Flamepath Integrity: Why the Enclosure Is the Safety Device

The Enclosure Is Designed to Leak

Two things must be true when the internal explosion occurs.

First, the housing must survive it. Internal pressure spikes to several times atmospheric in a fraction of a second, and the walls, cover, and fasteners have to contain it without permanent deformation or rupture.

Second — and this is the part nobody explains — the burning gases must be allowed out, but not allowed to ignite anything. A perfectly sealed enclosure would simply become a bomb. Instead, the pressure vents through deliberately engineered gaps between mating metal surfaces: the flange joint where the cover meets the body, the threads where a conduit entry screws in, the spigot where a shaft passes through.

As the hot, burning gas is forced through those narrow gaps, three things happen at once. The gas transfers heat rapidly to the large surface area of metal on either side. It expands and loses energy. And the gap is narrow enough to quench the flame front outright. What emerges on the outside is hot gas — but hot gas below the autoignition temperature of the surrounding atmosphere.

That machined gap is the safety device. Not the steel. Not the gasket. The gap.

Which is why a technician who “improves” a flame path by cleaning it up with emery cloth, or seals it with silicone, or paints it for corrosion protection, has disabled the one feature the whole certification rests on.

Where the Numbers Come From: MESG

The permitted gap isn’t arbitrary. It derives from a measured property of each gas called the Maximum Experimental Safe Gap (MESG) — determined in a standardized apparatus that ignites a mixture in an inner chamber and measures the largest flange gap through which the flame will not propagate to an outer chamber.

MESG is what defines the gas groups:

This is why hydrogen service is so demanding. Hydrogen’s flame propagates through a far narrower gap than propane’s, so a IIC enclosure needs dramatically tighter tolerances — and why a housing certified for Group D is not usable in Group B, even though it looks identical and passed the same pressure test.

Note the direction of the logic: an enclosure isn’t “more explosion proof” because it’s built heavier. It’s suited to a narrower group because its flame paths are held to tighter dimensions.

Three Variables, Not One

Flame path design is a relationship between three things, and the standard’s tables (IEC 60079-1, read alongside IEC 60079-0) express it as a matrix rather than a single number.

Gap (i). The clearance between mating surfaces. Typical permitted values for flanged joints are in the range of a few tenths of a millimetre — commonly 0.15 to 0.40 mm depending on group, volume, and path length. For scale: a sheet of copier paper is roughly 0.10 mm and a human hair around 0.07 mm. Your safety margin on a Group IIB joint may be two sheets of paper.

Path length (L). How far the gas has to travel through the gap. Longer path, more heat transfer, more quenching — so a longer path permits a larger gap. Minimum path lengths step up with enclosure volume.

Volume (V). Bigger enclosure, more mixture, more energy released, higher pressure and more gas to vent. So larger volumes demand longer paths and tighter gaps.

There’s a consequence of the path-length rule that field personnel routinely violate: anything that shortens the effective path is a defect. A bolt hole drilled through a flanged joint interrupts the path, and the effective length is then measured from the inner edge to the hole, not across the full flange. This is why you never drill, tap, or modify a flameproof flange to mount a bracket, a nameplate, or a cable cleat. What looks like a convenient piece of metal is a certified dimension.

Threaded Joints

Threaded entries are flame paths too — the helical route between male and female threads is the quenching path, and its effective length is a function of engagement.

Hence the familiar requirement of at least five full threads engaged, reflected in NEC 500.8(E) and in the IEC construction requirements, which also impose minimum axial engagement lengths that increase for the more demanding gas groups.

Field failures cluster here:

• Cross-threading. The first two threads gall on assembly, the fitter forces it, and the joint now has three good threads pretending to be five.
• Thread sealant used as a seal. PTFE tape and hardening compounds are not flame path components. Where a compound is permitted at all, it must be a non-hardening type applied per the manufacturer’s instructions — and the flame path still has to be a metal-to-metal helical path, not a filled gap.
• The wrong entry device. A reducer, an adapter, or a cable gland not certified for the enclosure and gas group turns a certified entry into an uncertified one. The threads may fit perfectly. Fit is not certification.
• Unused openings left with the wrong plug. A hardware-store plug in a flameproof entry is one of the most common findings on any first-time site audit.

Surface Finish, and Why Paint Is a Defect

Flame path faces are machined, and the standard caps their roughness — IEC 60079-1 specifies a maximum surface roughness of Ra 6.3 µm on flame path surfaces.

The reason is straightforward once you picture it. A “0.20 mm gap” between two rough surfaces isn’t a uniform 0.20 mm. Peaks touch while valleys open up local channels considerably wider than the nominal dimension, and those channels run in whatever direction the machining marks run. Surface roughness is effectively a distributed increase in gap, plus a set of preferential leak paths.

That leads to two rules that surprise people:

Never abrasive-clean a flame path. Emery cloth, wire wheels, and scrapers cut into the surface and destroy the finish. Cleaning is done with a non-metallic scraper if necessary and a solvent, followed by inspection.

Never paint a flame path. Paint isn’t a permitted flame path material, it changes the effective gap, it flakes and contaminates, and a painted joint can’t be gap-checked. Corrosion protection on flame paths is achieved with a non-hardening, non-setting grease of the type the manufacturer specifies — applied thinly, never packed into the gap.

What Damage Actually Does

Here is the field-relevant part. Not all damage is equal, and the distinction is geometric.

Radial scratches are the dangerous ones. A scratch that runs across the joint face, from the inside edge toward the outside edge, creates a channel that spans the entire flame path. Gas travelling down it gets much less quenching contact than gas squeezed through the nominal gap. A single radial nick from a dropped cover or a screwdriver used to lever off a jammed lid can compromise the joint.

A circumferential mark — one that runs parallel to the edges, around the face — is far less serious, because gas crossing it still has to traverse the full path length. It’s still a defect to be assessed, but the physics differs.

This is why covers get lifted straight off and set down face-up on clean cardboard, and why levering a stuck flameproof cover with a screwdriver is a genuinely destructive act rather than a shortcut.

Over-torque. Excess bolt tension bows the cover between fastener positions. The joint closes hard at the bolts and opens at mid-span, producing exactly the local gap increase the design cannot tolerate — while the fitter’s impression is that it’s tighter than ever. Over-torque also yields fasteners and can crack cast housings at the bosses.

Wrong fasteners. Flameproof bolts are specified by property class and length because they are structural components holding back the reference pressure. A lower-grade replacement from the site store may look identical and fail at a fraction of the load. Too short and there’s inadequate thread engagement in the boss; too long and it bottoms out before the cover ever clamps.

Missing fasteners. Very common, and it is not a “one bolt out of twelve” proportional loss — it’s a local gap opening at exactly the point where the joint is now unsupported.

Corrosion. Pitting on a flange face is dimensional damage. Worse, corrosion product occupying the joint can force the faces apart, opening the gap while the assembly appears tight.

The Type Test You’re Relying On

All of this traces back to a type test that your specific model passed once, years ago, in a laboratory.

That test determined the enclosure’s reference pressure by igniting worst-case mixtures inside it, then subjected the housing to a static overpressure test at a multiple of that reference pressure, then ran non-transmission testing to confirm that an internal explosion would not ignite the surrounding atmosphere — deliberately using enlarged gaps to build in a safety factor.

Two implications follow, and both are routinely missed:

1. The result belongs to the tested configuration. Change the internal free volume — by fitting a large heat sink or a driver assembly that wasn’t there — and you may change the reference pressure the enclosure was proven against. The same logic governs the temperature classification.
2. The margin is finite and already spent. The safety factor in the non-transmission test was consumed by the standards writers to cover manufacturing tolerance and reasonable service wear. It is not a reserve you can draw on to excuse a damaged joint.

What Inspection Actually Requires

IEC 60079-17 defines three grades of inspection — visual, close, and detailed — and the distinction matters more than most sites realise. Close inspections do not normally require the enclosure to be opened. Only a detailed inspection involves opening the housing and, where appropriate, physically verifying flame path condition and gaps.

That grading is a deliberate risk trade. Opening a flameproof enclosure is itself a risk event: it exposes flame paths to handling damage, it puts fasteners through another torque cycle, and it depends entirely on the workmanship of whoever closes it again. The standard’s structure reflects a judgment that you should look inside as often as necessary and no more often than that.

Where damage is found, IEC 60079-19 governs repair, overhaul and reclamation — including the accept/reject criteria for flame path defects and the marking obligations that follow a repair. Reclaiming a flame path is specialist work with defined tolerances, not a job for a site machine shop with a surface grinder. Where a conversion is contemplated instead, the listing question and the certified-kit pathway come first.

Proven in Place

Now the argument that follows from all of the above, stated honestly because the honest version is stronger than the sales version.

A flameproof housing that has been installed for twenty-five years and passes detailed inspection is a known quantity. Its flame paths have been verified in situ. Its threaded entries have been undisturbed for decades and have proven they don’t leak or loosen. Its fasteners are the right ones because nobody has had a reason to replace them. Its conduit seals are intact, and its gaskets have not been through a fresh sealing event. Whatever installation errors it might have had were either found long ago or have been demonstrably tolerable through a quarter-century of thermal cycling and vibration. That history has real value — worth weighing against the fact that the lamp and ballast platform inside it may not survive the decade.

A brand-new certified housing arrives with perfect factory flame paths and zero service history in your location. Its certification is sound; its installation is unproven. Every joint is fresh: fresh thread engagement, fresh torque, fresh seals, fresh workmanship — all executed at elevation, under a permit clock, possibly by a contractor crew who won’t be back. Installation workmanship is a well-recognized source of Ex non-conformity, and every new joint is a new opportunity for it — which is why the initial inspection record and the asset register matter as much as the certificate.

That’s a real argument for minimizing joint disturbance, and it’s a safety argument before it’s an economic one.

But state the limits, or the argument doesn’t survive contact with a skeptical engineer:

• A retrofit still opens the main cover joint. It is fewer disturbed joints — cover only, rather than cover plus entries plus mounting — not zero, though that difference drives most of the cost and schedule gap. The cover flame path is still exposed to handling damage and must be inspected, greased correctly, and torqued to specification on reassembly.
• Proven in place assumes it actually passes inspection. A housing with pitted flanges, seized fasteners, or a cracked boss is not a proven asset. It is a liability with a long service record, and the correct answer there is replacement, priced as the event it actually is.
• Twenty-five years of undisturbed threads can also mean twenty-five years of galvanic seizure. Proven-in-place cuts both ways: the joint you don’t want to disturb may also be the joint you can’t disturb without damage.

Bottom Line

The enclosure is not a box that keeps gas out. It is a precision machine that lets an explosion happen and cools the exhaust below ignition temperature on the way out, using gaps measured in fractions of a millimetre, path lengths fixed by the standard, and surface finishes held to a machining tolerance.

Everything follows from that. Don’t paint it, don’t abrade it, don’t drill it, don’t lever it, don’t over-torque it, don’t substitute the fasteners, and don’t open it more often than the inspection regime requires.

And when someone tells you a housing is “just an enclosure” — that any equivalent box will do — you now know precisely why that’s the most expensive sentence on the site.

Standards referenced: IEC 60079-1 flameproof enclosures “d” · IEC 60079-0 general requirements · IEC 60079-17 inspection and maintenance · UL 60079-17 · IEC 60079-19 repair and overhaul · UL 844 luminaires for hazardous locations · NFPA 70 / NEC Article 500 · NEC free access · OSHA 1910.307 · IECEx certification scheme · ATEX Directive 2014/34/EU