Short answer: coat the module, but mask the shield can before the first pass. Mask the vent holes and the corner seams with peelable masking, keep the silicone off the antenna trace and its connector, and lay the film thin at 50-100 µm. The pressure worry is a non-issue: air sealed in a can over a 15 to 40 C day-night swing gains about 8.8 kPa. That loads a 15 × 15 mm lid with roughly 2 N, a load its soldered rim carries without moving. The real risk is electrical. Silicone sits at a dielectric constant of 2.5-3 against air's 1.0, so a film on an RF trace or a printed antenna shifts the match and costs link margin.
What do you need to coat a shielded RF module?
You need four things: a silicone conformal coating with a UV tracer, peelable masking liquid or masking dots for the can openings, lint-free swabs and a solvent clean, and a 365 nm UV lamp to inspect the film when it cures.
Silicone is the right chemistry for a board that lives outdoors: it stays elastic across the daily 25 K swing and cures at room temperature. IPC-CC-830, the qualification standard for conformal coatings, lists 50-210 µm for the silicone class, and this job wants the lower half.
- Brush-on or aerosol silicone coating, SR class, with UV tracer
- Peelable masking latex, masking dots, or low-tack polyester tape for the keep-outs
- Isopropyl alcohol, lint-free swabs and a way to dry the board; adhesion is won at the wash, and cleaning a board so the coating adheres is a skill of its own
- A 365 nm UV inspection lamp
- A fixed test spot for the link check: same distance to the gateway, same antenna orientation, before and after
Which areas of the module stay uncoated?
Mask five areas: the vent holes in the can lid, the open corner seams, the antenna trace and its matching parts, the antenna connector, and any crystal or resonator that sits exposed on the board — everything else gets the full film.
The physics: air has a dielectric constant of 1.0, and datasheet values for silicone conformal coatings cluster at 2.5-3. A film over a controlled-impedance trace raises the effective permittivity around the line, so the impedance drops and the electrical length grows; over a printed antenna it shifts the resonant frequency. On a sub-GHz link the effect on a 50 Ω line stays small, but on the antenna area it can eat a few dB of margin. Vapor-deposited parylene-type films are the option vendors document for high-frequency circuits, because they land thin and uniform without loading the RF; at the bench, masking is faster than any rework.
Potting the whole module is the heavier alternative — read what potting does to a board before you pour one.
| Keep clear | Why | Mask with |
|---|---|---|
| Lid vent holes | They keep the can vented, in the module's own build and in service | Peelable latex over each hole |
| Corner seams | Open seams let the film wick inside the can | A thin dam of peelable mask along each seam |
| Antenna trace and matching parts | A film at εr 2.5-3 detunes the line | Low-tack tape cut to shape |
| Antenna connector | Coating in the contacts raises resistance | Masking dot or tape cap |
| Exposed crystal or resonator | Added dielectric and mass shift the frequency | Masking dot |
Will coating over the can's holes trap pressure and crack it?
No — the numbers are small: a can sealed at 15 C and warmed to 40 C gains 25 K over 288 K of absolute temperature, about 8.8 kPa of extra pressure at sea level, and a 15 × 15 mm lid feels only about 2 N.
The arithmetic at constant volume: ΔP = P × ΔT ÷ T = 101 kPa × 25 K ÷ 288 K ≈ 8.8 kPa, about 0.09 bar or 1.3 psi. Force is pressure times area: 8,800 Pa × 0.015 m × 0.015 m ≈ 2 N on a 15 × 15 mm lid, and about 5.5 N on a 25 × 25 mm one. Cooling the same 25 K the other way pulls with about 8.1 kPa. Neither load threatens a soldered rim. The swing sets the pressure, so it scales linearly; elevation lowers it, and the same swing at 2,000 m adds about 6.9 kPa; can size only scales the force.
Pressure is not the reason to keep the holes open, though. The reason is moisture: a can with clear openings breathes with the box and dries again, while a can sealed under film holds whatever condenses inside it, where you cannot see or reach it. Mask, cure, peel — the vents go back to work. If you want the interior air measured while you decide, a humidity sensor inside a sealed enclosure reads it directly.
How do you mask, coat and cure the board?
Work in this order: record the link, clean, dry, mask, one thin pass, flash, a second pass, cure overnight, peel the mask, then inspect under UV — a day of bench work plus a 24 h cure.
- Disconnect power, pull the board, and log the link baseline: RSSI and SNR at a fixed spot with the antenna in its working orientation. This number decides pass or fail at the end.
- Wash off flux residue with isopropyl alcohol, rinse, and dry the board fully; an hour at 60 C is a common shop bake. Film over flux or trapped moisture blisters within weeks.
- Mask the five keep-outs: peelable latex over each lid hole and along the corner seams, a masking dot or tape cap over the connector, tape cut to shape over the antenna trace.
- Lay pass one at 50-100 µm. Light strokes, and keep the brush or nozzle from driving film into the seams.
- Flash off 10-20 minutes between passes, or the coating datasheet's flash time, and add a second pass only where the first came up thin.
- Cure about 24 h at 25 C, or the datasheet's heat schedule when the job needs the board back sooner.
- Peel the masking and confirm every hole and seam is open again under a magnifier.
- Inspect the film under the 365 nm lamp, then re-run the link test against the baseline.
How do you check the coating before the board goes back outside?
Pass the board on three checks: uniform film under UV with no pools in the seams, every can opening clear, and link metrics within about 2 dB of the pre-coat baseline at the same spot.
A silicone puddle in a corner seam cures into a thick fillet, and a fillet that size lifts after enough 25 K cycles; find pools now and peel them while the film is fresh. The 2 dB rule is a shop-floor line: sub-GHz link budgets drift a little day to day, so a baseline held within about 2 dB says the RF stayed untouched, and a bigger drop sends you back to the antenna area with a swab. Reassemble with the gasket seated all round and the latch snug, because the film covers a small lapse, never an open door.
Which mistakes cost you the radio or the rating?
Four mistakes do the damage: spray straight over the can and fill the seams, coat the RF path, lay the film too thick, or seal the box afterwards with an unsealed hole — each one costs you either decibels or the enclosure rating.
- Spraying straight over the can. The film wicks into the vent holes and corner seams, plugs the vents and puts dielectric material at εr 3 inside the cavity — the exact failure the mask prevents.
- Coating the RF path. Trace, matching parts and connector each shift or resist; coating connectors without fouling them is its own job.
- One heavy pass. A 200 µm puddle pools in the seams and lifts; two passes at 50-100 µm stay put.
- Coating damp or fluxy boards. Blisters and adhesion loss inside weeks.
- Sealing the box with an open hole. Every new cable entry gets a gland matched to the cable OD; an unsealed 6 mm drill point drops the whole assembly below its rating, the same way vent holes cost a sealed enclosure.
How much water protection should the enclosure carry?
Let the box carry the hose and the film carry the leftovers: a 316L stainless enclosure manufacturer-rated NEMA 4X / IP66, not UL listed, stops hose-directed washdown, while the 50-100 µm coating only covers a gasket lapse or a latch left open.
| Layer | Stops | Its limit |
|---|---|---|
| 316L box, manufacturer-rated NEMA 4X / IP66, not UL listed | Hose-directed water and washdown | The assembled rating falls to the weakest part: one unmatched gland sets it |
| Glands matched to cable OD | Water wicking down the cable | A gland one size off leaks at the cord grip |
| Silicone film at 50-100 µm | Splash, condensate, a small lapse | Jet and immersion work stays with the box |
| Desiccant pack with indicator | Vapor that walked in during the day | Swap it when the indicator turns |
The night side matters as much as the spray. Air that leaks in during a 40 C day at 50% RH carries a dew point near 28 C. When the box cools to 15 C after dark, that vapor condenses on the walls. Glands matched to the cable and a dry desiccant keep the interior air dry in the first place.
Decide by site: fresh-water sprinklers on most sites — a 304 box holds. Chlorinated or brackish spray, coastal salt air, or hard-water deposits left by evaporation — go 316L, because 304 pits under chloride deposits even at room temperature. Occasional splash indoors — a Type 12 box plus the film is plenty. Constant hose-down outdoors — Type 4X / IP66 at minimum, with the coating as the second layer. If masking gets impractical, weigh conformal coating or a sealed inner enclosure instead; the wider case for the film is in when a conformal coat helps outdoors. A one-off 316L box built to your gland layout and cutouts is delivered in 10 business days from drawing approval, and the custom stainless enclosures page is the starting point.
What Type 4X and IP66 actually test sits in the NEMA 4X vs IP66 guide, if you need to defend the choice to an inspector.
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 |
| 12 × 10 × 6 in | 316L stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 316L stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 10 × 8 × 6 in | 304 stainless | 10 x 8 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Which coating chemistry suits a board that swings 15 to 40 C every day?
Silicone: it stays elastic across the swing and most room-temperature grades fully cure in about 24 h at 25 C. Its dielectric constant of 2.5-3 against air's 1.0 is why the RF keep-outs stay masked whichever chemistry you pick. Acrylic reworks faster but flexes less, and parylene needs a coating vendor with deposition equipment.
Does the daily 15 to 40 C swing hurt the rest of the board?
Industrial-grade parts are rated from -40 to 85 C, so a 25 K board-level swing sits well inside their range. Solder-joint fatigue shows up after thousands of swings at stiff joints, and thin silicone cushions the parts it covers; that is one more reason for two light passes instead of one heavy pour.
How thin should the silicone film be over a shielded module?
Aim for 50-100 µm: thick enough to bridge scratches and edge gaps, thin enough that nothing pools into the corner seams, where a 200 µm fillet can lift with the daily swing. IPC-CC-830 lists 50-210 µm for the silicone class, so stay in the lower half. Two light passes beat one heavy one.
Does a thin conformal film survive a hose jet or immersion?
A 50-100 µm film stops splash, condensate and slow seeps; jet pressure is another world. IPX6 is powerful jets, about 100 L/min through a 12.5 mm nozzle, and a Type 4 hose test is in that class, so the box carries the manufacturer-rated NEMA 4X / IP66 and the film stays the backup layer.
