failed pot drift? R200 -> R11000 (factor of 55 drift)

[coppercone2]

I was repairing something and I came across a potentiometer listed as 200R , with 200 Ohms written on the casing, that measures as 11kOhm.

The potentiometer still mainly works as it should, that is it goes from 0 to 11K, but it should go 0 to 200.

Other then like.. a bit of a stability problem on the low end, it works OK in the high Z range. I overlooked it like 50 times because I neglected to look at the value, until I determined by reading schematics that someone had replaced or put the wrong potentiometer into the circuit, because it explained the behavior, but it seems that the potentiometer is correct.I just spot tested it at both extremes and naturally assumed it was 10k pot, because it went from 0 - 10k without a hitch on a dmm.

How unusual is this? I don't know if I ever saw a potentimeter that basically stayed a potentiometer but mutated in value by 55 times. Like its carbon comp, but damn, 55x increase in value... I expect it to be out of spec by 100-200% or be broke, but functional @ 55x is a bit odd IMO.

This one happened to be setup as a trim rheostat adjusting balance on a leg, so you can imagine my confusion to the circuit behavior! I feel like I don't see these high error factors often. I was becoming more and more certain that some evil shit was happening with the wiring or interconnects.

[Nusa]

If the pot is correct as is, that suggests it's just a labeling problem that they knew about at the time. The fact you're spinning in circles now is irrelevant.

[coppercone2]

Its not correct ,it does not work, you can't perform the adjustment.

its a HP part number with a engraving (dual stage potentiometer), hard to imagine they would screw up that bad and send a product out the door with the wrong pot. It does not pass calibration. It's kind of wonky at 0-1k though, even if you adjust it slowly. I think it drifted. All the other potentiometers read OK though.

I saw really out of spec resistors before, but I always figured a pot would break if that happened, not maintain a confusing level of linearity. And its the main front panel pot too.... I suppose the carbon composition material drifts out of spec, so long its uniform across the substrate, it will just increase in value but maintain linearity. Usually people say like 100% change on bad parts over 50 years, this is 5500%... the usual "way out of spec' thing people say has been way blown out of the water, to the moon.

so it has
1) hand written technicians notes in the chassis, indicating it worked
2) correct HP part # on both potentiometers that are fit together. The other one is fine BTW.

[m k]

Is it open, can you see any lateral cracks inside a coal material, have you tried pressing it?

Pressing a carbon stack was a standard power control thingy back in the day.

[coppercone2]

nah I need to order a new one, when I do I will take it apart.

But this seems.. linearly distributed if its cracks, that is, the adjustment range seems to cover 0-12k quite linearly. I thought a bad pot would 'jump' suddenly to a high value after it traverses a crack.

And what do you mean by pressing it, to open the potentiometer, then press on the material?

[m k]

Yes, if material is damaged so that it's still more or less linear one can think that pressing one part will change the total resistance.
On the other hand, it's pretty thin and if raw material is paste the shale thing is not very plausible.

Maybe it's actually a 10k pot but coated with 200R material and now that coating has changed its address.

[coppercone2]

I did not know they were coated, that seems like it would limit the power dissipation rating.

[coppercone2]

if its fractured, the only thing I can see someone doing is pulling a liquidy nanosilver epoxy into the substrate under vacuum and then pressure treating it, then lapping it. How is taping it down reliable, how do the fractures make a solid again? At best it seems you fill things with rubble

[m k]

Yes, maybe coating was a word too thin, but the final connection is the tip of a viper, not necessary very big area either.

If new device as a whole is available then good, but what if it is not.
Maybe that coal compound in paste form is waiting for a buyer somewhere.

I guess making one's own trimmers would be something to bang one's suspenders.
But how to adjust, having very low or high resistance granules here and there is not very desirable.

When you open it do it above a bowl so nothing will escape and twist the verdict.
If the thing is a scrap you can try bending the viper, maybe other parts of a surface have other values.

[coppercone2]

I finally got around to the 'saucer separation' procedure.

 I hooked it up to a dmm and very gently ran over the carbon with a round drill bit shank.

After a bit of wiping, the resistance jumps to 30k. After a bit more it jumped to 50k.

The stuff on the surface I guess totally failed.

I bet it was heavily used since its the meter needle trim adjustment.

Luckily enough room for two potentiometers next to each other. I can live with it.

I recommend if you have a degraded potentiometer, do not be tempted to wipe the surface, but I am guessing at that point its just failed, I never had problems with other pots before give them a gentle q tipping.

My theory is that there is a bond failure on some old pots. Or it was just very heavily used, not surprizing if this is true, the tobacco STANK in this one, I assume someone might have been smoking pipe and fiddling with it for a few decades.

[wasedadoc]

I repaired a vocoder which had a front panel amplitude pot for each of the 14 frequency bands. When I tried to calibrate, many of the amplitudes were much too large and could not be brought down enough. The chips feeding the pots were current output type with the signal voltage developed using the 100k pot track as load resistor. All 14 measured much higher than 100k, many more than 1M. Had to replace them all.

[coppercone2]

I have seen carbon pot mechanical control systems before. I wonder what kind of shennaigans happen if those start drifting like that.