Transverse Tensile Specimens from Welded Joints

A welded joint is not one material. It is parent metal on both sides, weld metal in the middle, and a heat affected zone between them where the parent metal was cooked by the arc without ever melting. Those three regions have different hardness, different grain structure and different strength, and they are joined in series along the load path.

That is why a welded joint gets its own tensile test rather than inheriting the numbers from the plate it was cut from. The transverse tensile test pulls across the joint and asks the only question that matters to a welding procedure: where does this assembly break, and at what load.

What the Transverse Tensile Test Is For

ISO 4136 specifies the sizes of test specimen and the procedure for carrying out transverse tensile tests, in order to determine the tensile strength and the location of fracture of a welded butt joint. It applies to metallic materials in all forms of product, with joints made by any fusion welding process.

Two outputs, and the second is the one people underestimate. Tensile strength is a number. Fracture location is a verdict. A joint that reaches the required strength but breaks in the weld metal is telling a different story from one that reaches the same strength and breaks in the parent plate well away from the fusion line, and a procedure qualification reads both.

The specimen shall be taken transversely from the welded joint in such a way that, after machining, the weld axis remains in the middle of the parallel length of the test specimen. That single sentence carries most of the preparation work, because it means the machining has to be positioned against a feature that is partly buried and not perfectly straight.

Finding the Weld Axis Before Cutting Anything

On a ground or dressed plate the weld is not always where it appears to be. Fit-up wander, a cap that is offset from the root, and a second pass laid slightly to one side all move the true centre of the joint away from the visible bead. If the blank is set up on the visible bead and the weld axis ends up offset in the finished specimen, the joint is still tested, but the parallel length no longer carries the weld symmetrically and the fracture location becomes harder to interpret.

The practical answer is to establish the joint centreline on the blank before it goes into the machine, then set the program to that line rather than to the plate edge. It is a small step that costs a minute and protects the only thing the test is asked to report.

Before the cutter moves

Two Ways to Set Up the Same Blank

Scroll sideways to see the whole diagram

The bead is drawn offset in both panels because that is the situation being guarded against. Fit-up wander, a cap laid off the root and a second pass to one side all move the visible bead away from the joint centre, and only the right hand setup survives it.

Machining Through Three Hardnesses in One Pass

This is where preparing a welded specimen departs from preparing a plain one. A cutter crossing the specimen meets parent metal, then the heat affected zone, then weld metal, then the mirror image on the way out. On a hardened or high strength steel those regions differ enough that a feed which suits one of them will not suit the others.

The consequences show up as tool behaviour rather than as an alarm. Chatter appears as the cutter enters the harder region, the finish changes across the width of the specimen, and tool wear concentrates where the hardness step is. Left alone, that produces a specimen whose surface quality varies along the very section that is supposed to be uniform.

Across the joint

What the Cutter Meets in One Traverse

Scroll sideways to see the whole diagram

The specimen is machined so that the weld sits symmetrically in the reduced section. Everything about the setup exists to keep that true after material has been removed, not before.

The same reasoning we set out for machining Inconel, tool steel and hardened alloys applies here with an extra complication: on a welded coupon the hard region is not the whole part, it is a band in the middle, so the cutter is entering and leaving it on every pass.

Why the Preparation Route Decides What You Are Testing

Abrasive cutting is fast, and on a welded coupon it is the wrong kind of fast. It puts heat into a surface whose whole purpose is to represent the metallurgy underneath, and the heat affected zone of a weld is precisely the region least tolerant of a second thermal cycle from the preparation.

Controlled milling with adequate coolant removes material without adding a thermal history of its own, which is why it is the route we build machines around. It also leaves a surface whose finish can be specified and repeated rather than accepted, and surface finish is not cosmetic on a tensile specimen: it sets where a crack can start. Our article on surface roughness requirements works through the numbers.

Three things that quietly invalidate a weld coupon

The weld is off centre after machining. Setting up on the visible bead rather than on the established joint centreline moves the weld out of the middle of the parallel length, which is exactly what the standard asks you to preserve.

The surface carries preparation damage. Heat or smearing from an aggressive route sits on the region under test. The specimen then reports the preparation as much as the joint.

Finish varies across the width. When the cutter behaves differently in the weld metal than in the parent plate, the reduced section is not uniform, and the fracture location becomes an argument rather than a result.

Where the Coupon Comes From Matters as Much as How It Is Cut

A qualification coupon is welded specifically to be destroyed, and the sampling plan decides what the specimens can prove. Specimens taken close to the start of a run capture the conditions at the beginning of the weld, which is not always the same as conditions in the middle. Multiple specimens across a joint are taken because a weld is not homogeneous along its length, and one coupon is a sample of one.

None of that is the preparation shop’s decision, but it is the reason the preparation has to be repeatable. If four specimens are cut from one plate to compare positions along a joint, any variation the machining adds is indistinguishable from the variation the welding produced, which defeats the purpose of taking four.

Preparation equipment

Machines for Welded Plate Coupons

Welded butt joint specimens are flat coupons, so they are prepared on the flat specimen line, with the finish work done separately where the specification calls for it.

TensileMill CNC flat tensile specimen preparation machine for welded plate coupons

Flat Specimen Preparation

CNC milling of flat coupons to a programmed geometry, which is what keeps four specimens from one plate identical to each other.

Flat coupons

Programmed to the joint centreline
Repeatable across a batch

Flat tensile sample preparation machines

Longitudinal Polishing

Directional finishing along the specimen axis, for specifications that set a surface requirement on the reduced section.

Finish

Along the loading direction
Repeatable rather than judged by eye

Automatic longitudinal polisher system

Reading the Result Afterwards

When the specimen breaks in the parent metal away from the joint, the joint was stronger than the plate it connects, which is the outcome a procedure is usually aiming for. When it breaks in the weld metal or along the fusion line, the joint is the weak link, and the number matters less than the location.

The second output

What the Fracture Location Tells You, and the Case That Lies

Scroll sideways to see the whole diagram

The first three are results. The fourth is the reason the preparation has to be repeatable: it looks exactly like the second or the third on the report, and nothing downstream can tell them apart.

That is the practical reason a laboratory records where the fracture occurred and not only the load at which it happened. It is also why the specimen has to be prepared so that the fracture location means something: if the reduced section is not uniform, a break can be caused by the preparation and read as a property of the weld.

Transverse tensile work sits next to the other destructive tests on a qualification coupon, and the specimens for those come off the same plate. Preparing them all to a single programmed geometry is the difference between a set of results that compare and a set that merely coexist. Our guide to preparing tensile specimens to ASTM A370 covers the plain material case that the welded one builds on.

Next step

Send us the joint, not just the plate thickness

Tell us the parent material, the thickness range, the joint configuration and the standard your qualification cites, and we will tell you which flat preparation configuration handles it and what the setup looks like for keeping the weld centred in the parallel length.

Request a quotation

Related machines

Equipment That Does This Work