AWS D8.9 Resistance Spot Welding Testing of Automotive Sheet Steel

What the test does

D8.9 is not a single test. It is the automotive industry's collected set of laboratory procedures for finding out how a given sheet steel behaves when it is spot welded, and it contains several tests that are run together to answer that question: weldability lobe determination across a range of welding current, electrode life or endurance testing, hold-time sensitivity, weld hardness, shear tension strength and cross tension strength.

The mechanical parts of it are what reach a universal testing machine. A shear tension specimen is a lap joint containing one weld, pulled along the plane of the sheets. A cross tension specimen is two strips welded in a cross and pulled apart through the thickness. Both are loaded to failure and both are reported with the failure mode.

What it measures, and why it matters

The point of D8.9 is not a single number but a window. Welding current is stepped up while weld time and electrode force are held constant, welds are made at each current, and the resulting strengths and nugget sizes are plotted. The lower bound of the usable range is the current at which the nugget first reaches minimum acceptable size; the upper bound is the current at which expulsion begins. The gap between them is the weldability lobe, and its width is what tells a body engineer whether a steel can be welded reliably on a production line where nothing is perfectly consistent.

Hold-time sensitivity is the other result that matters commercially. Some advanced high-strength steels lose cross tension strength badly when the electrodes are held on the weld after the current stops, because the nugget quenches into hard martensite through the copper. A steel that passes on shear strength alone can fail this, and it is the sort of finding that changes which grade goes into a part.

Coupons, and how many of them

The mechanical coupons are conventional. What is not conventional is the quantity: a lobe study is a population, not a set of five.

Shear tension coupon

A lap joint containing one spot weld

Cross tension coupon

Two strips welded at right angles

Material

Automotive sheet steel, coated or uncoated

Coupon dimensions

Specified in the document, sized so failure occurs at the weld

Coating

Left on — it is part of what determines nugget growthPractice

Population for a lobe study

Several welds at each current level across the whole usable rangePracticePlus the coupons consumed by nugget sizing. Electrode endurance testing runs to several hundred welds on its own.

Weld the coupons on the schedule being assessed

DakThe document evaluates a material and a schedule together. Coupons welded on a convenient setting rather than the one under study answer a question nobody asked.

Test speed

Variable

Welding current, stepped across the range

Held constant

Weld time and electrode forceA lobe drawn at one electrode force cannot be compared with a lobe drawn at another. Both are correct and they describe different processes.

Hold time

Varied deliberately in the hold-time sensitivity test

Sheet range the schedules are written for

0.5 mm to 4.0 mm; the weld property tests for 1.2 mm and heavier, the endurance test for coated steels at or below 500 MPa and under 1.2 mmClause 1.1 of the 2022 edition, which also recommends a minimum of two endurance tests per material. The 2012 edition gave 0.6 mm to 3.0 mm.

Mechanical test rate

Fixed by the document, but the figure is in the purchased text — the free previews of the 2012 and 2022 editions both stop at the end of clause 1It sits in the clause on testing the shear and cross tension specimens. No independent source reproducing it was found, so read it from the standard. Where a laboratory runs the ISO counterparts instead, the rate belongs to those documents, and which one produced a strength figure has to be stated beside it.

Reported

Strength and failure mode at each condition, plus nugget size

What the machine must be capable of

Shear tension of a spot weld in automotive sheet fails in the low tens of kilonewtons at most; cross tension fails lower still. A 50 to 100 kN frame covers the mechanical work, and clean resolution at the bottom of the range matters more than capacity. D8.9M's own accuracy requirement cannot be read: both free previews end at clause 1, and clause 2.3, where the ASTM normative references sit, is not in them — so no class is published here, for the same reason the pull rate is not.

Shear tension needs flat grips with serrated faces and packing shims of the sheet thickness in both jaws, so the two sheets are pulled in line rather than being progressively peeled. Cross tension needs a dedicated jig that bolts to the four arms of the cross specimen and applies the load normal to the weld. Data rate has to be high enough to capture a peak that arrives without warning.

The pull rate is the one machine setting D8.9 does not publish: it sits in the clause on testing shear and cross tension specimens, and the free previews of both the 2012 and 2022 editions stop at the end of clause 1. Where a laboratory runs the ISO counterparts instead, the rate comes from those documents, so which one produced a figure belongs beside it. What D8.9 does fix tightly is the welding side, because that is where its answers come from: current stepped across its range while weld time and electrode force are held constant, hold time varied deliberately in the hold-time test, and the whole scheme written for sheet from 0.5 mm to 4.0 mm — the weld property tests intended for 1.2 mm and heavier, the endurance test for coated steels at or below 500 MPa and under 1.2 mm, with two endurance tests per material the recommended minimum.

The rest of D8.9 — welder, current monitor, electrode force gauge, the metallography behind nugget sizing — is welding-shop equipment rather than testing-machine equipment, and a laboratory quoting D8.9 capability should be clear about which parts it actually performs.

Grips and fixtures for this method

TJ-144

Heavy Duty Hydraulic Grips

For the shear tension coupons: flat serrated faces wide enough for the full coupon, with consistent clamping force across a long current series where the current is meant to be the only variable.

Specifications

Self-identifying

Load Cells

A cell sized for the joint. Shear tension on automotive sheet fails in the low tens of kilonewtons and cross tension lower still, so a cell chosen for the frame throws away resolution across the whole lobe.

Specifications

How the mechanical work runs

1. 01Fix weld time and electrode force for the series.
2. 02Step the welding current across the range from undersized nugget to expulsion.
3. 03Weld several coupons at each current level.
4. 04Size the nuggets by peel or by section to find the lower bound.
5. 05Note the current at which expulsion begins to find the upper bound.
6. 06Test shear tension coupons with packing shims in both grips.
7. 07Test cross tension coupons in the bolted cross-tension jig.
8. 08Record peak force and failure mode for every coupon.
9. 09Plot strength and nugget size against current.
10. 10Repeat at the hold times required for the sensitivity assessment.
11. 11Report the lobe, not a single strength figure.

Quoting one shear tension force as an AWS D8.9 result is the commonest misuse of this document. The document produces a window; a single number taken out of it carries none of the information the study was run to get.

Calculations

What the report has to contain

• Reference to AWS D8.9M and the edition
• Steel grade, coating type and mass, and sheet thickness
• Electrode material, geometry and dressing regime
• Weld time, electrode force and hold time
• Welding current at every level tested
• Nugget size at each level, and how it was measured
• Shear tension and cross tension force at each level
• Failure mode for every coupon
• The plotted lobe with its lower and upper bounds
• Which parts of the document were performed, and which were not

What goes wrong in practice

Reporting a single strength figure as though it were a D8.9 result is the commonest error, because the document's answers are curves and windows rather than numbers. Missing packing shims on the shear tension specimen, uneven bolt torque on the cross tension jig, and a logger too slow for the peak are the mechanical faults, and they are the same ones the ISO methods suffer. Comparing a lobe drawn at one electrode force with a lobe drawn at another is a subtler error and a common one.