Sound of potentiometer materials?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I keep hearing this claim on multiple forums but have yet to see a citation.
Their is a book that someone wrote that did spectral analysts of caps resistors, pots, etc. he said that 2nd harmonic content was very hi in resistors, and if used in a gain stage it is definitely a color.
 
That is a quite interesting claim.
Indeed, carbon resistivity changes with the electric field, as it tends to orientate the molecules, but this is an isotropic effect, meaning it has the same effect whatever the polarity. As a consequence, the effect on signal is mainly 3rd-order distortion.
No study of resistor voltage coefficient shows an asymetric behaviour. Actually they all show the presence of 3rd-order.
However, potentiometers have a significant difference with resistors; where resistors have two symmetrical metal-carbone junctions at the ends of the track, which preserves symmetry, a potentiometer has an additional junction at the wiper. Could it be that this junction has enough rectifying properties to produce this 2nd harmonic? Then distortion would depend on the wiper position, with a maximum at mid-rotation, and minima at both ends. This is assuming the wiper junction is of identical nature as the end junctions.
Not sure, but when replacing the Rs in one of my opamps, the 2nd was very much higher, maybe more so because of the fact that it was in the first stage where the gain was.
 
Not sure, but when replacing the Rs in one of my opamps, the 2nd was very much higher, maybe more so because of the fact that it was in the first stage where the gain was.
show us the data?

I am aware of some cheap resistors that expressed measurable nonlinearity (distortion) due to voltage coefficient when exposed to very large signal swings (like in the feedback network of high power audio amp). At normal line level signals even these cheap resistors were measurably clean.

I can imagine that for a high voltage tube circuit, voltage coefficient in small resistors could again become significant (but I am not the tube guy here).

JR
 
This is getting well away from small signal audio stuff but...where resistors are subject to high voltage surges/spikes/arcs they can change value a bit. Ie testing as normal after high voltage "event". This is tested where I work as it's variable and not always specified. For context I'm talking "High Voltage" resistors subject to tens of kVs.
 
Let's not go down the rabbit hole on wire. I've dealt with plenty of cable types and I hear differences and I'm not getting into any arguments about wire in this thread. We're talking about pots here, let's stick to that. Puleeez.
MM
 
Yep. I keep seeing that claim, but I need to see the actual test results. :)
Voltage coefficient is a well-documented parameter, for exampole here
https://forum.digikey.com/t/the-effect-of-voltage-coefficient-of-resistance-vcr/12229for a little more details, google
VCR Characteristics of High Voltage Thick Film Resistors
I haven't done any measurements, but Ivan at Cosmos has made a number of tests for his ADC, finally chose to use several resistors in series at some points in order to reduce the electric field.
Clearly, since the effect is related to the electric field, the longer resistors have a better performance. His goal was to measure THD at -150dB, so for typical audio use, it may not be strictly necessary.
 
This is getting well away from small signal audio stuff but...where resistors are subject to high voltage surges/spikes/arcs they can change value a bit. Ie testing as normal after high voltage "event". This is tested where I work as it's variable and not always specified. For context I'm talking "High Voltage" resistors subject to tens of kVs.
not extremely high voltage but while I was tilting at windmills designing my outlet tester, to meet a UL insulation criteria I had to deliver high impedance at 500v. SMD resistors in small package size failed to deliver their rated resistance at 500V, in larger SMD packages they worked adequately.

All kinds of parts can behave funny in extreme circumstances. The good news is 99.9% of our world is easy.

JR
 
I have measured distortion in pots. There are significant difference between different pots. But its all way down with new parts. I needed two Radiometer CLT-1's https://linearaudio.net/sites/linearaudio.net/files/CB-to-ES-V1-ref-2-IMG_0001.pdf to do the tests and the residual on these is -170 dB. An Alps, Noble or Bourns conductive plastic can get below -150 dB (10V in). The Taiwan Alpha carbon barely made it to -100 dB. Whether these are significant is clearly a judgement call but the better parts do last longer. Carbon comp resistors measure poorly on the instrument as well. It was invented to sort out passive components that would fail prematurely, important for applications like undersea repeaters and spacecraft. Not for audio.

Pots can be noisy if the gains, impedances and levels are not optimized. A 500K pot at -6 dB is a 125K Ohm source or 50 nV/rtHz or 7 uV over the audio band added to the source noise. Not to mention the impact of the input C of the next stage on the bandwidth.
 
Dave Hill of Crane Song did a test with SMT resistors and some of them showed significant distortion. Resistors shouldn't be so easily discounted.
 
Dave Hill of Crane Song did a test with SMT resistors and some of them showed significant distortion. Resistors shouldn't be so easily discounted.
problems with too small smd resistors is pretty well known... this thread is about pot resistance element media...

having a -170dB measurement residual is better than any test bench I ever touched.

To repeat, in my experience pot resistive elements have not been dominant sources of distortion in SKUs I was responsible for designing, but I have been out of the trenches for a couple decades.

JR .
 
Thats the moulded pots , they came in a sealed version also with a close fitting back , I think they may have been re-branded under various names , like radio spares.
There good pots , high resistance values will shift over time , more than likely up a bit , usualy not a problem in any case ,
Cleaning and re-lubing the old metal shaft ones is possible, they often become very stiff , theres a small metal c clip at the base of the shaft , the nylon shaft doesnt have the same trouble ,

I have a few examples on hand I'll check the manufacturers stamp and run a meter across it and see what it shows , in this case a little old analog VOM meter would shows us best whats going on ,although of course most modern digital meters have a bar graph .

1666952621557.png
 
Heres how it measured up

%
rotation Ohms
0 19
5 2k
10 3.97k
20 7.5k
30 12.6k
40 22.6k
50 46.5k
60 71.0k
70 110k
80 295k
90 492k
95 454k
100 556k

I attached a o-100% scale and chicken head dial for the test , pot mechanical mechanical rotation slightly exceeds the scale at either end , not much going on in the final 5% , all the action is at the low of the dial .
Maybe someone could measure a few other modern pots to see how the curve looks just for comparison .

The pot I tried is in good condition , it has a slightly granular feel which is very very nice ,I suspect its one of the ones with the carbon wiper /moulded carbon track .
I have another few on the test bench now ,
 
Some dust particles needed to be cleaned away with a dry tissue , no need for solvents at all ,
small drop of light machine oil in the bushing and its as good as new ,
The inner wiper also has a carbon tip although its obscured in the scan .

These things are what you find in Dave Reeves era Hiwatt amps ,



1666959754577.png
 
Last edited:
I’m not sure whether he tackled other components, but Steve Bench in San Diego once did an excellent study of the sound of different capacitors.
 
@Jen

I’ve been looking at dual pots lately and the ability to get 5% accuracy from Bourns in cermet across a variety of pot series is attractive. So I can understand your interest and the tradeoff considerations.

Might be worth looking at long life options in carbon, too - I see those in 9mm from BI and Tocos IIRC. There’s certainly a lot more that you can do with a P092 (concentric, center tap, switching) or the Tocos equivalent than a P260 or 51/53 dual. Even Bourns 5/8” pots have their limitations.

I’ve taken to looking at my mimimum resistances (y’know, for any control that doesn’t go down to infinity), and my slugging resistors, as the main tools for accuracy channel-to-channel or frequency-to-frequency, and i find workarounds for differential signals so that i’m never using more than one control element, even if it takes more op amps to do so. Because it’s pretty much a non-starter in a stereo control or SVF to use log pots, I just default to CP linear and think about how to make it work with the surrounding architecture. Makes things more complicated but guarantees that you can use the same pot type across multiple applications in a given project. The BI P260D, for example, in 5K 21-detent linear, has a whole lot of applications for me.

FWIW i have a 24 channel summing mixer with cermet Spectrols from the early ‘90s (148/149, i forget), and you can hear movement. I believe its designers felt that the controls were pretty much set-and forget, and since there are linear faders elsewhere in the chain, that having center be center or center be XdB down without trimming was more important than the ability to rotary pan or fade silently.

[EDIT] I missed the application….cut/boost center tap. Is it specifically related to Vishay P11? I haven’t found any other series that offers a center tap option in CP and cermet.
 
Last edited:
@Jen

I’ve been looking at dual pots lately and the ability to get 5% accuracy from Bourns in cermet across a variety of pot series is attractive. So I can understand your interest and the tradeoff considerations.

Might be worth looking at long life options in carbon, too - I see those in 9mm from BI and Tocos IIRC. There’s certainly a lot more that you can do with a P092 (concentric, center tap, switching) or the Tocos equivalent than a P260 or 51/53 dual. Even Bourns 5/8” pots have their limitations.

I’ve taken to looking at my mimimum resistances (y’know, for any control that doesn’t go down to infinity),
That is called hop off resistance and an important parameter for high gain mic preamps. Ideally you want low single digit hop off resistance and tens of K ohms at full scale. Last century at Peavey we tooled up custom pots for gain circuit with low single digit ohms min and 25K+ max . This required overlaying something like four screened resist layers to accomplish that much range smoothly.
and my slugging resistors, as the main tools for accuracy channel-to-channel or frequency-to-frequency, and i find workarounds for differential signals so that i’m never using more than one control element, even if it takes more op amps to do so. Because it’s pretty much a non-starter in a stereo control or SVF to use log pots, I just default to CP linear and think about how to make it work with the surrounding architecture. Makes things more complicated but guarantees that you can use the same pot type across multiple applications in a given project. The BI P260D, for example, in 5K 21-detent linear, has a whole lot of applications for me.
Way back when I designed a 4 pole SVF based loudspeaker crossover. I got decent tracking from a 4 ganged Alps but still had my technician slug outlier sections to be matched at 50% rotation.
FWIW i have a 24 channel summing mixer with cermet Spectrols from the early ‘90s (148/149, i forget), and you can hear movement. I believe its designers felt that the controls were pretty much set-and forget, and since there are linear faders elsewhere in the chain, that having center be center or center be XdB down without trimming was more important than the ability to rotary pan or fade silently.
customers don't like scratchy pots or clicks/pops from switches with DC. We use more capacitors in the signal path than we'd like to keep spurious control noises low.

JR
[EDIT] I missed the application….cut/boost center tap. Is it specifically related to Vishay P11? I haven’t found any other series that offers a center tap option in CP and cermet.
 

Latest posts

Back
Top