Bond, Development and Lap Length

A bar is worth what the concrete can grip

Reinforcement carries tension, and that tension has to get into the bar from the concrete around it. The transfer happens along the bar's SURFACE — mechanically, through the ribs bearing on the concrete between them — so a bar ending at a point where it is still fully stressed has nowhere to put its load and simply pulls out.

Development length is the length of embedment needed for the concrete to deliver the bar's full yield force. It is not a detailing nicety. A bar stopped short of it is a bar that is not doing what the design assumed, and the failure is sudden because bond failure is brittle.

So every bar has two lengths: the length over which it is needed structurally, and the length beyond that over which it is being anchored. Detailing drawings show the second because nobody can infer it.

Why bigger bars are disproportionately expensive to anchor

Follow the algebra above. The force a bar carries goes with the square of its diameter; the surface available to transfer that force goes only with the diameter itself. Divide one by the other and the required length goes with the DIAMETER.

The practical consequence is easy to miss when substituting bars. Replace four bars with two of twice the diameter — the same steel area, the same capacity — and each of the two now needs twice the anchorage length that each of the four did. In a shallow member, or at a beam-column junction with nowhere to go, that substitution turns a detail that fitted into one that does not.

It also explains the preference for more, smaller bars in congested zones, and why the largest bar sizes are essentially unusable in thin elements: the development length exceeds anything the geometry offers, whatever the strength calculation says.

Bond does not fail by sliding — it fails by splitting

The picture of a bar sliding smoothly out of a hole is wrong and it matters. The ribs on a deformed bar bear on the concrete ahead of them and push it outward radially, so the concrete around the bar is in RING TENSION. When it fails, it fails by splitting along the bar — a crack running down the cover face, following the line of the reinforcement.

That is why cover and bar spacing appear in every development length formula, and why they are not conservatism. A bar near a free face has little concrete to resist the splitting, so it develops a shorter bond stress before the cover breaks off; a bar deep in a section with confinement around it develops far more.

Transverse reinforcement — links, stirrups, ties crossing the bar — works by holding the splitting crack closed, which is why the codes allow a shorter development length where it is present. It is the same mechanism as a hoop round a barrel, and it is the reason a lap inside a well-linked column behaves quite differently from the identical lap in a plain slab.

Hooks, and why they are a different mechanism

A standard hook develops a bar in markedly less length than a straight embedment, and it does so by a different physical route: the bend BEARS on the concrete inside the curve, adding a direct mechanical anchorage on top of whatever bond the straight portion provides.

The trade is space and congestion. A hook needs a bend radius, a tail beyond the bend, and clear room for the bar to be bent at all — so hooks solve a length problem by creating a fit problem, usually in exactly the joints that were already congested. Headed bars are the modern alternative: a plate forged or threaded onto the bar end that provides the bearing without the bend, at the cost of a manufactured component.

The bearing inside the bend also concentrates stress on a small patch of concrete, which is why bend radii are specified as minima against bar size rather than left to the bender. A tighter bend develops less and risks crushing the concrete inside the curve.

Where it fails: laps, top bars and coatings

A LAP SPLICE is not a development length. Two bars overlapping transfer force from one to the other through the concrete between and around them, which is a harder demand than a single bar transferring into a mass of concrete — so a lap is typically longer than a development length rather than equal to it. Laps also want staggering: putting every splice in a member at the same section creates a plane where all the transfer is happening at once, and codes penalise the length accordingly.

TOP-CAST BARS are weaker in bond and the reason is physical. Fresh concrete below a horizontal bar settles and bleeds as it sets, leaving a thin zone of weaker, water-rich material directly under the bar. Bars with a substantial depth of concrete cast below them therefore need a longer development length — the codes apply an explicit factor for it, and it is not small.

EPOXY COATING reduces bond, because the coating is smoother than bare steel and the ribs bear less effectively. Lightweight concrete reduces it too, because the aggregate itself is weaker in the splitting mode described above. Both carry their own multipliers, and both are routinely forgotten when a detail is copied from a normal-weight, uncoated drawing.

The alternative: couple it mechanically, and test it

Where a lap will not fit, the alternative is a MECHANICAL COUPLER — a threaded or swaged sleeve that joins two bars directly and bypasses the concrete entirely. It removes the congestion a lap creates, it is testable as a component, and it costs more per joint. It is the standard answer in heavily reinforced zones, in precast connections, and wherever a lap would have to be staggered through a section that has no room.

For anchorage into existing concrete — a starter bar drilled and bonded into a slab that is already cast — the mechanism is not bond to cast concrete at all but adhesive to a drilled hole, which belongs to the post-installed anchor family and has its own failure modes and its own qualification testing.

In both cases the honest check is a test rather than a formula: couplers are proof-tested to a proportion of bar strength, and post-installed anchorage is qualified by an evaluation report against the exact base material, hole condition and installation procedure. A calculated length is the design tool; the test is what confirms the assumption held on site.

Post-tensioning: the force you lock in is not the force you jacked

A post-tensioned tendon is stressed by a jack and then released into wedges that grip it at the anchorage. Those wedges BITE as they take the load, and the strand draws back into the anchor by a small amount — typically only a few millimetres.

That slip releases some of the elongation the jack put in, and with it some of the force. Crucially the slip is a roughly FIXED distance set by the anchorage hardware, while the elongation it is subtracted from is proportional to the tendon's LENGTH. So the proportional loss is the slip divided by the extension.

Which makes anchor seating almost irrelevant on a long tendon and severe on a short one. The same two or three millimetres that costs a fraction of a per cent on a fifty-metre tendon can cost a large share of the prestress in a short one — and it is the reason short tendons are stressed from both ends, over-stressed within code limits to compensate, or avoided in favour of a longer layout.

It is also only one of several losses. Friction along the duct and wobble in its alignment reduce the force away from the jacking end; elastic shortening of the member takes more as it compresses; and creep, shrinkage and steel relaxation continue taking it for years. Seating is simply the one that happens in the instant the jack lets go, and the one whose behaviour reverses with length.

The verification on site is the ELONGATION MEASUREMENT. Recorded extension is compared against the calculated value within a tolerance, and a discrepancy means the friction assumption, the duct condition or the strand's modulus is not what was assumed — which is why the record of measured extensions is the acceptance document for post-tensioning rather than a gauge pressure alone.