Why do pots have channel imbalances at low volumes?

[Racheski]

This is a question carried over from the L30 thread.

 

[majingotan]

Because it's not the Alps RK50 that they put in there. Only pot (not DACT nor stepped attenuators) I know that has absolutely no channel imbalance at any volume position

 

[restorer-john]

At low volumes, the signal input voltage is dropped over virtually the entire length of the resistance tracks. Making two tracks (L/R) identical, especially when they are attempting to create a semi-logarithmic relationship in the material (resistance vs rotation) is difficult.

It's amazing they work as well as they do.

 

[KPH]

If there is channel imbalance in a headphone amp and the amp also has preamp, will you have imbalance when you connect powered speakers and control the volume with the amp?

 

[RayDunzl]

If there is channel imbalance in a headphone amp and the amp also has preamp, will you have imbalance when you connect powered speakers and control the volume with the amp?

If the same pot attenuates the line outputs, then, presumably, yes.

 

[pozz]

If there is channel imbalance in a headphone amp and the amp also has preamp, will you have imbalance when you connect powered speakers and control the volume with the amp?

It is more important for headphones or IEMs, whose higher sensitivities means that small imbalances will be more noticeable.

 

[JohnYang1997]

Because it's not the Alps RK50 that they put in there. Only pot (not DACT nor stepped attenuators) I know that has absolutely no channel imbalance at any volume position

Firstly there will be channel imbalance. You just haven't got enough sample size. Secondly, they don't have 2kohm pot or 5k pot for our use. Let alone 4 gang 2k pots. Plus the size is impossible to put in the small chassis.

 

[solderdude]

http://www.practicaldevices.com/techdocs/Pot_Channel_Matching.pdf

It basically is caused by small mechanical tolerances between wipers and 'resistance tracks'.
The smaller the pots dimensions the harder it is to get it perfectly aligned for 2 or more tracks.
When dimensions are bigger tolerances matter less.

 

[dasdoing]

I have a TC Electronic Konnekt 24D interface. it's analouge volume pot on 1+2 was driving my crazy for years. I now use it's digital pot on outputs 3+4

 

[egellings]

I worked at a company that makes pots of all kinds (not high power ones, though, like rheostats), among other things. The reason for the imbalance at low volumes on dual logarithmic audio pots (volume controls) is that at low volume settings, one degree of shaft rotation causes very little change in resistance. At higher volume settings, that same one degree of shaft rotation causes a much larger change in resistance. It is difficult to accurately screen on the resistor ink uniformly through the entire range of rotation, and tiny errors on the low end of the pot cause much larger volume variation than the same amount of error would at a higher resistance portion of the track would, because the same error represents a larger percent change at the low end than it would at a higher setting of the shaft. I experimented with laser trimming the pots to even this out, but the process was time consuming and not worth it to the company. With a linear taper pot, this problem does not exist.

 

[Berwhale]

With a linear taper pot, this problem does not exist

My Onix OA21s amplifier has linear pots for this reason. However, I was informed by a friend (who worked for Onix) that the ladder mechanism that allowed the volume to be controlled via a rotary control was fragile and that they ended replacing the linear pot with a rotary one for this reason.

 

[egellings]

Stepped attenuators get rid of the problem entirely. Another approach was to use a linear pot with a shunt resistor from the tap to ground. That sorta was loggish, and it got rid of the balance problem, if you did not mind the odd audible 'feel' of the control at higher volumes. Another way around the problem with a regular volume control is to reduce the signal going to the pot so that you have to turn it up higher for normal listening use, thereby getting out of the low setting range.

 

[restorer-john]

Alps RK-27s are excellent. They don't wear out, they are smooth, track well and depending on how much you pay, the tolerances can be really tight between gangs.

Too big apparently for all this miniature HiFi gear, but I'd say this, if you can't fit a decent pot into the casework, it's too small in the first place.

 

[Berwhale]

Hmmm, if it's much easier to avoid channel imbalance with a linear pot (and assuming that these are much cheaper then stepped attenuators), why do we not see linear pots on our headphone amps? Perhaps it's to do with ergonomics and our ability to apply fine motor control to rotational motion (with the fingers) vs linear motion (with our forearms).

 

[Berwhale]

Interesting read: https://sound-au.com/project01.htm

 

[restorer-john]

Hmmm, if it's much easier to avoid channel imbalance with a linear pot (and assuming that these are much cheaper then stepped attenuators), why do we not see linear pots on our headphone amps? Perhaps it's to do with ergonomics and our ability to apply fine motor control to rotational motion (with the fingers) vs linear motion (with our forearms).

By linear, we mean the resistance law on the track taper has a linear relationship to rotational angle, whereas a volume pot has a faux logarithmic track. It's not linear like a fader style pot on mixing console- both linear and log taper pots can be rotary or slide.

 

[Berwhale]

By linear, we mean the resistance law on the track taper has a linear relationship to rotational angle, whereas a volume pot has a faux logarithmic track. It's not linear like a fader style pot on mixing console- both linear and log taper pots can be rotary or slide.

Thanks, I understand the difference between a log and linear pot. However, I had imagined that, given the symmetry of the carbon tracks on the same PCB for each channel, it would be easier to channel match over the whole range with a linear fader style pot. However, after reading the article I linked to above, I realise that this is not radically different to matching the tracks on a the two PCBs in a dual gang rotary pot.

I have read that (as a manufacturer) it's possible to pay Alps for pots with better channel matching (it's mentioned on the JDS Labs blog). Am I correct in assuming that this means someone at Alps tests and picks pairs of PCBs with attenuation matched over the whole rotational range?

 

[majingotan]

Firstly there will be channel imbalance. You just haven't got enough sample size. Secondly, they don't have 2kohm pot or 5k pot for our use. Let alone 4 gang 2k pots. Plus the size is impossible to put in the small chassis.

I agree. The only two amps I've heard with Alps RK50 were the BHSE amp driving Stax SR009 and the Sony DMP-Z1 from a local dealer driving the sensitive CA Andromeda IEM. Even at the lowest registers of my threshold of hearing, at 6:30 position, I did not detect a channel imbalance with that 4-gang volume pot. Wish the dealer had an AP 555B to measure how well the RK50 pot within the Z1 performs in channel balance

 

[Racheski]

I worked at a company that makes pots of all kinds (not high power ones, though, like rheostats), among other things. The reason for the imbalance at low volumes on dual logarithmic audio pots (volume controls) is that at low volume settings, one degree of shaft rotation causes very little change in resistance. At higher volume settings, that same one degree of shaft rotation causes a much larger change in resistance. It is difficult to accurately screen on the resistor ink uniformly through the entire range of rotation, and tiny errors on the low end of the pot cause much larger volume variation than the same amount of error would at a higher resistance portion of the track would, because the same error represents a larger percent change at the low end than it would at a higher setting of the shaft. I experimented with laser trimming the pots to even this out, but the process was time consuming and not worth it to the company. With a linear taper pot, this problem does not exist.

Thank you for this answer, super helpful! This is the layman explanation I was looking for, and now I have the context to better understand the prior more technical posts above.

 

[egellings]

Linear pots are not used for volume controls because of what you experience if you use one for that function. If you start at volume of zero and set the knob to say, 1 on a full scale of 10 with a linear taper control, the sound would already be nearly unbearably loud, and by time you got to 5 or 6, there's be little perceived increase in volume as the control was advanced all the way to 10. The effect is due to the logarithmic relationship of how we perceive sound loudness to how loud the sound actually is. The logarithmic nature of our hearing lets us perceive a huge range of sound intensities, from a soft rain to a thunderbolt, because the log function acts much like a compressor would. We can hear the soft sound and not get overwhelmed by the loud one.