Offboard Pots/Shielded Cable: Ground to Chassis?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ncoak

Well-known member
Joined
Jul 5, 2010
Messages
93
Location
Oakland, CA
when wiring an offboard pot in the audio path with shielded cable, is the accepted practice to tie one end of the shield of each wire to chassis ground? both ends? or neither?

building the gyraf calrec and all my pots/switches are offboard. the wires will be very short, but there are many of them and i want to avoid potential noise issues down the road. i've read a few takes on this, but am unsure which is best for maintaining a pristine audio path.

any thoughts?
 
I've found it best to ground one end. Ground it to the same place that the tube/transistor is grounded to for that particular gain stage.

If you live by the phrase "ground follows audio" you'll usually be safe. Randomly grounding to the chassis will sometimes get you into trouble.
 
You could put a vactrol onboard, and use the pots to control the LED in the Vactrol (Opto-isolator).

Then shielding is a non-issue.
 
scott_humphrey said:
I've found it best to ground one end. Ground it to the same place that the tube/transistor is grounded to for that particular gain stage.

If you live by the phrase "ground follows audio" you'll usually be safe. Randomly grounding to the chassis will sometimes get you into trouble.

great, thanks for the tip.

vactrol sounds interesting, not sure i want to open that can of worms just yet though!
 
I would NOT do this for a line level device.  The cans are already grounded since they are fastened to the metal chassis.  You can rough-up the inside lip of an anodized front panel if you like a really tight build.

I solder cans only in the case of guitars or FX pedals built in plastic chassis.

Mike
 
it depends a lot on the circuit. Shielded cable adds several pF of C from each conductor to the shield. I recall one mic preamp gain control pot that needed to be mounted over a foot away from the active circuitry. I ended up generating an active (opamp) low impedance common mode shield driver.

Regarding the ground or float for a shield, if everything is treated properly wrt differentials either way should work, if passing signals around single ended you need to deal with keeping references clean.

JR

   
 
ruffrecords said:
Best practice is to connect the shield to zero volts signal at the receiving end only.

Cheers

Ian

not be dense, but in the case of a potentiometer, which end is the receiving end? all 3 shields to the can? or all 3 to a common ground point on the board?

thanks much for all the tips.
 
ncoak said:
ruffrecords said:
Best practice is to connect the shield to zero volts signal at the receiving end only.

Cheers

Ian

not be dense, but in the case of a potentiometer, which end is the receiving end? all 3 shields to the can? or all 3 to a common ground point on the board?

thanks much for all the tips.

Assuming it is a level control it will connect the output of some amplifier via the pot to the input of another amplifier. The input amplifier is the receiving end so you connect the screen to the 0V signal (not chassis) at this end and to the cold end of the pot at the pot (the signal wire being connected to the wiper). You then take a screened cable from the pot - with the signal wire being to the top of the pot and the screen to the bottom (cold) of the pot back to the send (output) amp. You do not connect the screen to anything at the sending end.

Cheers

Ian
 
hmmm...that makes sense. i'm wondering if i'm overthinking this a bit, is shielded wire even necessary in this application? these are only going to be 1.5 inch runs or so, but there are quite a few as the calrec has 8 pots (4 dual gang) and 5 toggles:

http://www.gyraf.dk/gy_pd/calreq/calreq.pdf

sodderboy said:
I would NOT do this for a line level device.  The cans are already grounded since they are fastened to the metal chassis.  You can rough-up the inside lip of an anodized front panel if you like a really tight build.

I solder cans only in the case of guitars or FX pedals built in plastic chassis.

Mike

could you explain this a bit further? do you mean you would not use shielded cable, or not connect any to ground?

thanks everyone  for your insights.
 
ncoak said:
hmmm...that makes sense. i'm wondering if i'm overthinking this a bit, is shielded wire even necessary in this application? these are only going to be 1.5 inch runs or so, but there are quite a few as the calrec has 8 pots (4 dual gang) and 5 toggles:

http://www.gyraf.dk/gy_pd/calreq/calreq.pdf

The Calrec has many pots used for EQ rather than level. It is probably more important for these to avoid stray capacitance than to be screened from interference so as thay are very short runs I personally would not screen them.

Cheers

Ian
 
ruffrecords said:
The Calrec has many pots used for EQ rather than level. It is probably more important for these to avoid stray capacitance than to be screened from interference so as thay are very short runs I personally would not screen them.

good point about stray capacitance...by my count there are 69 connections to be made in total - from what i understand they're all in the audio path so i want to make sure they're as clean as possible.

is interference generally not a concern at such short lengths, even with so many?

of course it would make things a lot easier to just mount them on the board, but the parts required aren't readily available in that format.

thanks for all the insights, it's been very instructive.
 
I've found it best to ground one end. Ground it to the same place that the tube/transistor is grounded to for that particular gain stage.

If you live by the phrase "ground follows audio" you'll usually be safe. Randomly grounding to the chassis will sometimes get you into trouble.

What can work well is to ground [ one side only ] to the psu decoupling cap for that particular tube stage
[ if multiple tubes and decoupling ]  separating the psu grounds from signal grounds regarding hi imp & gain things
like gtr amps .  Certainly you don't want a random grounds anywhere , but a few clip leads and careful about the
volume when testing it , good luck
 
As a practical matter here is what I have learned.  These are newbie level lessons but help me with thinking about design and construction of high quality audio, and they have served me well.  Please forgive these if this sounds like "baby talk" to you.

1) Worry about interference on HIGH IMPEDANCE or HIGH RESISTANCE circuits.  Think of it this way;  when the wire has high resistance... the surrounding air (and any signals therein) start to look like a good path for an electron (this is from the "think like an electron" school of thought").

2) Length matters, wire is wire - There is nothing inherently different between a circuit board trace, a wire, and a resistor lead, or the wound foils in a capacitor or the pins on a potentiometer.  The differences are the presence of shielding (ground planes, twisted pairs, shielded cable) and capacitive coupling (between traces and ground plance, pairs of twisted wire, shielded cable, etc) or coils that can create inductive effects (like the wound foil in a film capacitor).  So as you consider your design, try to minimize the length of high impedance circuits, and you will minimize the effect of both interference, and unintentional filters (a high enough impedance creates a filter in the audio range even with very low capacitance/inductance). So within your EQ for instance, some pots will likely be on higher impedance circuits in order to achieve frequency control with smaller inductors and capacitors.  If you were really worried, you could focus your attention there.

3) Cable impedance is easy to calculate, (spec'ed on quality cables), and are measured in pf per foot.  It is also pretty easy to measure if you have a DVM that measures capacitance.  Just take a bit of cable (use a long bit to get accuracy) suspend a few feet of it in the air twisted or however you intend to run it, and measure the capacitance between the open pairs.  In the case of shielded cable, measure between the shield and the conductor.  In the case of multi conductor cable, there are different capacitances between conductors, and between conductors and the shield, dependent upon how they are wound, and the insulation thickness etc.  Then divide the capacitance you measure by the length in inches to get the capacitance of your 1.5" cable.  That will be a very very very small capacitance.

4) In design, it is nice to keep an eye on everything in the search for perfection, but boring old case size is often more effective that sophisticated shielding and such.  Distance is natures mu metal!  In general the force of the field decreases as the square of the distance http://en.wikipedia.org/wiki/Inverse-square_law#Light_and_other_electromagnetic_radiation.  This means that if you can move your transformer or inductor twice as far from the 1.5" wire, that you will decrease interference by a factor of 4!  If you have a wire running 1/8" by some inductor, and move it to 1" away, by routing the cable better (wire ties and tie downs) you should reduce the effect by a factor of 64! (8 squared).  If you have to make the wire a little longer to do that, you win big in the tradeoff.

5) Easy things to consider.  1) Separate the power supply to a different case, and run only DC into the box.  Failing that... if you don't want to do it.... Make the box deeper or wider, and move the transformer to the far corner, keep AC wiring SHORT on both sides of the transformer and twist the pairs of AC conductors carrying opposite voltages (Twist the mains pair, and the secondary pairs).  Keep both of those twisted pairs short, keep the power inlet and the power supply near the transformer. Use a toroid.  Raise the toroid off the bottom of the case a little if possible. 3) Filter the AC line coming into the case, to avoid a high frequency entry point.  4) Pay attention to grounding, in a consistent way (read up on grounding read the Bill Whitlock papers on the Jensen site) 5) Twist AC and balanced pairs without their ground... and unbalanced conductor with it's ground. Pay attention to studio wiring too, connect equipment with balanced cables, and shorter is better (3 foot not 50 foot).

1.5" wire is very short in the scheme of things.  In my shop I would probably twist it as appropriate but I am willing to bet that if you take 4 steps back and say where is my biggest problem going to come from... your target should be elsewhere.

Forgive me if I have added uselessly to the noise, there are far more experienced folks here than I.  And I have been reading on the site that people want us newbies to shut up. But these ideas have served me well.
 
thanks for taking the time to type that out - some very helpful information. it's tough finding sources which address the practical application side of things, so these types of overviews are invaluable to  'just now learning' folks like myself.

see, i've already learned enough since i started this thread to realize it was a pretty ignorant question to ask in the first place!  8)

cheers
 
Back
Top