Why Blowing Things Off With Compressed Air Is a Regulated Act

Why this matters

Blowing chips, dust or swarf off a part with an air hose is the most ordinary act in a shop, and it is one of the few acts in this trade that a federal standard addresses by name and permits only conditionally. The conditions are three, they apply at the same time, and satisfying two of them is not partial compliance. It is the same violation. That matters commercially as well as legally: a shop doing air system work walks past a dozen blow-off stations on every visit and can tell a customer, accurately and in one sentence, which ones would not survive an inspection and what the fix costs in hardware rather than in argument.

Two hazards before anything else. Never direct a stream at skin or clothing, including your own, and do not put a hand in front of an outlet to check the flow; air can enter through a break in the skin and travel in the bloodstream. And an open or poorly-engineered nozzle is loud enough to be a real exposure, so wear hearing protection under a program meeting 29 CFR 1910.95 for any sustained work beside one.

The rule, and both of its homes

29 CFR 1910.242(b), in the general industry standards' hand and portable powered tools subpart, says compressed air shall not be used for cleaning purposes except where reduced to less than 30 psi, and then only with effective chip guarding and personal protective equipment.

29 CFR 1926.302(b)(4), in the construction standards' tools subpart, carries the same restriction for construction work. A field-service shop can fall under either depending on what the job is, and it is worth knowing which one is yours before you quote a fix, because the surrounding requirements around each differ even though this sentence does not.

Read the sentence as a structure rather than as a number. It is a prohibition with a three-part exception: pressure, guarding, protection. Everything below is about what each part actually requires and how to check it.

Part one: what "less than 30 psi" is measuring

This is the condition that is misread most often, in both directions.

It is not the header pressure. It is not the regulator setting on the drop. OSHA has long read the limit as the pressure at the point of blockage: the static pressure the nozzle can develop when its outlet is dead-ended or obstructed, which is the condition under which the stream could be driven into a body. A nozzle supplied at 90 psig that relieves through side ports so that a blocked tip cannot develop more than a low static pressure satisfies this condition, and that is exactly what an engineered safety nozzle is for. A plain open tube supplied at 25 psig also satisfies it, and does the job badly and loudly.

So the question is never "what is the shop running?" It is "what does this tip develop against a blocked outlet?" A shop that answers by pointing at the header gauge has not answered.

The failure mode of getting this backwards is expensive in the wrong direction: a plant turns its whole system down to below 30 psig to comply, every tool in the building loses power, someone puts it back up on Monday, and the blow-off stations are now non-compliant with no record of why the pressure moved.

Part two: effective chip guarding

The stream throws material, and the standard's word is effective, which is a performance requirement rather than a hardware list. What is being guarded is the path from the work to a person, and there are two people to think about: the operator, and whoever is walking past.

In practice, effective guarding means the debris is contained or deflected before it reaches either. A downdraft bench, a three-sided enclosure at the station, a screen between the station and the aisle, or a shielded nozzle arrangement can all be effective. A pair of safety glasses on the operator is not chip guarding; it is the third condition, and treating it as the second is the single most common way a station fails on paper.

The test to apply on a walk: stand where the next person along would stand and ask what stops a chip getting to them. If the answer is that the operator points the nozzle carefully, there is no guarding.

Part three: personal protective equipment, matched to the route

The standard says personal protective equipment without listing it, which means the selection comes from the hazard assessment and from the PPE standards that govern each route.

• Eyes and face. Flying particles trigger eye and face protection under 29 CFR 1910.133, and for a blow-off station where material comes back at the operator that normally means a face shield over primary eye protection rather than glasses alone, because a face shield is secondary protection and is not a substitute for the glasses under it.
• Hearing. Where the station's noise puts an operator over the action level, hearing protection under a program meeting 29 CFR 1910.95 applies, and an engineered nozzle usually reduces the noise as a side effect of reducing the flow.
• Breathing. This is the route that gets dropped. Blow-off makes an aerosol of whatever it removes. If the debris is respirable, eye and hand protection do nothing about it, and the control is respiratory protection selected under a program meeting 29 CFR 1910.134, or better, a method that does not make the aerosol at all.

The gate, run against two tasks

Task one: chips off a machined aluminium part at a bench station. The nozzle is an engineered type that relieves when blocked, verified below 30 psi at the tip against a dead-end. The bench is a three-sided enclosure with a screen facing the aisle, so debris does not leave the station. The operator wears a face shield over safety glasses. All three conditions hold at once, and the act is permitted. The remaining question is not compliance but whether it is a sensible use of air at all, which the article on inappropriate uses in this library takes up.

Task two: dust off a formwork surface before a pour, on a construction site. Fit the same engineered nozzle. Build the same screen. Issue the same face shield. All three conditions of 1926.302(b)(4) now hold, and the act is still not acceptable, because the dust is concrete and masonry dust and blowing it into the air is a respirable crystalline silica exposure governed by 29 CFR 1926.1153 in construction, and by 29 CFR 1910.1053 for the same material in general industry. Those standards drive work toward wet methods and local exhaust and away from dry dispersal, so the compliant answer here is a different method entirely: wet sweep, vacuum with appropriate filtration, or a tool with integrated dust collection.

That is the whole point of running one gate against two tasks. 1910.242(b) and its construction twin are a conditional permission, not a grant of authority. Satisfying every word of them tells you that this standard does not prohibit the act. It tells you nothing about whether some other standard does, and the material you are blowing is what decides that.

Verifying the dead-end pressure without lying to yourself

You cannot take a nozzle's compliance on faith and you cannot read it off the drop's regulator. Measure it.

Assemble a gauge to a fitting that threads onto or seals against the nozzle outlet, rated above the supply pressure, and mechanically restrained so nothing can be ejected. Never block a tip with a rag, a glove or a finger. Supply the nozzle through its normal regulator at the pressure the station actually runs, and read the static pressure with the outlet fully blocked.

Then read your gauge's plate before you read your gauge, because this is a measurement made near a hard legal threshold and the instrument's error basis decides whether your reading means anything.

Say the gauge's plate states 1 percent of full scale, which is illustrative and you must read your own. On a 0 to 200 psig gauge that is a bound of plus or minus 2 psi at every reading, including at 28 psig, where 2 psi is about 7 percent of the reading. A manufacturer's stated accuracy of that kind is a worst-case bound, so it is reported as a bound and not as a plus-or-minus interval you can average away: a reading of 29 psig on that gauge is consistent with a true value of 31 psig, which is over the limit, and you cannot report compliance from it.

Put a 0 to 60 psig gauge on the same tap and the same 1 percent of full scale is a bound of plus or minus 0.6 psi, about 2 percent of a 28 psig reading. Now 29 psig is bounded at 29.6 and the finding stands up. Nothing about the nozzle changed. The rule that generalises: when a measurement is going to be compared against a fixed threshold, pick an instrument whose span brackets the threshold, because a percent-of-full-scale bound does not shrink as the reading does and a wide-span gauge spends its entire error budget on a small reading.

Record the reading, the gauge span, the stated basis, the supply pressure at the drop and the date, and put it with the station rather than in a file. A nozzle that was verified two years ago and has since been replaced with whatever was in the drawer is a station with a document and no compliance.

The uses the rule never permitted at any pressure

Two acts sit outside this exception entirely, and reducing the pressure does not bring them inside it.

Cleaning a person. Blowing dust off clothing, hair or skin is not a cleaning task the pressure reduction makes acceptable. The hazards are air entering through broken skin, driving debris into the eye at close range, and moving contamination from clothing into the breathing zone of the person wearing it. The control is a vacuum with appropriate filtration or a change of clothing, not a lower setting.

Clearing a blockage or a chip nest by pushing air into a confined cavity. The stream does not remove the material so much as launch it, and the direction it leaves is not the direction the nozzle is pointing. Mechanical removal, then vacuum.

References

• 29 CFR 1910.242(b), general industry, hand and portable powered tools and equipment; and 29 CFR 1926.302(b)(4), construction, tools: the paired basis for the three-condition permission and for knowing which Part governs your job.
• 29 CFR 1910.133, eye and face protection, general industry: the basis for face and eye protection selection where flying particles are present at a blow-off station.
• 29 CFR 1910.95, occupational noise exposure, general industry: the basis for hearing protection where nozzle noise reaches the action level.
• 29 CFR 1910.1053, respirable crystalline silica, general industry, and 29 CFR 1926.1153 for construction: the basis for treating blow-off of concrete or masonry dust as a regulated exposure that the 242(b) exception does not reach.
• See related: What an Inappropriate Use of Compressed Air Looks Like; and this library's articles on pressure measurement and instrument error.