ASTM E290

What the test does

A coupon — flat strip, round bar or hexagonal section — is laid across two hardened supports or rollers and a mandrel or plunger is pressed into its mid-length, bending it to a specified angle or around a specified inside radius. Nothing is pulled apart and no curve is plotted: the bend is made, the load released, and the convex outer face examined for cracking. Four degrees of constraint are recognised — a guided bend, with the plunger forcing the piece down between the supports; a semi-guided bend, bent in contact with a mandrel to an angle judged while still under load; a free bend, with the ends brought together and nothing touching the concave face; and a bend-and-flatten, in which the legs are driven into contact along their length.

What it measures, and why it matters

The result is a condition rather than a number: whether the outer fibre survives an imposed strain intact. The acceptance criterion — how many cracks, of what size, visible to the unaided eye — belongs to the product specification, not to E290, which supplies only the mechanics of getting there.

That makes it a direct proxy for the press brake: material that cracks at the specified radius on the bench will crack at that radius in production. Because the outer fibre is worked hardest, the test is unusually sensitive to defects a tensile coupon averages away — inclusion stringers, banded structure, embrittled or decarburised surface layers, coarse grain in a heat-affected zone — and it is a fast lot check on heat treatment.

Specimen

Rectangular, round and hexagonal sections are all covered, normally at full product thickness so the as-delivered surface is the one placed in tension. Sheared edges are usually dressed, since a burr starts cracks that say nothing about bulk ductility. Orientation to the rolling direction, the position the piece is cut from and the number of pieces per lot come from the product specification, not the test method. No conditioning is called for; pieces are tested as received in ambient air.

What the machine must be capable of

This is a fixture-led method. Bend force follows the section and the span, not the standard: a 3 mm strip needs only a few kN, while a 25 mm plate coupon on a short span can call for 100 kN to 600 kN, which is why bend jigs are commonly bolted into frames of 300 kN to 600 kN. Stroke matters as much as capacity — enough plunger travel to take the legs into contact for a bend-and-flatten.

E290 publishes no bend rate. Force is applied continuously through a single bend, and A370, which cites E290 for its bend clause, treats the speed of bending as ordinarily unimportant; a slow approach is still the sensible habit, because an operator watching for first cracking cannot watch a fast one. No extensometer is required and there is no strain target. Whether Practices E4 is invoked for force accuracy could not be confirmed, so no accuracy class is asserted here; the load reading is not the reported result in any case.

The fixture is the test. Mandrel diameter, support span and roller radius come from the product specification, and the wrong mandrel changes the outer-fibre strain the coupon actually sees — a result obtained with a convenient mandrel is not a result. Supports must be hard, parallel and unable to shift under load. No temperature or humidity limits are stated.

What goes wrong in practice

Edge cracking is the most common false failure. A crack starting at a sheared or burred edge and running inwards is a preparation defect, and logging it as convex-face cracking condemns sound material.

Friction at the mandrel is the next. An unlubricated or galled mandrel drags on the outer fibre, adding tensile strain the geometry never intended, and leaves score marks that are then argued over as cracks.

Springback catches out semi-guided and free bends. The angle or inside radius has to be read while load is still applied; measured after release the bend looks shallower than it was, and a piece that should have gone further is passed.

Finally, supports that creep apart under load, or a span set too short to begin with, clamp the coupon and turn the bend into stretch forming — a different strain history, and a quietly invalid test.

Related and equivalent standards

ISO 7438 is the counterpart and covers the same physical act, but it is prescriptive where E290 defers: it fixes an ambient temperature band, defines the support span by formula, and recognises a slightly different set of bending devices. The two agree only when the product specification has pinned the geometry.

E190, the guided bend test for ductility of welds, is regularly mistaken for the counterpart. It is not — it is the companion document for weldments, and the two are used side by side. A370 reaches into E290 for the bend clause applied to steel products.