What Hand Signals Are For and When They Stop Working

Why this matters

A hand signal is a command with no expiry. The operator does what the last signal told them to do and keeps doing it until a signal says stop, so the moment a signal person disappears behind a column the crane is not paused, it is still running the last order. This card follows one courtyard set where that gap lasted four seconds and moved a suspended load about eight feet with nobody commanding it, and it names the four preconditions the hand signal protocol depends on, one of which nobody had written down.

The stop signal comes first

Before any of the mechanics: under 29 CFR 1926.1419 for construction work, anyone may give an emergency stop signal and the operator must obey it. Not only the designated signal person. That rule exists because the person who sees the problem is frequently not the person who is authorised to talk.

The two things every worker on a site with a crane on it needs to know are the stop signal and where the swing path is. Everything else in this card is for the people running the lift.

The operator's half of the same rule matters as much: if the operator cannot see the signal person, or does not understand the signal given, the operator stops. A signal that has to be interpreted has already failed, and guessing is how a load ends up somewhere nobody asked for.

A hand signal is a command with no timeout

This is the mechanism the rest of the card hangs on and it is worth stating plainly. The Standard Method signals for hoist, lower, swing and travel are continuous commands. The operator begins the motion on the signal and holds it until a stop signal, a change of command, or the operator's own judgment intervenes. There is no dead-man behaviour in the protocol, and there is no default to stop when the signal source goes quiet.

That is not a flaw in the system, it is what makes it usable: a signal person cannot re-issue a swing command four times a second. But it means the failure mode of losing the signal person is motion continuing, not motion stopping, and crews consistently expect the opposite. Ask around a crew what happens if the signal person vanishes mid-swing and a fair number will say the crane stops. It does not. The operator stops it, which takes as long as it takes the operator to notice.

Voice signals under 29 CFR 1926.1421 are built differently for exactly this reason. That section requires each voice signal to contain three elements in order: the function and direction, the distance or speed, and the function stop. The stop is inside every command rather than being a separate later command, so a lost radio ends a motion that was already bounded instead of leaving one running.

The four preconditions

Hand signals work when all four of these hold, and the protocol has no graceful degradation when one fails.

One designated signal person, and only one giving signals at any time. 1926.1419 requires that only one person give signals at a time, aside from the universal stop. Two people signalling is not redundancy, it is ambiguity, and the operator's correct response to ambiguity is to stop.

Continuous, unobstructed line of sight in both directions. The signal person sees the operator; the operator sees the signal person. The signal person also has to be able to see the load or the point it is going to. Sun angle, backlighting against a bright sky, rain, dust and low light all break this without breaking anything you can point at.

An agreed signal set, posted. 1926.1419 requires the type of signal to be agreed before operations begin, hand signals to follow the Standard Method in Appendix A to Subpart CC unless a non-standard signal is genuinely needed and agreed in advance, and 29 CFR 1926.1422 requires a hand signal chart to be posted either on the equipment or conspicuously in the vicinity of the hoisting operation. The posted chart is what makes the agreement checkable by somebody who arrived after the pre-lift meeting.

A position the signal person can hold. Outside the fall zone, outside the swing path including the counterweight tail swing behind the machine, with a clear view, and reachable without walking backwards. This is the precondition nobody writes down, and it is the one the case below breaks.

The case: a chiller into a courtyard

A mobile crane set on the street, lifting a packaged chiller over a two-storey wing and down into an interior courtyard. Hand signals throughout, one designated signal person on the street, chart posted on the machine, pre-lift meeting held. The pick and the initial hoist went exactly as planned.

The set did not. From the street position, the last part of the swing carried the load behind the building wing, and the courtyard landing point was never visible from where the signal person was standing. The plan said the signal person would "move to maintain visibility of the load."

That sentence is what failed. Moving to see the load meant walking around the corner of the wing, and walking around the corner meant losing sight of the operator. The signal person did the thing a person does under pressure: kept walking sideways and backwards, arm up, trying to hold both.

The operator lost the signal person behind the corner of the wing, held the swing command that was still in effect, and stopped the machine when nothing came back.

What was true: the four seconds

The crew timed the gap afterwards from the machine's own slew and their video: about four seconds between the last visible signal and the operator stopping.

The swing rate in use was roughly 3 degrees per second, a moderate rate for a set. Four seconds at that rate is about 12 degrees of rotation.

• Load radius: 38 ft
• Arc travelled by the load: 38 ft x 12 degrees in radians, which is 38 x 0.2094 = 8.0 ft

The load moved about eight feet with nobody commanding it. It did not hit anything, because the courtyard was open at that point in the arc. Eight feet earlier in the swing it would have been over the roof.

The second number is the one that made the crew go quiet.

• Distance from the centre of rotation to the rear of the counterweight: 14 ft
• Arc swept by the counterweight in the same 12 degrees: 14 ft x 0.2094 = 2.9 ft

The signal person, walking backwards to hold two sightlines that could not both be held, had come inside that arc. The counterweight passed close enough that they felt it. Nobody was struck, and the reason nobody was struck is where the crane happened to be in its rotation, which is not a control.

Two mechanisms, pointing the same way. The line of sight failed at the corner of the wing, which stopped new commands arriving. The command already in effect kept the crane swinging, which moved the load and the counterweight. Neither of those is a mistake anybody made in the moment; both are properties of the protocol that were live from the moment the plan said "move to maintain visibility".

The switch that should have happened before the load moved

The fix is not a faster signal person and it is not a slower swing.

A second signal person cannot simply pick up mid-swing. Only one person may give signals at a time, so an unmanaged handoff creates the exact ambiguity the rule exists to prevent. Where a handoff is genuinely needed it happens with the load stopped: the load stops, the handoff is confirmed out loud on a channel both parties share, the second signal person confirms they have the operator and the load, and only then does motion resume.

The real answer here was to switch protocols before the lift. Any lift with a segment the signal person cannot cover from one position is a radio lift, decided at the plan, not a hand-signal lift with a walking signal person. 29 CFR 1926.1420 sets the requirements when signals go by radio, telephone or other electronic transmission: the device must be tested on site before operations begin, transmission must be through a dedicated channel except where multiple cranes share one and the operator can reliably identify which signals are theirs, and the operator's reception must be hands-free. Combine that with the three-element voice protocol at 1926.1421, agreed between the operator, the signal person and the lift director before operations begin, and the segment behind the wing becomes ordinary rather than heroic.

And the plan gets a switching condition, in advance. "Switch to radio when the load passes the parapet" is a decision. "Move to maintain visibility" is a wish, and a wish is what a person under pressure will try to satisfy with their feet.

Where each method actually fails

Hand signals and radio fail differently, which is the argument for having both rather than choosing one.

Hand signals fail on sight. Distance beyond the point where a hand is legible, backlighting against a bright sky, sun in the operator's eyes, rain, dust, dusk, and any structure between the two people. They also fail on crowding: a signal person standing among other workers in similar high-visibility clothing is hard to pick out of a group, which is why the designated signal person is normally distinguished by something the operator can find instantly.

Radio fails silently. A dead battery, a channel someone else keyed over, a squelched receiver, and the dead spots inside a steel structure. The dangerous property is that a radio failure looks identical to nobody speaking, so the operator has no cue at all unless the protocol includes periodic acknowledgment. That is one reason to keep hand signals as the backup wherever the signal person is visible: the two methods fail on unrelated conditions.

Both fail the same way when a crew gets comfortable. Signals shrink. A full-arm swing becomes a wrist flick by mid-morning, a stop becomes a raised finger, the operator learns them, and then a relief operator arrives who has not. The posted chart is the reference the crew is supposed to be converging on rather than drifting from.

What to check before the first pick

Every item below is done with the machine stationary, the boom stowed or still, and everyone outside the swing radius and the tail swing arc.

• Name the signal person out loud, to the operator, and confirm the operator can pick them out visually from where they will be sitting.
• Confirm the signal type is agreed and that the hand signal chart is posted on the equipment or conspicuously nearby, as 1926.1422 requires.
• Walk the signal person's position for every segment of the lift. Stand where they will stand and check both sightlines from there. If any segment needs them to move, that segment gets a switching condition rather than a plan to walk.
• Test the radio on site before operations, both directions, from the actual positions, not from the tailgate.
• State the switching condition as an event anybody can observe, and say who calls it.
• Confirm everybody on site knows the stop signal, including trades who are not part of the lift, because the rule that anyone may give it is worth nothing if nobody else knows it.
• Mark the tail swing. It is behind the operator, it is invisible from most of the site, and it is the part of a mobile crane that most reliably surprises people on foot.

References

• 29 CFR 1926 Subpart CC, cranes and derricks in construction, including 1926.1419 signals general requirements, 1926.1420 radio and other electronic signals, 1926.1421 voice signals and their three required elements, 1926.1422 the posted hand signal chart, and 1926.1428 signal person qualifications
• Appendix A to 29 CFR 1926 Subpart CC, the Standard Method hand signals
• 29 CFR 1926.1425, keeping clear of the load, for the fall zone the signal person must stay out of
• ASME B30.5, mobile and locomotive cranes, in the edition adopted by your employer's program, your contract or your authority having jurisdiction, which is where the lift director role is defined
• See related: What the Qualified Rigger and Signal Person Roles Actually Mean; How to Write a Lift Plan Somebody Else Can Execute