Why a Leading Edge Is Treated Differently

Why this matters

A personal fall arrest system that is correct on a rooftop davit can kill the same worker at a deck edge, wearing the same harness and the same lanyard, on the same afternoon. Nothing about the equipment changed. What changed is where the anchorage sits relative to the worker's back D-ring, and that single geometric fact moves three quantities at once: how far the worker falls before anything begins to stop them, what the lifeline is dragged across on the way, and whether they arrive at the bottom travelling sideways into a column. Before any of that arithmetic, one rule: never connect to an anchorage at or below your feet with equipment whose own label does not say it is rated for anchorage at that level.

What makes an edge a leading edge

29 CFR 1926.500(b) defines a leading edge as the edge of a floor, roof, or formwork for a floor or other walking surface, such as a deck, which changes location as additional sections are placed, formed or constructed. The defining property is motion. The edge is where the work is, it moves with the work, and it is why a fixed guardrail cannot follow it.

Construction work at a leading edge is addressed at 29 CFR 1926.501(b)(2). Where conventional protection is genuinely infeasible or would create a greater hazard, 29 CFR 1926.502(k) permits a written fall protection plan, and leading edge work is one of only three activities that option reaches. That is a demonstration the employer carries, not a preference anyone exercises casually.

The practical version for service trades: any time your anchorage is at deck level, roof level, or below your feet, and any time your lifeline crosses a structural edge on the way to it, you are in this article's territory whether or not the edge is moving.

The gate: where is the anchorage relative to the D-ring

Ask one question before anything else. Is the anchorage above the dorsal D-ring, level with it, or below it?

The answer sets free fall, which sets everything downstream. With a lanyard of length L between the anchorage and the D-ring, and the anchorage a height h above the D-ring:

free fall = L - h

That is the whole relationship, and it holds for a lanyard of fixed length with the anchorage directly overhead, no lateral offset. When the anchorage is below the D-ring, h is negative and the term adds rather than subtracts, which is why a deck-level anchor produces a longer free fall than a 6 ft lanyard has any right to produce.

29 CFR 1926.502(d) caps free fall at 6 ft and also requires that the worker not contact a lower level, whichever is less. It caps deceleration distance at 3.5 ft and the maximum arresting force at 1,800 lb with a body harness. General industry sets its counterpart criteria at 29 CFR 1910.140. Those caps are a property of the system, and the geometry decides whether your system can meet them where you are standing.

A note that has to be made once in this library: a fall protection anchorage is not a concrete fixture anchor. The anchor articles here cover fasteners rated for a static hanging load in a defined base material. A fall protection anchorage is rated on a different basis entirely, at 5,000 lb per employee attached or an engineered design with a safety factor of at least two under a qualified person's supervision. Never tie a lifeline to a fixture anchor because the word matched.

The clearance stack, and what a published figure already contains

        anchorage
            |
            |  free fall: lanyard length, plus the
            |  D-ring height above the anchorage
            |  when the anchor sits below the D-ring
        ----+----  arrest begins here
            |  deceleration distance
        ----+----
            |  harness stretch and D-ring shift
        ----+----
            |  height of the worker below the D-ring
        ----+----
            |  safety margin
        ====+====  lower level, or the first obstruction

Five terms, and each one belongs to a different owner. Free fall is set by your geometry. Deceleration distance is capped by the regulation and published by the device maker, whose figure is usually smaller. Harness stretch and D-ring shift are published by the harness maker. The worker's height below the D-ring is measured on your own crew, in gear, boot sole to the D-ring, and rounded up, because rounding it down shortens a required clearance. The safety margin is stated in whatever chart you are reading.

Here is where people double count. A manufacturer's published "required fall clearance" figure is not a free fall number. It is a stack that already contains deceleration, harness stretch, the worker's height and a margin, computed for one specific anchorage geometry. Adding those terms again on top of the published number does not make you safer, it makes your arithmetic wrong in a way that hides which term actually governs.

The correct move is a re-basing, not an addition. If your anchorage geometry differs from the one the chart was published for, you use the chart for your geometry, which the maker publishes separately for overhead and foot-level anchorage. You do not take the overhead number and bolt a correction onto it. And there is one thing a published figure does not contain that you cannot fix by adding to it. If the anchorage is not above the work, the fall arcs, and an arc is not a longer vertical fall. It changes which surface the body reaches first and whether the body arrives there travelling sideways. Move the anchorage over the work, or the calculation does not describe the event.

Worked example: same worker, two anchorages

One tech, one harness, dorsal D-ring measured at 5.0 ft above the boot sole in work boots, rounded up from the tape.

Case A, overhead anchorage. A roof davit puts the anchorage 2.0 ft above his D-ring. He is using a 6 ft shock-absorbing lanyard.

Lanyard length L                                 6.0 ft
Anchorage height above D-ring h                  2.0 ft
Free fall = L - h                                4.0 ft
Cap in 29 CFR 1926.502(d)                        6.0 ft
Result under the cap

Deceleration comes from the lanyard maker's published elongation, and the required clearance comes from the maker's overhead-anchorage chart, used as published because it already contains the four terms below free fall.

Case B, deck-level anchorage, same lanyard. He moves to an open edge and clips to a deck-level anchor at his feet, 5.0 ft below his D-ring.

Lanyard length L                                 6.0 ft
Anchorage height above D-ring h                 -5.0 ft
Free fall = L - h = 6.0 + 5.0                   11.0 ft
Cap in 29 CFR 1926.502(d)                        6.0 ft
Result over the cap by 5.0 ft

Read what that says. The problem is not that he needs more room below him. The problem is that this lanyard cannot be used at this anchorage at all, at any clearance, because the free fall it produces exceeds what the standard permits and the arresting force that goes with an 11 ft free fall is not what the energy absorber was tested to manage. It is an equipment-class failure, and no amount of care with a tape measure fixes it.

Case B, corrected. He swaps to a self-retracting device rated for anchorage at or below the D-ring and for contact with an edge. Under ANSI/ASSP Z359.14-2021, in the edition your employer's fall protection program or your contract adopts, those are Class 2 devices, and the classification carries an edge-contact rating that Class 1 devices do not. The maker publishes the minimum edge radius the device was tested over and whether separate edge protection is required. ANSI binds through that adoption, never on its own; the regulatory floor remains 1926.502(d).

Now the clearance, with each line labelled as contained or added:

Maker's published required clearance for
  foot-level anchorage: contains deceleration,
  harness stretch, worker height, margin use as published
Swing-fall drop: NOT contained in that chart add
  anchorage offset 10.0 ft horizontally
  line length at lock-up 20.0 ft
  extra drop = 20.0 - sqrt(400 - 100)
             = 20.0 - 17.32                     2.7 ft (rounded up)
Vertical drop along the arc, NOT a corrected requirement    2.7 ft

And the term that is not a clearance term at all. Having dropped that extra 2.7 ft along the arc, he arrives at the bottom of the swing moving horizontally:

v = sqrt(2 x 32.2 ft/s2 x 2.68 ft) = 13.1 ft/s

About 9 mph, sideways, into whatever is in the arc. That figure treats the line as inextensible and the anchorage as a point, and ignores what the device absorbs, so it runs slightly high, which is the conservative direction for an impact hazard. It holds for a device that locks quickly and then swings; a slower lock changes both the drop and the arrival speed.

The clearance stack has no term for this. It measures vertical distance to a lower level. A worker can pass every line of that stack and still strike a column, a wall or the edge of a stair opening at the bottom of the arc. That is why the leading-edge answer is always to move the anchorage toward being overhead and in line, and only then to compute the clearance.

The failure mode is a crew that solves this by finding more clearance. The building has 40 ft below the edge, the published figure and the swing correction add up to less than that, and everyone signs off. They have satisfied the arithmetic they did and left untouched the two things the arithmetic never covered: the lifeline dragging across a bare steel edge, and the arc that ends at a column 8 ft away.

How to verify you got this right

Verify the geometry first and the numbers second, because the numbers are only meaningful once the geometry is fixed.

Stand where the work happens and look at the anchorage. If you are looking level or down at it, stop and check the label on your connecting device for a foot-level anchorage rating. If the label does not say it, the device is not permitted there and the conversation is about a different anchorage, not a different calculation.

Then check the offset by walking, not by estimating. From the anchorage, walk to the furthest point you will work, and look back. If the line to the anchorage is noticeably off the vertical plane you would fall in, you have a swing and it needs its own line in the calculation. The maker usually publishes a maximum offset angle for their device; that number governs and it is theirs, not yours.

Then look along the path the lifeline will take on the way to the anchorage. Every edge it crosses is an edge it will be dragged across under load, and a bare rolled steel edge or a cut deck edge is a cutting tool at arrest forces. Either the device is rated for that edge radius or the edge gets protection, and which one it is goes on the job plan.

Finally, record the measured D-ring height for the crew rather than using a generic figure. It is the one term in the stack you own outright, it varies by more than people expect, and it is the term that gets estimated exactly when it matters most.

References

• 29 CFR 1926.500(b) for the definition of a leading edge; 29 CFR 1926.501(b)(2) for leading edge work; 29 CFR 1926.502(k) for the fall protection plan option and the three activities it reaches.
• 29 CFR 1926.502(d), personal fall arrest system criteria for construction: 6 ft free fall limit, 3.5 ft maximum deceleration distance, 1,800 lb maximum arresting force with a body harness, and 5,000 lb per employee attached for anchorages. General industry counterpart at 29 CFR 1910.140.
• ANSI/ASSP Z359.14-2021, self-retracting devices, which defines Class 1 and Class 2 devices and the edge-contact rating, in the edition adopted by your employer's fall protection program or your contract.
• See related: universal-why-suspension-trauma-makes-rescue-a-timed-event for the clock that starts once the stack has done its job, and universal-what-fall-restraint-is-and-why-it-is-the-better-answer for the system that avoids all of this.