Roller Seam Welding: How A Continuous Seam Is Actually Made

Roller Seam Welding: How a Continuous Seam Is Actually Made

A seam weld looks like a line, and it is easy to assume it is made by a different process from a spot weld. It is not. A seam welder is a spot welder whose electrodes happen to be wheels and whose workpiece happens to move, and almost everything about how the seam behaves follows from that single substitution.

This article is about how the seam is actually formed, what makes it continuous rather than a row of marks, where the process wins against a series of individual spots, and where it quietly loses.

Our seam welding cell: the roller head sits inside the guarded enclosure, with the force cylinder above it. Panels are fed through the wheels rather than lowered onto them.

The Machine Is a Spot Welder on Wheels

Two copper rollers replace the two electrode caps. Force is applied through the roller holders by a cylinder above, current passes from one wheel through the work into the other, and the part is driven between them at a controlled speed. Where a spot welder stops, presses and releases, a seam welder does all three continuously while the work travels.

That changes the controlling variables. On a spot welder the schedule is current, force and time per spot, and the index distance between spots. On a seam welder the equivalents are current, force, wheel speed and the pulse rate of the current, because the current is not continuous: an inverter seam welder fires in short pulses as the work passes, so the two settings that matter most are how often it fires and how fast the work moves.

The travel speed is usually expressed in metres per minute and the pulse rate in hertz. Divide the speed by the pulse rate, with the units made consistent, and you get the pitch between consecutive weld nuggets. That pitch is the number that decides whether the result is a seam or a row of marks pretending to be one.

Force is the setting that is easiest to forget because it does not appear in the schedule as a number the operator watches. It is applied by a cylinder through the roller holders, and on a seam cell it has to be held while the work moves, not just while the current fires. Where it is switched rather than ramped, the first and last centimetres of a seam are welded at a lower force than the middle, and the two ends of a long seam are exactly where a joint is most likely to be weak, marked, or both. Ramping the force so that it is already on before the first pulse and stays on after the last is a small change in the control sequence that removes a recurring argument about the ends.

What Makes It Continuous

A seam is continuous when consecutive nuggets overlap. If the pitch is larger than the nugget diameter, the joint is a series of spots with unmeted gaps between them: it will hold load along the line but it will leak, and under a peel test it will open like a zip. If the pitch is small relative to the nugget, the nuggets merge into a continuous band and the joint behaves as a seam should.

The practical consequence is that overlap is a design decision, not an operator preference. A leak-tight joint needs overlap by a comfortable margin, which means either a slower travel speed at a fixed pulse rate or a higher pulse rate at a fixed speed. Both raise the heat input per millimetre, and both make marking and distortion more likely, which is the trade the process forces on you.

On the machine the pitch is not displayed as a number. It is the result of two settings that are displayed, which is why a seam schedule written as speed and hertz is more useful than one written as a percentage of power.

The Wheels Are the Consumable

The roller is a copper alloy wheel, water cooled through its holder, and like an electrode cap it changes shape with use. The rim flattens slightly, the edge rounds, and the contact band widens, which spreads the current over more area and reduces the current density without anything on the machine reporting it.

The upper roller of the same machine, a copper wheel on a water-cooled holder. Seam width is set by the rim of this wheel, not by the current.

Two things follow. The first is that wheel dressing belongs in the production schedule for the same reason tip dressing does on a spot welder: not as housekeeping, but as a parameter. The second is that the seam width you get is the width of the contact band, so a wheel that has flattened produces a wider, shallower seam than the one the schedule assumed.

Dressing the wheel changes the geometry in a second way that is easy to miss. Every time the rim is trued up the wheel gets slightly smaller, so at the same programmed speed the rim surface travels a little slower than it did when the schedule was written, and the pitch shifts toward more overlap. That is the safe direction to drift in, but it is still a drift, and on a machine that runs the same wheels for months it is worth recording the wheel diameter next to the schedule so that the change is visible before the seam tells you about it.

Wheel marks on the visible face are therefore unavoidable, and on a cosmetic panel they have to be designed around rather than eliminated. On an internal part they are irrelevant. It is worth deciding which one you have before the parts are cut.

Where a Seam Beats a Row of Spots

The clearest case is leak tightness. A fuel tank, a radiator, a water jacket or an enclosure that must keep liquid in cannot be made with discrete spots, however closely spaced, because each gap between spots is a potential leak path. A continuous seam closes the path by construction.

The second case is stiffness along a flange. A crash beam or a door reinforcement gains its behaviour from a continuous connection between the two sheets, not from a series of point connections, and the difference shows up in how the section deforms rather than in static strength.

The third is handling. On a long straight seam the part moves at a constant speed and the operator does not reposition it, so the cycle is dominated by loading rather than by the welding itself. Where the seam is long and straight, that is often the strongest argument of the three.

Where It Loses

Complex geometry is the first limitation. The work has to pass between two wheels, which means the joint has to be reachable from both sides along its whole length, and a curve in the wrong plane is enough to make the process impractical. A spot welder with a long arm can reach into places a roller cannot.

The second is coating. Zinc and primer contaminate the wheel surface, and a contaminated wheel marks and welds unevenly. On coated sheet the maintenance interval shortens and the seam becomes less consistent than the same seam on bare steel.

The third is the nature of the failure. A single bad spot in a row of forty is a local defect. A section of seam that did not overlap is a continuous leak, and it is much harder to find afterwards and much more expensive to repair.

What to Specify

When a seam job is specified, these are the values that decide whether the machine can do it, and the ones worth putting in writing rather than leaving to commissioning.

The overlap figure is worth agreeing explicitly, because it is the one number that separates a structural seam from a decorative line, and it is invisible in the finished part until someone cuts a section.

Requirements vary by destination market. Always confirm against the current edition of the applicable standard and with your certification body.

PDKJ continues to provide customized laser welding and resistance welding solutions based on different materials, workpiece structures, welding requirements, and production applications.

Related Products

• Medium Frequency Inverter Seamless Spot Welding Machine
• Spot Welding Machines

Our assembly area in Dongguan, where the seam machines are built and tested.

Contact PDKJ

Send us your part drawing or a sample, and our engineers will run a welding feasibility trial on your own material and recommend the equipment configured for your application.