms, that way i could make the gain of the booster vary from a maximum value set by the ratio of Rfet and Rcollector. and i would just need to make a simple LFO to drive the FET.
would that work or is it impossible to achieve?

Well, it has some hope of working, but it gets really complicated. The AC gain of a FET circuit (and a bipolar one) does approach Rc/Rs, but changing just Rs also changes the DC biasing, and radically. It's not a "pure" gain change.

You can easily get into a position where the AC gain has indeed changed, but it can't respond to signal bigger than a few microvolts because it's running on the edge of saturation or cutoff because the bias voltage is no longer good for bigger swings. Or it won't respond to any input signal at all because it's biased completely into saturation or cutoff.

One way out of this issue is to use a JFET as a voltage variable resistor, but capacitor couple it to another amplifier, so the JFET resistance change can change only the AC gain, not the DC conditions. This is how the EA tremolo works, as an example. You do suffer some more limited range in gains than you might by varying only the source resistor; but that's mitigated by it being practical to do.

[balkanizeyou]

I have a similar problem - I wanted to build a pedal consisting of 6 different overdrive/distortion/fuzz pedals with a 2P6T rotary switch to choose between them. All the circuits would have a gain, tone and volume pot, so for my convenience I'd like to have only three potentiometers in my box, each controlling gain, tone and volume of all 6 circuits at the same tame.

For the simplicity let's assume that the gain pot works as a variable resistor in all circuits, tone is wired as a normal potentiometer and the volume control isn't really that much of a problem, because I can set a master volume control after all the rotary switch.

So, if I understand correctly, I need a microcontroler with two A/D converters so my "real" pots could provide variable voltage to the uC. Then the uC would use 12 digital outputs (ok... I guess attiny45 won't do lol) to drive the digipots - or maybe I just need two outputs, each for each analog pot?

But my questions is - how would a digipot work as a for example feedback resistor - would it introduce much noise to the circuit? Same thing goes with tone control. I have never worked with digipots so I don't have any idea how they work in an analog circuit.

[Frances Rhodes]

i was thinking of replacing the emitter resistor in a simple booster circuit by an FET for instance, and make a cheap VC gain control.
since the resistance of an FET goes from around 100 ohms to several M ohms, that way i could make the gain of the booster vary from a maximum value set by the ratio of Rfet and Rcollector. and i would just need to make a simple LFO to drive the FET.
would that work or is it impossible to achieve?

Well, it has some hope of working, but it gets really complicated. The AC gain of a FET circuit (and a bipolar one) does approach Rc/Rs, but changing just Rs also changes the DC biasing, and radically. It's not a "pure" gain change.

You can easily get into a position where the AC gain has indeed changed, but it can't respond to signal bigger than a few microvolts because it's running on the edge of saturation or cutoff because the bias voltage is no longer good for bigger swings. Or it won't respond to any input signal at all because it's biased completely into saturation or cutoff.

One way out of this issue is to use a JFET as a voltage variable resistor, but capacitor couple it to another amplifier, so the JFET resistance change can change only the AC gain, not the DC conditions. This is how the EA tremolo works, as an example. You do suffer some more limited range in gains than you might by varying only the source resistor; but that's mitigated by it being practical to do.

ok, i see. the emitter can be increased but has to stay "reasonably low" so the transistor can proceed its task.
sorry about all the questions, it's been a long time since i "studied" basic electronics... i didn't learn much and forgot most of it
i built an amz mosfet booster a while back, it has a 100µ bypass cap coupled to the variable resistor, i'll try to do something similar with a 1M pot instead of the jFET to see if it may work and then try something inspired by the ea tremolo
thanks again!

[Frances Rhodes]

I have a similar problem - I wanted to build a pedal consisting of 6 different overdrive/distortion/fuzz pedals with a 2P6T rotary switch to choose between them. All the circuits would have a gain, tone and volume pot, so for my convenience I'd like to have only three potentiometers in my box, each controlling gain, tone and volume of all 6 circuits at the same tame.

For the simplicity let's assume that the gain pot works as a variable resistor in all circuits, tone is wired as a normal potentiometer and the volume control isn't really that much of a problem, because I can set a master volume control after all the rotary switch.

So, if I understand correctly, I need a microcontroler with two A/D converters so my "real" pots could provide variable voltage to the uC. Then the uC would use 12 digital outputs (ok... I guess attiny45 won't do lol) to drive the digipots - or maybe I just need two outputs, each for each analog pot?

But my questions is - how would a digipot work as a for example feedback resistor - would it introduce much noise to the circuit? Same thing goes with tone control. I have never worked with digipots so I don't have any idea how they work in an analog circuit.

it's just a thought but i guess if you put big rotary switches there, you could use them to activate relays (6 circuits and 2 pots each so 12 of them) to "switch in" the pots to the desired circuit...

[Jdansti]

Would an NTE7072 Dual DC Controlled Potentiometer not work?

http://www.nteinc.com/specs/7000to7099/pdf/nte7072.pdf

[Frances Rhodes]

Would an NTE7072 Dual DC Controlled Potentiometer not work?

http://www.nteinc.com/specs/7000to7099/pdf/nte7072.pdf

this looks interesting...
not sure i can understand everything but i'll give it a look!
thanks

[rumbletone]

I've done some experiments, and I'd recommend trying the LDRs first, partly because it's cheap and easy to try. If you don't need it to go to '0', which is often the case (e.g., if you want it to go from unity-ish to boosted), and the taper may be less sensitive than on, say, the time knob of a delay. The other thing is that you can also try different LEDs, different CLRs on the LEDs, and different proximity/angles between the LED and the LDR to vary the taper and the min/max. Depending on what expression device is controlling it, you can use 2 LDRs as a voltage divider and have the expression controller 'pan' between 2 LEDs controlling the respective LDRs (one gets brighter as other gets darker), which allows further variation/control.

It certainly involves quite quite bit of test and guess, but I've had at least few pedals where it worked surprisingly well for adding expression control of what was otherwise a pot, and because it's opto-isolated from the external controller there are fewer noise/etc. Issues to worry about. 

[Transmogrifox]

oops.    double post.
See below.

[Transmogrifox]

Would an NTE7072 Dual DC Controlled Potentiometer not work?

http://www.nteinc.com/specs/7000to7099/pdf/nte7072.pdf

this looks interesting...
not sure i can understand everything but i'll give it a look!
thanks

It's a pair of VCA's a cross-fader.  Probably useful for the majority of things you do with a pot, but it is NOT a variable resistor.  Anything that depends on a real resistance value (like a filter cut-off frequency) can't benefit from this unless you re-design the circuit to work with a standard VCA.  If you want a VCA that can cross-fade between 2 inputs then this would be a useful chip.

For example, the Vox/Crybaby circuit could benefit from one of these because the the sweep frequency is only dependent upon gain.  In that case you would only use 1/2 the IC.

For a tremolo it would work because you don't need a specific resistance value for the pot.  This could be a really nice IC for a stereo tremolo where you want 2 VCA's that track each other really well.

It wouldn't work as a replacement for, say, automating or remote controlling rate on an LFO where resistance of the pot sets the rate. 
Wouldn't work as an LDR replacement in a Mutron (neutron) type of filter without redesigning the way the filter works  (in this particular instance it would be relatively straight-forward).
Wouldn't work as a replacement for FETs in a FET phaser.
Wouldn't work as a replacement for LDR's in an LDR-controlled phaser or envelope filter.

That chip can't be used as a general-purpose pot replacement.  It has to be applied as a 2-input VCA that cross-fades between the 2 inputs.

However in examples above it may not be so hard to modify the circuit to get the intended function out of VCA circuits.