Hello. I have done this handy table that allows me to choose the right resistor's ratio when i need to change the potentiometer's curve characteristic. Also, this will allow me to buy linear pots of various values regardless their curve characteristic.

Beside this, I'm trying to achieve another curve but I can't quite understand how i can do it:
Turning the knob, i need to have the value totally unchanged until 12 O'clock , and then the value decrease.

This might sound weird but i would use it in an audio circuit where a dual gang potentiometer blends to signals together but one of the 2 signal remains unchanged for the first half.
None of the curves in the table attached is close to what I'm looking for
Any idea ?

 

Turning the knob, i need to have the value totally unchanged until 12 O'clock , and then the value decrease.

Don't see how that's possible with a standard pot.

 

There are Pots specifically custom designed for Stereo-Balance applications.
I've never seen them listed along with standard Pots,
probably because there is no demand for individual sales in quantities of less than ~10,000 pieces.
.
.
.

 

This is not gonna happen with a generic, analog pot.
You could use an MCU to read a normal linear pot, then use an equation or lookup table to change the curve to whatever, then output the signal to a digital pot chip.

 

If you've just got to have this functionality,
it can be done 100% Analog with a couple of Voltage-Controlled-Amplifiers, or "VCAs",
or,
2 Voltage-Controlled-Attenuators.
Using the Attenuators is the best route in my opinion.

Audio Balance-Controls are more of a "gimmick" than anything actually useful.
On the other hand,
a "Mono-Button" can be very handy considering some of the atrocious Audio quality on the Internet.
.
.
.

 

Alternatively, you can approximate this behaviour with a negative input impedance circuit.

 

Hello. I have done this handy table that allows me to choose the right resistor's ratio when i need to change the potentiometer's curve characteristic. Also, this will allow me to buy linear pots of various values regardless their curve characteristic.

Beside this, I'm trying to achieve another curve but I can't quite understand how i can do it:
Turning the knob, i need to have the value totally unchanged until 12 O'clock , and then the value decrease.

This might sound weird but i would use it in an audio circuit where a dual gang potentiometer blends to signals together but one of the 2 signal remains unchanged for the first half.
None of the curves in the table attached is close to what I'm looking for
Any idea ?

I am doing similar, and will be using nickel paint. I am using a regular (linear) audio pot with 1/4 inch shaft. I am dissasembling it to get to the carbon resistive element. For your application I would add the nickel paint from the start of rotation end to the 12 o'clock position along the outer edge of the carbon (not where the slider runs, but close to it). Assuming the pot has the correct resistance at the 12 position, this will work. The slider will see a very small resistance from itself <radially> to the nickel, then a short to the end of the nickel at the 12 position, and then the pot's normal resistance from there to the end. Once the wiper is at the 12 position, it will behave as a normal pot with decreasing resistance.

 

They are called MN taper potentiometers and can be found on ebay and other places.

 

Thanks, Spiceboy, now I don't have to fuss with std pots!!

 

I was thinking of the Audio Taper pots too and an option, but maybe not for this application...

 

You can practically get an acceptable result like by adding a resistor to the potentiometer. Try to find the blog of R.G. Keen ( I can't at the moment) or follow the following link of a YouTube video explaining with details the analysis and showing the result we can expect from the idea.

(In French, from 5:50, the potentiometer is R1 + R2, the added resistor is Rb, the obtained behavior, strongly not-linear, at 6:01)

 

Oh, I think you can add a resistor from the wiper to one side or the other depending on the direction of turning you want, and if you pick the correct value, as you turn the pot in the direction of the resistor, it becomes more dominant as a parallel value and doesn't effect the change, as you get away from the center, or parallel value in the pot, the pot becomes more dominant. I think that's what I remember seeing one time, I may be wrong, I have been before and I can prove it!

 

You may find some interesting ideas in the attached ZIP-file.
( it's an old webpage )

 

Hello. I have done this handy table that allows me to choose the right resistor's ratio when i need to change the potentiometer's curve characteristic. Also, this will allow me to buy linear pots of various values regardless their curve characteristic.

Beside this, I'm trying to achieve another curve but I can't quite understand how i can do it:
Turning the knob, i need to have the value totally unchanged until 12 O'clock , and then the value decrease.

This might sound weird but i would use it in an audio circuit where a dual gang potentiometer blends to signals together but one of the 2 signal remains unchanged for the first half.
None of the curves in the table attached is close to what I'm looking for
Any idea ?

Hi,

You could get some more interesting curves by placing another resistor between the pot arm and the node you have as output.

There are a couple ways to do what you require.

First, you can use a comparator to detect the halfway point and then have it switch in a network.

Second, with some pots you can take them apart and access the inside element. You'd have to cut it at the halfway point and possibly remove the lower (or upper) section. That when it is open circuit until it gets to the remaining resistive element and then the resistance will start changing.

If you don't want to modify the pot, then you have to find a way to force it to be very nonlinear which means a special circuit to detect the halfway point and cause a change of some sort that changes the resistance. It could be as simple as adding an analog switch that switches on at the halfway point.

I dont know if you can use diodes or a zener, but you might get close by connecting diodes to the output or something. That way as the voltage increased, there would be little output until you get to a certain point, if that is what you want.
If you want a voltage that stays constant and then decreases as you pass the halfway point, then you might be able to use a low dropout regulator that regulates the output to a certain voltage, then as the voltage goes below the regulation point the output will start to drop out. Regulators are notorious for doing that. You'd need a 2.5v regulator i guess, and maybe the TL431 would fit the bill. It's a decently accurate shunt regulator. You'd have to limit current to it while in the regulation mode.

 

Could you simply add a wire or 1 ohm resistor between the span you wish to zero? Of course, the pot value would have to be higher to start with.

 

Just for clarification ..........
Do You want a "Balance-Control" for a Stereo-Audio-Pre-Amplifier ?
Or, is this a specialized Circuit that must behave exactly as described in your initial Post ?

Your initial specifications are not required for a Stereo-Balance-Control.
You don't need a special Pot,
You don't even need a Dual-Pot,
and a standard Linear-Taper-Pot is the preferred part.

How many dbs of difference do You want between the Left and Right Channels ?
~6-dbs should be more than adequate,
unless You require the ability to
completely mute one Channel or the other with this Balance-Control.
.
.
.

 

They are called MN taper potentiometers and can be found on ebay and other places.

This is exactly what i need. Fantastic and thanks for the hint.
this link gave me more informations about MN tapers.
Now let's see how is easy to find them!

@LowQCab thanks for the link. Very useful informations there (as in all ESP pages).

The Pot will need to be at least a Dual-Pot because is a Stereo application (Quad-Pot would be really the best since it would allow complete indipendent attenuations, while a Dual-Pot will probably require a circuit that blends 2 signals sharing the wipers of the Pot).

In particular, the circuitry is an Effect "Return" control.
The "Return" will normally increase one signal from 0% to 100% of the Shaft rotation on one gang while the other part will decrease the "Dry" signal after 50% of the rotation.

If the Dry remains unaffected until 50% of the rotation, the signal won't drop at all until that point and this is handy if I want to hold the Pot at 12 O'clock and play with the relative "Send" instead of the "Return".
Example: I'm playing and controlling the effect's "Return" while "Send" is fully open... at some point i leave the "Return" Pot at 12 O'clock. Now I manually attenuate the "Send". IF the "Return" pot's curve is normal, the Dry signal will have some attenuation (even if not so much noticeable), IF the "Return" pot's curve is like the one I'm looking for, then it will be 1:1 with with the unaffected Dry.
This does make a difference because when i will bring back the "Return" shaft's position at 0% I will experience again an annoying change in volume.
A requirement is that the Return Pot can stay open at 50% without dropping the volume of the Dry signal.

Hope it makes sense and it should explain that is not really a standard Balance-Control or Left-Right mix.


Since I just discovered the MN tapers and it looks like it will solve the issue, I will accept that curve behaviour also for the "Return"

Also, I initially talked about Linear Pots because I could achieve curves that were quite similar (but not enough) to what I was looking for. Linear Pots are not a requirement.

@MrAl that's very interesting. The comparator idea would probably work. The rest of the suggestions would work but not in case of an audio signal i believe. Am I wrong?
Opening a knob is unfortunately sometimes not possible on some pots.

 

You're making this far more complex than it needs to be.
.
.
.

.

 

This is exactly what i need. Fantastic and thanks for the hint.
this link gave me more informations about MN tapers.
Now let's see how is easy to find them!

@LowQCab thanks for the link. Very useful informations there (as in all ESP pages).

The Pot will need to be at least a Dual-Pot because is a Stereo application (Quad-Pot would be really the best since it would allow complete indipendent attenuations, while a Dual-Pot will probably require a circuit that blends 2 signals sharing the wipers of the Pot).

In particular, the circuitry is an Effect "Return" control.
The "Return" will normally increase one signal from 0% to 100% of the Shaft rotation on one gang while the other part will decrease the "Dry" signal after 50% of the rotation.

If the Dry remains unaffected until 50% of the rotation, the signal won't drop at all until that point and this is handy if I want to hold the Pot at 12 O'clock and play with the relative "Send" instead of the "Return".
Example: I'm playing and controlling the effect's "Return" while "Send" is fully open... at some point i leave the "Return" Pot at 12 O'clock. Now I manually attenuate the "Send". IF the "Return" pot's curve is normal, the Dry signal will have some attenuation (even if not so much noticeable), IF the "Return" pot's curve is like the one I'm looking for, then it will be 1:1 with with the unaffected Dry.
This does make a difference because when i will bring back the "Return" shaft's position at 0% I will experience again an annoying change in volume.
A requirement is that the Return Pot can stay open at 50% without dropping the volume of the Dry signal.

Hope it makes sense and it should explain that is not really a standard Balance-Control or Left-Right mix.


Since I just discovered the MN tapers and it looks like it will solve the issue, I will accept that curve behaviour also for the "Return"

Also, I initially talked about Linear Pots because I could achieve curves that were quite similar (but not enough) to what I was looking for. Linear Pots are not a requirement.

@MrAl that's very interesting. The comparator idea would probably work. The rest of the suggestions would work but not in case of an audio signal i believe. Am I wrong?
Opening a knob is unfortunately sometimes not possible on some pots.

Hi again,

Oh you mean you are working with an audio signal?

 

You're making this far more complex than it needs to be.

I believe that the circuit you posted is not going to give the response i look for if I would use a standard linear Pot but i should use a dual-MN taper instead. In that case I think It's gonna be good!

Hi again,

Oh you mean you are working with an audio signal?

yes, sorry for my bad english. I need to blend 2 audio signals