Machine Guarding in Plain Language

If a moving part can reach a person, something has to stand between them. Most guarding citations come down to that one sentence.

Machine guarding under 29 CFR 1910.212 produces some of the most severe injuries in manufacturing — amputations, crushes, degloving. It also produces some of the most preventable ones, because the entire standard reduces to a single idea: if a moving part can reach a person, something has to stand between them.

Three hazard areas

Point of operation. Where the machine actually does its work — the cut, press, form, bend, shear. This is where the worst injuries happen because it is where hands have to be nearby.

Power transmission. Belts, pulleys, chains, sprockets, flywheels, gears, couplings and shafts that move power from the motor to the work. Covered under 1910.219.

Other moving parts. Reciprocating, rotating and transverse components — feed mechanisms, auxiliary parts, anything that traverses.

Programs commonly get the point of operation right and then leave a chain drive open behind a machine because nobody works there. Nobody works there until someone sweeps there.

What a guard has to do

A guard is not just a barrier. To meet the intent it must:

• Prevent contact with the hazardous area
• Be secured so it cannot be casually removed or defeated
• Create no new hazard — no pinch points, no sharp edges, nothing that can shear
• Not interfere with the work, or it will be removed within a week
• Allow lubrication and routine service without full removal wherever possible

That fourth point is the one engineers underestimate. A guard that makes the job meaningfully harder has a short life expectancy on a production floor.

Types of safeguarding

• Fixed guards — a permanent barrier, the most reliable option and the first choice
• Interlocked guards — opening the guard cuts power and stops the machine
• Adjustable guards — accommodate varying stock size
• Self-adjusting guards — move aside as the stock passes, then return
• Presence-sensing devices — light curtains and safety mats that stop the machine when the field is broken
• Pullbacks and restraints — physically keep hands clear of the point of operation
• Two-hand controls — both hands must be occupied for the cycle to run

Presence-sensing devices deserve care. They must be positioned so the machine stops before the hand reaches the hazard, which depends on the stopping time of the machine. A light curtain mounted too close is decorative.

The removed-guard problem

The most common serious finding on an audit is a guard that was removed for maintenance and never went back. It is rarely defiance. Someone needed access, took the guard off, got called to another job, and the machine ran a full shift open.

Two controls fix most of this:

• Guard removal is tied to the lockout procedure — the guard goes back before the lock comes off
• Operators are trained and expected to refuse to run a machine that is missing a guard, and are backed by supervision when they do

The second only works if a worker who shuts down production is treated as having done the right thing. If they get pressure instead, you have taught the whole floor a different lesson.

Anchor the machines that need it

Machines designed for a fixed location must be secured to prevent walking or moving during operation. Bench grinders are the classic example, along with tolerances people forget: the work rest adjusted to within one-eighth inch of the wheel, and the tongue guard within one-quarter inch.

A five-minute walkthrough

Take one pass across your floor and look for:

• Any rotating part you could touch without moving something first
• Guards secured with zip ties, wire, or nothing at all
• Interlocks bypassed with tape, magnets or a jammed switch
• Light curtains mounted too close to the point of operation
• Bench grinder gaps out of tolerance

Anything found in that pass is visible to an inspector too.