PM Sensors in Dusty Sites: Protecting the Inlet and Reading Through Dust
Share
- The normal failure mode in a dusty site is slow upward drift: dust films the optics and the sensor reads a background that is not in the air at all.
- The dangerous failure is the opposite — readings that fall during the dustiest events, because a blocked inlet starves the sample flow exactly when the data matters most.
- The cheapest check costs nothing: after rain, a healthy sensor should fall toward background within an hour or two — a floor that rises month by month is the earliest drift warning.
- Choose a sensor with a serviceable inlet and a fan flow alarm, and clean the inlet and chamber on a schedule set by the site, not by the calendar.
- Remember the limit: a scattering sensor trends and screens dust well, but it is not a reference sampler — do not present its numbers as if it were.
Table of contents
The quarry's dust monitor had been running for three weeks when the numbers stopped making sense. On a still, damp morning — the kind of morning when the haul road is wet and nothing is moving — the PM10 reading sat stubbornly high, higher than it had been on a dry, windy afternoon with trucks running. The sensor was not broken. It was choking on the very dust it was supposed to measure, and the data had been drifting away from reality for days before anyone noticed.
That is the normal failure mode of a particulate sensor in a dusty site. A PM sensor does not measure dust the way a thermometer measures temperature. It counts particles, one by one, and turns the count into a mass concentration — and everything in that chain, from the inlet to the laser to the maths, is vulnerable to dust. If you run a PM sensor at a quarry, a cement plant, a grain terminal, a construction site, or a desert weather station, the question is not whether the environment will attack the sensor. It is whether you have designed the installation so the sensor keeps telling the truth anyway.
How a laser-scattering PM sensor actually measures dust
Most outdoor PM2.5/PM10 sensors work by laser scattering, also called light scattering. Air is drawn through a small chamber by a fan. A laser beam crosses the chamber, and when a particle passes through the beam it scatters light in all directions. A photodetector catches some of that scattered light, and the brightness of the flash is roughly proportional to the particle's size. The sensor counts flashes, sorts them into size bins, and converts the counts into a mass concentration — micrograms per cubic metre — using an assumed particle density and a standard particle-size distribution.
Three things about that chain matter in a dusty environment. First, the conversion from count to mass is an assumption, not a measurement. The sensor assumes particles have a typical density and shape; mineral dust from a quarry is denser and more irregular than the urban aerosol the factory calibration assumed, so the same count can mean a different mass. Second, the fan and the inlet define what actually reaches the chamber. If the inlet is partly blocked, or the fan is struggling against dust loading, the sample flow drops, fewer particles arrive, and the reading falls — or the flow becomes uneven and the reading becomes noisy. Third, the laser and the detector live in a chamber that dust can coat. A film of fine dust on the optics scatters the beam even when no particle is passing, and the sensor starts reading a background that is not in the air at all.
On top of all that sits the size cut itself. PM10 means particles with an aerodynamic diameter of 10 micrometres or less; PM2.5 means 2.5 micrometres or less. In a dusty site the PM10 channel carries the load, and it is also the channel most affected by inlet design, because big particles are heavy: they do not follow air around corners, they fall out of slow flows, and they jam inlets that fine particles would pass through.
Why dusty sites break the readings in specific ways
The most common failure is a slow upward drift. Dust accumulates inside the sensing chamber and on the optics, and the baseline creeps up. The giveaway is the clean-air test: on a wet, still morning, or after rain, a healthy sensor in a dusty site should fall close to the local background.
The second failure is the opposite: readings that fall when dust rises. A blocked or partially blocked inlet starves the chamber, so during the dustiest events the sensor reports less than it should, exactly when the data matters most. This is the dangerous one, because it looks like good news: a manager who sees low readings during a dusty shift concludes the dust controls are working, when in fact the sensor has stopped sampling properly.
The third failure is noise. A fan labouring against a clogged inlet, or a chamber with dust rattling around in it, produces spiky readings — sudden jumps and drops that no real plume explains. The data looks like a seismograph instead of a trend, and averaging hides the problem instead of fixing it, because the average of bad samples is still a bad number.
And then there is the failure nobody checks: the sensor measuring the wrong air. A PM sensor mounted downwind of a haul road measures the road, not the site boundary. Placement is part of the measurement, and in a dusty site it is the largest part.
Choosing a PM sensor for a dusty site
Start with the question the data has to answer. A sensor that supports a compliance report needs a documented inlet, a stated measurement principle, and a cleaning and calibration routine you can actually follow; a sensor that only needs to trigger a dust-suppression sprinkler can be simpler, because a rough threshold is enough. Do not buy a sensor designed for a clean office and expect it to survive a quarry — check whether the enclosure, the inlet and the fan are built for outdoor, dusty service, and whether the manufacturer documents a cleaning interval at all.
Look for a few practical features. A sensor with a serviceable inlet — one you can open, inspect and clean without tools you do not have on site — will get maintained; one that needs the unit dismounted and sent away will not. A fan with a tachometer or a flow alarm is worth more than a slightly better specification on paper, because it tells you when sampling has degraded instead of letting you discover it in a monthly report. An enclosure rated for outdoor use matters, but the rating is not the whole story: an IP-rated box still needs its inlet kept clear and its cable gland sealed.
Be honest about what a low-cost scattering sensor cannot do. It cannot distinguish dust types, it cannot match a reference gravimetric sampler, and its absolute accuracy in heavy mineral dust will be worse than in the urban air it was calibrated in. If your permit requires reference-grade numbers, a scattering sensor is a screening and trending tool, not the compliance instrument — and buying it as if it were the compliance instrument is how a site ends up arguing with a regulator over numbers the sensor was never able to produce.
Installation: protecting the inlet and the sample
The inlet is the whole game. Mount the sensor so the inlet faces open air, not a wall, a roof edge, or the lee side of a structure where dust settles and eddies. Keep it clear of the dust sources you are trying to measure around rather than inside: at least a few metres from a haul road, a conveyor discharge, or a crusher, unless the point of the sensor is to measure that source directly.
Fit an inlet hood or rain cap if the sensor's design allows one, and check that it does not become a dust trap itself. Angle the inlet downward where the design permits — a downward-facing inlet sheds rain and large particles that would otherwise fall in — but do not improvise an inlet modification the manufacturer does not support, because changing the inlet geometry changes the size cut and invalidates the calibration.
Keep the sensor out of direct exhaust and vibration. Mounting it on a vibrating conveyor frame shakes dust loose inside the chamber and shortens the fan's life; mounting it beside a generator exhaust coats the optics with soot, which is a different failure from dust and much harder to clean.
Verification: proving the sensor still tells the truth
The single most useful check costs nothing: watch the sensor after rain. A good rain washes the air, and a healthy sensor should fall toward background within an hour or two. If it stays high, something is wrong inside the sensor — a dirty chamber, a failing fan, or a coated optic. Do this check deliberately and record it; a series of after-rain readings is a drift history, and a floor that rises month by month is the earliest warning you will get.
Once a quarter, or monthly in a very dusty site, do a side-by-side check: put a second sensor — even a portable handheld unit — next to the fixed sensor for a day and compare. They do not need to agree exactly; they need to track each other. If one rises and the other does not, or the gap between them grows over the day, one of them is lying, and the fixed sensor is the usual suspect.
Check the fan and the inlet on the same visit. Listen to the fan — a fan that has slowed sounds different, and a fan that has stopped is obvious if you listen. Look into the inlet with a torch: a visible ring of dust, a cobweb, an insect nest, or a bent inlet tube are all things you can fix in a minute. And log what you find. A cleaning log turns "the sensor has been drifting" into "the sensor drifts about this much between monthly cleans,"
Troubleshooting table
| Symptom | Likely cause | Check |
|---|---|---|
| Readings creep up over weeks, never fall to background after rain | Dust film on optics or inside the sensing chamber | Clean the chamber and optics per the manual; compare against a portable sensor before and after |
| Readings fall during the dustiest events | Blocked or partially blocked inlet starving the sample flow | Inspect the inlet, check the fan sound and any flow alarm, clear obstructions |
| Spiky, noisy readings with no real plume to explain them | Fan struggling against dust loading, or loose dust inside the chamber | Check fan operation and clean the chamber; check the mount for vibration |
| PM10 high but PM2.5 flat, or the two channels move independently | Coarse dust dominating, or inlet geometry changed by damage or a modification | Inspect the inlet for damage or improvised changes; confirm the sensor is measuring the intended air |
| One sensor disagrees with a neighbour that used to track it | One of the two has drifted — usually the one that lives in the dust | Side-by-side comparison with a clean portable unit; clean and re-check the suspect sensor |
Checklist: PM sensors in dusty sites
- Mount the inlet in open air, clear of haul roads, discharges and sheltered eddies; choose the height for the question being asked (exposure vs boundary).
- Angle the inlet down where the design permits and keep a rain cap from becoming a dust trap.
- Keep the sensor off vibrating structures and away from exhausts; mount it solidly and leave it reachable for cleaning.
- Do the after-rain check regularly: a healthy sensor falls toward background within an hour or two of rain.
- Clean the inlet and chamber on a schedule set by the site, not by the calendar — monthly in heavy dust, quarterly at least.
- Run a side-by-side check with a portable unit every quarter; track the gap, not just the absolute reading.
- Listen to the fan and look into the inlet on every visit; log cleanings and comparisons so drift becomes a history.
- Remember the limit: a scattering sensor trends and screens dust well, but it is not a reference sampler — do not present its numbers as if it were.
A RS485 smart air quality sensor (PM2.5 / PM10), mounted with a clear inlet, cleaned on a schedule the site can actually keep, and checked against rain and against a portable reference, keeps telling the truth in the dust — and a sensor that is never cleaned, never compared, and never checked after rain is a display that reports whatever its chamber has become.
Our engineers can help you pick the right model, check compatibility, and get your installation right the first time.
RS485 Smart Air Quality Testing Sensor (PM2.5 PM10)View product →
Related Blogs
The Site Got a Noise Complaint but Your Sensor Showed Nothing: Compliance Monitoring on Construction Sites
A Weather Station Kit, Unboxed: What's Inside and the Assembly Order That Saves a Day
The Classroom Feels Stuffy but the CO2 Sensor Says Fine: Ventilation Rates from CO2 Data
Orchard Sensors Keep Dying in Deep Soil: Placing a 5-in-1 Probe for Deep Roots and Drip Lines