Potentiometer Questions

[mike7877]

Other than balance between channels, which properties of sound are affected by potentiometers?

I'm thinking things like THD - can they add noticeable harmonic distortion?
And are these things worse at higher attenuation levels?
What about bandwidth? Is bandwidth reduced by them? If so, how much?
Pots are also rated for maximum power - if you were to measure one at 1% of its max power and then 80%, would there be a difference in any of these measurements?

I've looked around and haven't been able to find these answers, just possible explanations/causes for degradation, but not so much which degradation...
I think someone who's done measurements on amps, maybe while making one, might know?

 

[RayDunzl]

A potentiometer is a variable resistor.

What it does depends on the circuit to which it is applied.

Could be volume, balance, tone control, etc.

Within proper design constraints, the problems you mention should be minimal.

 

[Scytales]

I think that by principle, contact resistance of the wipers with the potentiometer substrate may also be a source of distortion. An inconsistent source, because of the variable nature of the moving contact. There are also the issue of wear and tear or ingress of dirt which may cause comparable issues.

I also think that as potentiometers are relatively big component and the geometry of their conducting paths are not necessarily optimal, they may also pick-up interferences.

As far as those problems are concerned, the quality of the mechanical construction and of the materials used matters.

A not so humble component to dealt with.

All in all, that does not means that those issues are detrimental. The impairment (if any) introduced by a potentiometer can be assessed objectively to verify if those impairments are above or below audibility threshold or target specifications, or are concerning for the reliability of a product.

 

[Roland68]

Other than balance between channels, which properties of sound are affected by potentiometers?

I'm thinking things like THD - can they add noticeable harmonic distortion?
And are these things worse at higher attenuation levels?
What about bandwidth? Is bandwidth reduced by them? If so, how much?
Pots are also rated for maximum power - if you were to measure one at 1% of its max power and then 80%, would there be a difference in any of these measurements?

I've looked around and haven't been able to find these answers, just possible explanations/causes for degradation, but not so much which degradation...
I think someone who's done measurements on amps, maybe while making one, might know?

Simply use potentiometers equipped with resistors (THT or SMD), and you only have one switch contact and one resistor in the signal path. I've been using them for years, and you can get them for around €10.
Alternatively, relay-based volume control starts at around €40.
Another advantage is perfect channel balance.

 

[egellings]

Other than balance between channels, which properties of sound are affected by potentiometers?

I'm thinking things like THD - can they add noticeable harmonic distortion?
And are these things worse at higher attenuation levels?
What about bandwidth? Is bandwidth reduced by them? If so, how much?
Pots are also rated for maximum power - if you were to measure one at 1% of its max power and then 80%, would there be a difference in any of these measurements?

I've looked around and haven't been able to find these answers, just possible explanations/causes for degradation, but not so much which degradation...
I think someone who's done measurements on amps, maybe while making one, might know?

Potentiometers usually have small voltages imposed upon them so thermally induced noise does not rise above what there would be at room temperature. The resistor will not generate THD when operated at those low power conditions because the resistor's nonlinearity, if any, will be very low to nonexistent when operated at levels that potentiometers usually see. Bandwidth will be affected by any stray capacitance in the potentiometer's circuit, and will be volume setting-dependent due to an RC rolloff generated by the pot's resistance and the stray capacitance. Resistors have a temperature coefficient that results in their changing resistance value when heated or cooled. You could see a difference in the pot's resistance at different temperatures. Since the pot is usually operated as a voltage divider in a volume control application, the resistance ratio of the resistances on either side of the slider contact will not change. The total resistance might change, but the ratio, which sets the volume, won't.

 

[mike7877]

A potentiometer is a variable resistor.

What it does depends on the circuit to which it is applied.

Could be volume, balance, tone control, etc.

Within proper design constraints, the problems you mention should be minimal.

Oops lol, when I shortened the original post, I took out the sentence explaining its location...
I'm speaking about it carrying the audio signal in a preamp.

I know there are quality pots, I'm not looking to find the best pot to use, this is more of an academic look at their weaknesses. I've gathered that pots are generally safe to use and can be high performance - I'm wondering more, at what point in the device is the highest measured distortion added (however high (or low.) it is), and how does position affect this measurement

...

All in all, that does not means that those issues are detrimental. The impairment (if any) introduced by a potentiometer can be assessed objectively to verify if those impairments are above or below audibility threshold or target specifications, or are concerning for the reliability of a product.

I agree, those are the likely sources, I'm more wondering which distortions they introduce.
I shortened an originally longer OP - things like does distortion change at 5% attenuation vs 80%?

Potentiometers usually have small voltages imposed upon them so thermally induced noise does not rise above what there would be at room temperature. The resistor will not generate THD when operated at those low power conditions because the resistor's nonlinearity, if any, will be very low to nonexistent when operated at levels that potentiometers usually see. Bandwidth will be affected by any stray capacitance in the potentiometer's circuit, and will be volume setting-dependent due to an RC rolloff generated by the pot's resistance and the stray capacitance. Resistors have a temperature coefficient that results in their changing resistance value when heated or cooled. You could see a difference in the pot's resistance at different temperatures. Since the pot is usually operated as a voltage divider in a volume control application, the resistance ratio of the resistances on either side of the slider contact will not change. The total resistance mat change, but the ratio, which sets the volume, won't.

Hmm, do you know if the capacitance is the same on both channels of a stereo pot at its various settings?

 

[DVDdoug]

Pots have the advantage that they never go to full-volume all by themselves!

I have had stereo pots with audible left-right mismatch, and probably everybody's had a "scratchy-noisy" old pot. I've never heard noise or distortion from a properly working pot.

But my car stereos and AVR have rotary encoders and they've never jumped to full volume automatically. I guess that's only a problem with some stand-alone DACs. (I think the car stereos reset to low volume ever time the ignition is turned-on.)

 

[Roland68]

But my car stereos and AVR have rotary encoders and they've never jumped to full volume automatically. I guess that's only a problem with some stand-alone DACs. (I think the car stereos reset to low volume ever time the ignition is turned-on.)

This is precisely the point that most people don't understand.
The problem with volume control via the DAC chip is that the default setting in the DAC chip is always 0 dB, i.e., full volume. It should be clear what can happen if an error occurs here and the device or the DAC chip is reset.

Almost every AVR (and other devices) has digitally controlled ICs with analog volume control. The initial setting is not 0 dB like with DAC chips, but always the lowest volume.

These problems are also not known with DACs with an independent digital volume control upstream of the DAC chip, such as RME ADI-2 devices, or with an independent analog volume control directly upstream of the output, such as the Topping DX9.

 

[Sokel]

DACs have an additional problem been connected to a PC/Laptop/whatever.
Some of them can expose their 0dB by something so simple as a BSOD, win update, etc.

 

[egellings]

Oops lol, when I shortened the original post, I took out the sentence explaining its location...
I'm speaking about it carrying the audio signal in a preamp.

I know there are quality pots, I'm not looking to find the best pot to use, this is more of an academic look at their weaknesses. I've gathered that pots are generally safe to use and can be high performance - I'm wondering more, at what point in the device is the highest measured distortion added (however high (or low.) it is), and how does position affect this measurement



I agree, those are the likely sources, I'm more wondering which distortions they introduce.
I shortened an originally longer OP - things like does distortion change at 5% attenuation vs 80%?




Hmm, do you know if the capacitance is the same on both channels of a stereo pot at its various settings?

I can't see why the capacitance of 2 identical channels of an amp or preamp circuit would vary enough to make a difference worth worrying about. If two different models of amplifier are used for the two channels, then there could be an issue.

 

[Punter]

A potentiometer is a variable resistor, how can it introduce distortion? The only way I can think of is if there's a poor electrical connection between the wiper and the resistive material. This wouldn't cause harmonic distortion but noise and an interruption to the signal. There seems to be some sort of agreement on "cermet" potentiometers for audio applications. This type certainly appears in professional equipment routinely. My experience with them is that they seem to exhibit longer service life than carbon pots which is important when you might have several hundred of them in a professional mixing console.
Bournes make excellent cermet potentiometers

 

[AnalogSteph]

A potentiometer is a variable resistor, how can it introduce distortion?

The carbon variety is basically a carbon film resistor, which are not the last word in temperature coefficient, and the track width in a log pot is not constant so you can expect different sections to heat up differently when current goes through them. If it were linear, the changes an in unloaded pot would cancel.

Even so, I've never heard of substantial distortion being measured in a pot. Them usually being 10-100k (with 50k being common on hi-fi gear) tends to keep power dissipation down, I guess.

 

[solderdude]

Other than balance between channels, which properties of sound are affected by potentiometers?

I'm thinking things like THD - can they add noticeable harmonic distortion?

NO... they can't. You want proof ?
If it were the L30-II would not have such low distortion and that one uses a cheap pot, the test signal passes through that pot and the amp.

And are these things worse at higher attenuation levels?

No they aren't.
The only measurable (that can even become audible) effects are:
Symmetry in 'gangs' (tracks), especially true for semi-log (all log potmeters are essentially semi-log) potmeters near the 'bottom' of the volume range and sometimes at the point where the track resistance changes per distance (to create the semi-log)

Another effect is low-pass filter which can be frequency dependent when high impedance potmeters are directly followed by a high capacitance (think long and high-capacitance cables directly behind a 'passive attenuator'.
This is not the case in any amp. At best the potmeter 'sees' an input capacitance (pF) of the circuit behind the potmeter but that is so little it won't affect bandwidth.
The bandwidth limiting factor is the amp circuit itself or in some cases a deliberate filter at the input (against RF). Never the potmeter itself.

What about bandwidth? Is bandwidth reduced by them? If so, how much?

See the above remarks.
Potmeters all can do well above 1MHz. It is the load capacitance between wiper and ground that creates the low pass.

Pots are also rated for maximum power - if you were to measure one at 1% of its max power and then 80%, would there be a difference in any of these measurements?

That would only be interesting for L-Pad and low impedance potmeters in series with power consuming gear.

An example... even a very log value (for audio) 1k potmeter connected directly to a 3V DAC output dissipates 0.009W when a constant 3V would be applied.
In 10k pots we are looking at <1mW 'peak' power levels. Given how large the surface area of the resistance track is even the smallest potmeters can handle thousands of times higher 'power' levels.

I've looked around and haven't been able to find these answers, just possible explanations/causes for degradation, but not so much which degradation...
I think someone who's done measurements on amps, maybe while making one, might know?

Well... a lot of potmeters eventually become 'scratchy' and may get 'dead points' due to dust/corrosion etc.

When one wants a 'better' volume control (better L-R balance or a more log-type) control there is a stepped one (switches, relays) or IC potmeters.
The relay and IC potmeters requires power supply.

 

[voodooless]

This is precisely the point that most people don't understand.
The problem with volume control via the DAC chip is that the default setting in the DAC chip is always 0 dB, i.e., full volume. It should be clear what can happen if an error occurs here and the device or the DAC chip is reset.

Almost every AVR (and other devices) has digitally controlled ICs with analog volume control. The initial setting is not 0 dB like with DAC chips, but always the lowest volume.

These problems are also not known with DACs with an independent digital volume control upstream of the DAC chip, such as RME ADI-2 devices, or with an independent analog volume control directly upstream of the output, such as the Topping DX9.

Don't blame hardware for firmware issues.. There are probably a dozen or so of registers you have to set in a DAC chip before you get music out. Initializing volume at a low enough level when the DAC is in volume control mode should be trivial.

 

[mike7877]

NO... they can't. You want proof ?
If it were the L30-II would not have such low distortion and that one uses a cheap pot, the test signal passes through that pot and the amp.
View attachment 453602



No they aren't.
The only measurable (that can even become audible) effects are:
Symmetry in 'gangs' (tracks), especially true for semi-log (all log potmeters are essentially semi-log) potmeters near the 'bottom' of the volume range and sometimes at the point where the track resistance changes per distance (to create the semi-log)
View attachment 453598
Another effect is low-pass filter which can be frequency dependent when high impedance potmeters are directly followed by a high capacitance (think long and high-capacitance cables directly behind a 'passive attenuator'.
This is not the case in any amp. At best the potmeter 'sees' an input capacitance (pF) of the circuit behind the potmeter but that is so little it won't affect bandwidth.
The bandwidth limiting factor is the amp circuit itself or in some cases a deliberate filter at the input (against RF). Never the potmeter itself.


See the above remarks.
Potmeters all can do well above 1MHz. It is the load capacitance between wiper and ground that creates the low pass.



That would only be interesting for L-Pad and low impedance potmeters in series with power consuming gear.

An example... even a very log value (for audio) 1k potmeter connected directly to a 3V DAC output dissipates 0.009W when a constant 3V would be applied.
In 10k pots we are looking at <1mW 'peak' power levels. Given how large the surface area of the resistance track is even the smallest potmeters can handle thousands of times higher 'power' levels.



Well... a lot of potmeters eventually become 'scratchy' and may get 'dead points' due to dust/corrosion etc.

When one wants a 'better' volume control (better L-R balance or a more log-type) control there is a stepped one (switches, relays) or IC potmeters.
The relay and IC potmeters requires power supply.

Thanks for the detailed reply, very useful and very much what I was looking for.

The reason I was wondering about bandwidth was for imaging reasons - phase. Other than balance, slightly different phase seems to be the only parameter that could possibly explain the altered imaging effect (usually inferior) that I sometimes experience through pots vs relay switched resistor networks

To explain the potential for >1mW and the possibility for localized heating to be a factor, one of the scenarios I was envisioning an integrated amplifier circuit with a minimized stages to potentially 2 with a 1-2kohm pot. I didn't share that in the OP (it started very very long and I trimmed it a little too much lol)

 

[mike7877]

A potentiometer is a variable resistor, how can it introduce distortion? The only way I can think of is if there's a poor electrical connection between the wiper and the resistive material. This wouldn't cause harmonic distortion but noise and an interruption to the signal. There seems to be some sort of agreement on "cermet" potentiometers for audio applications. This type certainly appears in professional equipment routinely. My experience with them is that they seem to exhibit longer service life than carbon pots which is important when you might have several hundred of them in a professional mixing console.
Bournes make excellent cermet potentiometers
View attachment 453359

Thanks for the info. To explain further, I'm more investigating where the small amounts of distortion would be coming from and what effects pots can have within the circuit to know what's optimal. I'm aware 99+% of the time that switching out the pot for a carbon resistor of the value it's set to wouldn't have a measurable effect on THD, even in pretty good circuits