Why a Scaffold Is a Designed Structure Rather Than an Assembly

Why this matters

A scaffold that has been modified in the field is out of service until a competent person clears it, and clearing it is done from the ground and from levels already verified, not by climbing onto the modified bay to look at it. Under 29 CFR 1926.32(f) that competent person has authority to stop the work, and a modified scaffold is the case that authority exists for.

Frame scaffold goes together like a kit. Slot, pin, brace, plank, repeat. Because assembly requires no calculation, the object it produces reads as an assembly, and an assembly is something a crew feels entitled to adapt.

It is not an assembly. 29 CFR 1926.451(a)(6) requires scaffolds to be designed by a qualified person and to be constructed and loaded in accordance with that design, and for standard frame systems that design arrives as the manufacturer's load table. The claim this card carries is that a field modification does not weaken the design, it replaces the design while leaving the old basis in everyone's head - and the losses that follow are not proportional to how small the change looked. The following case runs three ordinary modifications through arithmetic and finds that the one the crew considered most reversible was worth more than the other two together.

Day one: the basis everyone agreed on

A three-lift frame scaffold along a building face. Erected off the manufacturer's table for that frame model, bays fully braced, planked across four boards per bay, guardrails on the open sides, ladder access, base plates on mudsills.

The table gave an allowable leg load for that configuration. Two things about that number, because both come back later.

It is an allowable, and the factor 29 CFR 1926.451(a)(1) requires - the scaffold and each component supporting its own weight plus at least 4 times the maximum intended load - is already inside it. The table is the output of the design, not an input you factor again.

And it is a number for that configuration. Not for that frame, that model, or that manufacturer. The configuration is the argument to the table, and a modification changes the argument.

Day two: one brace comes out

A duct run needed to pass through the second lift. The crew pulled one intermediate cross brace, moved the duct, and did not put the brace back because the next section of duct was coming the following morning.

What a brace does is set the unbraced length of the legs it connects. A compression member's elastic buckling capacity varies with the inverse square of its unbraced length, holding the cross-section, the material and the end conditions constant. That is the whole relationship and the held-constant list is the important half of it.

Line 1. Braced, the leg's unbraced length is the lift height between brace points. Call it 5 feet.

Line 2. With that brace out, the leg is unbraced over two lifts, so its unbraced length is 10 feet.

Line 3. The capacity ratio is 5 over 10, squared: one quarter. The leg retains about 25 percent of its braced buckling capacity, a loss of about 75 percent, from removing one part that a person can carry in one hand.

Line 4, the qualifier that has to travel with line 3. That inverse-square relationship is the elastic buckling case, and it holds while the member is slender enough to buckle before it yields. A short, stocky leg reaches material yield first and loses less. Which regime a given scaffold leg sits in is a property of that frame that the manufacturer knows from testing and a crew cannot determine by looking. So treat the quarter as the working assumption in the direction that protects people, and treat any belief that your legs are stocky as something the manufacturer would have to confirm.

Line 5, what the crew believed. That a brace is a stiffener, that the scaffold felt no different to walk on, and that a missing part costs roughly its own share. It felt no different because nothing had loaded it to the point where buckling governs. Buckling is not a symptom that builds. It is an instability, and the warning ahead of it can be a fraction of a second.

Day three: a plank goes past the end

The corner detail needed about 2 feet more reach than the last bay gave. The crew slid one plank out past the end frame and worked off the overhang.

A plank spanning a 7-foot bay with a person at mid-span develops a bending moment at mid-span of the load times the span divided by four: 1.75 times the load.

Slide that plank 2 feet past the end support and stand on the tip and the bending moment at the support is the load times the overhang: 2.0 times the load.

The ratio is 2.0 over 1.75, about 1.14. A 2-foot overhang on a 7-foot bay produces about 14 percent more bending in the plank than standing at mid-span of a full properly supported bay does. The overhang looked like a small extension of a big span. It was the worse of the two positions.

Then the second effect, which is the one that actually happened. For a plank with a tip load 2 feet beyond support B and 7 feet back to support A, the reaction at A is the load times 2 divided by 7, acting upward on the plank - that is, the far end lifts. About 29 percent of the applied load appears as uplift at the far end. An unsecured plank whose far end lifts by 29 percent of the load rotates about the near support, and the person on the tip goes with it.

That is what happened on day four. A tech stepped onto the overhang, the inboard end of the plank came up, and he went down onto his knees inside the guardrail with the plank swinging. Nobody was hurt. The guardrail did its job, which is worth stating plainly, because the guardrail was the only part of that scaffold still doing what it was designed for.

Day four: the hoist

The third modification never got used, and it is included because it is the one crews most often see as harmless. A rope and pulley was lashed to the top guardrail to bring material up.

Three loads that the frame table never contained:

• The vertical weight of the material, at a point on a rail that is a fall protection component, not a load-carrying one.
• A dynamic increase on starting and stopping the lift, which is not a small percentage and which depends entirely on how the rope is handled. A hand-hauled load that is snubbed and dropped can peak well above its static weight.
• A horizontal component whenever the rope is not vertical, applied at the top of the scaffold, which is the worst place to apply a horizontal force on a structure whose stability is a moment balance about its base.

29 CFR 1926.451(a)(6) puts a scaffold's design with a qualified person, and adding a hoist to a scaffold is a design change. Certain scaffold classes and heights within 29 CFR 1926.452 require design by a registered professional engineer outright. Neither of those routes is available retroactively, after the rope is already tied.

Putting the three next to each other

The crew, reviewing it afterwards, ranked the modifications by how much work each took to make: the hoist first, then the cantilever, then the brace, which took ten seconds and one pin.

The arithmetic ranks them the other way.

Note what the third row says and does not say. It is not that the hoist is the smallest effect. It is that a scaffold's own documentation contains no basis for evaluating it, which is a different and worse category than a number you can look up.

Note also that these do not simply add. The brace removal changed the capacity of the legs; the cantilever changed the load path into the frame near the same corner; the hoist would have added a horizontal force at the top of the same modified section. The design's job was to keep all of those consistent with each other. Once three people each changed one thing, nobody held the whole.

What re-establishes a basis

Not a look. A modified scaffold has three honest routes back into service, and the shop picks one before anyone stands on it:

Restore the original configuration exactly. Brace back in and pinned both ends, plank pulled back within its designed bearing, nothing hanging off it. Then the manufacturer's table applies again because its argument is true again. This is the cheapest route and it is available in most cases.

Get the modification designed. A qualified person under 29 CFR 1926.451(a)(6), or a registered professional engineer where the scaffold class or height requires it, evaluates the change and issues a basis for the modified structure. This is the route where the modification is genuinely needed for the work.

Take it out of service and change the access method. Where neither of the first two is available, a lift or a differently configured scaffold is a decision, not a defeat.

What is not a route: a competent person deciding the modification looks acceptable. The competent person's authority under 1926.32(f) runs to identifying hazards and stopping work. Producing a new structural design is a qualified person's job, and 1926.451(a)(6) says so.

What changes this answer. Systems where reconfiguration is part of the design - system scaffolds with published tie and brace patterns, or suspended scaffolds with their own engineering - still have a design basis; the difference is that the basis covers more configurations, not that it covers all of them. The question is always whether the configuration you now have is one the design speaks to. And on any scaffold, employees must be trained by a person qualified in the subject under 29 CFR 1926.454, which is where a crew learns that a brace is not a handrail with a diagonal in it.

References

• 29 CFR 1926 Subpart L, Scaffolds (construction): 1926.451(a)(1) on capacity and the 4-to-1 factor already contained in a manufacturer's allowable, 1926.451(a)(6) on design by a qualified person, 1926.452 for the additional criteria and engineer-design requirements that attach to specific scaffold types, and 1926.454 on training.
• 29 CFR 1910 Subpart D, Walking-Working Surfaces, the general industry home for platform and scaffold duties. State which Part your job falls under.
• 29 CFR 1926.32(f), the definition of a competent person, including the authority to stop work.
• Scaffold manufacturer's load tables and erection instructions, which constitute the design for standard frame systems and are configuration-specific.
• See related: universal-what-a-scaffold-has-to-have-before-anyone-stands-on-it for the pre-use condition set and the foundation question.