Short answer: a steel raceway around a grounding electrode conductor takes bonding bushings only when the raceway is not physically continuous from the cabinets to the electrode. A short nipple made up through listed fittings into two back-to-back enclosures is one unbroken piece of metal, so your standard connectors satisfy the rule and 0 bushings go on the order. A sleeve pushed through a wall with the conductor loose inside attaches to nothing, so you bond each end to the conductor: 2 bonding bushings and 2 jumpers. The NEC writes the rule for ferrous metal — steel EMT, IMC and rigid steel conduit — because unbonded steel around the conductor works like a choke and raises the impedance of the very path that has to carry fault and surge current. The worked example below puts a 200 A disconnect back to back with a stainless panel and shows both calls.
What the NEC asks of a steel raceway around a grounding electrode conductor
A grounding electrode conductor runs from the grounded terminal at the service to the electrode itself: the water pipe, building steel or a ground rod. It is not the equipment grounding conductor that follows the circuit, and the NEC treats it under its own rule. When you run it through ferrous metal, that rule puts three demands on the metal: the raceway must be electrically continuous from the cabinet where it starts to the electrode it serves; it must be securely fastened to the ground clamp or fitting that holds the conductor; and where the metal is not physically continuous, each end bonds to the conductor itself. One more line extends the reach: bond every intervening ferrous raceway, box and enclosure between the cabinet and the electrode.
Notice what the rule never names: a fitting. It asks for a physically continuous steel path and leaves the hardware choice to you. Standard set-screw and compression connectors landing the raceway in the cabinets create that path on their own, which is why the same nipple that mounts your boxes satisfies the rule. A sleeve floating in a wall cannot, because it lands in nothing. The enclosure grounding and bonding guide sets out the wider framework this rule sits in.
Physically continuous or not: the test that decides the fittings
Physically continuous means the steel runs unbroken and lands in equipment at both ends where the conductor passes through. You get there with the fittings the nipple needs anyway, each making metal-to-metal contact with its enclosure. A sleeve slipped through a wall or a slab attaches to nothing, and the code answers that with end bonds. The calls, in one table:
| Arrangement | The call | What you add |
|---|---|---|
| Nipple through listed connectors into both enclosures | Physically continuous | 0 bonding bushings; the standard fittings carry it |
| Sleeve in a wall, attached to nothing | Not physically continuous | 2 bonding bushings, 1 per end, jumped to the conductor |
| Ferrous pull box on the run | The box joins the bond list | Bond the box into the conductor path, 2 joints |
| Aluminum conduit, any length | Outside the ferrous rule | 0 added bonds under this rule |
Three field details move the answer. An eccentric or off-center knockout can leave the raceway sitting on paint or air instead of metal; the bonding jumper for an eccentric knockout article shows the fix when a knockout will not make contact. A ferrous pull box on the conductor's path takes bond points of its own. And when the raceway is asked to serve as the grounding path itself rather than just surround the conductor, different rules take over; rigid conduit as grounding conductor with NEMA 4X boxes covers that case.
Why unbonded steel chokes the conductor
The conductor earns its keep when surge and fault current move through it, fast and heavy. Lightning-protection literature puts a typical first stroke near 30 kA, and the panels on such services are commonly marked with 10 kA or 22 kA short-circuit current ratings. Every extra ohm and microhenry in that path works against you.
Plain steel is ferromagnetic, with a relative permeability in the hundreds to thousands (magnetic-materials handbook values) against 1 for air and aluminum. A steel tube around your conductor is therefore a one-turn winding on a magnetic core. Current in the conductor drives flux around the tube wall, and that loop adds impedance in series with the conductor at exactly the point where the code wants the least. Bond each end of the sleeve to the conductor and the loop shorts out; the flux cancels, and the conductor performs as if it ran bare.
Our panels sit partly outside this picture. 304 and 316L austenitic stainless reads close to 1 on the same permeability scale; the induced-currents article on ferrous and magnetic stainless maps which alloys are which. The rule's word is still ferrous, and iron is in the name, so bond a stainless sleeve's 2 ends too and the argument closes at 2 bushings.
A 200 A disconnect back to back with a 24 x 20 x 8 in panel
Here is the call with real hardware on it:
- Service: 200 A, 120/240 V single phase (example input).
- Inside panel: a 24 x 20 x 8 in 304 stainless wall-mount enclosure mounted on the wall directly behind the disconnect (catalog size).
- Nipple: 6 in [150 mm] of 2 in EMT, the make-up through a 6 in wall plus two connectors (illustrative input; sized here for the feeder conductors a 200 A service runs).
- Conductor: 6 AWG copper grounding electrode conductor to the water pipe, through the nipple and exposed beyond it (illustrative input for the example).
- Terminations: two standard steel set-screw connectors, metal-to-metal into both enclosures.
The rule in words: the ferrous raceway around the conductor must be electrically continuous from the cabinet to the electrode, and it bonds each end to the conductor wherever that physical continuity breaks.
Count the chain from cabinet to cabinet: disconnect enclosure, connector, 6 in of EMT, connector, panel enclosure. That is 4 metal-to-metal interfaces and 1 unbroken raceway, with 0 intervening ferrous boxes. The chain is physically continuous, and the only steel around the 6 AWG conductor is that continuous nipple, so the rule adds 0 bonding jumpers and 0 bonding bushings. The standard fittings carry the call.
Flip one input and the answer changes. Swap the nipple for a 4 in [100 mm] sleeve in the masonry wall, attached to nothing: continuity fails at 0 fittings into equipment, so the rule bonds each end, which is 2 bonding bushings and 2 jumpers landing on the 6 AWG conductor. Add a ferrous pull box to that run and the box bonds as well, 2 joints of its own. Each skipped bond comes back the same way: a correction notice, about an hour of rework and a second trip.
The threshold, stated once: steel physically continuous through listed fittings means standard fittings only and 0 added bonds; any break in the steel path — a floating sleeve, an off-center knockout, an unbonded ferrous box — means bond each end to the conductor. Keep the half that never relaxes in view too: the clamp at the water pipe and the fittings at the cabinets stay tight, because a loose locknut deletes the continuity the call rests on.
The safety point that belongs here: a service panel stays current-carrying while the service is on, so de-energize and verify absence of voltage before you open either enclosure to work on the conductor.
What changes on a real panel order
A painted carbon-steel box needs paint back to bare metal at every bonding contact; NEC 250.12 requires paint, lacquer and enamel removed from threads and contact surfaces at grounding and bonding connections. Our 304 and 316L stainless boxes come bare, so each connector lands metal on metal as delivered. They are manufacturer-rated NEMA 4X / IP66, not UL listed; NEMA writes the type definitions but does not test, certify or endorse enclosures.
Where the design lands the conductor on a ground bar, keep it separate from the equipment grounding conductor bundle; backplane bonding versus ground bar explains why the two stay apart. Where a bonding jumper is called for, the NEC sizes it through its bonding-jumper tables; the sub-panel bonding jumper article works the two sizing rules with numbers. And where the conductor enters a box wall on its own, the grounding conductor entry article covers the connector and the bond.
One boundary keeps the inspection short. A separate NEC rule bonds the metal raceways that carry service conductors, and inspectors often look for a bonding bushing or a threaded hub on those runs. The nipple in the example carries load-side feeders and the grounding electrode conductor, so the ferrous-raceway rule above is the one that decides.
What to put on the drawing
- Nipple: 2 in EMT, 6 in [150 mm] make-up, landed with a set-screw connector into each enclosure; bonding bushings under the ferrous-raceway rule: 0.
- If the wall detail forces a sleeve: 2 bonding bushings, 1 per end, each jumpered to the 6 AWG grounding electrode conductor.
- Conductor: 6 AWG copper, kept out of steel wherever the route allows, clamp at the water pipe tight.
- Question-killer alternative: run the conductor exposed on the wall surface or through aluminum conduit; 0 ferrous surround, 0 added bonds.
- Nipple build details: hubs, spacing and support live in extend a junction box off a panel with a conduit nipple.
- Box: pick from stainless wall-mount enclosures or send your cutouts through custom enclosures; the order lands in 10 business days from drawing approval.
Order the nipple made up rather than sleeved, and the bonding-bushing question closes itself: 0 bushings, standard fittings, and a steel path the inspector reads at a glance.
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 |
|---|---|---|---|
| 24 × 20 × 8 in | 304 stainless | 24 x 20 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 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 |
Questions
The inspector asked for bonding bushings on a continuous nipple — do I add them?
You can, and it never hurts: bond each bushing's jumper to the grounding electrode conductor and the run keeps everything it had. It costs about an hour to pull one connector per end on a 6 in nipple. Under the ferrous-raceway rule, though, a nipple made up through listed fittings is physically continuous, so 0 bushings are required there; the ask usually traces to a sleeve or a service-side rule.
Does aluminum conduit around a grounding electrode conductor need end bonds?
No added bonds under the ferrous-raceway rule. The NEC writes it for ferrous metal, and aluminum sits at relative permeability 1, so there is no magnetic choke around the conductor to short out. A 2 in aluminum nipple between the same two cabinets needs only the fittings the nipple itself requires. Steel and stainless sleeves still take 2 bonding bushings, 1 per end.
Can the grounding electrode conductor run exposed instead of in raceway?
Yes, and it removes the question entirely: 0 raceway, 0 bushings. A 6 AWG copper grounding electrode conductor may run exposed along the building surface where it stays clear of damage, and it needs protection only where a hit or abrasion is likely. Where it must cross a wall, keep it out of ferrous metal or bond the sleeve's 2 ends.
What breaks the physical continuity of the raceway?
Three things show up most: an eccentric or off-center knockout that leaves the raceway without metal-to-metal contact, a ferrous pull box that joins the run between the cabinets, and a raceway that stops and restarts instead of running through. The knockout and the restart mean end bonds at 2 points; the pull box takes 2 bond points of its own.
