Unprotected PCB outdoor lifetime: what limits it

You can run that board bare. A clean, uncoated PCB covers 50 fair-weather hours at about 20 °C with no measurable corrosion. Corrosion and electrochemical migration need a sustained moisture film under bias, and fair-weather daylight supplies none. The failure you can actually meet in a 30 minute mission is condensation. Any surface below the air's dew point fogs within minutes, and a water film on a 40 GHz radiating element detunes the circuit on contact. So the plan is scheduling, not sealing. Check the dew point for the mission window and hold every surface 3 °C above it. Cover the board with an expanded-polystyrene foam dome kept 10 mm [0.4 in] off the radiators. Save the enclosure for the day the weather stops being optional; the sections below carry the numbers.

How long does an unprotected PCB last outdoors?

A clean bare board covers its 50 fair-weather hours at 20 °C with margin. Atmospheric corrosion needs a sustained moisture film under bias, and fair-weather daylight keeps every surface above the dew point, so corrosion has no wet hours to work with.

All three ageing paths — oxidation of bare metal, corrosion under a film, electrochemical migration between biased tracks — are counted in wet hours. Clean air above the dew point gives them nothing to work on. One hundred 30 minute missions at 20 °C and 60 % RH add zero wet hours. The corrosion ledger closes at zero by arithmetic, not by luck. Bare copper still tarnishes chemically within days, but a tarnish a few nanometres thick sits under the pad finish. ENIG or HASL pads stay bright through months of this treatment. When corrosion does show after a genuinely wet deployment, it clusters at contaminated, biased nodes, where you find and measure it under the microscope.

What fails first: corrosion or condensation?

Condensation fails a bare board first. A water film over a 40 GHz patch detunes and attenuates it within seconds, because water's relative permittivity is about 80 at low frequency and still in the tens at millimeter wavelengths, with heavy loss.

Ranked by speed:

• Detuning and loss on the radiators. Water over a patch pulls the resonant frequency and absorbs power; the effect lands with the film and leaves when it dries.
• Leakage across biased nodes. A wet film between rails raises leakage and noise; on a clean board this reverses fully on drying.
• Electrochemical migration. Water, bias and ionic contamination together grow metal filaments between tracks over weeks of continuous wet bias, and that damage is permanent.
• Dust as an accomplice. Dust alone does nothing; a dusty board that dews later holds an ionic film, so wipe the board down after a dusty session.

The permittivity is why the antenna effect is instant: the wave sees a lossy dielectric where it expected air. Keep any conformal coating off the radiators — a dielectric laid over a patch tuned bare pulls the same frequency in a milder way. Let the foam dome do the weather work without touching the pattern.

What is the dew point on a mission day at 20 °C?

At 20 °C the dew point runs 9.3 °C at 50 % RH, 12 °C at 60 % and 16.4 °C at 80 %. The mission rule is simple: hold every surface at least 3 °C above the dew point before the cover comes off.

These are Magnus-formula values for 20 °C air, with a = 17.62 and b = 243.12 °C. Sample arithmetic at 60 %: ln(0.60) = −0.511; 17.62 × 20/(243.12 + 20) = 1.339; the sum is 0.828, and Td = 0.828 × 243.12/(17.62 − 0.828) = 12.0 °C. The 3 °C margin absorbs forecast error and the offset between sensor and skin.

Worked example: the fog you bring with you. Forecast for the window: 20 °C, 60 % RH, dew point 12.0 °C. The rig spent the night in a closed car that settled at 10 °C, which is 2 °C under the dew point, so every face fogs within minutes of meeting the warm air. The fix is free: keep the assembly in its insulated case until the skin reads 15 °C, or give it 30–60 minutes in still air. A foil-and-foam board reaches air temperature in tens of minutes.

The same physics runs in reverse inside a chilled chamber, and warm humid air meeting any cold surface is the make-up air condensation problem familiar from building services. The NEMA and IP ratings guide and the condensation in outdoor enclosures guide carry the fuller dew-point method with heater and drain sizing.

Can fair weather still dew the board overnight?

Yes, once evening humidity at 20 °C passes about 75 % RH. A clear calm sky pulls an exposed face 3–5 °C below the air, so a face at 15 °C sits under a 16.4 °C dew point and dew forms with no cloud in sight.

Dew forms when a surface drops below the dew point, and a clear calm sky cools every exposed face 3–5 °C below the air by radiation alone. Set the surface to 20 − 5 = 15 °C and solve the Magnus formula back: ln(RH) = 17.62 × 15/258.12 − 1.339 = 1.024 − 1.339 = −0.315, which is 73 % RH. Call the crossover 75 % for margin. Evenings drier than that cannot dew an uncovered board; evenings wetter than that can.

Practical consequences:

• Leave the rig outdoors between missions and the sky, not the forecast rain, is the threat; bring it in or cover it at dusk.
• The foam dome doubles as a blanket: 25 mm of foam slows the radiant chill and holds the skin near air temperature.
• Cloud and wind shrink the 3–5 °C cooling, so an overcast evening at the same humidity is the safer one.

What cover stays transparent at 40 GHz?

Low-density foam: at εr about 1.05 it reflects 0.015 % of the beam at each face, against 7 % for a plain polycarbonate sheet. A foam cup or bowl over the board changes the pattern far less than any solid plastic window.

Reflection at a flat sheet, normal incidence, is ((√εr − 1)/(√εr + 1))² of the power. Foam at εr 1.05: √1.05 = 1.025, (0.025/2.025)² = 1.5 × 10⁻⁴, so 0.015 % per face. Polycarbonate at εr 3.0: √3.0 = 1.73, (0.73/2.73)² = 0.072, so 7.2 % bounced at each of the two faces. That gap, not stiffness or cost, is why foam wins.

Radome practice for 18–40 GHz specifies an average dielectric constant of about 2.1 or less, so the top two rows clear the bar and the bottom two do not. PTFE, PVC and ABS are the usual RF-transparent plastics. At millimeter wavelengths keep the wall thin and measure the pattern with the cover fitted before you commit. For this job the foam cup or bowl wins on every count: electrically invisible, cheap, a shield for a surprise shower, and at 25 mm of thickness it is also the dew blanket from the last section.

How much clearance does the cover need above the antenna?

Keep 10 mm [0.4 in] of air between the foam and every radiating element. λ at 40 GHz is 7.5 mm, the far field of a 5 mm patch starts near 7 mm, and the 10 mm rule clears both with margin.

Two boundaries set the number. Wavelength: λ = c/f = 3 × 10⁸ m/s ÷ 40 × 10⁹ Hz = 7.5 mm. Far field: r ≈ 2D²/λ, so a 5 mm patch reaches it at 2 × 25/7.5 = 6.7 mm. Hold the foam beyond both and it reads as a thin, low-εr slab; 10 mm does that for any single element up to about 6 mm across.

Arrays are the exception, because D is then the whole aperture and 2D²/λ grows fast. A 50 mm aperture puts its far field near 670 mm, which no cup will clear. Pick one of the two honest options: measure the pattern with the cover on and accept the detune if it holds, or run the antenna out in the clear and leave the foam over the rest of the board.

When does a bare board need a real enclosure?

Move to an enclosure when missions run in rain, dust or months of exposure. A custom stainless enclosure rated NEMA 4X / IP66 by the manufacturer (not UL listed) keeps the electronics dry while the antenna works through a matched window or an external gland.

At 40 GHz a metal box is a mirror. Skin depth in stainless is δ = √(2ρ/ωμ): with ρ ≈ 7 × 10⁻⁷ Ω·m and ω = 2π × 4 × 10¹⁰ = 2.5 × 10¹¹ rad/s, δ ≈ 2 µm. A 1.5 mm wall stands about 700 skin depths deep, and the beam does not leave. The antenna either fires through a window or lives outside:

• Matched window. A dielectric sheet reflects least at half-wave thickness, t = λ/(2√εr): for polycarbonate, t = 7.5/(2 × 1.73) = 2.2 mm. The factory cuts that aperture to your drawing; the custom stainless cutouts page lists what the drawing needs.
• External antenna. Coax or waveguide runs from the sealed box to an antenna in free air. The gland at the wall becomes the part that sets the rating, so use the gland the maker documents for the enclosure type, and size the cutout with the gland and cutout guide.

NEMA 250 defines Type 4X by hose-directed water and corrosion resistance, and dew is not in that test; the IP rating chart separates what a rating covers from what it skips, and the NEMA 3R vs NEMA 4 guide shows what each outdoor type adds. Inside any sealed box, manage the microclimate with a silica gel pack for short programs or a 10–20 W anti-condensation heater once missions run all day. When the life target stretches from 50 hours to years, plan for drains, heaters and quarterly gasket checks from the start. For the box itself, a custom stainless enclosure to your aperture and gland layout is one piece, delivered in 10 business days from approval of the drawing.

What do you check before each mission?

Four checks, two minutes: forecast dew point at least 3 °C below every surface, board and cover visibly dry, bias-to-ground insulation in the hundreds of MΩ on the meter, and 10 mm [0.4 in] of clearance between foam and radiating elements.

Run them in order and stop at the first failure:

1. Dew point. Read the forecast for the exact mission window; every surface must sit at least 3 °C above it.
2. Warm-up. Rig, cover and connectors at outdoor temperature for 30–60 minutes before uncovering; a cold-soaked board meeting warm air fogs instantly.
3. Dryness. Board, foam and connectors visibly dry; wipe off the dust from the last session.
4. Insulation. Rail-to-ground on the meter reads in the hundreds of MΩ on a dry board; under about 10 MΩ, dry it again before power — bench rule of thumb.
5. Cover. 10 mm [0.4 in] of air at every radiator, foam intact and dry inside.
6. Close-out. Back in the case with a silica gel sachet before the rig cools, lid shut.

Check 2 is the one that saves the mission. The return loss of a fogged 40 GHz front end wanders for exactly as long as the water stays, and 30 minutes of patience costs less than any part on the bench.

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.

Questions