Carbon monoxide: the one gas a direct-reading meter can settle
This page explains what the rules and common practice call for. Every site is different, and what a survey or a filing needs at your facility is something we work out with you after we see it.
Carbon monoxide (CO) is the one substance on this site where a direct-reading meter, which shows the concentration on the spot, gives data good enough for your OSHA record. It is also the one substance where waiting three weeks for a laboratory report is the wrong choice.
OSHA has a fully validated method that uses a direct-reading electrochemical sensor (OSHA 1026). It is not a screening tool or a “for indication only” instrument. It is a validated method.
That changes the cost and the timing of a CO survey completely, because you can find the source, change something and measure again the same afternoon.
For almost every other substance we sample for, the meter you can afford to carry is a screening tool, and the defensible number comes from a filter or a tube analyzed in a laboratory. CO is the exception, and it is worth knowing why before you assume other substances work the same way.
Exposure limits
The limits are in parts per million (ppm). OSHA's permissible exposure limit (PEL) is a time-weighted average (TWA), while NIOSH's recommended exposure limit (REL) pairs a TWA with a ceiling limit, a level that must not be exceeded at any time.
| Limit | Averaging time | |
|---|---|---|
| OSHA PEL | 50 ppm | 8 hours. No ceiling limit, no short-term exposure limit (STEL) |
| NIOSH REL | 35 ppm TWA · 200 ppm ceiling limit | Up to 10 hours / ceiling |
| Immediately dangerous to life or health (IDLH) | 1200 ppm | Not applicable |
Only NIOSH has a limit for short excursions. Its 200 ppm ceiling is the number to watch during brief events, such as a forklift idling near a dock door or a generator started in a partly enclosed space. OSHA's 8-hour TWA will absorb a short spike without moving much.
Where we find it
- Propane and gasoline forklifts run indoors. This is the single most common industrial CO exposure, and it gets worse in winter when the dock doors close.
- Propane floor burnishers and scrubbers. Retail, warehouse and food distribution, where the work is usually done by an overnight crew that rarely gets sampled.
- Portable generators. Construction, restoration, disaster response, and the deadly case of a generator in an attached building.
- Welding and thermal cutting, especially with processes shielded by carbon dioxide (CO₂) and in confined work.
- Confined space entry. Tanks, vessels, vaults, sewers and any space downwind of an engine.
- Ice rinks. Propane and gasoline ice resurfacers. This is a well-documented, recurring exposure that reaches spectators as well as workers.
- Boiler rooms and fuel-fired equipment. Leaking flue gas, backdrafting and poorly tuned burners.
How we sample for it
| Primary method | OSHA 1026. A direct-reading electrochemical sensor that logs its readings |
| Alternate method | OSHA ID-210. Aluminized gas bag, GC-DID (air collected in a bag, then analyzed in a laboratory) |
| NIOSH method | NIOSH 6604 |
| Biological monitoring | Carboxyhemoglobin, the form CO takes in the blood. It measures the dose a person absorbed, not the concentration in the air |
| Sample type | Personal sample in the worker's breathing zone, for the full shift, with the data log retained |
The OSHA direct-reading method has one exclusion worth knowing: do not use it where hydrogen, acetylene or ethene are present. Those gases interfere with the electrochemical sensor, which rules the method out in the very places you might reach for it. Battery charging areas give off hydrogen and oxy-fuel work gives off acetylene, so method selection here is not a formality.
What makes a direct-reading CO result defensible is the documentation, not the meter. Three things matter:
- Calibration and bump test records dated around the survey, with the gas concentration and lot number. A bump test is a quick check that the sensor responds to a known gas. A reading from an uncalibrated meter does not stand up as a record.
- Readings logged often enough to show what you care about. A log at 1-minute intervals shows you the forklift passing, while a 15-minute average smooths it into nothing.
- The data log itself, retained and reported. Much of the value of CO monitoring is in the readings over time, where you can see the shift start, the door closing and the second truck arriving. A report that gives only an 8-hour TWA discards the most useful thing the survey collected.
That sensor interference matters in the field too, beyond method selection. Near battery charging, you may see a CO reading that does not line up with any combustion source, and it deserves a second look before anyone spends money on ventilation.
A blood or breath test answers a different question
A carboxyhemoglobin test, on blood or breath, measures what a person absorbed, while air monitoring measures what was there to breathe. The two differ for legitimate reasons: respirator use, how hard the person works, and smoking. Smoking raises a person's baseline carboxyhemoglobin level, whatever the workplace is like.
Both tests are sound, and neither replaces the other. A program that uses biological monitoring as its exposure record has no way to tell which piece of equipment caused the result.
Why meters win here and lose almost everywhere else
CO is a stable gas that does not react under normal room conditions, and it is present at concentrations well within the range of a low-cost sensor. There is essentially one sensor chemistry for it, and it works reliably.
Other substances are harder. Beryllium has to be measured at 0.2 micrograms per cubic meter of air (µg/m³), hexavalent chromium can change form on the filter before it reaches the laboratory, and isocyanates need a reagent that binds them on contact. None of those can be read off a handheld display.
In practice, a CO survey gets you answers faster than any other survey on this site, and it can offer something the others cannot: a before-and-after comparison in the same visit. Move the propane burnisher's route, prop a door open or re-tune a burner, then measure the change while the crew is still there.
→ What an air monitoring survey costs in 2026
When you may not need this
If your only concern is whether a space is safe to enter, that is a confined space or hot work question, and your own calibrated four-gas meter answers it. You do not need a consultant to tell you a vault reads 300 ppm.
A survey is worth ordering in three cases:
- You need an 8-hour exposure record for a job title, and it has to stand up to an inspector or an auditor.
- Workers have symptoms and no one has found the source.
- You are about to spend money on ventilation and want to know whether it will work.
The third case is the most common. CO usually comes from a small number of specific machines, and identifying which ones costs less than ventilating the building.
Plan a carbon monoxide survey
Ron Fleming has done this work himself since 1993. You talk directly to the person who does the job. No account managers. No call center. Our prices are well below the competition. Tell us what is going on and we will send a written scope of work and a price, at no cost.
Talk with us. No cost.