routing tracks for volume potentiometer

[JAY X]

Hi all!

Maybe it is a neurotic question...

Say you have  3 pads for 3 pin molex connector  on the top layer pcb, for the 3 wires of a potentiometer...

Regarding crosstalk, ¿would it be better place the ground (thermal pad) in the middle , and have the Input/output tracks in the sides of the connector?

JAY X

 

[ruffrecords]

Ground in the middle will minimise crosstalk across the connector but I suspect that is not the major source of crosstalk. There is likely to be more capacitive coupling in the cables leading to/from the potentiometer.

Cheers

Ian

 

[JAY X]

Hi ian!

Ok, So,  twisting the wires coming from or going to the potentiometer, may help reduce capacitive coupling?

 

[JohnRoberts]

Crosstalk there is generally inconsequential.. I would put the wiper in the middle, but don't expect it matters....

JR

 

[joaquins]

Hi ian!

Ok, So,  twisting the wires coming from or going to the potentiometer, may help reduce capacitive coupling?

More likely to increase it... I don't think Xtalk is the name for something leaking into the same signal path, maybe just leaking. 4 wire connection (kind of differential) as JR usually says may help more in that aspect than caring about few pF on the wires (I guess you are not running the fader 6 feet away from the board)

JS

 

[ruffrecords]

Hi ian!

Ok, So,  twisting the wires coming from or going to the potentiometer, may help reduce capacitive coupling?

Twisting the wires will probably increase the crosstalk by increasing the capacitive coupling.  Anyway, you should be using screened cables to and from the pot. Sometimes you can get away with using a single twin screened cable but this does mean there could be crosstalk between the the two wires inside the screen. Ideally you should use a single screened cable to the pot and another single screened one back from the pot.

Cheers

Ian

 

[JAY X]

Hi  again!

Ideally you should use a single screened cable to the pot and another single screened one back from t

So this means to have a 4 pin molex on the pcb, with ground/output ---> to the pot input,  and  ground/input ---> to the pot output.
right? ???

Jay X                                                                                                                 

 

[ruffrecords]

Hi  again!

Ideally you should use a single screened cable to the pot and another single screened one back from t

So this means to have a 4 pin molex on the pcb, with ground/output ---> to the pot input,  and  ground/input ---> to the pot output.
right? ???

Jay X                                                                                                               

It means connecting ground/output to the bottom and top of the pot using a screened cable. Then a separate screened cable connected to the wiper and the bottom of the pot and with just the wiper lead connected back at the connector (the screen at that end being left unconnected).

Cheers

Ian

 

[JohnRoberts]

Sorry this is probably over-inspection but I like use a 4 wire connection . Pot bottom, wiper, and top on separate leads. Shield attaches to pot bottom at pot, and ground at the circuit board.  So pot current flows through the cable shield to ground, while the pot bottom wire is not carrying any current, just a pot 0V reference voltage.

Receiving the pot wiper differentially with wiper relative to pot 0V will deliver superior pot kill. Like wise the signal can be sent to the pot differentially relative to pot 0V. 

You may notice I am big on using differentials coming and going.

I expect crosstalk within pot wiring to be secondary to other possible signal corruption vectors.

JR

 

[JAY X]

Hi All,

After your suggestions attached are two layouts, the first one (left) as i usually do, and at the right the new layout  following your suggestions...maybe i did not understood all the explanations... :

 

[JohnRoberts]

I can't see all the connections on version two...

JR

 

[JAY X]

Hi John!

maybe because of the constrast with the green back...hmm... tomorrow i will post a more contrasted version....

jay x

 

[joaquins]

You shouldn't connect the ground on the return (the one which goes wight he out) to the ground plane, you should use it as input signal for a differential amplifier, together with the out signal.

JS

 

[ruffrecords]

I am a little confused. Is this a little PCB on its own that just holds the pot?

Cheers

Ian

 

[JAY X]

Hi!

Yes, it is part of a single pcb  that holds 2 potentiometers and 2 switches and 2 leds.

I just draw the potentiometer part to make it clear. The potentiometer is placed between  a low impedance opamp  output stage and the input of an opamp fader amplifier.

 

[JAY X]

another picture more detail...

 

[JAY X]

and just the pot board

 

[JohnRoberts]

OK now I understand my confusion... you are using the ground plane for the pot 0V.... not how I would do that. I was looking for actual traces.

My preference is to actually use two traces for the pot 0V. A fat one for the ground and current , a more normal sized one for the reference voltage. Kind of like a sense line in precision meter (google Kelvin contact).

Using the ground plane is brute force hoping everything is 0V, I prefer to use differential circuits to finesse errors from trace voltage drops or whatever.

Since the pot is between two active op amps stages you could built differentials into both send and receive to reference between sundry local grounds.

JR

 

[JAY X]

Hi,

I don`t understand very well the solution related with kelvin contact...and implementing differentials, maybe with an opa1632...buff...

Attached is a way of shielding  the conductors to a potentiometer, with two shielded wires ( gnd + out on a shielded wire, and gnd + input on another shielded wire) and both shields connected together at both ends... (henry ott).

 

[joaquins]

The shield of the return from pot is a 0V signal, not a ground. At the PCB you should connect it to the input of the next stage, not to the ground.

JS