Industrial safety equipment manufacturer · Qinhuangdao, China

Fall arrest or fall restraint: the on-site decision

A travel restraint system is designed so the fall never starts; a personal fall arrest system is designed so a fall in progress ends safely. Both definitions fit in one sentence. Choosing between them is a site decision, and the two failures auditors keep finding are the same ones: a restraint rig stretched into what is really arrest, and an arrest system hung where restraint was possible all along.

A travel restraint system is designed so the fall never starts; a personal fall arrest system is designed so a fall in progress ends safely. Both definitions fit in one sentence, and 29 CFR 1910.140(b) gives both. Choosing between them is a site decision, and the two failures auditors keep finding are the same ones: a restraint rig stretched into what is really arrest, and an arrest system hung where restraint was possible all along.

Two systems, two intentions

29 CFR 1910.140(b) defines a personal fall arrest system as "a system used to arrest an employee in a fall from a walking-working surface", consisting of a body harness, anchorage and connector, with the connection made by a lanyard, deceleration device, lifeline or a combination. A travel restraint system is "a combination of an anchorage, anchorage connector, lanyard (or other means of connection), and body support that an employer uses to eliminate the possibility of an employee going over the edge of a walking-working surface."

The verbs carry the difference. Restraint eliminates the fall; arrest survives it. A restraint system that holds still has a worker on the surface. An arrest system that fires has a worker suspended, needing rescue within minutes, with clearance, swing and arresting force all designed in beforehand. That is why the choice is not a matter of preference: one system removes the fall, the other manages it.

The five criteria that actually decide it

CriterionRestraint fits whenArrest is needed when
Edge exposure the worker never needs to approach or reach past the edge the task requires reaching over or stepping across the edge
Anchor geometry a sound anchor exists behind the worker, at or above waist height, reachable without exposure no rear anchor exists at height; only roof-edge or overhead points are available
Clearance below largely irrelevant, because nobody goes over the edge total fall distance, including deceleration, fits above the nearest obstruction
Movement pattern work stays in a fixed zone along one edge the work path crosses edges or changes level
Rescue posture no suspended-worker rescue to plan prompt rescue is planned and rehearsed; suspension time is part of the hazard

Two of these are decided by the site and cannot be argued around. If the task requires leaning past the edge, restraint is gone, because the definition's whole point — eliminating the possibility of going over the edge — no longer holds. If no competent anchor exists behind the worker, restraint is usually gone too, and the remaining options are engineered anchors or an arrest system rigged to a verified point. Where both hold, restraint deserves real consideration: it removes the fall instead of surviving one, and it does not need clearance or rescue to work.

Which system does the edge decide for?

Take a flat roof with photovoltaic arrays set back two metres from the edge. The work never crosses the edge line, so a tether to a rear anchor keeps the worker inside the zone and arrest never enters the day. Now move to the same roof for gutter replacement: the task requires reaching past the edge, the tether has to be lengthened to let the worker get there, and the moment it permits reaching past the edge it has stopped being restraint. The harness did not change; the job did. The system is chosen per task, not issued per site.

How the choice goes wrong on site

A restraint line quietly stretched into arrest. A two-metre tether adjusted "just long enough to reach" is the most common version. Once the tether allows the worker over the edge, there is no deceleration device behind it, no clearance calculation and no rescue plan — only a worker going over with equipment that was never rigged for what happens next.

An anchor below the feet, in either system. Restraint from a foot-level anchor lets the worker swing past the edge; arrest from a low anchor develops a pendulum and forces the system was not positioned to manage. The rigging distance between the attachment and the worker's feet decides this, and from the ground it is invisible — which is why it keeps happening (see low anchorage: the invisible violation).

Free fall distance left uncounted. Where arrest is rigged, 29 CFR 1910.140(d)(2)(ii) requires that the employee cannot free fall more than 1.8 m or contact a lower level. A lanyard anchored at foot level can consume the whole allowance before the fall begins.

Mixed crews at the same edge. A subcontractor works in arrest, the main crew works in restraint, and nobody reconciles the interface where one crew's tether ends and the other's lanyard begins. The mix is legal; the unmanaged boundary is not safe.

What monitoring looks like in each system

The two systems fail in different places, so the monitoring that helps also differs.

Restraint's weak point is configuration drift: a tether lengthened between tasks, an anchor swapped for a convenient pipe, a hook clipped to a rail instead of the rated point. Monitoring watches whether the connection is actually made and whether the attachment sits below the wearer — the two states that turn restraint into an unmanaged fall.

Arrest's weak point is the state of the connection at the moment of work: hook closed on the rated point, attachment at the dorsal or front D-ring rather than carried, no second hook dragging. A multi-state harness monitor (SBM3000 series) watches wear state, hook closure and attachment height relative to the wearer, raises a local alarm the moment a state is wrong, and files the event with a timestamp. It takes no part in load bearing and does not act on the system — it observes and reports; the arrest physics stay in the harness, lanyard and anchor.

In both systems, the record matters afterwards: which hooks were open, when, on which crew and at which anchor — evidence that feeds the anchor and training decisions that follow an audit.

Questions buyers ask

Can one harness serve both systems?
Yes. The harness is the common component; the system is how it is rigged. A monitor does not need to know which system is in use — it watches connection states either way, which is useful when one crew runs both in a day.

Is restraint always the safer choice?
Where the task genuinely stays inside the zone, restraint removes the fall rather than surviving one. Where edge exposure exists, restraint is not available — and writing restraint into the paperwork while the work crosses the edge leaves the real exposure unmanaged.

Does a monitor replace inspection?
No. Pre-use inspection of the system still applies (29 CFR 1910.140(c)(18)) and covers webbing, hardware and labels — see the harness inspection checklist. Monitoring watches use, not wear; the two answer different audit questions.

About Yunku Technology

Yunku Technology builds the SBM3000 multi-state smart safety harness: wear detection, hook-connection detection and anchor-height discrimination with edge alarm logic and timestamped records, described on the multi-state safety harness page. The capability boundary is deliberate — what a smart safety harness does and does not do is set out on the definition page.

Tell us your site conditions — we will size the system for you

Share your industry, the work-at-height, confined-space or perimeter-control tasks involved, and any applicable standard you must comply with. Our engineering team will come back with a configuration and a budget range.