i need a voltage controlled potentiometer

[JanJansen]

Hi, i need a voltage controlled potentiometer,
a digital potentiometer does not work for me, i need it smooth without steps and with voltages higher then 5v.

I have this LM13700, can i use it as a voltage controlled potmeter ?

And what would be the schematic to use 2 transistors ?, 1 NPN and 1 PNP maybe to simulate the potmeter center pin ?
btw : i need it in 50K

thanks

[Yansi]

Please specify more specifically what you are trying to achieve.  You seem to be trying to do a stunt that may not be needed at all for the application.

Please describe the circuit you need this in, preferably include schematics too.

LM13700 can definitely be used as a VCA, but it all depends on the application.

[amspire]

If you can describe the application, you will get much better suggestions.

Otherwise, we have to ask many questions to narrow down what you need. Accuracy, distortion, peak voltage, peak current at low ohms, frequency, DC offset, stability, and so on.

It may be that in your application, there are far better ways to achieve the result you want then a voltage controlled resistor or potentiometer. For example, if you want the potentiometer to give you a voltage controlled gain, you may be able to use an IC with voltage controlled gain instead.

Active circuits that simulate passive components can work for some very specific applications, but they don't come close to matching resistors, inductors and capacitors in all respects.

A simple voltage controlled resistance would be a light dependant resistor controlled by a LED. Not accurate, but it works and it is very cheap.

[JanJansen]

Ok, i need to replace 2 potentiometer in this schematic, both should do the same trick.

In the schematic its the CUTOFF FREQ. and the ENV.MOD pot,
http://www.martianarts.net/web/images/stories/ourgear/roland_tb303_schematics.gif

here a better schematic maybe :
http://www.ladyada.net/media/x0xb0x/mainboard_beta.png

That its type A potentiometers dont matters, i will change the curve to logarithmic in software.
I also need to replace the stereo resonance pot, there i plan to use the LM13700 since it need audio passing thru.

[JanJansen]

By the way for the ENV.DECAY i used a simple 2N3904 transitor, works the same as the 1M potmeter.

A simple voltage controlled resistance would be a light dependant resistor controlled by a LED. Not accurate, but it works and it is very cheap.

A LDR ?, yes i still have to buy it to try, then i need a dark enclosure also or something, dont they sell ready to use parts ?,
also i need the center pin from the potmeter.

Active circuits that simulate passive components can work for some very specific applications, but they don't come close to matching resistors, inductors and capacitors in all respects.

They still have to invent something like that.

[Yansi]

They do not, as there are many solution for voltage controlled circuits, including VCA, filters.

Why do you need to replace a mechanical potentiometer with a voltage controlled one in the first place?

[JanJansen]

To make my synthesizer MIDI controlled.

[ogden]

To make my synthesizer MIDI controlled.

MIDI is not "smooth without steps" and it is not voltage controlled.

You can use 8-bit digital potentiometer (256 taps):

http://www.microchip.com/design-centers/analog/data-converter/digital-potentiometer

[amspire]

You have two ganged potentiometers. They are both involved in feedback loops in an amplifier. This is one of those cases where simple resistors can behave much better then active circuits. In the case of the left hand gang in the original circuit, the potentiometer is basically acting as a voltage divider. A LM13700 circuit could work, or you could use a voltage multiplier IC like the AD8338 which actually has a log response in the voltage control input (12mV/dB gain).

The right gang is more difficult as it is a resistance as part of a peak detector circuit that is very intimately tied in to the quirky way that circuit works. Trying to change it will probably change the effect to some degree. It is modulating the frequency of the voltage controlled filter based on the peak Accent waveform. The actual resistances of the potentiometer and the wiper position are part of the current circuit.

I would seriously suggest looking at the 8 bit digital Microchip potentiometers odgen suggested as they probably will work in that circuit and you can get 50K versions..

Otherwise, the simplest solution may be to use a motor controlled dual gang potentiometer, if that works for you. They are available, or you could use the existing pot connected to one of the $2 geared stepping motors.

[JanJansen]

You have two ganged potentiometers. They are both involved in feedback loops in an amplifier. This is one of those cases where simple resistors can behave much better then active circuits. In the case of the left hand gang in the original circuit, the potentiometer is basically acting as a voltage divider. A LM13700 circuit could work, or you could use a voltage multiplier IC like the AD8338 which actually has a log response in the voltage control input (12mV/dB gain).

The right gang is more difficult as it is a resistance as part of a peak detector circuit that is very intimately tied in to the quirky way that circuit works. Trying to change it will probably change the effect to some degree. It is modulating the frequency of the voltage controlled filter based on the peak Accent waveform. The actual resistances of the potentiometer and the wiper position are part of the current circuit.

I will take a look at this AD8338, and yes this is a tricky one.
Maybe there is some way with 2 transistors ?, 1 NPN and 1 PNP, with in between the center potmeter pin ?,
i dont know exact how these NPN/PNP complementary transistors work, and what you can do with 2 complementary transistors.
For the decay i used a transistor, and gives some more range then wanted, there i can select the range within software, make a 7 bit range out of 8bit DAC or 10 bit PWM signal,
then i dont have the to short and to long decay tones to match the original synthesizer, ofcourse still accessible with another control number.

Otherwise, the simplest solution may be to use a motor controlled dual gang potentiometer, if that works for you. They are available, or you could use the existing pot connected to one of the $2 geared stepping motors.

A motor controlled potmeter, this is really funny, i bet they also make noise,
i am sure there will be a better way, i,m not going for motor potmeters, anyways thanks for the hint, its really funny.

MIDI is not "smooth without steps" and it is not voltage controlled.

You can use 8-bit digital potentiometer (256 taps):

MIDI will be smooth with a capacitor on the DAC or PWM control voltage.
The digital potentiometer will give audible steps, its break before make and not smoothed between the steps, also it works on 5 volt, and cant take any higher voltages because of the gates inside.

[ogden]

Maybe there is some way with 2 transistors ?, 1 NPN and 1 PNP, with in between the center potmeter pin ?,

FETs are voltage controlled resistors:

https://www.vishay.com/docs/70598/70598.pdf

[JanJansen]

I have some FETS, only no negative supply voltage, how would i connect them then ?

[ogden]

I have some FETS, only no negative supply voltage, how would i connect them then ?

Add negative supply : ) Other option would be to use NMOS.

[JanJansen]

NMOS is N channel MOSFET ?
Any suggested part and connection diagram ?
Are MOSFETS always very big ?, i need the smallest one.
I only have JFETs not MOSFETS, new for me, should do the same i readed.

[amspire]

Otherwise, the simplest solution may be to use a motor controlled dual gang potentiometer, if that works for you. They are available, or you could use the existing pot connected to one of the $2 geared stepping motors.

A motor controlled potmeter, this is really funny, i bet they also make noise,
i am sure there will be a better way, i,m not going for motor potmeters, anyways thanks for the hint, its really funny.

Motor controlled potentiometers are pretty quiet. They were huge in the 1990's when remote controls for audio HiFi amplifiers were the rage and they were not using digital control. They were quiet but slow.

Professional mixer consoles use slider potentiometers with motors. The idea is that the audio engineer could dynamically change settings during a mix and record the slider positions. Then the engineer can replay the mix and all the sliders move automatically. I think a common way to do this was to have an extra potentiometer track in the slider to sense the position. If you want to search for these, they are often called Motorized Faders.

My concern for that circuit is that the potentiometer you want to change is involved in the Voltage Controlled Filter circuit, the Voltage Controlled Amplifier circuit and the Accent control. If you fully worked out what is happening, there may be simple solutions to replacing the control, but otherwise any attempt to replace it with an amplifier circuit may change the behaviour of the device. The voltage controlled filter circuit is interesting. They are using diodes (transistors with base-collector shorted) as variable resistors. They form a chain of RC filters probably resulting in something like a 20B/octave lowpass filter and the frequency is controlled by the current through the diodes (controlled by Q10). The resonance control is feeding back the output of the filter to the input so that it rings and the control is also modulating the lowpass cutoff frequency.

This was probably designed with a massive amount of trial and error to get just right. That is why changing the circuit is definitely not trivial.

Doesn't mean you don't try.

[ogden]

NMOS is N channel MOSFET ?
Any suggested part and connection diagram ?
Are MOSFETS always very big ?, i need the smallest one.
I only have JFETs not MOSFETS, new for me, should do the same i readed.

Right. N channel MOSFET.
Suggested part could be 2N7002.
Connection diagrams are all over the internet.
Believe me, you don't need smallest one. SOT-23-3 shall be fine for you. You really shall learn parametric search of mouser.com.

[BrianHG]

You can use photo-fet optocouplers:
http://www.mouser.com/ds/2/149/h11f1m-185284.pdf

But, I would go with more modern I2C controlled pots as mentioned above if I had to stay in the analog realm.

I would rather use a single 4$ DSPic for your app.  Have it's ADC sample optional analog pot inputs, send to MIDI.  Use it's RS232 for working with the MIDI and use it's DAC to synthesize the output from ADC or MIDI source to a headphone amp.

No VCO, no VCA, no mixer, do all of this in software as these things are simple to code today if you are already capable of programming your own MIDI controller.

[amspire]

But, I would go with more modern I2C controlled pots as mentioned above if I had to stay in the analog realm.

I have a suspicion that that analog circuit probably has a Resonance control that is extremely touchy and that would mean 8 bit resolution is nowhere near enough. It could also mean a very stable and repeatable solution is wanted.

[obiwanjacobi]

I have looked into this myself for a potential project many years ago (I wanted to automate guitar pedals/stomp boxes). The best/cheapest option, I found, is to go with a FET based solution.
Use the opto-isolated variant (as suggested or other) if speed is not very important (in audio it usually is not) and you want to keep the target circuit and controller circuit separated, which is always a good idea IMO.

You need to put a small micro in between the MIDI input and the FET driving outputs in order to correct for small differences in FET sensitivity (and other component variations) and to be able to adjust the curve to be either linear or logarithmic. Adding some extra brains will also give you more MIDI features to play with (invert, offset, curve etc) in the future.

My design used two FETs to mimic a potentiometer. The micro calculated how to drive each of the FETs to get a certain resistance value for the pot-center-tab. Depending on how closely the FETs were match, could differ a little bit. Use low-tolerance components for the driving circuit.

I faked the DAC you will need (to drive the FETs from software) with a powersupply so I have no advise on that. Will you need a high precision DAC or will a R-2R network do? See if you can find a micro that has on board DACs (not very common, but existing) - or a couple of cheap DAC ICs.

[2c]

[Rerouter]

For higher voltages / currents, you can use a dac with an external reference voltage pin that allows down to 0V, The using an op amp and a pass element, you can make a constant resistance circuit.

[JanJansen]

First of all i am using the PIC16F1708, it has 20 pins, 1 x 8 bit DAC and 4 x 10 bit PWM.

Second you cannot make this in digital, else i was not making this.
I also made a virtual analog synthesizer, the thing i found out is the digital taking more milliamps.

@ obiwanjacobi : how did you connect the FETs ?, any schematic for me ?, i dont have negative power supply, a pedal also dont have negative supply.

[InterestedTom]

http://www.ti.com/lit/ds/symlink/lm13700.pdf

p16 figure 28

If you are getting marked for this it doesn't matter whether your supervisor/teacher sees this page, they will know you don't understand how it works, you are just copying others work.

Instead of panicking and being pushy about exactly what to do, use your head, and read the datasheet for the part in question.

I know that this is only one half of a potentiometer but hopefully you will think of a simple way to derive a second Control Voltage from your DAC. This way you will have thought of something for yourself and you will learn how to design electronics.

The other suggestions will work just as well, probably better: your PIC probably has at least one I2C or SPI peripheral, that would be used to control the 8bit potentiometer. 8bit pot vs 8bit DAC is pretty even, but the LM13700 specifies 0.5% distortion at best, why would you risk the existing sound of your synthesizer getting worse? Try setting a single turn pot to 1/256 using a multimeter, then realise that 8bit is pretty damn smooth even with steps. You can filter out resistance steps too, smooth them out a bit at the very least...

One thing to be aware with both circuits I have mentioned will have a limit on input voltage that may be lower than the voltages in your synthesizer. But different part numbers will have different limits, read the datasheets...

[JanJansen]

http://www.ti.com/lit/ds/symlink/lm13700.pdf

p16 figure 28

Ok, can it be made on single supply ?

Are the input and output where the calculation is in between ?
The calculation is : RX = 2R / GM RA
what should i do to get 50K ?, i,m not a professor or schooled.
Ok the control voltage = Vc, only they are connected to 2 circles, what is that ?

[JanJansen]

by the way i have this idea :
using a PWM from 0 to 5 volt, connect that somehow to the variable voltage and the 12 volt, how can that be done if that is possible ?

Try setting a single turn pot to 1/256 using a multimeter, then realise that 8bit is pretty damn smooth even with steps. You can filter out resistance steps too, smooth them out a bit at the very least...

I noticed when building a MIDI controller, thats why i use a 10bit PWM for the tuning, 10bit PWM into hexinverter on 5,333v then smoothed in rail to rail opamp also on 5,333v,
there you have it i have done something myself.

I know that this is only one half of a potentiometer but hopefully you will think of a simple way to derive a second Control Voltage from your DAC. This way you will have thought of something for yourself and you will learn how to design electronics.

The problem is the DAC in the PIC16F1708 is not totally linear, so it has to be done with resistors somehow, i will ask in the forum how to do that, i have done enough already and am always busy with stuff that dont need external brain processing.

If you are getting marked for this it doesn't matter whether your supervisor/teacher sees this page, they will know you don't understand how it works, you are just copying others work.

Yes i use the forum first before thinking and reading and searching, saves so much stress.
Please dont start a square wheel discussion ( again ), square wheels dont exist.
Do i really need to re-invent the wheel ?, maybe a square wheel will exist then.

[obiwanjacobi]

@ obiwanjacobi : how did you connect the FETs ?, any schematic for me ?, i dont have negative power supply, a pedal also dont have negative supply.

No schematic, sorry - I experimented a lot with several different options at the time, building and testing parts and pieces. I just remember FETs being the most successful in those experiments. Like I said I think I faked driving the FETs with a separate (bench) power supply at the time. But the problem was not -not having a negative power supply- the project was going to power the pedals (PCBs) and not the other way around, so I anticipated a beefy power supply - also allowing to 'over clock' (supply more than the usual 9v) some pedals that could handle that - gives you more headroom and some pedals react well to that - its a tone thing.

Could you not use a negative dc-dc convertor to generate a negative rail from the positive supply you have?

Tip: MIDI has double precision (14-bits) commands you could use if you need more resolution.