Comment tester l'air d'un espace confiné: Règles de l'OSHA, Ordonnance sur les tests de gaz & Meilleures pratiques

Le premier espace de permis que j'ai jamais signé était une fosse de vannes en fonte, trois mètres de profondeur, coincé entre une route très fréquentée et une conduite de gaz. The entry supervisor handed me a four-gas monitor that, when I checked the records, hadn’t been bump tested in nine days. I asked about it. He shrugged and said, “It was fine yesterday.”

We pumped the space anyway, took the readings, climbed down. Nothing went wrong that day. I’ve never forgotten the shrug, cependant. Most monitoring programs don’t fail in the equipment, and they don’t fail in the standard. They fail in the gap between “we have a monitor” et “we verified what it’s telling us.”

If you’re here because your crew works in tanks, égouts, vaults, silos, or pits, you already know confined spaces kill quietly. What you might not have is a straight answer to what compliant confined space gas monitoring looks like on a normal workday.

Here’s the short version. Test from outside before anyone enters, in the right order, at every level, and keep monitoring while people are inside. Do it with a monitor you’ve verified that same day.

Confined space gas monitoring is the practice of measuring and continuously tracking the atmosphere inside a permit-required confined space, typically for oxygen, gaz combustible, and toxic gas, before and during entry as required by OSHA 29 CFR 1910.146.

Why Confined Space Gas Monitoring Can’t Be Skipped

Atmospheric hazards are the leading cause of confined space fatalities, and they’re almost always invisible. A tank can look perfectly harmless and sit at 16% oxygène. A sewer can smell like nothing and hold enough hydrogen sulfide to drop a worker in seconds. Methane can pool in the upper half of a vault while the air at waist height reads clean. You can’t see, smell, or hear any of it, which is why the monitor is doing the real work.

OSHA estimates roughly 2.1 million workers enter permit-required confined spaces in the United States every year. The agency’s confined space standard exists because entering an untested space is so often fatal. Confined space deaths tend to run in a chain: one worker goes in, a second goes in to help, sometimes a third. Rescuers account for a big share of confined space fatalities, which tells you how fast these situations go wrong.

OSHA’s answer is blunt. Before an employee enters a permit space, the internal atmosphere shall be tested with a calibrated, direct-reading instrument. Judgment calls don’t qualify, and neither does “it was fine yesterday.” A test, with a working instrument, before entry.

Everything else in this article is about doing that one test right.

What Gases Do You Monitor in a Confined Space?

OSHA’s Appendix B to 1910.146 spells out the testing order: oxygen first, then combustible gases, then toxic gases. The order isn’t a formality.

Oxygène (O₂). L'air normal est 20.9% oxygène. For confined space entry, the accepted safe band is 19.5% à 23.5%. Below 19.5% you’re in oxygen deficiency territory, where judgment slips first and consciousness goes second. Au-dessus de 23.5% you have an oxygen-enriched atmosphere, where materials that normally just smolder decide to burn violently. Oxygen is tested first for a practical reason: most combustible gas sensors need enough oxygen to produce a true reading, so a dead or skewed oxygen reading poisons everything after it.

Gaz combustibles. Flammable gas readings are expressed as a percentage of the lower explosive limit, or LEL. Most confined space programs treat 10% LEL as the ceiling for entry. Above that, the space gets ventilated and retested before anyone goes in. Méthane, propane, vapeur d'essence, hydrogène: they all show up on the LEL channel. If you want the mechanics of how LEL is measured, our combustible gas detector guide goes into the details.

Toxic gases. This is where the space’s history decides what you measure. The standard four-gas setup covers carbon monoxide and hydrogen sulfide. If your space sits near ammonia refrigeration, chlorine dosing, or diesel exhaust, you need to add those sensors too. Here are the common toxics with their exposure limits. Treat this table as a starting point, not the final word: limits change, and your program should always check current OSHA and NIOSH values.

Gaz	Typical source in confined spaces	OSHA PEL	IDLH
Monoxyde de carbone (CO)	Combustion, generators, welding	50 ppm (8-hr TWA)	1,200 ppm
Hydrogen sulfide (H₂S)	Sewage, petroleum, decomposition	20 ppm ceiling	100 ppm
Ammoniac (NH₃)	Refrigeration, fertilizer	50 ppm (8-hr TWA)	300 ppm
Chlore (Cl₂)	Traitement de l'eau, chemical plants	1 ppm ceiling	10 ppm
Sulfur dioxide (SO₂)	Industrial emissions	5 ppm (8-hr TWA)	100 ppm
Dioxyde d'azote (NON₂)	Diesel exhaust, welding	5 ppm ceiling	20 ppm

Now the part people skip. Gases don’t mix evenly inside a confined space. Methane and other light gases ride the top. Hydrogen sulfide and heavier vapors sink to the bottom. Carbon monoxide hangs around wherever it pleases. A reading taken at chest height tells you nothing about the top of the tank or the floor of the pit, which is why proper testing samples the upper, middle, and lower levels of the space, every single time.

OSHA Requirements for Confined Space Gas Monitoring

The regulatory core is 29 CFR 1910.146, and Appendix B to the standard gives the procedures for atmospheric testing. Here’s what the standard actually requires, in plain language.

Testing before entry. The atmosphere must be tested before any employee enters, with a calibrated, direct-reading instrument, and the results have to be recorded on the entry permit.

Evaluation testing vs. verification testing. Evaluation testing happens first, before the space is prepared, to find out what hazards are actually present. Verification testing comes after ventilation and preparation, to confirm the space now meets the safe entry conditions written on the permit. They are two different steps. Treating them as one is a common mistake, and it’s the kind of mistake that shows up in an audit as “atmosphere not verified.”

Continuous or periodic monitoring. If conditions can change while people are inside, the standard expects you to keep monitoring. Soudage, nettoyage chimique, material movement, even a shift in ventilation can move the atmosphere mid-job. If there’s any chance conditions change, the entrant stays connected to a live monitor, not just to a clipboard.

Retest after ventilation. Ventilating a space doesn’t automatically make it safe. The atmosphere has to be retested after ventilation, before re-entry, every time. I’ve seen crews ventilate for twenty minutes, take a single reading at the manway, and call it good. That reading covered about a foot of a twelve-foot tank.

Documentation. Readings go on the permit, next to the acceptable entry levels. If it’s not written down, from an inspector’s point of view it didn’t happen.

You don’t need to memorize subsection numbers to run a compliant program, but you do need the standard’s text on hand. Link the full 1910.146 standard and its Appendix B, which covers atmospheric testing procedures, into your procedure documents so the crew can check them.

How to Test a Confined Space: Step-by-Step Procedure

Here’s the procedure I use on site. It maps directly to OSHA’s Appendix B, and it’s the part I’d print out and tape to the permit board.

1. Confirm the space type. Is it a permit-required confined space or a non-permit space? The permit carries the hazard list and the acceptable entry conditions. If there’s no permit program in place, stop here and build one before anything else.
2. Verify the instrument that morning. Bump test or full calibration check before the first use of the day. A monitor that fails the test doesn’t go in the space, no exceptions.
3. Test from outside. Nobody enters until the initial test is done from outside the space. Use a pump-style monitor with a sample line or probe lowered into the space. This is not optional.
4. Test in order. Oxygen first, then combustibles (Lel), then toxics. That order exists for a reason: every reading after oxygen depends on the oxygen reading being real.
5. Sample every level. Top, middle, bottom. If the space is deep, sample at roughly four-foot intervals as you work down. Stratification is the rule, not the exception.
6. Give the sensors time to respond. A monitor pulled out after two seconds reads the air at the opening, not the air at the bottom. Watch the numbers settle. Long sample lines need extra time for the air to actually reach the sensors.
7. Write it down. Record the readings on the permit before entry, with the time. This is the documentation step, and it’s non-negotiable.
8. Keep monitoring during entry. Continuous monitoring while people are inside, and retest after ventilation changes, breaks, or anything that could shift the atmosphere.

That’s the whole procedure. It’s not complicated, and the spaces that get people hurt are the ones where someone decided one of the eight steps didn’t apply to them.

Which Confined Space Gas Monitor Should You Use?

The equipment question is easier than most people make it, as long as you start from the hazard list instead of the product catalog.

The standard tool is a four-gas monitor: oxygène, Lel, monoxyde de carbone, et du sulfure d'hydrogène. If that term is new to you, our article on what a 4 gas monitor is covers the basics, including what each sensor actually measures. Four gases cover what most confined space programs need, and for a lot of teams that’s the whole fleet.

The bigger decision is pump versus diffusion. A diffusion monitor samples the air right at the sensor, which is fine for a worker wearing it on a harness. A pump monitor draws air through a sample line, which is how you test a space from the outside before entry, and how you sample at depth without dropping a person in. For confined space work, a pump model should be the default. We build both: le Otywell G60 four-gas detector covers the standard four gases with a twenty-hour run time and a 300,000-record data log, et le S311 pump-type gas detector is built specifically for pre-entry sampling through a sample line.

The sensors inside matter too. Catalytic bead sensors handle combustibles but need oxygen to work. Electrochemical sensors cover most toxics and are what you’ll find in the CO and H₂S channels. Infrared sensors handle combustible gas in oxygen-starved atmospheres where a catalytic bead would fail. Photoionization detectors catch VOCs and other volatile compounds. Notre guide to the four gas detector sensor types explains the trade-offs in detail.

When do you need more than four gases? Ammonia in a cold storage plant. Chlorine near water treatment. VOC or specific toxics in chemical work. If your space can contain something beyond O₂, Lel, CO, and H₂S, you need a monitor that measures it, and we’ve covered when a confined space detector needs more than four gases in a separate guide.

The full selection logic, including wireless alarm linking and fall detection for spaces where the attendant can’t see the entrant, lives in our confined space gas detector buying guide. The short version: match the sensors to the hazard list, make sure it can sample from outside, and don’t buy on price alone.

One more equipment thought. If a space sees constant work, some facilities add fixed gas detection to cover it around the clock, with the portable units handling entry and personal protection. The trade-offs between fixed and portable gas detection are worth reading before you build that out.

Bump Testing, Étalonnage & Maintenance

This is the part where most programs quietly fall apart.

Bump test before each day’s use. A bump test exposes the sensors to a known concentration of gas and confirms the monitor actually detects it and actually alarms. It takes under a minute with a bump gas cylinder or calibration station. If the monitor doesn’t alarm, it doesn’t go in the space, et “it alarmed yesterday” doesn’t count.

Full calibration on a schedule. Calibration adjusts the sensor readings against known gas concentrations, which corrects the drift that happens with normal use and exposure. How often depends on the manufacturer, the sensor type, and how hard the instrument works. For life-safety confined space entry, many programs calibrate monthly or quarterly, and always after a failed bump test or suspected sensor damage.

Keep the records. Calibration logs, bump test results, and sensor replacement dates are the paperwork that keeps you out of trouble in an audit. More importantly, they’re how you know the monitor told the truth. A sensor that’s been in service for four years past its rated life is a liability, not a tool.

I’ve watched crews treat a beeping monitor as a malfunction and pull it out of the space to “fix” it. Nine times out of ten the monitor was right and the atmosphere was the problem. Trust the instrument, but only after you’ve verified it that morning.

Confined Space Gas Monitoring FAQs

What is the OSHA oxygen level for confined space entry?

The accepted safe range is 19.5% à 23.5% oxygène. Below 19.5% is oxygen deficiency; above 23.5% is an oxygen-enriched atmosphere.

What is a safe LEL in a confined space?

Flammable gas should stay below 10% of the lower explosive limit for entry. Above that, ventilate and retest before anyone goes in.

How often should a confined space gas monitor be calibrated?

Bump test before each day’s use. Full calibration per the manufacturer’s schedule, typically monthly to quarterly for confined space entry, and after any failed bump test.

Quels gaz un 4 détecteur de gaz?

Oxygène (O₂), gaz combustible (Lel), monoxyde de carbone (CO), et du sulfure d'hydrogène (H₂S).

Can you test a confined space before entering?

Oui, and you must. Use a pump-style monitor with a sample line from outside the space, and test oxygen, combustibles, and toxics at multiple levels.

Do you need continuous gas monitoring in a confined space?

If conditions can change during the job, yes. Soudage, chemical use, and ventilation changes can all shift the atmosphere while people are inside.

Build a Program, Not a Checklist

A confined space program isn’t a binder on a shelf. It’s the monitor you verified that morning, a crew that can rattle off the testing order, a supervisor who actually writes the readings down, and an attendant who keeps watching after the ladder goes in.

If your team needs the equipment side sorted, the confined space gas detector buying guide walks through choosing the right detector based on how your crew actually works, whether the attendant can see the entrant or not. If you want to talk through your specific spaces, Otywell builds portable and fixed gas detection systems for confined space work, and we’re happy to help you spec the right setup for your crew and your hazards.

Test from outside, en ordre, at every level, and keep testing while people are inside. Do that and you’ve covered most of what actually kills people in these spaces.