How to control multiple POTs with one Master POT??

[ibnul7]

Hey guys, I was just wondering how I would go about controlling 3 potentiometers around a circuit with a single (master) potentiometer. I am making a sine wave generator and I need a way to change 3 different pots, but with the same value (ie. 10k, 10k, 10k). I was thinking of using MOSFETs or BJTs as variable resistors by controlling their gates somehow, but how do i do it?

Thanks

Here is the cct



- - - Updated - - -

Also, i would like the resistor values to be from 6ohms to 10k.

 

[godfreyl]

You can use a JFET as a variable resistor for signals less than about 1/2 a Volt. The trouble is that their parameters are so loosely defined that you'd end up spending all day measuring individual devices and trying to find three that match.

Another approach would be to use LDRs (light dependent resistors). They're not very well matched either, but you could get reasonable tracking by adding trimpots to adjust the low and high resistance of each LDR. Yech.

Plan C is to use digital potentiometer chips. No problem with that except that the resistance isn't continuously variable, it increments in discrete steps.

Or you could just use a 3-gang potentiometer. You can't buy those, but you can buy seperate wafers, shafts etc and make your own.

----------------------------------------------

However all of those solutions involve too much hassle, expense and complication for my liking, which suggests that any circuit requiring three pots is a bad idea.

Another slight problem with your circuit is that the three variable resistors have to be well matched. If they're not, then the gain of the circuit changes. Presumably you were going to make Rf or Rg adjustable so you can set the gain to a value that gives stable oscillation with minimum distortion. However you don't want to have to re-adjust the gain every time you change frequency because the three pots don't track well.

----------------------------------------------

I suggest using a circuit like the one below instead. It only needs two potentiometers (VR2 and VR3) to set the frequency so you can use a normal stereo potentiometer.

As an added bonus, the outputs of the four opamps are in quadrature, with equal amplitude, which may come in handy.

Because it's based on all-pass phase shift filters, any mismatch in the capacitor or potentiometer values just causes a change in the relative phase of the outputs, rather than affecting the gain.

VR1 is a trimmer to adjust the gain. In the pic below, some of the waveforms are clipped because I set the gain too high in the simulation. It could be a lot better but if you want the lowest possible distortion VR1 should be replaced with some sort of AGC control.

Also, i would like the resistor values to be from 6ohms to 10k.

That's a bit of a tall order. The normal approach is to use the potentiometers to adjust the frequency over a range of about 10 to 1, and use a switch to select different pairs of capacitors for different decade ranges.

 

[rusty81]

You can easily buy something like that:

 

[godfreyl]

Oh! Where? Can you give a link?

 

[rusty81]

Here: **broken link removed**

or here: https://www.potentiometers.com/

or here: https://www.galco.com/buy/Precision-Electronic-Components/KKKU1031S28

 

[godfreyl]

Thanks. I got nowhere searching on the RS, Digikey and Mouser websites.

 

[FvM]

You can easily buy something like that

Should be better phrased as "are made on custom order from specialized manufacturers. With some luck you might find a surplus part that fits your requirements."

As previously reported, there are trim potentiometers available that can be stacked on a shaft. https://www.edaboard.com/threads/288599/
The 6 ohms to 10 k range mentioned in the original post requests however for a log curve which is less easily available as more than dual gang.

As a generally remark, it seems reasonable to design electronic circuits considering the availability of critical components instead of realizing lately that a design isn't feasible. In so far, we won't design a variable frequency oscillator requiring a 3-gang pot if circuits based on dual gang exist.

 

[godfreyl]

The 6 ohms to 10 k range mentioned in the original post requests however for a log curve which is less easily available as more than dual gang.

Ideally you want an anti-log* law pot so the frequency increases when you turn the knob clockwise. Even more of a PITA to find.

* or reverse-log, or whatever they're called

 

[rusty81]

Should be better phrased as "are made on custom order from specialized manufacturers. With some luck you might find a surplus part that fits your requirements."

eBay is plenty of these n-gang potentiomenters, 15-20 years ago was widely used without problems, i don't see all this weirdness on find the simple solution for a problem like that.

Don't use a Gatling for kill one ant

 

[ibnul7]

Thanks for everyone's support I think i will go with the adaptation to the quadrature oscillator (actually called bubba oscillator) suggested by godfreyl. I don't think i wanna ditch out more than $5 for a gang pot

Here is a wide variety of oscillators with feedback and stability explained by TI
https://www.ti.com/sc/docs/apps/msp/journal/aug2000/aug_07.pdf

- - - Updated - - -

Actually nvm godfreyl. I had originally replied that your suggested circuit was a bubba oscillator, but your one only has 2 poles that shift it by 90deg each :/ that may be a problem for frequency drift

 

[Analog Ground]

Also, i would like the resistor values to be from 6ohms to 10k.

Only 6 ohms and 10nF will probably be too much of a capacitive load for the op amps. 6 ohms is not enough to prevent loss of phase margin in the amplifier. The whole thing may oscillate at a higher frequency than intended at low settings of the pot(s).

 

[FvM]

I don't know for the shown circuit, but the classical dual integrator topology is another oscillator that can be well tuned with two variable resistors.
https://www.edaboard.com/threads/277482/
https://www.edaboard.com/threads/175550/

 

[godfreyl]

Actually nvm godfreyl. I had originally replied that your suggested circuit was a bubba oscillator, but your one only has 2 poles that shift it by 90deg each :/ that may be a problem for frequency drift

Yes, it's definately not a Bubba oscillator. Unfortunately I have no idea what it's called, if it even has a name.

Interesting point about frequency stability. TBH, I hadn't actually thought about that before. I notice that Texas document you linked has this to say:

Three equal cascaded RC filter sections have a higher dφ/dt, and the resulting oscillator has improved frequency stability. Adding a fourth RC section produces an oscillator with an excellent dφ/dt, thus this is the most stable oscillator configuration.

Out of curiousity I simmed a few different oscillators to compare their loop gain and phase shift vs frequency. Interestingly, the phase shift vs frequency of the one I posted is identical to that of the Bubba oscillator. (the green curve is completely covered by the blue curve)



- - - Updated - - -

the classical dual integrator topology is another oscillator that can be well tuned with two variable resistors.

That reminds me - there's an interesting variation of it in the book "A practical introduction to electronic circuits" by Martin Hartley Jones. The author claims low distortion (<0.1%), and explains the non-obvious bits as follows:

The 4.7MΩ resistor R7 applies a small amount of "negative damping" to ensure that oscillation starts quickly when the circuit is switched on. Amplitude limiting is provided by Zener diodes D1 and D2 and the divider R8, R9, which brings in positive damping when the output amplitude rises above approximately 4V peak. Thus both output (1) and output (2) give a very steady 8V peak-to-peak.

Disclaimer: My copy of the book is over 30 years old. I've no idea what's in the newer editions.

 

[ibnul7]

Thanks once again for the replies. BTW what program do you use to simulate these circuits? I've tried using LTSpice to simulate these circuits, but I can't get a descent oscillation to start.

 

[godfreyl]

I use the free version of SIMetrix-SIMPLIS, but I have tried LTSpice as well.

To get the oscillation started in a simulator, you need to specify an initial condition somewhere, or inject a small pulse.

In real life, circuit noise is enough to get it started, (or the jolt from switching the power supply on)

- - - Updated - - -

Aha, I've still got LTSpice installed. From the help file:

.IC -- Set Initial Conditions

The .ic directive allows initial conditions for transient analysis to be specified. Node voltages and inductor currents may be specified. A DC solution is performed using the initial conditions as constraints. Note that although inductors are normally treated as short circuits in the DC solution in other SPICE programs, if an initial current is specified, they are treated as infinite-impedance current sources in LTspice.

Syntax: .ic [V(<n1>)=<voltage>] [I(<inductor>)=<current>]

Example: .ic V(in)=2 V(out)=5 V(vc)=1.8 I(L1)=300m

 

[BradtheRad]

Your post #1 schematic is the phase shift oscillator with 3 capacitors. This variation has op amps after each RC cell, to improve stability.

It is a tricky circuit to get going. Gain needs to be adjusted at just the right amount, so that oscillations sustain, yet not get out of hand.

The potentiometer settings do not have to be equal. The RC time constants do not have to be equal. Oscillations will occur at the frequency which results in a phase shift of 180 degrees.

To fine tune the frequency, you only need to adjust one resistor value (one pot adjust). Gain may be affected to some extent.

BTW what program do you use to simulate these circuits?

I used Falstad's animated interactive simulator:



If you click the link below, it will open the www.falstad.com/circuit website, load my schematic into his Java program,and run it on your computer.

https://tinyurl.com/oltek7y