The Cylinder That Fell in the Shop

Why this one is worth reading

Nobody was hurt, and that is the only reason this is a near-miss write-up instead of an injury report. A full nitrogen cylinder holds well over 2,000 psi in a steel shell weighing more than a hundred pounds. If the valve shears off when it falls, the cylinder becomes an unguided projectile that can go through a block wall, and the cap is the part designed to stop that from happening. In this case the cap was off, the valve held, and the shop got to learn from it for free. The account below runs in the order it should be run: first the five minutes after the bang, then the reasons the cylinder was where it was.

The first five minutes

A ten-person HVAC shop, 7:20 am, receiving area. A helper turning a loaded manual pallet jack near the side door clipped a nitrogen cylinder that was standing loose, leaning against the wall beside the door frame. It went over and the valve end hit the concrete. Everyone in the bay heard it.

What the shop lead did, in order, and why each step came before the next:

1. Got people out of line with the cylinder. Everyone was moved away from both ends of the cylinder and away from the line of the valve outlet. A cylinder that is going to fail at the valve will do it along its axis, and a damaged valve can let go a few seconds after the impact, not only at it.
2. Listened, then looked. With the bay quiet, the lead listened from about 10 feet away for any hiss. None. Then, standing to the side of the valve rather than in front of the outlet, the lead checked the valve visually: handwheel bent at the rim, a fresh gouge on the valve body, outlet threads undamaged.
3. Checked for a slow leak. A spray of leak-detection solution on the valve stem, the packing nut and the outlet. No bubbles after a full minute. Had there been a leak that the handwheel would not close, the instruction was: open both bay doors, clear the building, and call the gas supplier's emergency line from outside. Nitrogen is not toxic or flammable, but it displaces oxygen, and in a closed room a large release can leave too little to breathe without any smell or warning.
4. Stood it up without touching the valve. Two people, gloves on, lifted from the shoulder and the body, never by the valve or handwheel, walked it upright, and chained it to the nearest wall bracket. A cylinder is never lifted by its valve, and a damaged one least of all.
5. Took it out of service. A tag reading "DAMAGED VALVE - DO NOT USE" went on the neck, the shop manager called the supplier, and the cylinder went back on the next delivery as a damaged return. The shop did not try to straighten the handwheel or test the valve on a regulator. A valve that has taken a blow can have internal damage that only shows under use, and repairing a cylinder valve is the owner's job, which here is the supplier.

Only after the cylinder was chained and tagged did anybody ask why it had been standing by the door.

What the floor showed

The shop manager walked the receiving area and the cage with the lead that afternoon and wrote down what was there, before anybody tidied up.

• The cage was full. It had 10 chained positions and all 10 were occupied.
• 4 of the 10 caged cylinders were empty. None of them was tagged as empty. Two were nitrogen, one oxygen, one acetylene. Nobody could tell full from empty without putting a regulator on each.
• 2 cylinders were outside the cage. The nitrogen that fell, which was full, and an empty oxygen cylinder lying flat on the floor behind the cage, out of sight.
• No restraint near the door. There was no wall chain or strap anywhere in the receiving area, so a cylinder put down there had nothing to hold it.
• Caps in a bucket. Six valve protection caps were sitting in a bucket on top of the cage. Nobody knew which cylinders they had come off.

That is 12 cylinders on hand that morning against 10 spaces, and 5 of the 12 empty: the 4 in the cage and the one on the floor.

Why the cylinder was by the door

The obvious story was carelessness: someone set the cylinder down and walked away. That was true and it explained nothing, because the next question was why a careful tech had nowhere else to put it.

The trace went back two weeks. The supplier delivers on a swap basis: the driver takes empties and leaves fulls. The driver only takes empties that are staged at the pickup spot and clearly marked. Because nothing in the cage was marked, the driver on the last two deliveries took nothing back and dropped the ordered fulls. Each delivery added cylinders without removing any. The cage filled, and the next full nitrogen had to go somewhere. The tech who unloaded it leaned it by the door intending to sort it out after the morning dispatch, and never did.

So the fall had three causes, each of which on its own would not have produced it:

• No way to tell empties from fulls, which broke the swap and overfilled the cage.
• No restraint outside the cage, so the overflow stood loose.
• A pallet jack route through the overflow spot. The side door is the route from receiving to the parts room. The cylinder was in the path of the one piece of equipment that moved through that spot every morning.

The cap

The cap on the fallen cylinder had been taken off the day before by a tech who put a regulator on it to confirm it was full before loading a van, then decided to take a different cylinder and left this one uncapped. It was not malice or laziness; there was no rule saying the cap goes back on, and the caps lived in a bucket. A valve protection cap is a heavy steel shroud screwed over the valve on the cylinder's neck ring, and it exists for this exact event. With the cap on, the valve would probably not have been touched at all. OSHA's welding and cutting standard, 29 CFR 1910.253, requires valve protection caps to be kept in place on cylinders designed to take them unless the cylinder is connected for use, and 29 CFR 1910.101 requires in-plant storage and handling of compressed gas cylinders to follow the Compressed Gas Association's pamphlet P-1. Read the current text of both for the exact wording.

What the shop changed

The fixes were matched to the three causes and the cap, one for one.

• A chained cage with two marked sides. The cage was rebuilt to 14 positions with a painted line down the middle: FULL on one side, EMPTY on the other, every position with its own chain at about two thirds of cylinder height. Oxygen moved to its own small cage more than 20 feet from the acetylene, which is the separation 29 CFR 1910.253(b)(4)(iii) requires unless a noncombustible barrier at least 5 feet high with at least a half-hour fire rating stands between them.
• An empty is tagged the moment it is empty. The tech who finishes a cylinder closes the valve, puts the cap on, hangs an "MT" tag and moves it to the empty side before leaving the shop. On delivery day the empties are on that side and the driver takes them.
• Caps on, every time it is not connected. The bucket is gone. A cylinder that is not on a regulator has its cap screwed on, including when a tech checks pressure and decides not to take it.
• A wall chain at receiving. Two strap points by the side door, for a cylinder that has to be set down for a minute. It is still not storage; it is a place to hold one upright until it is caged that same hour.
• A cap on how many may be on hand. The full side has 7 positions. The shop orders only when at least 2 are free, and never more fulls than there are free positions, so a delivery cannot arrive with nowhere to go. Ordering against "we might run out" is how the old cage reached 12 cylinders in 10 spaces.

The weekly check, as the shop now runs it

Done Friday by the shop lead, before the Monday delivery, and filed with the monthly safety walk sheets.

CYLINDER CAGE WEEKLY       Date ______   By ______
POSITIONS  full side 7   empty side 7   O2 cage 4
COUNT      full ___   empty ___   O2 ___
OUTSIDE ANY CAGE (should be 0)             ___
EVERY CYLINDER CHAINED                     Y / N
CAP ON EVERY CYLINDER NOT CONNECTED        Y / N
EVERY EMPTY TAGGED MT, ON EMPTY SIDE       Y / N
O2 CAGE 20 FT FROM FUEL GAS                Y / N
ACETYLENE STANDING VALVE END UP            Y / N
FREE FULL POSITIONS (order only if 2+)     ___
NOTES _______________________________________

The first Friday it ran, it found one cylinder with no cap and one empty on the full side, both from the same van restock. Both were fixed on the spot. The point of a weekly count is that small drift gets caught while it is one cylinder, not twelve.

What would still make this go wrong

• A delivery on a day nobody is at the shop. If the driver arrives at 6 am with no one there, fulls get left wherever the driver can reach. The shop now books deliveries for a window when the parts lead is in.
• A different cylinder type. A refrigerant recovery cylinder or a disposable refrigerant jug does not follow this cage's rules; those live on their own shelf (see related: Refrigerant and Recovery Cylinders in the Shop). Mixing them into the nitrogen cage would bring back the "is it full or empty" problem in another form.
• A move. The next building's receiving door and pallet jack route will be different, and the cage location gets decided by the route, not by where there is a blank wall.

References

• OSHA 29 CFR 1910.101 (compressed gases, general requirements, referencing CGA P-1) and 29 CFR 1910.253 (oxygen-fuel gas welding and cutting, cylinder storage and separation).
• Compressed Gas Association P-1, Safe Handling of Compressed Gases in Containers, in the edition OSHA's standard references.
• See related: The Compressed Gas Cylinder Storage SOP; Storing Compressed Gas Cylinders at the Shop.
• See related: The Monthly Shop Safety Walk SOP.