Electrical cabinet earthing: design, failure modes and the fix

An electrical cabinet leaves the coating booth with a continuous, even, 75 µm film over every external and internal face. That film is the reason the cabinet will last fifteen years in a plant, and it is the reason the first continuity test at commissioning reads 4.2 Ω between the earth stud and the door. Nothing is broken. The earth path was designed out by a coating specification that never mentioned masking.

Electrical cabinet earthing is usually treated as an electrical discipline and specified on the electrical drawing. In a fabricated enclosure it is a coating and sheet metal discipline: which faces stay bare, how they are masked before pretreatment, how the stud is fixed to a 2 mm panel, and how the reading is taken and recorded. The electrical drawing says "earth stud" and the coating drawing says "powder all over"; the gap between the two is where the fault lives.

This guide covers the earth path through a floor-standing cabinet: the stud, the frame, the door, the gland plate and the mounting plate. It sets out what each joint has to achieve, the numbers that decide whether it does, the failure modes that show up in the field, and the masking plan to put on the drawing. The standard behind the practice is covered in IS 3043 for earthing; this page is about building it.

Where does an electrical cabinet earthing path break?

It breaks at one of five joints, and the failure is almost always a coated face rather than a loose bolt. The chain runs from the incoming earth bar to the stud on the frame, from the frame to the door, from the frame to the gland plate, from the frame to the internal mounting plate, and from the frame to any separate instrument panel. Every one of those is a metal-to-metal contact that the pretreatment line and the powder booth will happily cover.

Start with the physics of the film. Polyester powder has a volume resistivity above 10¹³ Ω·cm and a dielectric strength around 30 kV/mm, so an 80 µm film does not break down at the few volts a continuity tester applies. A 20 mm diameter face covered by that film is not a high-resistance joint; it is an open circuit with a stray capacitance. The tester reads whatever leakage the film and the surrounding metal allow, which is why the numbers are erratic: 4 Ω on one cabinet, 40 Ω on the next, 0.4 Ω on a third where a screw thread happened to cut through.

A sound joint is a different order of magnitude. A bolted M8 connection between two bare steel faces, tightened to 15 N·m, measures 0.2–2 mΩ. Against the 0.1 Ω acceptance value that is 50 to 500 times inside the limit. This matters for diagnosis: a reading of 4 Ω is not a joint that needs tightening. It is a joint that does not exist, and no amount of torque will fix a film of powder between the faces.

What the earth path has to carry

The enclosure bond is asked to do two jobs, and only one of them involves fault current. The first is protective: it holds every exposed metal part at earth potential so that a fault inside the cabinet trips the protection rather than energising the door handle. The second is functional: it gives instrumentation, VFDs and communications a common reference, and it shunts interference that would otherwise travel on a signal cable.

The fault job is done by the protective conductor, not by the enclosure. Take a 25 kA prospective fault cleared in 0.1 s. If that current returned through a bond of 0.1 Ω, the voltage across the bond would be 2.5 kV, which is not a touch voltage any cabinet survives. The adiabatic check sizes the conductor: S = √(I²t)/k, and with I²t of 6.25 × 10⁷ A²s and k of 143 for copper, the minimum is 55 mm². That is the PE conductor's job, and it is why the earth bar in a cabinet is a copper bar and not a 6 mm² strap.

The enclosure bond then has a smaller job: stay continuous, survive vibration and corrosion, and never be the only path for a fault current the PE conductor should be carrying. A 6 mm² braided strap is the usual size for a door, and it is chosen for fatigue life across the hinge rather than for ampacity. Where the enclosure itself is the protective conductor, which happens in some machinery enclosures, the adiabatic check applies to the steel and the section has to be sized for it; that is a design decision, not a shop decision.

The masking and bonding limits for the drawing

The table is the set of values we build to for a floor-standing or wall-mounted electrical cabinet with a powder finish.

The stud, the panel and the load path

An M8 earth stud on a 2 mm CRCA panel is a small welded connection with a large job. The weld nugget is about 10 mm across, so the sheet has to tear around a 10 mm plug before the stud leaves. The shear area is π × 10 × 2, or 63 mm², and at 0.6 of the 240 MPa yield that is a 9 kN tear-out. A fault current of 25 kA through the stud for 0.1 s produces an electromagnetic force on the order of a few hundred newtons, so the mechanical connection is not the weak point.

The weak point is the coating around it. A stud that is masked on its face but not around its base leaves a ring of powder where the lug sits, and the lug then contacts only the stud thread. That path runs through the thread, which is masked only if a plug was fitted before coating, and through the nut. It measures 0.3–3 Ω depending on how much powder stayed in the thread, and it is the most common reason a cabinet fails its first test.

The masking plan should be explicit and dimensioned. Mask a 20 mm diameter circle concentric with the stud, both the face and the first 2 mm of the shank, and plug the thread. Mask both faces of every gland plate landing over the full gasket width plus 5 mm. Mask the mounting plate studs and the door strap points. Everything else gets coated. That plan is a drawing note with a symbol, not a verbal instruction to the booth.

A tapped boss in a pressed insert is an alternative where a stud would show on a visible face. The insert has to be welded or riveted into the panel before pretreatment, and the insert itself has to be masked, because the powder will bridge a 6 mm hole in a 1.5 mm panel. Where the panel is aluminium, use a stainless stud or a plated insert: an aluminium stud in a steel frame is a galvanic couple, and the joint corrodes to a white powder within a year in a humid plant.

The door, the hinge and the braided strap

A cabinet door is the largest removable metal part on the enclosure and the one most often found floating. The hinge is a mechanical joint with a pin in a bore, and both surfaces are coated. Even a masked hinge bore leaves the pin coated, and a stainless pin in a coated bore is an insulator with a mechanical job.

Bond the door with a 6 mm² braided strap, 200 mm long, from a masked stud on the door to a masked stud on the frame, positioned so that the strap flexes rather than folds as the door swings. A strap that is too short work-hardens at the crimp and breaks after a few thousand cycles; a strap that is too long loops into the door swing and gets caught. Two hundred millimetres on a 1,800 mm door is the usual compromise.

The strap also fixes the reading. With the strap fitted, the continuity from the main stud to the door is the strap resistance plus two joints, which is a few milliohms. Without it, the reading depends on how many hinges are fitted and how tight they are, which is why two identical cabinets can read 0.4 Ω and 40 Ω on the same day.

The gland plate is a gasket joint and an earth joint

A gland plate has to do two things at once: seal against the enclosure and bond to it. The gasket needs a flat, coated or bare landing; the bond needs bare metal on both faces. The usual compromise is to mask the landing, fit the gasket on the masked face, and bond the plate with a separate stud and strap so that the gasket never has to carry current.

That separation is worth the extra part. A gland plate bolted through a gasket has no reliable metal-to-metal contact at all, and a cable fault that raises the plate to line potential finds its return path through the mounting bolts, which are usually the smallest fasteners on the cabinet.

Corrosion, galvanising and the joints that fail quietly

A bare earth pad is a corrosion site. Once the coating is removed from a 20 mm circle, that circle is the only part of the cabinet without a barrier, and in a humid or chemical plant it will rust unless it is protected. The usual answer is a thin film of petroleum jelly or a zinc-rich touch-up around the pad, applied after the continuity test and before the lug is fitted. Grease is not a coating and does not insulate the joint; a zinc-rich paint does, so it goes around the pad and not across it.

Galvanised enclosures are a different case. The zinc layer is conductive, so a galvanised cabinet can be bonded without masking, but the zinc also corrodes preferentially at a scratch and forms a white oxide that is a poor conductor. A galvanised cabinet that reads 0.05 Ω when new can read 2 Ω after a year outdoors. Specify a masked stainless stud and a plated lug, and the joint survives.

Aluminium is the worst case for a bolted bond. The oxide layer forms in seconds and is an insulator, so an aluminium enclosure needs a masked pad that is abraded and greased at assembly, or a plated insert. A stainless steel stud in an aluminium frame is acceptable with a dielectric barrier at the interface; a bare aluminium-to-steel joint is not.

Testing continuity so the number means something

Test with a low-resistance ohmmeter that injects at least 200 mA, which is the method described in IEC 61557-4. A multimeter on its 200 Ω range injects a fraction of a milliamp and will read a coated joint as continuous, because the film's leakage is enough to move the display.

Take the reading from the main earth stud to the furthest point in the chain: the door, the gland plate, the mounting plate and any separate instrument panel. Record it against the cabinet serial, with the ambient temperature and the instrument's calibration date.

Where the customer asks for a fault-current test rather than a continuity test, the test current and duration come from the assembly standard, not from the fabricator. IEC 61439-1 covers the assembled low-voltage switchgear and controlgear assembly, including the protective circuit and the routine tests; the fabricator supplies a cabinet with a continuous protective circuit, and the panel builder proves the assembly.

What Ganesh runs for earthed enclosures

Enclosures and rack frames are cut on the 12 kW fibre laser with the 40 ft × 16 ft bed, formed on the 200 T, 5-axis brake with the 4,450 mm bed, and welded at the 30+ manual MIG stations or on the three CLOOS cells. Earth studs and tapped bosses are fitted before finishing, the masking plan is applied as a drawing-controlled operation before pretreatment, and the parts run through the seven-tank line into the 18 ft × 14 ft powder booth.

Continuity is measured at assembly and again before dispatch, and the reading goes on the inspection record with the DFT map and the cross-hatch result, under ISO 9001:2015. The same route builds 19-inch rack frames, PDU brackets, blanking panels and wall-mount racks, where the bonding points are part of the rack drawing rather than an afterthought. The data centre rack range lists the family, the paint shop capability covers the pretreatment and booth, and the racks and enclosures hub collects the related guides.

We do not certify the completed assembly to IEC 61439-1 and we do not carry out the dielectric or fault-current tests that belong to the panel builder. What we supply is an enclosure with a documented, continuous protective circuit and a masking plan that survives the coating line.

What to put on the drawing

1. Earth stud positions, thread, material and the bare ring diameter.
2. The masking plan, as a hatched view or a symbol with a note, covering studs, gland landings, strap points and mounting plates.
3. The door and panel bond: strap size, length, termination and the stud it lands on.
4. Continuity value and test method, with the instrument class.
5. Gland plate size, gasket type and the flatness of the landing.
6. The finish system, DFT and whether internal faces are coated.
7. Dissimilar-metal joints and the plating or barrier specified at each.
8. The record to be supplied: continuity per cabinet, DFT map and adhesion result.

Send the drawing with the earth stud layout and the gland plate schedule. The quotation will state the masking plan, the coating system and the continuity record it includes, so the cabinet that arrives is the cabinet that was specified.

Related reading: IS 3043 for earthing covers the standard behind these numbers, and degreasing electrical enclosures before coating covers the pretreatment stage that comes before masking. Send the drawing through the RFQ route.