Sealing penetrations in a submersible electronics enclosure

A submersible electronics enclosure fails at its penetrations, not at its walls. The short answer: use a manufacturer-rated cable gland or bulkhead feedthrough for every wire that crosses the wall, deal with the stranded-wire leak path deliberately, and pressure-test the assembled box before it goes in the water. At 30 m (100 ft), the outside pressure is about 3 bar (44 psi), and any silicone-dabbed hole will find that out for you.

What the depth does to your penetrations

Depth is the first design decision. IEC 60529 tests IP67 at 1 m (3.3 ft) for 30 minutes; deeper ratings such as IP68 are set by the manufacturer for the depth and duration it states. Ratings are manufacturer ratings, not UL listed, and a rating earned at 1 m tells you nothing about 30 m.

At roughly 30 m (100 ft) every sealed volume sees about 3 bar (44 psi) of external pressure. That pressure squeezes the box, compresses the door gasket sideways, and pushes water into any gap with enough force to weep past a loose gland. Flat plastic sides bow inward at this depth, which relaxes the gasket pressure around the door. A rigid metal box holds its shape; a thin consumer drybox may not.

Design the penetrations for the depth, not for splash: pick glands and connectors with a stated depth rating at or beyond your maximum, and derate for salt water and temperature.

Choose the penetration type before you drill

You have three workable ways through the wall, in order of preference:

• Rated cable gland. A compression gland clamps the cable jacket and seals it. It is the standard answer for power and multi-conductor cable, and the same rules apply as in shallow water: gland size matches the cable's actual jacket diameter, not its nominal size. See install a cable gland without losing the rating for jacket-to-gland sizing on data cable.
• Bulkhead feedthrough or sealed connector pair. Solid pins through the wall, sealed on the inside face, with the mating half outside. This beats a gland whenever you need frequent connect and disconnect, and gold-plated contacts resist corrosion in seawater. Where the entry is fixed and never serviced, a true pressure feedthrough is the right part; the reasoning is covered in why a pressure feedthrough is a bushing, not a cable gland.
• No penetration at all. For low-rate data, inductive coupling or optical windows through the wall remove the leak path entirely. Power still has to cross, so plan at least one gland.

What does not work: bolt-style contacts through drilled holes bedded in silicone caulk or adhesive. Caulk is not rated for depth, has no defined compression, and creeps under sustained pressure. A through-bolt also carries water on its threads unless both faces are sealed, which is exactly the kind of joint that passes a bench test and fails at 10 m.

The stranded-wire leak path

This is the penetration people miss. Stranded cable is a bundle of copper in a plastic jacket, and the gaps between the strands form capillary paths. Water that gets past the gland's outer jacket seal wicks along the cable, between the insulation and the copper, and drips into your enclosure on the other side of the seal. Some cable makers fill those interstices to stop it; most general cable does not.

Three fixes, any one of which is enough:

• Use cable with filled interstices, specified as water-blocked.
• Use a gland that clamps and seals both the outer jacket and an inner barrier, or add a secondary boot with heat-shrink adhesive over the gland's cable side.
• Land the stranded run on a bulkhead connector inside a small sealed chamber, so solid pin or flexible, filled lead wire does the final crossing.

If your gland is sealing a multi-pair data cable, the RJ45 gland sizing guide explains why an oversized hole around a small cable never seals, whatever depth you target.

Install each gland in steps

1. Confirm the cable's actual jacket diameter with calipers and pick a gland whose clamping range matches it.
2. Cut the wall hole to the gland's specified panel hole size, deburr it, and confirm the gland's sealing washer sits on a flat, smooth face.
3. Assemble the gland with the cable inserted, and torque the compression nut to the maker's value so the insert grips the jacket.
4. Route the cable so tension on the outside never pulls on the gland; add a strain-relief loop or clamp within a few inches (say 100 mm) of the entry.
5. For the door, check the gasket for cuts, clean the sealing faces, and close it evenly; the door gasket sealing guide covers gasket choice and compression.
6. Seal any unused hole with a metal plug and its own washer, never with tape or caulk.

Pressure-test before deployment

Bench-waterproofing lies. Assemble the complete box, then test it:

• Submerge with a dry paper towel inside, hold at your working depth or in a pressure pot at the equivalent, and check the towel.
• Pressurize from outside if you have a fitting; 3 bar (44 psi) simulates 30 m (100 ft).
• Dry the box, add a desiccant pack, and reopen after the first real deployment to look for condensation.

A damp towel or fogging points at the door gasket or a loose gland first. Condensation after cycling between cold water and warm air is a separate problem, covered in why condensation defeats a sealed enclosure and the humidity sensing approach.

Mistakes that cost the rating

• Drilling oversized holes and relying on sealant to fill the gap.
• Sealing a stranded cable directly at the wall without addressing the wicking path.
• Trusting an IP67 stamp at depth it was never tested for.
• Penetrating flat plastic panels with fasteners, which bow inward and unload the seal.
• Ignoring strain relief; one snagged cable works a gland loose in a few dives.
• Skipping the pressure test because the parts are all rated. The rating belongs to each part; the assembly is what leaks.

When to specify a real enclosure instead

If the box needs serviceable glands, a rigid wall at depth, and corrosion resistance in seawater, a consumer drybox is the wrong base. A stainless wall-mount enclosure holds its shape and its gasket compression under pressure, and a manufacturer-rated NEMA 6P / IP66-type 316L box is built for wet and submerged-adjacent service; ours are rated by the manufacturer, not UL listed. You can order a custom stainless enclosure with your gland holes cut before it is built, which is cleaner than drilling after, and the gland plate guide explains when a dedicated gland plate beats individual wall holes. The stainless size finder filters wall-mount and HMI lines by grade and size, and 316L vs 304 selection helps if salt water is in the picture. Whatever you specify, ask for the gland cutouts on the drawing so every hole arrives to size, sealed, and rated for the depth you actually work at.

Questions