Troubleshooting potentiometer problem

MrChips

Joined Oct 2, 2009
30,706
10μF electrolytic capacitor will be ok for testing for now.

189Ω or 200Ω is not much different. Even 100-1000Ω should work.
At this point we don't know what is giving you the symptoms that you have reported. Only further investigation will give us more clues.
 

MisterBill2

Joined Jan 23, 2018
18,167
The problem of disturbing the grid bias would get worse as the control was advanced, and matches the complaint about the level dropping a the control is advanced past half way. No clue as to what you would see on a scope.
 

KeithWalker

Joined Jul 10, 2017
3,063
Again, thank you all for your help and the generous explanation. Regarding the resistor and cradle, that is how I found the impedance above. I do not plan on using the headphone out for a variety of reasons. (If my math is correct, it is closer to 189 Ohms input impedance). And regarding the dc offset, I recently acquired an oscilloscope that registers dc coupling and there seems to be none showing up as the center of the 1 k signal sits at 0. I did find a 10 uf electrolytic cap in my tiny spare parts box. I read that one should use only polypropylene or metalized poly caps for input coupling…will the polarized cap be ok for test purposes?
No, don't use a polarized capacitor because there is no DC offset. The signal is feeding into a very high input impedance so you can use 0.1uF or even a little less. A value lower than that will not pass the lower frequencies but it would test if the potentiometer works correctly.
 

Thread Starter

garyh

Joined Dec 20, 2008
17
To update: I put a 1 uf poly cap in series with the input. There was no change in the pot behavior in circuit…resistance to ground increases to about 5k Ohms with CW rotation and then steadily returns to 0 Ohms. My next guess is to bypass all the low voltage connectors, switches, etc. to rule out any bad solder joints or other shorts. i should also post a schematic of the preamp which I will do presently
 

MisterBill2

Joined Jan 23, 2018
18,167
The only explanation for that resistance change is that both ends of the pot are tied to each other and to the common. That can happen if a solder bridge is formed between the top terminal and the frame of the pot.So the next step is an ohm meter check with all power off. One meter probe on the circuit common point and then the other probe checking each of the terminals. The control should be set about half way up to do this check.
 

KeithWalker

Joined Jul 10, 2017
3,063
Why would yo expect the resistance to the slider of the pot to measure any different from the way it does. You have about 200 Ohms connected between the high end of the pot and ground so that is what you measure when the slider is at the top. When it is half way round, you have two halves of the pot in parallel connected to ground.
The measured resistance is not the problem. How well does the pot control the volume with a series capacitor?
 
Last edited:

Dave Lowther

Joined Sep 8, 2016
224
So that you can picture what Keith is saying in #26 and because this wasn't obvious to me when I first came across it.
I assume you are doing the resistance measurement like this, i.e. with the iPod connected and no series capacitor between the iPod and the pot. I'm not sure how you currently have pin 2 of the pot connected to the amplifier, but assuming you have a series capacitor between pin 2 and the amp, and / or the input resistance of the amp is much greater than 20K then my explanation should be valid.
1625491467216.png
When the pot is in the mid position, as shown above, you will measure 10K (bottom half of pot) in parallel with 10.2K (top half of pot + iPod DC resistance), or about 5K (10K in parallel with 10K).
When you move the pot in either direction from the centre position one of those two 10K 'resistors' shown inside the pot will reduce in value. For example at 1/4 way from either end stop, ignoring the iPod 200R you will have [Edit corrected values] 7.5K 15K in parallel with 2.5K 5K, which is about 3.75K. If you move the pot all the way to the GND end you should measure about 0R, and at the other end about 200R. BTW: I'm assuming you have a linear pot, not a log pot. With a log pot the max value of 5K would not be at the centre position
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,167
OK, I had not considered that the source would still be connected when the meaurement wa made, but certainly the explanations in pots #26 and #27 are correct. But that does not, in itself, explain the drop in volume as the control is advanced. It does demonstrate the reason that a capacitor is required between the control wiper and the grid circuit.
 

Thread Starter

garyh

Joined Dec 20, 2008
17
This is great input. Thanks. So my first mistake has been that I put the 0.1 uf cap before the pot. I will try again today and edit Dave’s schematic to reflect cap after the wiper and measured voltages and resistances. I appreciate you all hanging in with me. Btw, it is a log pot, and the 5 k intermediate resistance occurs at about 1-2 o,clock…which would follow. I think.
 

Thread Starter

garyh

Joined Dec 20, 2008
17
Today I disconnected pot from chassis and removed connections to input selector and input/output jacks. Ipod is jumpered to pot, pot is jumpered to input on PCB. With o-scope attached to input and output on pcb, everything looks fine and the pot behaves as expected. The signal present at full-off is gone and the voltage increases in a steady fashion through the pot's range. I believe I owe you all an apology for a goose chase, probably averted by more experienced test practices. My guess is that the input selector or pot are somehow making connection to the chassis. Because the chassis is grounded to earth ground at one star point, as is the signal ground (same point), I still don't understand how increasing the volume could result in a louder-then-softer decibel level as I described before. Anyway, from here I will reassemble and test at every solder point. I would like to thank you all again, and notice how satisfying it is to reach out to a community and get such interest and support in return. Wonderful.
 

MisterBill2

Joined Jan 23, 2018
18,167
The symptom as described was of a low impedance between the "max" end of the variable resistor and the signal common, plu the signal in and out both being tied to the wiper terminal. Not saying that is how it was, but that such a connection would cause what you observed. Reality can be confusing at times, and this is far from the first time. There was a Sony reel-to-reel recorder/player model that would develop a power supply hum. Replacing the obvious components in the power supply would not fix it. Instead, replacing the volume control would cure the hum every time. That never made sense to me, but it fixed four of those machines.
 

KeithWalker

Joined Jul 10, 2017
3,063
The problem of "louder towards the middle of the pot" was caused by connecting it directly to the grid input. At mid position, there was ,maximum resistance to ground. As you varied the pot either way, you reduced the resistance and upset the bias voltage on the grid, reducing the gain..
 

MrChips

Joined Oct 2, 2009
30,706
The problem of "louder towards the middle of the pot" was caused by connecting it directly to the grid input. At mid position, there was ,maximum resistance to ground. As you varied the pot either way, you reduced the resistance and upset the bias voltage on the grid, reducing the gain..
That may be true but I don't know if it is the whole story.
The grid to cathode bias is set by the cathode to ground resistor. The voltage on the grid is at 0V through the 1MΩ resistor from the grid to ground.
I would think that it has more to do with the impedance to ground (a short to ground via the low output impedance of the source. (I would like to test this theory with my tube breadboard that I have.)
 

MisterBill2

Joined Jan 23, 2018
18,167
That may be true but I don't know if it is the whole story.
The grid to cathode bias is set by the cathode to ground resistor. The voltage on the grid is at 0V through the 1MΩ resistor from the grid to ground.
I would think that it has more to do with the impedance to ground (a short to ground via the low output impedance of the source. (I would like to test this theory with my tube breadboard that I have.)
The entire audio source voltage is developed acros that low source resistance and so it would be present at the end of the control resistance. So while the impedance will be low, the signal voltage would be a maximum.
 
Last edited:

KeithWalker

Joined Jul 10, 2017
3,063
That may be true but I don't know if it is the whole story.
The grid to cathode bias is set by the cathode to ground resistor. The voltage on the grid is at 0V through the 1MΩ resistor from the grid to ground.
I would think that it has more to do with the impedance to ground (a short to ground via the low output impedance of the source. (I would like to test this theory with my tube breadboard that I have.)
The cathode to ground resistor does cause a negative potential on the grid when current flows through it but that is not the whole story. With a very high value resistor connected between the grid and ground, the electron emissions from the cathode will result in a small grid current that flows through the resistor to give a volt or so of negative bias on the grid. It's called "grid current biasing". That will be shorted to ground if the resistor is made a much lower value, leaving only the cathode resistor to bias the grid.
That is what is happening in this case. You can verify it quite easily if you breadboard a circuit. Measure the anode voltage with a 1 Megohm resistor connected from grid to ground. Observe the change in anode voltage when you connect a lower value resistor (200 Ohms?) from grid to ground.
 

MisterBill2

Joined Jan 23, 2018
18,167
Based on what I see in post #31, it seems that there was either a wiring error or an unknown connection present that has been removed by taking the parts off of the chassis. So now it should be possible to reassemble the installation while carefully examining each step to assure that no unintended connections are made. Rather tedious but not so very difficult.
 

Thread Starter

garyh

Joined Dec 20, 2008
17
I have resoldered all the connections between input jack-input selector switch-pot-signal in on PCB and all is working...and doing so without the 1uF cap. Almost definitely it was a wiring/soldering issue. Your collective attention to address the problem may have made the difference between me putting in the time to mend things, and just quitting. I prefer the first choice for many reasons. BTW, not sure if this connects with the grid bias issue, but the cathodes are referenced to 1/4 plate voltage via a voltage divider and cap, which, I believe is in order to make sure the heater/cathode relationship in each triode is separate. I am attaching the schematic (minus PS). Thanks again, I hope to hunt you all down soon with my next amateur mistake.
 

Attachments

MrChips

Joined Oct 2, 2009
30,706
I tried to replicate your situation with a tube circuit and cannot get it to behave as you experienced.
I used DC coupling with the signal from a function generation with 600Ω output to the grid of the triode.

I have not studied what you have done to remedy your situation. It would appear to me that you did something wrong with your initial setup.
 

MisterBill2

Joined Jan 23, 2018
18,167
That circuit shown is often called a "cascode" connection. Having four stages direct coupled is quite a challenge. The reason is that having each stage at the correct bias point for linear operation depends on the previous stages being at the correct point.
Wiring errors and wrong part values are the main source of problems in new circuit construction.

Understanding how a circuit is supposed to operate is of great value in searching for the cause of incorrect operation.
 
Top