Shop Air Quality: Why a Cheap PM2.5 Monitor Misses Wood Dust, and What to Buy Instead

The cyclone spools down and the shop goes quiet. Outside it's the usual damp Seattle evening, cedar and fir doing their work across the fence line. Inside, the low light coming through the roll-up door picks out everything the dust collector didn't get: a slow drift of fines that will still be hanging there when I come back after dinner.

That drift is a different engineering problem from the one the collector solves. Capture at the tool is a ducting and velocity problem. What's left in the room afterward is an ambient problem, and the instinct is to go buy a small particle monitor and put a number on it. I've come to believe that instinct leads most shops to the wrong instrument, for reasons that are in the sampling literature rather than in anybody's marketing.

Wood Dust Is a Coarse Aerosol, and the Rules Are Written That Way

Start with what's actually in the air. A field study of personal samplers across seven US wood-products facilities — veneer, plywood, engineered flooring, door skins, shutters, cabinetry, all of it cutting, drilling, sanding and sawing — reported in the Annals of Occupational Hygiene via PubMed Central that "most airborne wood dust particles are >10 μm in size" and that "relatively little mass of wood dust is found in smaller particle sizes." That's 888 valid samples from 444 pairs, with no significant difference between any pairing of sampler type.

A furniture-factory study of pedunculate oak published in BioResources (NC State) puts a number on the split: the respirable share of the inhalable fraction ran between 12% and 31%, and stayed below 16% for the samples above 2 mg/m³. Belt-sander inhalable concentrations there ran 1.569 to 3.710 mg/m³.

The health framing follows the same size logic. IARC classifies wood dust as a Group 1 human carcinogen, with sufficient evidence for cancer of the nasal cavity, paranasal sinuses and nasopharynx, and notes that in humans "coarse particles (2.5–10 μm) deposit by impaction in the nasal region" — the same region where those cancers occur. OSHA's woodworking eTool lists oak, mahogany, beech, walnut, birch, elm and ash as hardwoods reported to cause nasal cancer in woodworkers.

The limits are written in those terms too. OSHA's PEL is 15 mg/m³ total dust (5 mg/m³ respirable) as an 8-hour TWA, the NIOSH REL is 1 mg/m³, and ACGIH sets a 0.5 mg/m³ TLV for western red cedar on the basis of asthma. Those are milligrams per cubic meter of dust caught gravimetrically on a filter and weighed.

What the Sensor Behind a Cheap Monitor Actually Sees

Consumer monitors in this class are built around laser-scattering modules, most often the Plantower PMS family. Two peer-reviewed papers settle what those modules can do.

Kuula et al., in Atmospheric Measurement Techniques (2020), evaluated six sensors against a TSI aerodynamic particle sizer and a Grimm spectrometer and found the PMS5003's valid detection range sits below 0.7–0.8 µm, concluding it cannot measure coarse 2.5–10 µm particles. The Nova SDS011 was "not suitable for the measurement of coarse-mode particles." Their headline finding: "none of the low-cost sensors adhered to the detection ranges declared by the manufacturers."

Kaur and Kelly, also in AMT (2023), took the PMS5003 into five coarse-dust events in the Salt Lake Valley against a federal-equivalent beta-attenuation monitor. The three-sensor average tracked reference PM10 with a slope of 0.0463 — it moved about 5% as much as the reference — with R² between 0.128 and 0.482. A true optical particle counter, the Alphasense OPC-N3, hit R² of 0.865 to 0.937 in the same events.

EPA's own long-term evaluation of six sensor models states the general case: "PM air sensors often measure particles only 0.3 to 1 µm with variable success," with mean bias error running from +13 to −18 µg/m³ across sites. EPA is equally direct about indoor use, saying there are "currently no widely accepted indoor performance criteria" for low-cost monitors and that the agency "does not generally conduct evaluations of low-cost air pollution monitors under indoor test conditions." Its published sensor performance targets don't help either; the Air Sensor Toolbox FAQ limits them to non-regulatory monitoring in "ambient, outdoor, fixed-site environments."

For a consumer data point: BreatheSafeAir tested the Temtop M10+ at $129.99 and found it doesn't report PM10 at all, and against an EPA-corrected reference it was "likely overreporting PM2.5 concentrations significantly." That's the cheapest monitor price I can verify.

So here's the honest position. I have no source for what fraction of a specific shop's wood dust a PM2.5 monitor registers; that's an experiment, not a finding. What the literature does support is narrower and still decisive: the sensor is blind to the size range that carries most of the mass, and a µg/m³ PM2.5 reading is not the same quantity as a mg/m³ inhalable-dust limit (1 mg/m³ is 1,000 µg/m³). Use the monitor as a relative trend line for the fine tail. Do not read it as a dose.

CADR Tops Out at 11 Microns

Clean Air Delivery Rate is defined by ANSI/AHAM AC-1. Allison Bailes at Energy Vanguard lays out the three test aerosols — smoke at 0.09–1.0 µm, dust at 0.5–3 µm, pollen at 5–11 µm — and the arithmetic, which is simply airflow times filtration efficiency: 600 cfm × 0.5 = 300 cfm CADR. Bliss Air notes the standard uses a 1,008 cubic foot test chamber and covers 0.10 to 11 micron.

AHAM Verifide's sizing rule is that CADR should be at least two-thirds of the room's area in square feet; Washington State DOH translates that to roughly 5 air changes per hour at an 8-foot ceiling, with 7.5 ACH if CADR equals full room area. EPA's home guide gives 195 cfm minimum for a 300 sq ft room.

Note where that leaves us. AHAM's largest test particle is 11 µm, and most wood dust mass is above 10 µm. CADR describes the fraction that stays airborne longest, which is arguably the fraction an ambient cleaner exists to handle, but it was never designed to characterize sawdust.

Nominal CFM on the Ceiling, Measured CADR on the Floor

None of the woodworking-brand product pages I have in front of me publish a CADR figure. They publish fan CFM and a micron claim. Prices below were retrieved 2 August 2026.

Unit	Price	CFM (speeds)	Filtration claim	Replacement filters
Rockler Dust Right Compact	$399.99 (Rockler)	250 / 350 / 500	"down to 1 micron", triple-stage	$29.99 + $39.99
Rockler Dust Right Full-Size	$399.99 (Rockler)	550 / 700 / 1,100	"down to 1 micron"	$29.99 + $39.99
Jet AFS-1000C	$549.99 (Rockler)	6 speeds, 450–1,000	"down to 1 micron"	$43.99–$91.29
Powermatic PM1200	$649.00 (Powermatic)	547 / 706 / 1,196	see below	not published

The PM1200's filtration claim is genuinely contested between sellers. Rockler's listing says it "traps up to 99% of all 5-micron particles and 85% of all 1-micron particles." Powermatic's own page lists a 1-micron filter and says the unit "reduces airborne particles by 99.7%." Those are not the same claim, and neither is a CADR.

Run cost per nominal CFM and you get roughly $0.36 for the Rockler Full-Size, $0.54 for the PM1200, $0.55 for the Jet, and $0.80 for the Rockler Compact. That's my arithmetic on published prices and published airflow, and airflow is not delivered clean air.

The DIY side has been measured properly. EPA's Office of Research and Development ran box-fan builds against wildfire smoke in a lab: 111 ± 1 CADR for a fan with one 1-inch MERV 13 filter, 156 ± 4 with a cardboard shroud added, 248 ± 15 with a 4-inch filter and shroud, 263 ± 22 with two 1-inch filters and a shroud, and 401 ± 31 for a four-filter Corsi-Rosenthal box with a shroud. EPA's own summary of that work: the shroud raises CADR by 40% "without any change in the cost or physical footprint," and DIY cleaners "were almost completely ineffective with dirty filters."

HouseFresh gives the cleanest head-to-head, same room and same PurpleAir Zen sensor for both: an $85 Corsi-Rosenthal box measured 198 cfm PM1 CADR and cleared a 728 cubic foot room in 30 minutes, at 40.5–46.7 watts and 49–60.7 dB; a $99.99 Levoit Core 300 measured 109 cfm and took 54 minutes. Clean Air Crew reports an Illinois Institute of Technology chamber test of a MERV 13 box at 166 CFM for 0.09–1 µm, 321 CFM for 0.5–3 µm and 464 CFM for 5–11 µm, which is the size-dependence in one line: these things get better as particles get bigger. Clean Air Crew also credits David Elfstrom with measuring a 47% airflow gain from a 13.5-inch shroud on Utilitech and Hurricane fans and 38% from a 15-inch shroud on Lasko Comfort fans. Jim Rosenthal at Tex-Air Filters, a filter seller and co-originator of the design, measured face velocity on one fan at one speed: 780 fpm bare, 320 fpm through a single 20×20×1 MERV 13, 400 fpm through a 2-inch, 460 fpm through a 4-inch. Those are feet per minute at the face, not CFM.

The Clearance Math, and What It Assumes

Milne et al. in the Canadian Journal of Anesthesia give the working relationship: ACH equals flow rate divided by enclosed volume, times 60, under cessation of the source and perfect mixing. From that, 99% clearance takes 276 ÷ ACH minutes. Applied to a 20 × 20 × 8 shop (3,200 ft³), using only measured CADR inputs, my arithmetic gives:

Configuration	CADR source	eACH	99% clear
Corsi-Rosenthal, 4× MERV 13 + shroud	EPA, 401	7.5	~37 min
Box fan, two 1" MERV 13 + shroud	EPA, 263	4.9	~56 min
CR box, measured PM1	HouseFresh, 198	3.7	~74 min
Box fan, one 1" MERV 13, no shroud	EPA, 111	2.1	~133 min

There is no honest row here for a ceiling unit, because I found no measured particle-removal rate for one. Jet's claim of 19.58 air changes per hour for the AFS-1000B in that same room is nominal airflow divided by volume with no efficiency term at all. Perfect mixing is also a fiction in a shop full of benches and machines.

P100 Is the Cheap Option, Not the Premium One

Under 42 CFR Part 84, as OSHA describes it, there are nine filter classes: N (not oil resistant), R (oil resistant) and P (oil proof), each at 95%, 99% and 99.97% efficiency, tested at 85 L/min against particles of roughly 0.3 µm median aerodynamic diameter. OSHA's 1910.134 assigns an APF of 10 to half masks, and Note 3 to that table puts filtering facepieces and elastomeric half masks in the same category. On paper, a disposable N95 and a rubber half mask are rated identically.

Measured performance says otherwise. He et al. (2015), testing 25 subjects against 10–400 nm particles, found geometric mean simulated workplace protection factors of 110 and 114 for two N95 filtering facepieces versus 4,571 and 9,420 for two P100 filtering facepieces, and 108 and 358 for N95 elastomeric half masks versus 11,046 and 12,605 for P100 elastomerics.

Now the prices, from Enviro Safety Products on 2 August 2026 (a retailer's listed sale prices, not a market survey):

Item	Class	Price
3M 7093B cartridge, 2 ea.	P100	$6.99
3M 2091 filter pair	P100	$8.89
3M 2097, P100 + nuisance OV	P100	$11.89
3M 5N11 filter pair	N95	$18.89
3M 8233 disposable, each	N100	$10.79
3M 8293 disposable, case of 20	P100	$238.79 (~$11.94/mask)

Read the fourth row against the second. On the same page on the same day, a pair of P100 filters costs less than half what a pair of N95 filters costs, and the P100 carries the far higher measured protection factor. If you already own a half mask, there is no economic argument for N95 media.

Where I'd Put the Money

Buy the P100 cartridges first. They cost less than the alternative, and against the 10–400 nm particles He et al. tested they carry the far higher measured protection factor.

For ambient cleanup, build the box fan with a newer fan carrying a UL or ETL mark, add the shroud, and change the filters early. EPA's finding that dirty filters made these things almost completely ineffective is the sentence I'd tape to the housing.

The monitor is the last thing I'd buy, and I'd buy it knowing what it is: a trend line for the fine tail, useful for telling you whether the room is getting better after you sweep, useless for telling you whether you're under a limit. Designing with intent means picking instruments that measure the thing you care about. For wood dust, the cheap one doesn't.