Crackling volume control when potentiometer mounted in aluminum enclosure

[knutolai]

Hi folks!

I'm having an issue with a misbehaving volume control that I can't seem to figure out. First things first, here's my circuit:

Quick rundown of the circuits features:
-The input signal connects to an opamp buffer going into a passive RC network into an inverting opamp volume control
-The MOSFET, diode, resistor and LED is part of a millennium style bypass circuit (see Valvewizards "engineers Thumb: http://www.valvewizard.co.uk/engineersthumb2.html)
-The 78L33 regulator and the three B10k potentiometers serve no function in regard to the audio circuit (they connect to headerpins that are not connected). The blue crosses on the pot wipers are "not connected"-markers.
-The circuit is housed in an aluminum enclosure and the Input and Output jacks are grounded in the enclosure (non-isolated jack sockets)

The circuit works perfectly as long as the volume potentiometer (A100k) isn't mounted in the enclosure. I've tried replacing the potentiometer without any improvement. All three potentiometers I've tried worked fine as long as the metal part of the pot "head" didn't touch the enclosure. I've had plastic knobs mounted on the potentiometer for all the tests. My power supply unit does not power any other circuits (no daisy chaining). I'm using this potentiometer model:


The symptoms; Turning the potentiometer from 0 to 70% produces large amounts of crackling noise. The input signal can be heard in the background, but has a lower volume than when the system works as intended. At around 70% rotation there's a large pop. From 70 to 100% there's hardly any crackle, but the background hiss is more intense than when the system works as intended.

I feel like this is the symptom of some electronics 101 issue that I just never have had to experience. Please let me know if you have any suggestions for things to try!

[GibsonGM]

The symptom kind of suggests that there is DC on the pot, or some path is completed by mounting it in the (grounded) enclosure, no?  *shrug*      You could poke with the DMM to see if that's got something to do with it (?)   
Leaky cap could perhaps account for it.   Just a thought.

[BluffChill]

Does it have a dust cap on like in the picture? If not, is the bottom of the pot touching some of the circuit board and shorting something?

[Rob Strand]

If it works when not attached to the enclosure I doubt it can be a DC issue or a damaged pot.
I'm assuming you are using *exactly* the same build, same PCB and parts in both cases ie. not something that worked on a breadboard and is now put on a PCB and in a box. 

At this point I'd put a 22pF cap across pins 6 and 7 of IC1B.  Located at IC1.  The reason is a pot mounted to a chassis creates stray capacitances from the enclosure to the circuit points at pins 6 and pins 7.  Moisture in the phenolic insulator can increase the problem.     It's hard to reason out but easy to try.  For the experiment try anything up to 100pF but you probably don't want more than about 33pF in the real thing.

Is your enclosure connected to circuit ground?

[knutolai]

Does it have a dust cap on like in the picture?

Pot has a dust cap. Good suggestion.

If it works when not attached to the enclosure I doubt it can be a DC issue or a damaged pot.
I'm assuming you are using *exactly* the same build, same PCB and parts in both cases ie. not something that worked on a breadboard and is now put on a PCB and in a box. 

At this point I'd put a 22pF cap across pins 6 and 7 of IC1B.  Located at IC1.  The reason is a pot mounted to a chassis creates stray capacitances from the enclosure to the circuit points at pins 6 and pins 7.  Moisture in the phenolic insulator can increase the problem.     It's hard to reason out but easy to try.  For the experiment try anything up to 100pF but you probably don't want more than about 33pF in the real thing.

Is your enclosure connected to circuit ground?

You assume correctly in regards to what I'm testing. The enclosure is connected to circuit ground (only) via the jack sockets. I'll test adding capacitors between pin 6 and 7 and report my findings tonight.

[antonis]

A pair of plastic washers and a small piece of heatshrink tube should earn you peace of mind..

[knutolai]

Adding a 22pF cap solved the issue. Thanks a lot! I have two other builds with a very similar volume control configuration where there's no crackling. I should probably add caps for those builds too as a precaution.

[EBK]

Another solution is to hand paint the words "crackle okay" on the face of the enclosure.   

[bluebunny]

Another solution is to hand paint the words "crackle okay" on the face of the enclosure.   

That could set a dangerous precedent:

"Hum okay."

"Inexplicable silence okay."

"Retinal damage okay."

  . . .                                                                                 

[GibsonGM]

At this point I'd put a 22pF cap across pins 6 and 7 of IC1B.  Located at IC1.  The reason is a pot mounted to a chassis creates stray capacitances from the enclosure to the circuit points at pins 6 and pins 7.  Moisture in the phenolic insulator can increase the problem.   

I've never heard of this - something new!  Is this the reason for those pF-range caps in this position in other things like the Guv'nor?   Always considered they were there for 'stability', but I hadn't heard the enclosure itself could add to INstability....cool!  I was thinking more like RF, self-oscillation.   Good call.

[EBK]

Another solution is to hand paint the words "crackle okay" on the face of the enclosure.   

That could set a dangerous precedent:

"Hum okay."

"Inexplicable silence okay."

"Retinal damage okay."

  . . .                                                                                 

Or, "SHO lasting longer than four hours okay." 

[EBK]

My gut tells me that this is a mechanical problem that merely becomes apparent when the wires to the pot are stressed in a certain direction.  Perhaps the resistive track is slightly separating from the phenolic substrate or a cold solder joint.  I'm having a hard time accepting the stray capacitance theory (but I may be wrong about that idea).

If you think it is the enclosure, then you should be able to verify by taking it out of the enclosure and making sure it doesnt crackle, then adding a ground wire to the case and retesting, which should give you your crackle back. 

Edit:  Thinking about this further, I would almost bet money on a bad solder joint as the cause of the crackle.  Adding the cap fixed the problem not because of its capacitance, but rather because adding it required the reflowing of the bad solder joint.

[GibsonGM]

My gut tells me that this is a mechanical problem that merely becomes apparent when the wires to the pot are stressed in a certain direction.  Perhaps the resistive track is slightly separating from the phenolic substrate or a cold solder joint.  I'm having a hard time accepting the stray capacitance theory (but I may be wrong about that idea).

If you think it is the enclosure, then you should be able to verify by taking it out of the enclosure and making sure it doesnt crackle, then adding a ground wire to the case and retesting, which should give you your crackle back. 

Edit:  Thinking about this further, I would almost bet money on a bad solder joint as the cause of the crackle.  Adding the cap fixed the problem not because of its capacitance, but rather because adding it required the reflowing of the bad solder joint.

That was/is my thinking too, FWIW.  I would be excited if a cap in the FB loop could stop such problems, as I've had them before (squeal in box, no squeal out of box, probably parasitics, NOT crackle tho), hence my post 3 above this.     

Contact with a component under the pot, or pushing on something to reveal a loose solder joint....nearly broken wire to  the pot or AT the pot, something miswired putting DC on the pot (ok, that IS a long shot!).   Something ground to the enclosure is almost always MY problem when this happens, anyway.

In my experience it is about 99.99% user assembly error, meaning I do it too  

[knutolai]

My gut tells me that this is a mechanical problem that merely becomes apparent when the wires to the pot are stressed in a certain direction.  Perhaps the resistive track is slightly separating from the phenolic substrate or a cold solder joint.  I'm having a hard time accepting the stray capacitance theory (but I may be wrong about that idea).

If you think it is the enclosure, then you should be able to verify by taking it out of the enclosure and making sure it doesnt crackle, then adding a ground wire to the case and retesting, which should give you your crackle back. 

Edit:  Thinking about this further, I would almost bet money on a bad solder joint as the cause of the crackle.  Adding the cap fixed the problem not because of its capacitance, but rather because adding it required the reflowing of the bad solder joint.

I had the same gut feeling, but if you read my original post it seems very unlikely this was the issue as I've tested several pots. Also my soldering skills are quite excellent 

HOWEVER! I had the circuitboard manufactured as part of a 100x100mm panelized PCB (consisting of several circuitboards that you have to break apart, cheap option for prototyping..). I was a bit stupid when I did the panelization resulting in me having to place more stress than usual on the boards in order to break them apart from each other. When I stacked the boards in a pile I noticed that the pile was a little uneven/unbalanced meaning that the boards were deformed ever so slightly from the process of breaking the boards apart. Maybe this potentially could have caused some traces to partially fracture?

The circuitboards are intended for a workshop I'll be hosting in April. I'm considering to build another one of the boards just to make certain.

Another solution is to hand paint the words "crackle okay" on the face of the enclosure.   

Genius! 

[thermionix]

you have to break apart

Razor saw.  Dremel.  (dust).

100x100mm

Nobody ever says decimeter.

[EBK]

I had the same gut feeling, but if you read my original post it seems very unlikely this was the issue as I've tested several pots. Also my soldering skills are quite excellent 

Yeah, I've been known to be a poor reader sometimes.   

HOWEVER! I had the circuitboard manufactured as part of a 100x100mm panelized PCB (consisting of several circuitboards that you have to break apart, cheap option for prototyping..). I was a bit stupid when I did the panelization resulting in me having to place more stress than usual on the boards in order to break them apart from each other. When I stacked the boards in a pile I noticed that the pile was a little uneven/unbalanced meaning that the boards were deformed ever so slightly from the process of breaking the boards apart. Maybe this potentially could have caused some traces to partially fracture?

Hmm.... Maybe?  I'm less comfortable with this explanation (due to my lack of experience), but you'd better test another board, just in case (pick the most warped one you've got).

[Rob Strand]

Is this the reason for those pF-range caps in this position in other things like the Guv'nor?   Always considered they were there for 'stability', but I hadn't heard the enclosure itself could add to INstability....cool!  I was thinking more like RF, self-oscillation.   Good call.

They are there for stability.  Stability in the sense of the negative feedback stability.
Without the caps there is a risk of oscillation.

Any capacitance on an opamps input pins to ground add a low-pass filter (a pole) in the feedback path and that can cause the feedback to become unstable.   You need these caps even where there's no wiring as the opamp has some input capacitance, you just don't need large cap values.  It becomes more of a problem when the resistors around the opamp are large in value (as the "parasitic" low pass filter filter frequency drops f3 = 1/(2pi RC).    When you have wiring and stray capacitance then it's that extra capacitance that pushes the low pass frequency down.

So that *is* why you see caps across the feedback resistors.   For each situation there's a certain minimum value of capacitance to make it stable.  If you are too close the minimum then extra capacitance from the layout and wiring can push you over again.   You can of course use feedback caps deliberately to form a low-pass filter.   Normally in that case the cap value will be so much higher than the minimum you forget you needed a minimum.

I didn't make this stuff up!

[EBK]

Yes, but crackling ≠ oscillation, right?

[GibsonGM]

I just never considered some sort of capacitance happening with respect to an enclosure...but when you put it that way, Rob, it does start to look logical.

I knew the caps were for stability, but didn't recognize them as a possible cure for that problem (issue with enclosure, which I HAVE had...not crackle).  Makes sense, along with the wiring and all.

Crackling is a new technical term - for what, we'll have to dig deeper   

[Rob Strand]

For an unbiased reference see,

http://www.ti.com/lit/an/slyt087/slyt087.pdf

Yes, but crackling ≠ oscillation, right?

When you move the pot the "oscillator" can stop and start.  The opamp might even enter and exit full clipping.  It's these transistions that cause the small glitches..