Rubber O-ring between locknuts on rigid conduit: does it seal

Short answer: the rubber O-ring squeezed between two locknuts neither seals nor bonds, and in a Class II Division 2 room both failures count. Pull the ring and fit a listed bonding hub made up wrenchtight, or keep the double locknuts for holding and add a bonding jumper sized from NEC Table 250.122: 10 AWG copper for a 60 A branch, 8 AWG at 100 A. The NEC bonding rule for Class II locations does not depend on locknut contacts to carry fault current, and no standard tests a field-cut rubber ring as a seal. The joint fails twice at once. The elastomer lets the conduit rock a few hundredths of an inch, so dust walks in. The fault path, meanwhile, runs through paint and rubber instead of copper. Here is what to check, what the code accepts and what to order.

What each symptom points to

Work down the signs on your machine in this order; the first two show up on any dusty line.

• Dust in a clean arc on the backplate under the entry. The ring has shifted, so the leak is the conduit joint, not the door gasket. Wipe the area, run the machine for a week and look again.
• The conduit rocks at the wall. Movement you can see by hand, and 0.02 in [0.5 mm] is already obvious, means the stack clamps through rubber. A hubbed joint shows none.
• An ohms reading where there should be none. Conduit-to-ground-bus resistance above 0.1 Ω means the path crosses paint or the elastomer. That 0.1 Ω figure is the pass mark ground-continuity tests use.
• Locknut flats that turn under light hand pressure. The joint was never pulled up, so the threads carry no clamp load at all.
• Rust or powder under the outer locknut on a coated wire trough. The coating crushed under the nut is the same break that kills the bond.

Why the joint fails, most likely cause first

Four things break this stack, in this order of likelihood.

1. No controlled squeeze. A designed face gland compresses the O-ring cord by about 20% of its section (typical O-ring gland design), and the squeeze numbers for a sealed enclosure face show how that share is set. A locknut pair gives whatever the thread run-out leaves, and every retightening changes it. Axial loading is the harder case to control, which the page on axial versus radial O-ring sealing on bulkhead fittings works through.
2. The ring twists and walks. Tightening drags the rubber around the conduit, and vibration then spirals it out of the gap a fraction at a time. Once it moves, dust crosses and the clamp load drops with it.
3. Coatings under the locknuts. NEC 250.12 has paint, lacquer and enamel removed from threads and contact surfaces at grounding and bonding connections. On a painted wire trough the locknuts sit on coating, so the raceway-to-box bond is paint pressing on paint.
4. A hole cut oversize. A field hole a couple of millimetres over the fitting size lets the ring extrude into the gap. The clamp load then lands on the conduit threads alone.

What the code accepts at the wall

Four rules decide this joint, and they point the same way.

• The NEC bonding rule for Class II locations wants bonding connections made by threaded bosses or hubs pulled up wrenchtight, by bonding bushings with locknuts, or by bonding jumpers. Locknut-bushing and double-locknut contacts are not depended on for bonding there.
• The NEC hazardous-location bonding rules say the same generally: a double-locknut contact alone is no dependable bonding path, so the joint gets a jumper or a hub.
• The NEC connection standard for raceways carrying over 250 volts to ground sets the bar these joints have to meet.
• NEC Table 250.122 sets the jumper size, and the grounding and bonding guide carries that path end to end.

Which hub bodies count as listed bonding connections is sorted out in the piece on listed conduit hubs for bonding. The accepted means look like this:

Whether the raceway itself can act as the grounding conductor at a NEMA 4X box is its own case; the piece on rigid conduit as a grounding conductor covers it. Whether a Class II Division 2 run needs a conduit seal at all is another one again; the Class II Division 2 sealing piece covers that. For how the manufacturer rating of the box itself reads against IEC codes, see the NEMA 4X and IP66 comparison.

Check the joint in ten minutes

Lockout the feeder and confirm the circuit is dead with a meter before you open anything.

1. Continuity. Measure ohms from the conduit at its far end to the box's ground bus. At or under 0.1 Ω passes; 1 Ω or more means the path runs through paint or rubber.
2. Rock test. Grip the conduit at the wall and push. Movement you can see is the rubber carrying the load.
3. Look at the entry face. The ring, washer and outer locknut sit exposed outside. A hub shows a gasket, a nut and a flat metal face.
4. Trace the dust. Powder arcs inside, directly under the entry, confirm the path even when the ring looks seated.

The fix, step by step

1. Lock out, and confirm the circuit is dead.
2. Back off the inner locknut and strip the ring, washer and outer locknut off the conduit.
3. Open the hole to the hub datasheet size; a 1 in hub typically mounts in a 1.375 in [35 mm] hole on common hub datasheets. Deburr both faces.
4. Fit the listed bonding hub: gasket to the outside wall, contact bar and screw inside, body pulled up wrenchtight.
5. Where the hole must stay as cut, keep the double locknuts for holding and add the bond separately: a bonding bushing at the conduit end, a jumper to the ground bus, landed on bare metal per NEC 250.12.
6. Treat the wire trough the same way at its end of the run.
7. Re-measure conduit to ground bus; at or under 0.1 Ω, the job closes.

For weather-exposed walls, the hub-versus-locknut decision gets its own comparison, outdoor top entries included.

Sizing the jumper: a 60 A branch in a 24 x 20 x 8 in box

Work the numbers once and copy the pattern; swap in the ratings from your own job.

• Box: 24 x 20 x 8 in outside (H x W x D) in 304 stainless, the catalog outline behind this link: 24 x 20 x 8 in NEMA 4X wall-mount enclosure.
• Overcurrent device: a 60 A breaker ahead of the raceway (example input; replace it with the rating on your job).
• Raceway: 1 in rigid metal conduit entering the side wall.
• Jumper run: 3 ft [0.9 m] from a bonding bushing at the conduit end to the ground bus.

Per NEC Table 250.122, a 60 A device ahead of the run calls for 10 AWG copper, 5.26 mm².

Two checks on the result:

• Resistance. Standard copper tables list 10 AWG at 1.21 Ω per 1000 ft. The 3 ft jumper therefore adds 3 x 1.21 / 1000 = 0.0036 Ω (3.6 mΩ), under 4% of the 0.1 Ω pass mark. The conductor is never the weak link here; the contacts are.
• Growth path. Keep 10 AWG while the device ahead stays at 60 A or less. The table moves to 8 AWG at 100 A and 3 AWG at 400 A.

Decision: a 10 AWG green copper jumper from the bushing's lug to the ground bus carries the bond. The hub or the locknuts carry the mechanical load. Both ends land on bare metal, and the final reading stays at or under 0.1 Ω.

Keeping the rating at the new entry

A wall entry keeps the box's rating only through a listed fitting of equal rating, fitted the way its instructions show, because the weakest part sets the assembly rating. Our wall-mount boxes are manufacturer-rated NEMA 4X / IP66, not UL listed, and NEMA does not test or certify enclosures. We build the enclosure and the cutouts, not certified Class II fixtures; the classified-location duty sits in the listed hubs, the bonded raceway and the installation around them. On a box you order with factory cutouts, every hole arrives at the size on your drawing with edges deburred, so draw each hole at its hub datasheet size. Send one hole per hub and the box is delivered in 10 business days, counted from your approval of the drawing, through custom stainless enclosures with factory cutouts.

When to re-specify the enclosure

• Dry indoor dust leaves the 304 stainless outline as it is. Wet dust, washdown or chlorides in the same room move you to 316L in the same outline: the 316L 24 x 20 x 8 in wall-mount enclosure.
• A crowded wall, with entries crowding the feet, the hinges or each other, steps up through the catalog outlines from 20 x 16 in to 60 x 36 in: wall-mount enclosures.
• Repeated bond failures across several boxes are a specification problem, not a repair problem. Put the line on every future order:

Enclosures in this article

Questions