Short answer: the smeared sealant comes off and the seam comes apart. Strip both faces to bright metal across 25 mm [1 in] each side of the joint, then rebuild it on a 2 mm [1/16 in] bond line of neutral-cure silicone or primed polyurethane rated for continuous water contact. A bead bridged over a damp, dirty or previously sealed joint sticks to the old layer, and that layer had already let go. Silicone also cures from the surface inward at about 2 mm per 24 h, so a 6 mm smear sat soft at its base for three days. Flaking paint on the bottom is a separate verdict on the box itself, and it often settles the repair-or-replace question before you lift a scraper.
What does the water inside my enclosure tell me?
The position of the water names the leak path: drips and streaks aligned with a vertical edge point at a seam, a wet floor under the box points at the bottom lap or the glands, and an even film means condensation rather than a leak at all. Read the pattern before you open a tube of anything, because water travels inside the wall and shows up far from where it entered.
- Puddle on the floor, walls dry. The bottom lap seam or a low cable gland; water ran down the outer wall and dropped off the lowest edge.
- Streaks from one corner. That corner's seam. Corners take shipping knocks and the most door flex.
- Drips in a line under the door edge. The door gasket, or a latch that has stopped squeezing it evenly along its length.
- One rust-coloured track from a single fastener. An unsealed through-hole; fastener lines are the most common single leak on a bolted seam.
- Splash marks on top of components. A missing gland insert, or a field-drilled hole that was never dressed and sealed.
- Everything evenly damp, no tracks. Condensation from day-night temperature swings. Sealing will not fix that; a small anti-condensation heater will.
Two strips of duct tape across the closed door gasket split the suspects in ten minutes. If the box stays dry with the tape on, the gasket is your patient and the seams are cleared.
Why did the sealant I smeared over the seams fail?
It failed for three reasons: it bonded to a contaminated surface, it bridged a joint that flexes instead of filling it, and it cured only on the outside, because silicone hardens from the surface inward at about 2 mm per 24 h.
Take the cure first, with your own numbers. A smear 6 mm [1/4 in] thick needs 6 ÷ 2 = 3 days to cure through, counted from the day you laid it. The cure front reaches the metal last, so the base of the bead — the only part doing any bonding — was the newest and softest material in the joint when the water returned.
Adhesion went the same way. Sealant wets whatever it touches, and your bead touched coolant film, dust and old cured polyurethane. Silicone over cured polyurethane forms no chemical bond, so the chain held only as well as the old layer, which had already let go. A joint is a chain, and it holds at the strength of its dirtiest face.
The bridge also covered a joint that moves. Each smear dammed water at its edges, the trapped water lifted the paint around the seam, and those lifted edges are your new leaks. You sealed where the water exited, too, and on a lap seam the exit usually sits well below the entry.
How do I find every leak before I repair anything?
Hose-test one seam at a time using the IEC 60529 jet-test numbers as your standard: a 6.3 mm nozzle delivering 12.5 L/min from 2.5 to 3 m [8 to 10 ft], at least 3 minutes per seam, with the door gasket taped off so it cannot mask a seam leak.
- Dry the inside and lay paper towels along every seam and under every gland.
- Run the jet on one seam for 3 minutes — 12.5 L/min × 3 min = 37.5 L on one joint — then check the towels before moving to the next.
- Mark each weep on the outside with a paint marker while it is still wet and visible.
- Test fastener lines as their own seams. A weep at one fastener means the hole, the whole line, or both need attention.
- Repeat the identical test on every repaired seam after full cure, and log the result seam by seam.
Glands and field-cut holes follow their own rules, and the lowest-rated part sets the assembly's rating: an IP54 gland on an otherwise IP66 box makes it an IP54 box. The details are in what rating a panel keeps after you cut it and mount devices. A garden hose with a thumb over it approximates the jet; a pressure washer does not, and it drives water past seals that hold fine in normal service.
Which sealant survives water and coolant on an enclosure seam?
Use a neutral-cure silicone rated for continuous water contact, or a one-part polyurethane with the maker's primer on stainless; both hold water-based coolant, and they differ in temperature range, abrasion and how long they take to cure.
| Property | Neutral-cure silicone | One-part polyurethane, primed |
|---|---|---|
| Service temperature | about -60 to 200 C | about -40 to 90 C |
| Continuous water contact | holds once cured through | holds after full cure |
| Water-based coolant | holds | holds |
| Scuff and abrasion | moderate | good |
| Primer on stainless steel | none on clean, abraded metal | yes, the maker's own |
| Time to chemical service | about 24 h per 2 mm of thickness | about 7 days |
| Reopening the seam later | peels off in one strip | needs mechanical removal |
Buy neutral-cure, and skip acetoxy silicone. Acetoxy releases acetic acid while it cures, and that acid corrodes bare steel, copper and brass — the wrong chemistry inside a panel. The temperature rows are elastomer-handbook values, the same ranges used for gasket materials.
Machine coolant is mostly water — a typical 20:1 dilution leaves about 5 % concentrate — so both families hold it once cured. Silicone suits seams you may want to reopen; polyurethane suits fillets that take scuffing. When the washdown chemistry on your line changes, sealant compatibility moves with it, and whether enclosure ratings survive a washdown chemical change walks that question.
For planning the job: a fillet 6 mm wide by 2 mm thick has 12 mm² of section, so one 300 ml [10 fl oz] cartridge dresses 300,000 ÷ 12 = 25 m of seam.
How do I rebuild the seam so it seals the first time?
Take the joint fully apart, strip both faces to bright metal across 25 mm [1 in] each side, and reassemble on a fresh 2 mm [1/16 in] bead so the sealant bonds to steel, not to paint or old sealant; that is the whole repair.
- Lock the panel out and prove it dead with a meter before you open it. Drain any coolant pooled against the seam.
- Unbolt the joint and separate the faces with a spreader. A screwdriver tip gouges the sealing faces and cuts water a channel.
- Peel the old bead away with a plastic scraper. Old silicone comes off in one strip if you pull it slowly at a flat angle.
- Abrade both faces to bright metal with 150- to 180-grit, 25 mm [1 in] each side of the joint, through paint as well as old sealant. Paint is a weaker bond layer than either sealant, which is why painted flange gasket faces are stripped to bare metal in flange work.
- Wipe both faces with a clean rag and isopropyl alcohol, then let them flash off and dry completely.
- Lay one continuous bead about 6 mm [1/4 in] wide on the inner face, unbroken at the corners, and bolt the joint back in sequence so the bead compresses to about 2 mm [1/16 in].
- Tool a fillet across the joint with a spatula within about 10 minutes, before the bead skins, so the sealant wets both faces.
- Cure 24 h for silicone — add 24 h for every extra 2 mm of thickness — or about 7 days for polyurethane before chemical service, then run the 3-minute jet test on that seam.
Can I use a sandable body filler instead of sealant?
A sandable body filler will not seal a moving seam: polyester filler is rigid, it carries no rating for continuous water contact, and the first flex of the panel cracks it, so it is a cosmetic overcoat over a sealed joint, never the seal itself.
The movement is small but real. Warm a 600 mm [24 in] panel face by 20 C — one washdown shift can do that — and the steel grows 12 × 10⁻⁶ × 600 × 20 ≈ 0.14 mm. Silicone stretches across that movement for years; filler cracks across it in weeks and passes water behind itself. If you want the joint smooth for cleaning, lay filler over a cured, jet-tested sealant fillet, sand it, then prime and paint it. The sealant underneath is still doing the sealing.
What does flaking paint on the bottom of the box mean?
Flaking paint along the bottom means water and coolant have been under the coating for a while, and it tells you the coating system is at the end of its life in that zone, not that one seam missed its sealant.
This is osmotic blistering, the standard failure of painted carbon steel in wet service. Water diffuses through the paint film, the salts and amines in coolant pull more water in behind it, and the film lifts. Scrape the zone back 50 mm [2 in] past the last flake, abrade, degrease, and rebuild with an epoxy primer under a compatible topcoat — repairing minor coating damage covers the sequence.
If coolant pools on that floor week after week, the honest fix is material, not paint. A 304 stainless floor with a #4 brushed finish shrugs off water-based coolant, and 316L earns its place once sanitizers or brine arrive — 304 vs 316L stainless steel for enclosures lays out the trade. Mind the chloride limit at about 60 C [140 F], where chlorides stress-corrosion-crack both grades. 304 also pits under a chloride deposit at room temperature, so rinse the floor rather than wiping it. Keeping coolant off the seams in the first place is cheaper still: containing machine-tool coolant at enclosure gaps shows where the gaps hide.
When is replacing the enclosure cheaper than repairing it?
Replace the box when three or more seams still weep after one proper repair, when flaking paint covers a third of the bottom, or when a washdown unit is coming that will hit this enclosure every shift.
| What you are looking at | The decision |
|---|---|
| One or two seams weep, paint sound | Rebuild those seams on a 2 mm [1/16 in] bond line, retest each with the 3-minute jet, keep the box |
| Seams plus gland leaks | Rebuild the seams and fit glands rated IP66 or better; the lowest-rated part sets the rating |
| Bottom paint flaking, coolant pooling weekly | Recoat now to stay running, and budget a 304 stainless replacement |
| Three or more seams still weep, or washdown is coming | Order a welded one-piece enclosure: 304 for coolant, 316L for sanitizers and salt air |
That last row is where the lap-seam game stops paying. A custom enclosure from us is folded and welded from one piece, so there is no seam to seal and nothing to re-do each year. You set the material, the width, height and depth, and every cutout. It is delivered in 10 business days from your approval of the drawing, anywhere in the world. Ratings on those boxes are manufacturer ratings — rated NEMA 4X / IP66 by the manufacturer; not UL listed — and NEMA does not test, certify or endorse enclosures.
Before you order, size the replacement with the enclosure selector and take the layout to custom enclosures. Read what washdown water temperature means for your enclosure as well, because a hot washdown changes the gasket conversation before it changes anything else.
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 |
|---|---|---|---|
| 12 × 10 × 6 in | 304 stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 16 × 14 × 8 in | 304 stainless | 16 x 14 x 8 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
Can I fix the leak from outside without taking the seam apart?
You buy weeks, not years. A bead bridged over a dirty, wet seam bonds to the old layer and lets go with thermal cycling, typically within a few weeks. The durable repair is the disassembled joint on a 2 mm [1/16 in] bond line with 25 mm [1 in] of bright metal each side, and it needs 24 h of cure before it sees water.
How long does enclosure sealant need to cure before washdown?
Neutral-cure silicone is fully cured in about 24 h at 2 mm of thickness, and every extra 2 mm adds about 24 h. Polyurethane needs about 7 days before continuous water and chemical contact. Run the 3-minute, 12.5 L/min jet test on each repaired seam before the first washdown goes through.
Is silicone or polyurethane better against machine coolant?
Both hold water-based coolant, which is about 95 % water at a typical 20:1 dilution. Silicone stays elastic from about -60 to 200 C and needs no primer on clean, abraded steel; polyurethane serves about -40 to 90 C, wears better against scuffing and wants the maker's primer on stainless. Pick silicone where you may reopen the seam.
Does resealing the seams change the enclosure's NEMA rating?
The type on the label belongs to the manufacturer's original design, and a field repair is only as good as your prep and retest prove — retest each seam with the 3-minute jet. Ratings such as NEMA 4X / IP66 are manufacturer ratings, not UL listings, and NEMA does not test, certify or endorse enclosures.





