Explosion Proof Cable Gland Installation Guide
A practical guide to installing explosion proof cable glands: barrier compound curing, entry thread engagement, tightening criteria and the inspection record that IEC 60079-14 expects.
Seguir leyendo
Ask an installer whether the armoured cable gland is earthed and you will get a confident yes. Ask how they know the armour satisfies the disconnection time, and the answer usually stops at “the tag is on and it reads low on a multimeter”. Those are two different questions, and only the first one has been answered.
This guide deals with the earthing of the armour specifically: what hardware completes the connection, where the resistance actually accumulates, whether the armour is permitted to act as the circuit protective conductor at all, and how to interpret the test result you write on the sheet. The general theory of earthing and bonding — including how it differs from EMC bonding — is covered in the cable gland earthing guide.
An armoured gland does one electrical job: it terminates the armour to the enclosure. That connection can be perfectly made and still be inadequate, because adequacy is a property of the armour cross-section and the fault current, not of the gland.
This is the distinction that separates a passing installation from a compliant one. The gland contributes a low-resistance joint; the cable determines whether that joint can carry the fault current for the time the device takes to operate.
The armour is clamped by the gland, but the current still has to get from the gland body into the enclosure’s earth system. There are three routes, and they are not interchangeable across enclosure types.
| Route | Where it fits | Strength | Weak point |
|---|---|---|---|
| Earth tag (banjo) under the locknut | General industrial enclosures with an internal earth bar | Visible, replaceable, accepts a fly lead of any practical size | Only contacts through the locknut and thread face; paint or a nylon washer in the stack breaks it |
| Integral earthing nut | Glands supplied with an earthing nut in place of the standard locknut | One component instead of three; less assembly-dependent | Fixed lead size; the nut must match the thread form |
| Dedicated earth boss or screw on the gland body | Ex-rated and heavier armoured glands | Direct metal-to-metal contact on the body; specified in the certificate | Part of the certified configuration — an added tag does not substitute for it |
On a hazardous-area installation the third route is not an option to be improved on. If the gland certificate specifies an earth boss, that is the bonding point, and moving the connection elsewhere departs from the tested configuration.

When an armour continuity reading is high, the instinct is to blame the tag. In practice the resistance is the sum of several interfaces, and the tag is rarely the largest term.
| Interface | What goes wrong over time | What to look at |
|---|---|---|
| Wire to cone and clamping ring | Wires not fully seated, corrosion under the clamp, a wire cut through | Even wire distribution, no bright-cut strands, no white or rust-coloured deposits |
| Clamp to gland body | Thread contamination, partial engagement, body not fully closed | Body seated against its stop, thread clean, no sealant in the joint |
| Body thread to enclosure | Paint or powder coat in the thread, insulating washer, loose locknut | Bare metal at the contact face, washer type permitted by the gland |
| Tag to locknut and wall | Tag fitted outside the locknut, oxidised contact face, tag painted | Tag position in the stack, contact faces cleaned |
| Fly lead and lug | Undersized lead, loose lug, corrosion at the bar | Lead cross-section, crimp quality, torque at the bar |
Because these are in series, one bad interface dominates the reading and hides the rest. Inspect the stack before measuring, and measure at the bar rather than at the gland body.
In many installations the steel wire armour is used as the circuit protective conductor. It is permitted only when its cross-section is adequate for the fault current and disconnection time, and that check is arithmetic, not a judgement call.
| Paso | What you need | Where it comes from |
|---|---|---|
| 1. Fault current | Prospective earth fault current at the point of the fault | Installation design, supply data, or measured loop impedance |
| 2. Disconnection time | Time the protective device takes at that current | Device time-current curve |
| 3. Material factor k | Factor for the conductor material and its initial and final temperature | Your wiring regulation’s table — commonly cited values are around 143 for copper and around 46 for steel armour, but use the table your standard edition publishes |
| 4. Adiabatic check | S ≥ √(I²t) / k, using the armour’s actual cross-section | Calculated |
| 5. Verdict | Armour adequate, or a separate CPC core required | Design decision recorded on the drawing |
The consequence surprises people: steel has a much lower k than copper, and the armour is usually far smaller than the phase conductor. A cable that looks generously armoured often fails the check, and the correct answer is a separate CPC core bonded at both ends rather than a thicker gland.
Two rules follow. The armour is a conductor, so it is bonded at both ends — the single-end rule belongs to shields, as explained in the SWA gland guide. And no choice of gland makes an inadequate armour adequate.
Selection is driven by what the enclosure offers and what the certificate says, in that order.
Whatever the route, the mechanical assembly must respect the entry: the sequence at the enclosure face is in the SWA gland installation guide.
The fly lead is a conductor in the fault path and is sized as one. Two habits cause most problems: a lead smaller than the CPC requirement, and a lead routed long and loose so its impedance — not the armour’s — becomes the limiting term.
Where several glands share an enclosure, each needs its own connection to the bar. Daisy-chaining fly leads puts one circuit’s fault current through another circuit’s tags.
A continuity reading proves a conductive path exists. It does not prove the path can carry fault current, nor that disconnection times will be met.
Write the measured resistance on the sheet and check the design against the earth fault loop impedance for that circuit, not against a fixed number in ohms.

Installation pages often state that armour continuity must be “less than 1 ohm”, or “less than 0.1 ohm”, as if these were acceptance criteria. They are not published limits in IEC 60364 or the UK wiring regulations; they circulate because they are easy to remember.
They also mislead in both directions. A 0.4 Ω reading on a long run can be normal and still satisfy the loop impedance requirement; a 0.05 Ω reading on a joint about to corrode open proves nothing about next year. What matters is whether the total earth fault loop impedance lets the device operate within the permitted disconnection time — calculated from the circuit, not read off the gland.
Practical acceptance therefore has three parts: a low and stable measured resistance at the termination, a calculated loop impedance that meets the disconnection requirement, and a record showing both were checked. Where the armour is also the CPC, add the adiabatic check above.
Armour earthing rarely fails on day one. It fails two or three years later, at an interface that was marginal when it was assembled.
For enclosures where wall type, gasket line or gland plate determines whether earthing survives, read the entry-side details before ordering. If you need armoured glands supplied with tags, shrouds and earthing nuts as a matched kit, send us the cable schedule and the enclosure type.
Yes, for the armour. It is a conductor providing a fault path, so it is bonded at both ends. The single-end rule applies to shields used for EMC, where a second bond can create a circulating current path — a different function with different rules.
Only if its cross-section passes the adiabatic check for the prospective fault current and disconnection time. Because steel has a low k factor and the armour is usually much smaller than the phase conductor, a separate CPC core is often required.
No. It is not a published limit in the main installation standards. The design check is whether the earth fault loop impedance lets the protective device operate within the permitted disconnection time; the measured resistance at the termination is a separate workmanship indicator.
Inside, over the entry thread with the locknut on top of it. Fitting the tag outside the locknut, or between locknut and wall with no clamping force, is the most common reason an armour continuity test fails on a new installation.
Not on its own. A plastic enclosure offers no earth path, so the armour bond must be taken through a metal gland plate or an internal bonding stud connected to the system earth. The tag only connects the gland to whatever it is bolted to.