Short answer: split the build in two compartments. Your particle sensor and humidity sensor live in a vented shelter that opens downward behind stainless insect mesh; your ESP32, its power supply and every interface board live in a sealed stainless box rated NEMA 4X / IP66, with one cable gland and 20 g of indicating silica gel. Condensation on interface chips, not rain, is what kills these builds. Separate the sample air from the electronics and the failure mode disappears.
Why the electronics die and the sensor survives
A typical failure runs like this: the particle sensor's inlet sits in a hole on the underside of a sealed case, insect screen over it, everything else inside. After a stretch of humid weather the monitor stops reporting. Dried out, the processor comes back; the interface chips around the sensor stay dead. The sample path carried moist air into the case, and the night sky finished the job.
Run the dew point numbers once. Air sealed at 25 °C and 60% RH has a dew point near 17 °C. A metal box radiates to a clear night sky and sits a few degrees below ambient, so a case that reached 25 °C in the afternoon can drop through 17 °C before morning. The condensate lands on the coldest, smallest parts you own, which are the interface chips. For the wider physics, read condensation control in an enclosure without rain; here the fix is structural.
The vented sensor chamber
This chamber keeps rain, insects and sun off the sensor heads while leaving the air path wide open, and it always opens downward. A fan-aspirated particle sensor of the PMS5003 class pulls its own sample through, so the chamber needs openings rather than a fan.
Take the geometry from weather-station practice. A 2024 radiation-shield study in the journal Measurement built its shield from stacked plates, with three middle plates shaped to steer airflow toward the sensor probes. Copy that idea: three horizontal plates above the sensor, each wider than the sensor body, edges open to the wind. The plates shade the heads and stop driven rain, while the open edges keep the sample fresh.
Cover every opening with stainless woven mesh at about 1 mm [0.04 in]. A PM10 particle is 10 µm across, so the openings are a hundred times wider and nothing you count is caught. Keep the sensor intake at least 2 in [50 mm] above the lowest edge, out of splash range.
For the chamber itself, an 8 x 6 x 4 in wall-mount box works, drilled with a vent pattern on the face that points down. Once drilled it has no enclosure rating, and nothing inside needs one, because the sensor heads are built to see outside air. Your humidity sensor lives here as well: it must see ambient air to read it, and the outdoor humidity sensor shelter comparison weighs the options. If you are still weighing a vent against a sealed box in general, the vent-versus-sealed comparison settles it by climate.
The sealed electronics compartment
Everything else goes in its own box, for this build a 10 x 8 x 6 in 304 wall-mount. These boxes are rated NEMA 4X / IP66 by the manufacturer; they are not UL listed, and the NEMA 4X versus IP66 guide covers what each mark means. In US installations, NEC 110.28 states that an IP rating is not a substitute for an enclosure type rating, so order by type when the panel goes to a US site.
Put the gland in a side wall; gland orientation on an outdoor junction box explains why the top face stays undrilled. Match the seal insert to your cable's outside diameter, and land the sensor harness the way outdoor sensor cable connections for a sealed enclosure shows, with a drip loop below the entry.
Desiccant closes the moisture budget. Indicating silica gel picks up about a quarter to a third of its own weight at 50% RH, the figure its datasheets carry. Twenty grams therefore holds roughly 5 g of water, while the 7.6 L [0.27 ft³] of air inside this box carries about 0.14 g of water at 25 °C and 80% RH. The pack covers the sealed air thirty-five times over, so the real load becomes door openings and cable jackets. Held below 50% RH, the air must cool from 25 °C to about 14 °C before anything condenses: 11 K of margin.
Keep vents, drains and fan-filters off this box. Each accessory carries a rating of its own, and an assembly drops to the rating of its weakest part. The sealed side stays sealed, and the vented chamber does the breathing.
Heat in a 10 x 8 x 6 in sealed box at 40 °C ambient
The inputs, each with its source:
- Electronics load, 3 W: 2.5 W worst case for an ESP32-class board, at 0.5 A and 5 V, the supply size its datasheets call for, plus about 0.5 W for the interface boards.
- Wall area, 0.24 m²: the outside surface of a 10 x 8 x 6 in [254 x 203 x 152 mm] box, from the catalog.
- Wall heat transfer, 3.7 W/m²K: the value this site uses for stainless walls in every heat calculation.
- Ambient, 40 °C [104 °F]: a hot summer day.
The rule in words: at steady state the walls carry the whole load, so the air inside rises by the load divided by the wall coefficient times the wall area.
The arithmetic: 3.7 W/m²K x 0.24 m² = 0.89 W/K of shedding. Then 3 W ÷ 0.89 W/K = 3.4 K of rise. Inside air settles at 40 °C + 3.4 K, about 43 °C [110 °F].
The decision: ESP32-class modules carry an 85 °C operating rating in their datasheets, so you sit 42 °C below the limit. Order the box with no cooling at all. This size holds a 20 °C rise up to about 18 W (3.7 x 0.24 x 20 ≈ 18 W); above that, step up a size or move to a sealed heat exchanger, because a fan-filter trades the rating for airflow.
Material, mounting and sun
Match the grade to the site. 304 stainless covers inland and most outdoor locations; order the same box in 316L within sight of salt water, near brine, or where winter road salt reaches the mounting. The stainless enclosure range filters by grade and size, and wall-mount boxes run from junction-box sizes up.
Mount the shelter with its opening down and the electronics box beside or below it. Keep the sensor intake about 12 in [300 mm] off the wall so it samples free air rather than the wall's dust layer, at 2 to 3 m height for the breathing zone and out of casual reach. The outdoor mounting guide carries the fundamentals. On the electronics box, sun load dwarfs 3 W of electronics, and the solar shield numbers show what shading saves. On the shelter, the stacked plates are the shade.
What to put on the order
| Item | Specification for this build | Notes |
|---|---|---|
| Electronics box | 10 x 8 x 6 in wall-mount, 304 stainless, or 316L near salt | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| Sensor shelter | 8 x 6 x 4 in wall-mount, vent pattern drilled on the face that points down | No enclosure rating once drilled; none needed inside |
| Mesh | Stainless woven, about 1 mm [0.04 in] openings | Passes PM10 and PM2.5 with nothing caught |
| Gland cutout | One 20 mm [0.79 in] hole in a side wall of the electronics box | Suits an M20 gland matched to the cable OD |
| Desiccant | 20 g indicating silica gel | Replace when the color flips |
| One-piece option | Custom box with an internal partition: vented half opens down, sealed half keeps its gasketed door and gland | Delivered in 10 business days from your drawing approval |
Both routes are delivered in 10 business days: the catalog boxes counted from your order, the partition build from your approval of the drawing. Custom enclosures take material, dimensions and every cutout from you, from one piece up, so the two-compartment layout can arrive as a single welded body with one mounting footprint.
Enclosures in this article
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 10 × 8 × 6 in | 304 stainless | 10 x 8 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 10 × 8 × 6 in | 316L stainless | 10 x 8 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 8 × 6 × 4 in | 304 stainless | 8 x 6 x 4 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 8 × 6 × 4 in | 316L stainless | 8 x 6 x 4 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Can the particle sensor sit outside with just a hood over it?
Yes. The sensor's own fan pulls the sample, so it needs rain and sun protection only: a hood that extends at least 4 in [100 mm] past it on the weather side, plus 1 mm [0.04 in] mesh against insects. Expect UV to embrittle a plastic hood within a few seasons; stainless holds up indefinitely.
How often does the desiccant need changing?
Change it when the indicator flips color. In a box whose door stays shut, most builds settle on every 3 to 6 months; check monthly through the first humid summer and after any door opened in damp weather. A 20 g pack holds about 5 g of water before it is spent.
Does the sealed box need a breather drain?
No. The desiccant does the moisture work, and a membrane breather admits outside air on every temperature swing, which is the exact air you excluded. An open drain or louver would also drop the sealed assembly below its rating, since an assembly follows its weakest part.
Will the mesh lower my particle readings?
No. PM10 particles are 10 µm across and the 1 mm [0.04 in] mesh openings are a hundred times wider, so nothing you count is caught. What does distort readings is a dust-loaded mesh: wash it when it looks grey, because a clogged screen starves the sensor's fan before it stops particles.
Can both compartments be one enclosure?
Yes, as a custom one-piece build: an internal partition splits a vented half, opening downward, from a sealed half with its own gasketed door and gland cutout. It is delivered in 10 business days anywhere in the world, counted from your approval of the drawing.
Does this build need a UL listed enclosure?
For a monitor you wire and maintain yourself, the manufacturer rating is what you buy: these boxes are rated NEMA 4X / IP66 by the manufacturer and not UL listed. Where a plant requires listed equipment, use a UL listed box for that installation; the vented shelter never needs a listing at all.
