Bond it. Tie the pot's metal case to your ground plane — or straight to the enclosure metal — at one point, and keep that bond path under about 25 mm [1 in]. A floating 20 mm [0.8 in] can sits at high impedance to everything around it, picks up ambient noise and feeds it to the nearest traces. Worse, it gives an ESD strike no defined exit, so the discharge picks its own route through your board. An 8 kV air discharge jumps about 2.7 mm [0.11 in] of air, so the gap between case and pins never keeps it floating anyway. What follows is the component-level decision: the bushing, the snap-in cutout, and the single bond point.
Should the metal case of a potentiometer be grounded?
Yes: bond the case at one point to your ground plane or to the enclosure metal, keep that bond path under about 25 mm [1 in], and the 20 mm [0.8 in] can stops being an antenna and becomes a shield.
The resistance track sits on an insulated substrate inside the can, so the bond touches nothing you measure. Wiper, end resistance, taper — every electrical parameter stays exactly as the datasheet says. What changes is stray behaviour: picked-up fields and static charge get a low-impedance way out instead of leaking through parasitic capacitance into your circuit.
Choose one bond point and mark it. Two bonds on opposite ends of the same can invite circulating current around the loop between them, which is the loop you avoid when you tie a board to its chassis once. The whole-board version of the argument is chassis or float: tying PCB ground to a metal enclosure; the part-level version has the same answer.
What does a floating metal body do to nearby signals?
It acts as a small antenna at high impedance: the floating can sits a few picofarads from every conductor around it, picks up mains hum, switching noise and RF, and couples the lot into traces running within about 10 to 20 mm [0.4 to 0.8 in].
Bonded, the same metal works as a shield instead of a source. Fields land on it and drain to ground rather than passing through to the pot's terminals. Published EMC studies of shielding, grounding and filtering report radiated-emission reductions up to 40 dB — a factor of 100 in field strength — from grounded shielding, and the same physics cuts pickup in reverse.
The ESD Association's fundamentals state the principle: an assembly stays safe at a potential well above zero volts as long as everything in it sits at the same potential. One bond puts the can on your plane's potential. That is equipotential bonding, which is a different job from earthing; ESD grounding vs protective earth separates the two.
Can a floating case stay floating in real life?
Almost never: dry air breaks down near 3 kV per millimetre [30 kV per centimetre], so an 8 kV strike arcs across about 2.7 mm [0.11 in], while the space between a pot's case and its solder pins is under 1 mm.
IEC 61000-4-2, the immunity test behind most industrial ESD specs, sets air-discharge levels of 2, 4, 8 and 15 kV and contact levels of 2, 4, 6 and 8 kV. Its contact-discharge calibration drives a first peak near 30 A within about 1 ns.
Put those together and your "floating" can never stays isolated: the strike hops the pin gap or the snap-pin gap no matter what you connect. From there the charge leaves through the nearest conductor, usually the wiper line and whatever input it drives. A defined bond gives that 30 A an exit that never crosses the board — shell to bond to ground, one hop.
How do you bond a panel pot in a metal enclosure?
Through the bushing: the nut clamping the pot's threaded bushing to a bare metal face bonds case to enclosure directly, and the enclosure's protective-earth conductor carries any strike out to supply ground.
This is the best bond a pot can get. The discharge path runs shell to front panel to enclosure to earth, and your PCB is never on the route. Our fronts come in brushed 304 or 316L stainless, bare and conductive as machined, so a nut and locknut on the usual 3/8 in [9.5 mm] panel bushing does the whole job.
On a painted, powder-coated or anodized face, add a star washer or internal-tooth lockwasher so the teeth bite through the coating to bare metal. The enclosure grounding and bonding guide carries the code rule behind it. NEC 250.12 requires paint, lacquer and enamel to be removed from threads and contact surfaces at grounding and bonding connections, and biting hardware is the fitting that makes removal unnecessary. One layer in, grounding a mounting plate inside a metal enclosure does the same job for the backplate.
How do you bond the case in a plastic enclosure?
With copper: connect the plated holes for the snap-in pins to your ground plane, or run a short tab or finger stock from the bushing to that plane, all within 25 mm [1 in] of the pot.
Length is the whole game, and the numbers show why. Take typical values as assumptions — a printed trace carries about 0.7 nH per millimetre (usual range 0.5 to 1 nH/mm), and an 8 kV strike drives a first peak near 30 A with a rise of about 1 ns:
- 25 mm bond path: 25 mm x 0.7 nH/mm = 17.5 nH, and 17.5 nH x 30 A / 1 ns is about 525 V across the bond while the peak flows.
- 100 mm detour to a far-corner ground: 100 mm x 0.7 nH/mm = 70 nH, and 70 nH x 30 A / 1 ns is about 2100 V — enough to arc roughly 0.7 mm (2100 V / 3 kV per mm) into the trace running beside it.
Those are illustrative figures, so swap in your own trace width and stack-up for a real design. The ranking survives any reasonable inputs: a short bond holds the spike near a few hundred volts, and a long one hands the discharge back its freedom.
The bond only moves the strike onto your plane, so give that plane a route to protective earth and keep suppression on the wiper line. ESD protection in plastic enclosures covers the rest of that stack.
When can you leave the case floating?
When three things hold at once: no signal traces run within about 20 mm [0.8 in] of the can, nobody can touch the metal body through a plastic or shielded front, and the product faces no radiated-emissions or immunity test.
A one-off bench instrument behind a plastic panel qualifies. A machine HMI with an operator's hand 50 mm [2 in] from a metal knob does not, and neither does anything booked into an EMC lab. Here is the decision by build:
| Your build | What the case gets |
|---|---|
| Bare brushed stainless front | Bushing, nut and locknut straight to the face |
| Painted, coated or anodized front | Star washer biting to bare metal; wired tab if the coating is thick |
| Plastic box, pot near the PCB | Plated snap-pin holes on the ground plane, path under 25 mm [1 in] |
| Plastic front, no plane within 25 mm | Local copper tab to the nearest plane point |
| Bench unit, plastic front, no EMC test | Floating is acceptable; keep traces 20 mm [0.8 in] clear |
One freebie from the pin layout: your pot's drawing states the square snap-in cutout and the pin positions but usually leaves pad shapes to you. A plated through-hole on at least one snap pin gives the bond and anchors the part against twist at the same time.
How do you put the bond on the drawings?
Call out three items: the square snap-in cutout at the datasheet's stated size and tolerance, a plated through-hole for at least one snap-in pin tied to the ground plane, and one assembly note naming the bond point.
Write the note in plain words — "case bonded to chassis at bushing" — so the panel shop knows a bare contact ring matters more than a pretty one. If the pot mounts through a front we punch, lay the square hole out in the cutout designer along with glands, operators and display openings; the cutouts, glands and HMI windows guide holds the sizing fundamentals. A front from the custom enclosure builder is delivered in 10 business days anywhere in the world, counted from your approval of the drawing, and the stainless arrives bare where the bushing lands.
Catalog dimensions to compare
Compare these catalog dimensions with your component layout. Open each enclosure for its rating, material and usable space, then check those specifications against your application.
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 12 × 10 × 6 in | 304 stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 304 stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 24 × 20 × 8 in | 304 stainless | 24 x 20 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Is the bushing nut enough, or does the case need a wired bond?
On a bare stainless or unpainted metal face, yes: the bushing and nut give a solid metal-to-metal bond, and a locknut keeps it tight. On a powder-coated or thick-painted front the coating insulates, so add a star washer that bites to bare metal or a short wired tab; NEC 250.12's remove-the-coating rule for grounding contact surfaces is the reason.
What ESD level should this bond survive?
Design for the level your spec calls: IEC 61000-4-2 lists 2, 4, 6 and 8 kV contact and 2, 4, 8 and 15 kV air, with 8 kV air the common industrial demand. The bond only has to carry the first peak, near 30 A at 8 kV, for a few tens of nanoseconds, so a nut, a plated hole or a 25 mm copper tab handles it comfortably.
Will soldering a ground tab to the pot damage it?
No. The can is a bigger thermal mass than the pins, so with a 350 C iron you hold 2 to 3 seconds instead of 1 and feed a little more solder; the joint ends up stronger, which helps a part people push and twist. Solder to the mounting tab or the plated snap-pin hole rather than the bushing threads.
Can the case bond and the pot's signal share one ground?
Yes, and that is the arrangement to use: bond the case to the same ground plane the pot's signal reference uses, at that one point. A case tied to a separate quiet ground invites circulating current between the two references, and a 20 mm can sitting between them couples the difference straight back into the wiper line.
