After a hose test, a washdown or a hard rain, the pass/fail question is the same: did water actually cross the seal? For most enclosures, a careful visual inspection is the accepted answer, and it is what rating tests themselves generally use. On a crowded panel it is subjective, so back it up with water-indicating paper or a dye trace at the seams, and use weighing or a pressure-decay check only when you need a number you can defend. Each method tells you something different, and each finding points to a specific weak point: the door gasket, the cable glands, or the mounting. If a NEMA type or IP code is in play, remember those are manufacturer ratings; our enclosures are rated NEMA 4X / IP66 by the manufacturer and are not UL listed.
What the symptoms point to
The location and form of the water inside tells you where the barrier failed before you run any second test:
- Water on the floor of the box, dry walls: entered low. Suspect the bottom cable glands, a missing drain path, or a conduit connection, not the door.
- Water tracks down the back wall: came from behind. Mounting screws through the wall or an unsealed wall interface are the usual path.
- Damp line around the door perimeter, dry interior: the gasket is weeping but not yet passing. Compression is uneven or the gasket has taken a set.
- Moisture on components with no liquid anywhere: likely condensation, not ingress. A sealed box breathes on temperature cycles; see the article on moisture inside an enclosure before you chase a leak that is not there.
- Rust streaks from a screw or seam: water has been entering slowly over months, not just this one test.
How the checks rank in practice
Causes of an ambiguous result, in order of how often each one is to blame:
- Visual inspection done too fast, under poor light, on a box that was already damp.
- Gasket compression lost at corners or at latch points.
- Glands installed with the wrong clamping range for the cable, so the seal grips nothing.
- Fastener penetrations through the enclosure wall leaking behind the box.
- Genuine door or body distortion, usually from a heavy backplate or a knock.
The first four are field-fixable. The fifth means re-specifying.
The visual check, done properly
Visual inspection is not wrong; a sloppy one is. Work it like this:
- Open the door slowly so you do not dislodge standing water at the seal.
- Look first at the floor and the lowest corner, then the back wall, then around each penetration.
- Use a bright directional light held at a low angle; a sheen of film shows water that a top-down look misses.
- Lay dry, absorbent paper towel along the seams before the test where you can, and inspect the paper, not just the metal.
- Photograph the interior before the test so you can argue about what was already there.
Judge against the standard's intent, which for enclosure types like NEMA 4X is that no water reach live parts, not that the interior be bone dry to the last molecule. Traces that pool, run, or reach a terminal block fail; a faint mist on a stainless wall that evaporates in minutes is a judgment call you should record either way.
Dye tracing at the seals
Dye is the cheapest upgrade over eyeballing. Add a fluorescent dye or plain food coloring to the spray water, run the test, then inspect the interior seams with a UV lamp in a dimmed area. Dyed water inside is an unambiguous fail, and the dye line shows you which seal it crossed. Used in reverse, dab dye at a suspected entry point on the outside, spray, and see whether it washes in. Dye has no accuracy problem, needs no instruments, and documents well with a photo. Its limit: it tells you where water went, not how much.
Weighing, and why it usually disappoints
The before-and-after scale reading sounds objective but rarely is. Water clings to the outer surface, sits in gasket grooves, and is absorbed by gasket material and paint without ever reaching the interior, so a stainless box with a neoprene gasket gains weight from wetting alone. To get a clean reading you must dry the exterior thoroughly and consistently, and a gram-level resolution only beats the noise on small enclosures. Even a true gain tells you water entered, not where. Use weighing only as a cross-check on small sealed boxes, never as the primary verdict.
Pressure decay as the shop method
For a repeatable number, seal the openings with blanking plugs, pressurize the enclosure to a few psi above atmospheric with dry air, and watch the gauge. A sealed box holds pressure; a leaking gasket or gland bleeds it down, and brushing soapy water on the seams shows bubbles at the exact leak path. This is how many panel shops verify gasket integrity before the box ever sees water. Two cautions: pressure tests the gasket and glands only, so it will not catch a mounting-hole leak behind the box; and never pressurize an enclosure beyond what its door and latches are designed to take. A gentle pressure check paired with a water test covers each other's blind spots.
Gasket and gland findings and the fix
Match the finding to the fix:
| Finding | Likely cause | Fix |
|---|---|---|
| Damp line all around the door | Gasket taken a set, or latch not compressing | Replace gasket; check latch torque; see foam or rubber gasket choices and the tongue and groove design |
| Leak only at corners | Uneven compression or a pinched gasket | Re-seat gasket, check for distortion; a formed-in-place groove beats a flat-flange squeeze |
| Water at one gland | Seal not gripping the cable | Match the gland to the cable diameter; the gland clamping range guide covers sizing |
| Water behind the box | Fasteners through the wall | Reseal and re-mount; the article on mounting screws without leaks gives the compliant patterns |
| Nothing wet, but fog on the window | Condensation cycle, not ingress | Follow the condensation angle in the moisture diagnosis article |
Stopping it from failing again
A box that passed a shop hose test can still fail a plant washdown, because washdown adds heat and chemistry that change gasket behavior; the article on washdown water temperature covers why. To keep a pass a pass:
- Torque or latch the door the same way every time; a half-closed latch is the most common self-inflicted leak.
- Keep every gland on the bottom or sides, never facing up; the gland orientation article shows what an upward gland collects.
- Add a drip loop on every cable entering from above, as covered in sealing an outdoor junction box.
- Inspect the gasket at every scheduled service, not only after failures; the corrosive-area inspection checklist works for wet areas too.
When to re-specify the enclosure
Resealing fixes gaskets and glands. It does not fix geometry or design, so re-specify when:
- The door or body is warped and the gasket cannot compress evenly anywhere along the seam.
- The gasket groove itself is damaged or corroded at the sealing face.
- The design depends on fasteners through the enclosure wall, and the site washes the back of the box directly.
- The interior layout leaves no way to keep water away from live parts even on a marginal leak.
That is the point to move to a custom stainless enclosure with factory-cut openings, welded mounting points instead of through-wall screws, and a gasket groove machined into the body, delivered in 10 business days from your approval of the drawing. If the driver is choosing the right box rather than fixing the one you have, start from the enclosure selector and let the verification methods here confirm the choice. For what the water jets themselves actually do during a rating test, the IP water test article covers the spray side, and the IP65 vs IP66 comparison explains which jets your box will face.
Enclosures for this job
Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 16 × 14 × 8 in | 304 stainless | 16 x 14 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 16 × 14 × 8 in | 316L stainless | 16 x 14 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 24 × 20 × 8 in | 316L 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 |
Questions
Is a visual inspection enough to pass a water ingress check?
For most enclosure volumes it is the accepted first method, and it is what most rating test reports rely on. Its weakness is subjectivity: a few drops on a crowded backplate are easy to miss, and the standard allows judgment about whether traces matter. Make it defensible with a bright light, dry absorbent paper wiped along the seams before you look, and a second person re-checking the areas around every penetration.
How does dye tracing find a leak path?
Run the water test with a fluorescent dye or food coloring added to the spray, then inspect the inside seams with a UV lamp. If dyed water shows inside, something crossed the seal. The reverse also works: apply dye at a suspected entry point and see whether it washes away or is carried inside. Dye is cheap, needs no special scale, and pinpoints which seal leaked rather than only that water got in.
Why is weighing the enclosure before and after unreliable?
A gram-level scale costs little, but the reading is confounded: water can cling to the enclosure surface, sit in gasket grooves, or be absorbed by gasket material and paint without ever reaching the inside. A stainless box with a neoprene gasket can gain weight from wetting alone. Weighing works only with careful drying of the outside first, and even then it tells you water entered, not where.
When should I ask for a replacement enclosure instead of resealing?
If the body is corroded at the seam, a gasket groove is damaged or deformed, the door is warped so the gasket cannot compress evenly, or the mounting holes leaked because the design puts fasteners through the wall, resealing will not hold. Re-specify with welded studs or an internal backplate, a different gasket material, or a fully welded body with factory cutouts, so the leak path is designed out rather than patched.
