top of page

ISO 7438
Bend Test of Metallic Materials

Universal Grip ISO 7438 metallic bend test tooling set

Downloads: ISO 7438 Metallic Bend Test Tooling Set – Universal Grip Data Sheet (PDF)

Written by Chris Thomas, Universal Grip Company · Reviewed by Frank Bertini · Updated October 24, 2026

ISO 7438 is the international standard for the bend test of metallic materials. A test piece of round, square, rectangular or polygonal section is bent in one direction, without reversing the load, until a specified angle is reached, and the outer (tension) surface is then inspected for cracks. It is a ductility and soundness test, not a strength test: there is no stress, no modulus and no force to report, only whether the material took the bend without cracking. Mills use it to release plate, sheet, strip, bar and wire; fabricators use it for incoming inspection; and the same tooling runs the related weld bend test of ISO 5173 and the rebar bend and rebend tests of ISO 15630-1.

This guide covers the current edition, ISO 7438:2020, and what changed; the three bending devices the standard allows (former and supports, V-block, clamp) and when each is used; the support-distance formula l = (D + 3a) ± a/2 with a worked example; the V-block geometry; the full specimen preparation rules for width, thickness reduction and edge rounding; temperature, speed and the three ways a bend may be completed; the angle-from-displacement calculation of Annex A; the acceptance rule and the plane-strain variant of Annex B; a numeric comparison with ASTM E290, ASTM A370, ISO 5173 and ISO 15630; and the bend tooling, fixtures and machines Universal Grip supplies for the test. The bend testing page gives the lab overview, the metals testing page covers the companion tensile tests, and the ISO standards hub lists every ISO method we equip.

In this guide: What ISO 7438 Covers · Editions and What Changed in ISO 7438:2020 · The Three Bending Devices · Support Distance and Former Diameter: a Worked Example · Test Piece Preparation: Width, Thickness and Edges · Temperature, Speed and How the Bend Is Completed · Step-by-Step Procedure · Acceptance: What Counts as a Crack · Common Errors and How Big They Are · ISO 7438 vs ASTM E290, A370, ISO 5173 and ISO 15630 · Equipment for ISO 7438: Tooling, Fixtures and Machines · FAQ

What ISO 7438 Covers

ISO 7438 specifies a method for determining the ability of metallic materials to undergo plastic deformation in bending. It applies to test pieces taken from metallic products as the relevant product standard specifies: hot-rolled and cold-rolled sheet and strip, plate, bars, sections, wire and forgings. Two things are explicitly outside its scope and have their own standards: tubes tested in full section go to ISO 8491, and welded joints go to ISO 5173. The test piece is bent without changing the direction of loading until a specified angle of bend is reached; the two legs stay in a plane perpendicular to the bending axis, and for a 180° bend the legs either touch each other or remain parallel at a specified distance set by an insert.

The result is pass or fail. Unless the product standard says otherwise, the test piece passes if, after bending to the specified angle, no cracks are visible on the outer surface without magnifying aids. Two conventions follow from that: a specified bend angle is a minimum, and a specified internal bend radius is a maximum. A material that survives a tighter radius or a larger angle than required has also met the requirement. Because nothing is calculated, the test is only as good as the geometry of the tooling and the preparation of the piece, which is why most of the standard is about formers, supports, widths and edge radii rather than about the machine.

Editions and What Changed in ISO 7438:2020

The current edition is ISO 7438:2020, the fourth edition, published in October 2020. It cancels and replaces ISO 7438:2016, and its one substantive technical change is a new normative Annex B describing a plane-strain bend test, in which a wide test piece (width at least ten times its thickness) is bent so that the outer fibre is in plane strain rather than the uniaxial-plus-free-edge condition of a narrow strip. The plane-strain variant is used for sheet-forming studies of ductility and fracture, and its acceptance rule differs: edge cracks within 2a of the edge are disregarded, while a crack extending further than 2a from the edge invalidates the test piece.

The Three Bending Devices

ISO 7438 allows three test arrangements, and the product standard or the purchase agreement normally says which one applies. All three bend the piece around a former of diameter D to the angle α; they differ in how the piece is supported and in the geometry that must be controlled.

Device 1: former and two supports (three-point bend)

The test piece rests on two supports and the former pushes down at mid-span. The supports and the former must be wider than the test piece (wider than b, or than the diameter for round pieces) and of sufficient hardness so they do not indent. The former diameter D comes from the product standard. The distance between the supports is fixed by the standard's Formula 1:

l = (D + 3a) ± a/2

and it must not change during the test, which means roller supports on a rigid beam, not supports that can walk apart under load. The standard adds a warning: if l is reduced to D + 2a or less, the piece can be clamped between the former and the supports and stretch-formed rather than bent, which changes the strain in the outer fibre and can mask or create cracks. This is the arrangement for plate, flat bar and sections, and the geometry most bend fixtures for universal testing machines reproduce with rollers on a graduated beam.

Device 2: V-block and former

The piece sits across a V-shaped die and the former pushes it into the V. The tapered faces of the block must make an angle of 180° − α with each other, so a 90° bend uses a 90° V and a 120° bend uses a 60° V. The edges of the block where the piece first contacts must be rounded to a radius of 1 to 10 times a, and the block must be hard enough not to deform. It is the practical device for sheet, strip and thin plate, for QC bends to a fixed angle, and as the first stage of a 180° wrap closed between platens. The Universal Grip ISO 7438 tooling set uses this arrangement: a bolt-down V-block die, a wedge-type former for tight-radius and V-bends, and interchangeable mandrels of different radii.

Device 3: clamp (vice) and former

One end of the piece is held in a clamp against the former and the free end is bent around the former, by a lever or by the crosshead. The standard's one geometric rule is that the clamp face must not reach or extend past the vertical plane through the centre of the former, because its position affects the bend radius and therefore the result. It is the device for thin sheet, wire and small sections and for full 180° wraps.

Support Distance and Former Diameter: a Worked Example

The product standard sets the former diameter, usually as a multiple of thickness: 1a, 2a, 3a or 4a are typical for structural steel plate, and larger ratios for high-strength or thick material. ISO 7438 then fixes the span from that diameter and the thickness. Take a 12 mm hot-rolled plate to be bent 180° over a former of 2a:

  • Former diameter: D = 2 × 12 = 24 mm.

  • Support distance: l = (D + 3a) ± a/2 = (24 + 36) ± 6 = 60 ± 6 mm, so the supports are set between 54 and 66 mm apart; 60 mm is the target.

  • Stretch-forming limit: D + 2a = 24 + 24 = 48 mm. Supports closer than 48 mm would clamp the piece against the former and the test is no longer a free bend.

  • Width: the plate is wider than 20 mm and a ≥ 3 mm, so the test piece is cut 20 to 50 mm wide, typically 40 mm.

  • Edge rounding: a = 12 mm falls in the 10 to 50 mm band, so edge radii may be up to 1.5 mm.

The same plate bent over a former of 1a (D = 12 mm) gives l = (12 + 36) ± 6 = 48 ± 6 mm and a stretch-forming limit of 36 mm. For a 3 mm sheet over D = 2a = 6 mm, l = (6 + 9) ± 1.5 = 15 ± 1.5 mm. Rollers that slide on a scale make this quick to set; fixed-hole supports usually cannot hit the tolerance for every former in a set, which is the reason adjustable spans are specified on bend fixtures for universal testing machines rather than drilled plates.

Capacity follows from the same numbers: F ≈ Re·b·a²/l for a fully plastic hinge. The 12 × 40 mm S355 plate at l = 60 mm needs about 34 kN, nearer 50 kN with work hardening; a 25 × 50 mm piece at 2a needs about 90 kN and can pass 100 kN by the end of the bend, which is why 100 kN is the natural standard capacity and thicker or high-strength plate moves to 200 to 600 kN frames.

Test Piece Preparation: Width, Thickness and Edges

Clause 7 of the standard is the part most often got wrong in practice. The rules exist because the bend test is a surface test: a sheared edge with cold-worked cracks, a flame-cut heat-affected zone or a sharp machined corner all start cracks that have nothing to do with the material.

  • Cut edges: material affected by shearing or flame cutting must be removed by machining. The standard allows leaving it only when the result is still satisfactory, in other words a pass with the damaged edge is a pass, but a fail must be repeated on a machined piece before the material is condemned.

  • Edge rounding (rectangular pieces): the long edges may be rounded to a maximum radius of 1 mm for a < 10 mm, 1.5 mm for 10 ≤ a < 50 mm and 3 mm for a ≥ 50 mm. There must be no transverse scratches or burrs on the tension face.

  • Width: if the product is 20 mm wide or less, the full product width is tested. If it is wider, the test piece is 20 ± 5 mm wide for a < 3 mm and 20 to 50 mm wide for a ≥ 3 mm. Annex B plane-strain pieces need b ≥ 10a.

  • Thickness of sheet, strip, plate and sections: the product thickness, except that products thicker than 25 mm are machined on one face to not less than 25 mm, and the unmachined (as-rolled) face is placed on the tension side of the bend.

  • Round and polygonal bars: tested as the product up to 30 mm diameter; from over 30 to 50 mm the diameter may be reduced to not less than 25 mm, and over 50 mm it must be reduced to not less than 25 mm.

  • Forgings, castings and semi-finished products: dimensions per the product standard or agreement; larger a/b ratios by agreement.

  • Length: the standard gives no formula; the length depends on the thickness and the device. In practice allow the span plus enough overhang to sit on both supports throughout the bend, and for 180° wraps enough leg length to be pressed flat.

Temperature, Speed and How the Bend Is Completed

The test is run between 10 °C and 35 °C, or at 23 ± 5 °C when controlled conditions are required. Low temperature matters for ferritic steels, whose bend ductility falls with their impact toughness, so a cold shop floor in winter can turn a marginal material into a failure that would pass at 23 °C.

Loading is applied slowly, to allow free plastic flow of the material. The standard does not fix a speed for routine testing, but in case of dispute the rate of former displacement is (1 ± 0.2) mm/s, which is 60 ± 12 mm/min. That is faster than most people assume; a servo-electric frame comfortably holds it, and a lab procedure that bends at 5 mm/min is compliant but not the referee condition. The force must be applied continuously without reversal, and the direction of loading must not change during the bend.

Three completion criteria are recognised, and the product standard picks one:

  • Bend to a specified angle α: the former is advanced until the angle between the legs reaches α, measured under load or after unloading as the product standard requires. With a three-point device the angle can be computed from the former displacement f (Annex A); with a V-block it is set by the block.

  • Bend to 180° with the legs parallel at a specified distance: after the initial bend over the former, the piece is pressed between two parallel platens with an insert of the specified thickness between the legs. The insert thickness, which fixes the internal radius, comes from the product standard or agreement.

  • Bend to 180° with the legs in direct contact: the same, without an insert, so the legs are closed flat on each other. This is the tightest bend and is used for very ductile material (deep-drawing sheet, soft wire, annealed copper).

Annex A relates the former displacement f to the bend angle for the three-point device from the support radius R, the former diameter D, the thickness a and the span l, with the vertical distance c = R + a + D/2 before the test. Programmable crossheads use it to stop at the target angle; check the angle after unloading too, because springback reduces it by a few degrees on high-strength material.

Step-by-Step Procedure

  • 1. Identify the requirement. From the product standard: test orientation, former diameter D (as a multiple of a or an absolute value), bend angle α, and completion mode (angle, 180° with insert, 180° closed). Record the material, cast number and direction.

  • 2. Prepare the piece. Machine off sheared or flame-cut edges, round the long edges within the limits above, reduce thickness or diameter if the product exceeds 25 mm (one face only), as-rolled face on the tension side.

  • 3. Measure. Thickness a and width b at the bend location; from them compute the span l = (D + 3a) ± a/2 for a three-point device, or select the V-block with faces at 180° − α.

  • 4. Set up the fixture. Fit the former of diameter D, set the supports to l (locked) or bolt down the V-block; former and supports wider than the piece and centred.

  • 5. Condition. Bring the piece and tooling to 10 to 35 °C (23 ± 5 °C if required). Centre the piece on the supports with the tension face down (away from the former).

  • 6. Bend. Apply the force slowly and continuously, without reversing, until the angle is reached; in a dispute, at 1 ± 0.2 mm/s former displacement. For 180° bends, finish between platens with or without the insert.

  • 7. Inspect. Examine the outer surface of the bend without magnification. Note any crack, its position and length. For Annex B pieces, disregard edge cracks within 2a of the edge and invalidate the piece if a crack runs further.

  • 8. Report. Reference to ISO 7438, test-piece identification (material, cast, axis direction), shape and dimensions, device used (supports and former, V-block, clamp), any deviations, former diameter, angle or completion mode, and the result.

Acceptance: What Counts as a Crack

In the absence of a requirement in the product standard, the piece is judged to have passed if no cracks are visible without magnifying aids on the outer surface after bending to the specified angle. Surface roughening, orange peel, Lüders bands and lustre changes are not cracks. Fine edge cracks on a shear-cut edge are the most common false failure; the fix is to machine the edges within the clause 7 limits before testing, not to argue about the crack. For plane-strain pieces under Annex B, cracks that start at the edge and extend no further than 2a from it are disregarded, while a crack that extends more than 2a invalidates the result.

Because the angle is a minimum and the radius a maximum, a lab may bend further than required and still report a pass, but not less or over a larger former. Where the product standard gives a radius, the former is 2r; where it gives a ratio such as 2t, use the actual test-piece thickness after any reduction.

Common Errors and How Big They Are

  • Supports too close. Setting l at or below D + 2a turns the bend into stretch forming; the outer-fibre strain rises and sound material can crack. For the 12 mm plate example the limit is 48 mm against a 60 mm target.

  • Wrong former. A former of 3a instead of 2a halves the outer-fibre strain at the bend (ε ≈ a/(D + a): 33 % at 2a, 25 % at 3a, 50 % at 1a), so a material that passes at 3a may fail at 2a. Always take D from the product standard and record it.

  • Sheared edges left on. Cold-worked, micro-cracked edges fail long before the material would; the standard requires them removed unless the piece passes anyway.

  • Sharp edges or transverse scratches. Edge radii above the 1 / 1.5 / 3 mm limits reduce the strain concentration unrealistically; scratches across the tension face start cracks.

  • Cold shop. Bending ferritic steel at 5 °C instead of the 10 to 35 °C window can turn a pass into a fail; the standard's temperature range is a requirement, not advice.

  • Confusing the test with flexural strength. ISO 7438 reports no strength or modulus. A page or procedure that quotes a bend 'strength' in MPa from an ISO 7438 test is describing a plastics flexure test (ISO 178) or a composite flexure test (ISO 14125), not this one.

ISO 7438 vs ASTM E290, A370, ISO 5173 and ISO 15630

The bend test exists in several standards families, and the tooling is shared, but the geometry rules differ in detail. The numbers below are what change when a customer moves between them.

  • ISO 7438 (metals, general): former D from the product standard; supports at l = (D + 3a) ± a/2; V-block faces 180° − α, edge radius 1 to 10a; width 20 ± 5 mm (a < 3) or 20 to 50 mm (a ≥ 3); thickness reduced to ≥ 25 mm above 25 mm; 10 to 35 °C; dispute rate 1 ± 0.2 mm/s; pass = no cracks without magnification.

  • ASTM E290 (metals, general): four constraint types: guided bend (mandrel or plunger between two supports at a defined spacing), semi-guided (bent to an angle or inside radius, measured under load), free bend (no transverse force and no contact on the concave side) and bend-and-flatten. Continuous bending with no reversal, as in ISO. Cracks within one thickness of the edge are not counted; corner cracks are ignored unless the product specification sizes them. Mandrel ratios come from the product specification; full-section pieces are allowed.

  • ASTM A370 (steel products): 180° bend around a pin whose diameter the product specification sets. ASTM A615 rebar pins at 90° for Grade 60: 3.5d for #3 to #5, 5d for #6 to #8, 7d for #9 to #11 and 9d for #14 and #18. The jig is the same three-point former-and-supports device; see the rebar and steel testing post.

  • ISO 5173 (welded joints): former diameter d = 4t for materials with A ≥ 20 % (steel, nickel alloys) and larger for less ductile material; support distance d + 2t + 1 ≤ l ≤ d + 3.5t (compare ISO 7438's D + 3a ± a/2); roller radius ≥ 20 mm; side-bend pieces 10 ± 0.5 mm thick; bend normally to 180°; acceptance per the welding application standard. Weld procedure and welder qualification bends belong here, not to ISO 7438, although the same tooling set runs both.

  • ISO 15630-1 (reinforcing bars): mandrel diameter and angle from the product standard (EN 10080 / BS 4449 give 4d for bars up to 16 mm and 7d above), with a rebend test: bend to 90° or more, age at 100 °C, then bend back by 20° and inspect. The pass criterion is again no cracks to the unaided eye.

Equipment for ISO 7438: Tooling, Fixtures and Machines

Universal Grip ISO 7438 bend tooling set

Our ISO 7438 set is a bend test tooling set for metallic specimens comprising a V-block die, a wedge-type former and a range of interchangeable mandrels of different radii. The specimen is bent over the chosen former into the die by the test machine, then inspected for cracks on the outer surface. From the data sheet: a bolt-down V-block die with mounting holes for secure location; interchangeable mandrels in several radii to suit bend-ratio requirements; a wedge-type former for tight-radius and V-bends; hardened, ground working faces for repeatable bend geometry; test pieces of round, square, rectangular or polygonal section; load capacity 100 kN standard, higher capacities on request; hardened tool steel; clevis pin or threaded adapter to suit the customer frame; compatible with electromechanical and servo-hydraulic universal testing machines. Applications listed on the sheet are steel plate, sheet and bar, aluminium and non-ferrous products, mill and fabrication quality control, and weld procedure and welder qualification, the last of which is run to ISO 5173 or ASME IX with the same tooling. When ordering, specify specimen thickness and width, the mandrel diameters, bend angle, required capacity and the test machine connection.

The V-block method is ISO 7438 clause 6.3: the die faces must be at 180° − α and the contact edges rounded to 1 to 10a, so tell us the bend angle and thickness range and the die is made to suit. For plate and bar that need the former-and-supports arrangement, our bend testing fixtures with rollers on a graduated beam set l = (D + 3a) ± a/2 directly and take the same mandrels; the small bend fixture covers thin strip and wire.

Machines and capacity

Using the plastic-hinge estimate plus 30 to 50 % headroom, a 50 kN benchtop UTM covers sheet and strip to about 8 mm and bar to about 16 mm in structural steel; a 100 kN benchtop machine covers plate to about 20 mm at 2a; and 20 to 25 mm plate, high-strength grades and 180° closes between platens move to a 200 kN floor-standing UTM, a 600 kN frame or a hydraulic machine. The crosshead must hold 60 mm/min for the dispute rate and the daylight must take the fixture plus the legs of a 180° piece. Tight-radius sheet-metal bends to VDA 238-100 use a knife-edge fixture covered on its own page.

ISO 7438 FAQ

What is ISO 7438?

The international standard for the bend test of metallic materials. A round, square, rectangular or polygonal test piece is bent in one direction over a former of specified diameter until a specified angle is reached, and the outer surface is inspected for cracks. It is a pass/fail ductility test with no stress or modulus; the current edition is ISO 7438:2020.

What is the support distance in ISO 7438?

For the former-and-supports device, l = (D + 3a) ± a/2, where D is the former diameter and a the thickness. A 12 mm plate over a 24 mm former gives l = 60 ± 6 mm. The supports must not move during the test, and spacings of D + 2a or less are warned against because they clamp and stretch-form the piece.

What are the three bending devices in ISO 7438?

A former with two supports (three-point bending, for plate and bar), a V-block with former (faces at 180° − α, edge radius 1 to 10a, for sheet and strip and for QC bends to a fixed angle) and a clamp with former (for thin sheet, wire and full 180° wraps). The product standard says which applies.

What is the former diameter for ISO 7438?

ISO 7438 does not fix it; the product standard does, usually as a multiple of thickness such as 1a, 2a, 3a or 4a. The outer-fibre strain is roughly a/(D + a): about 50 % at 1a, 33 % at 2a and 25 % at 3a, which is why the ratio changes the outcome.

What temperature and speed does ISO 7438 specify?

Testing at 10 to 35 °C, or 23 ± 5 °C under controlled conditions. The force is applied slowly to allow free plastic flow, and in case of dispute the former displacement rate is 1 ± 0.2 mm/s (60 ± 12 mm/min), continuously and without reversing direction.

What is the pass criterion in ISO 7438?

Unless the product standard says otherwise, the test piece passes if no cracks are visible without magnifying aids on the outer surface after bending to the specified angle. A specified angle is a minimum and a specified inside radius is a maximum. Under Annex B (plane strain), edge cracks within 2a of the edge are disregarded.

How wide should an ISO 7438 test piece be?

The full product width if it is 20 mm or less. Otherwise 20 ± 5 mm for thickness under 3 mm and 20 to 50 mm for thickness of 3 mm or more. Plane-strain pieces under Annex B need a width of at least ten times the thickness.

Does ISO 7438 allow reducing the thickness of thick plate?

Yes. Products thicker than 25 mm are machined on one face to not less than 25 mm, with the unmachined face placed in tension. Round bars over 30 mm may, and over 50 mm must, be reduced to a diameter of not less than 25 mm. Edge radii are limited to 1 mm (a < 10), 1.5 mm (10 to 50) and 3 mm (a ≥ 50).

How is a 180° bend completed in ISO 7438?

Either with the legs parallel at a specified distance, by pressing the pre-bent piece between platens with an insert whose thickness the product standard gives, or with the legs in direct contact, pressed flat without an insert. The initial bend is made over the former with one of the three devices.

What is the difference between ISO 7438 and ASTM E290?

Both are one-direction bend ductility tests with a no-cracks criterion. ISO 7438 fixes the three-point span at (D + 3a) ± a/2 and the V-block at 180° − α; E290 defines four constraint types (guided, semi-guided, free bend, bend-and-flatten) and disregards cracks within one thickness of the edge. Mandrel ratios come from the product specification in both.

Can ISO 7438 tooling be used for weld bend tests?

The same V-block, formers and rollers run ISO 5173 weld bends, but the geometry rules differ: ISO 5173 uses a former of 4t for materials with 20 % elongation or more, a support distance between d + 2t + 1 and d + 3.5t, 20 mm-radius rollers and 10 mm side-bend pieces. Weld procedure and welder qualification are run to ISO 5173 or ASME IX, not ISO 7438.

What machine capacity is needed for an ISO 7438 bend test?

Estimate F ≈ Re·b·a²/l and add 30 to 50 %: a 12 × 40 mm S355 plate over a 24 mm former needs about 35 to 50 kN, and a 25 × 50 mm plate at 2a about 90 to 130 kN. A 100 kN fixture and frame cover most plate to 20 mm; thicker or high-strength material moves to 200 to 600 kN floor-standing or hydraulic machines.

Need bend tooling, a V-block die, mandrels or a complete bend-test system for ISO 7438? Contact Universal Grip with your specimen thickness and width, the former diameters and bend angle your product standard calls for, and your machine connection, and we will quote the right set.

bottom of page