Resistance spot welding

A car body holds together on something like four to five thousand spot welds, each made in about a quarter of a second, and not one of them can be inspected by looking at it. The process is almost entirely invisible: everything that matters happens at the interface between two sheets, under an electrode, inside a fifth of a second.

A robot welding gun closing on the flange of a car body panel in a body shop, the copper electrode caps gripping the two sheets, with a shower of sparks and a bright flash at the weld and rows of earlier weld marks visible along the flange. — Rows of dimples along a flange. Each one is a weld nobody will ever see, and none of them can be judged from the mark it left.

Where the heat comes from

Current passes down through the electrode, the upper sheet, the interface, the lower sheet and out. Every part of that path has resistance, but the faying interface has the most — two rough surfaces touching at their high points only.

A sectioned view of a resistance spot weld in progress. Two sheets are clamped between water-cooled copper electrodes. The faying interface between the sheets is marked as the point of highest resistance, with a lens-shaped molten nugget forming there and the heat affected zone shaded around it. Arrows show current passing vertically through the stack and water flowing through channels inside both electrodes. Beside the section, a bar chart compares the resistance of the electrode-to-sheet contact, the bulk sheet and the sheet-to-sheet interface, with the interface much the highest. A second small diagram shows the same joint after welding, peeled apart, with a plug torn out of the lower sheet and its diameter marked as the measured nugget size. A third panel shows expulsion, with molten metal escaping sideways at the interface and the resulting nugget drawn smaller and containing a void. — The interface is the hottest point because it is the worst conductor. The electrodes are water cooled so that they stay the coldest, and the weld stays where it was aimed.

So the interface melts first, while the copper electrode faces — water cooled, highly conductive — stay solid. A lens-shaped nugget grows at the interface, entirely enclosed by solid metal. That containment is why no filler, no shielding gas and no joint preparation are needed.

Parameter	Raise it and	Practical limit
Current	Heat rises steeply — it goes as the square	Expulsion
Weld time	Heat rises in proportion	Cycle time, electrode wear, wide heat affected zone
Electrode force	Contact resistance falls, so heat falls	Too low and the interface arcs; too high and the sheet is marked

The counter-intuitive one is force. Squeezing harder flattens the asperities, contact area rises, resistance drops, and the same current now deposits less energy. Force and current are not independent settings, which is why a weld schedule specifies all three together.

The weld lobe, and why the middle of it matters

Plot current against time and two boundaries appear: below one, the nugget is too small; above the other, expulsion starts and molten metal is thrown out, leaving a small porous weld behind. Between them is the lobe — the usable window.

A weld lobe diagram with welding current on the horizontal axis and weld time on the vertical axis. A shaded band runs diagonally, bounded on the left by a minimum nugget size curve and on the right by an expulsion limit curve. A dot marks a robust production setting in the middle of the band, with arrows showing that electrode wear pushes the effective operating point leftward toward the undersize boundary over the life of a tip. Beside it, a second diagram shows three destructive test results side by side, a good weld with a round plug pulled from one sheet, an undersized weld with a small plug, and an interfacial failure where the joint has separated flat across the weld with no plug at all. A third panel shows shunting, with two adjacent welds and current dividing between the new weld and the existing one. — The window narrows as the tip wears, so the setting is chosen in the middle of the lobe rather than at the point that just works today.

Production settings sit in the middle of the lobe, not at the edge that just works, because in service the electrode face grows and roughens and the delivered current drifts. That drift is why welding controllers run stepper programs that raise current as the tip ages, and why tips are dressed on a schedule.

Expulsion is a common misreading. The sparks look like evidence that something energetic and therefore good has happened. They are evidence that metal has left the joint.

Coatings and high strength steels

Advanced high strength steels bring a different problem. Their higher carbon equivalent means the nugget cools to hard martensite, and a hard nugget has shear strength but little toughness. Instead of tearing a plug out of the parent sheet, the joint can crack flat across the interface — an interfacial failure, which is a reject even if the measured load was high. The answers are tempering pulses after the weld and testing in cross tension as well as shear. The underlying metallurgy is the same as in heat treatment of steels.

How a weld is actually judged

Test	What it gives	Where used
Peel / chisel	Nugget diameter of the torn plug	Shop floor, every shift
Cross tension	Toughness, detects interfacial failure	Validation, high strength steels
Macro section	Nugget depth, penetration, porosity	Setup and investigation
Ultrasonic	Non-destructive estimate of nugget size	Audit, in-line on some lines

The workhorse is destructive. Peel the joint apart, and if the weld is sound it tears a plug out of one sheet; the diameter of that plug is the weld. A common minimum is about four times the square root of the thinner sheet thickness in millimetres — roughly 4.4 mm on 1.2 mm sheet.

The surface indentation tells you almost nothing. A gun can leave a textbook dimple and no nugget at all.

Designing for the process

Access from both sides. Two electrodes have to reach the flange, which is why body structures are built up from open pressings before they are closed.
Flange width must hold the nugget, the edge distance and the tip. Too close to the edge and the molten pool bursts out sideways.
Pitch between welds must be large enough to limit shunting — an existing weld is a low resistance path in parallel, so it steals current from the new one.
Stack thickness ratio beyond about 3 to 1 pushes the nugget toward the thin sheet, and three sheets need a schedule of their own.

Those flange rules interact directly with the pressing rules in designing sheet metal parts — the flange has to be formable before it can be weldable.

What to take away

The interface is the highest resistance in the circuit, so that is where the metal melts.
Current and time add heat; more force reduces it by lowering contact resistance.
Squeeze, weld, hold — and the hold is what stops the nugget cracking as it shrinks.
The weld lobe is the window between undersize and expulsion; work in the middle of it.
Expulsion is a failure, not a success. The weld left behind is small and porous.
Zinc lowers interface resistance and destroys electrode faces — it drives tip dressing and steppers.
High strength steels can fail interfacially; test in cross tension, not shear alone.
A weld is measured by the diameter of the plug a peel test tears out, not by the dimple.
Shunting starves a weld placed too close to an existing one, so pitch is specified.
Flange width must suit the electrode and the edge distance, not just the nugget.