Large-format laser cutting for long sheet metal parts

A part 9 m long used to be a fabrication: three laser-cut sections, two splices and a straightening operation after welding. A 12,192 mm bed changes that decision, and large format laser cutting for long sheet metal parts is now mostly about the constraints that arrive with the single piece rather than about whether it can be cut. Thermal growth across 12 m, coil bow in the stock, slat support under a thin panel and the width of the sheet you buy all start to matter more than the cutting head.

This article works through the capacity, the arithmetic that decides single piece versus splice, the nesting yield on long stock, the edge quality call a drawing should carry and the point where the 4,450 mm brake quietly caps the length. Machine and material limits are listed for the plant, and the cost side sits in the laser cutting hub and in what the 12 kW machine enables.

What can a 40 ft bed cut in one piece?

The bed is 40 ft x 16 ft, which is 12,192 mm x 4,877 mm. A part up to roughly 12 m long and 4.8 m wide can therefore be cut without a joint, and the practical limit is usually the stock rather than the machine: sheet and coil are supplied in widths of about 1,250 to 1,500 mm for common gauges, and a 4.8 m width means a mill order or a welded blank.

That changes the design conversation. A 9 m trim panel, a 10 m canopy side, a 6 m rail cover or a 7 m fascia can be one piece, so the splice moves from the cutting drawing to the assembly drawing, where it should have a reason: handling, transport, a service joint or a change of material. If a splice exists only because the previous supplier could not cut past 3 m, delete it and re-check the flat pattern, because every splice adds a lap, fasteners, a step and a coating failure point along the joint.

How do you decide between one piece and a splice?

Compare the two options on four numbers: material yield, the cost of the joint, the risk of scrapping a long part, and the downstream operation. Cut one 12 m part and you buy a long blank, scrap the whole blank if the nest is wrong, and handle a floppy 12 m panel through every step. Cut two 6 m parts and you add a lap joint with, say, 14 rivets or bolts, a 25 mm overlap that consumes material, and a step of one thickness at the joint that shows through a painted face unless it is planned.

A working rule: a splice is worth it when the joint falls on a structural member or a hidden flange, when the part has to be coated in a booth shorter than the part, or when handling the single piece needs equipment the buyer does not have. A single piece wins when the part is visible, when flatness across the joint matters, or when the joint would sit in a wet area.

The arithmetic is simple enough to run at quotation stage. If the spliced version carries 14 fasteners at a fitted cost of USD 0.35 each, that is USD 4.90 per part, plus the drilling or punching stations, plus the inspection of a joint that no drawing toleranced. A single-piece blank that avoids all of it is usually cheaper even when the material costs slightly more, and it removes a field failure mode.

How much does temperature move a 12 m part?

Steel expands at about 11.7 micrometres per metre per degree Celsius. Over 12 m, one degree is 0.14 mm, so a part measured 25 degrees warmer than the reference temperature is 3.5 mm longer on the tape. On a 12 m part with a plus or minus 2 mm callout, the shop temperature is a larger variable than the machine.

Two responses work. The first is to inspect at a stable temperature, which means letting the part cool on a supported table rather than measuring it as it leaves the bed. The second is to write the tolerance for the size: ISO 2768-1 class m tolerates about plus or minus 2 mm on a dimension of 2,000-4,000 mm, and general tolerances widen with length for the same physical reason. A plus or minus 0.5 mm callout on a 12 m overall length is not a tight drawing, it is an unachievable one, and it invites a supplier to sign a deviation for every part.

The same effect explains why a long part can measure differently at the two ends of a shift, and why a first article cut at 08:00 and checked at 16:00 can pass on one dimension and fail on another. Record the part temperature on the report and the discussion ends quickly.

Cutting limits and tolerance calls

How does nesting change the cost of a long part?

Stock width decides yield before nesting skill does. Take a panel 2,900 mm x 1,450 mm in 3 mm CRCA. On a 6,000 mm x 1,500 mm sheet, two panels fit along the length with a 10 mm web and 25 mm trim, giving 8.41 m2 of part from 9.00 m2 of sheet, or 93% utilisation. On a 12,192 mm x 2,438 mm sheet the same panel fits four along the length but only one across the 2,438 mm width, giving 16.82 m2 of part from 29.72 m2 of sheet, or 57%.

The second sheet is nearly 300 mm narrower than two panels and loses the whole width twice over. A quote built on that stock looks expensive, and the fault is the sheet choice rather than the laser. When a long part is being priced, ask which stock width was assumed, because 57% and 93% utilisation is the difference between two suppliers quoting the same drawing at very different prices.

Nesting for long parts has a second layer: the remnant. A 12 m bed leaves a 2-3 m offcut on most jobs, and that offcut is worth money only if it is identified, labelled and stored for a part that uses it. Unlabelled remnants become scrap. How nesting reduces cutting cost covers the software side; the discipline of labelling remnants belongs to the shop.

What goes wrong on a long part?

Bow is the first problem. Coil and sheet carry residual stress, and when a long blank is cut free the stress releases and the part curves, often by 8-15 mm along 9 m. The curve is not a cutting error and it will not be fixed by cutting slower. The remedies are to cut from stock that has been stress-relieved, to leave tabs that hold the part flat until it is lifted, and to design a stiffening flange so the part does not have to be flat on its own.

The second is support. A thin 12 m panel sags between slats, and if the sag is present while the head passes, the focus distance changes along the cut and the edge quality varies from one end to the other. Slat condition and sheet support are set-up decisions, and a worn slat bed is the usual reason a long part shows a line of dross at one position.

The third is handling damage. A 12 m panel that has been lifted from the middle by two people is not the same part that was measured, and a crease is permanent. Long parts need a lifting plan and a stacking plan before the first one is cut.

How should the edge and the measuring plan be specified?

Laser edge quality is classifiable rather than arguable. ISO 9013 covers thermal cuts from 0.5 mm to 32 mm on laser, plasma and flame processes, and it defines the perpendicularity tolerance and the mean profile height Rz5 that separate one quality class from another. A drawing that cites a class for the edges that will be welded or seen, and leaves the rest to normal commercial finish, gives the inspector something to accept or reject against.

Measuring a 12 m part needs the same discipline. Support it on level tables at three or more points so it is not bowed by its own weight, let it reach a stable temperature, and measure length with a laser tracker or a taut wire and gauge blocks rather than a tape hooked over a burr. Record where along the length each dimension was taken: a 12 m panel that is 1 mm over at one end and 1 mm under at the other is not the same defect as one that is 2 mm over in the middle, and the correction is different.

What Ganesh runs for long parts?

The 12 kW fibre laser with its 40 ft x 16 ft bed is the machine that makes single-piece long work possible here, supported by two 6 kW and two 3 kW machines for shorter nests and thinner gauges. It cuts mild steel, CRCA, stainless steel, aluminium, brass and copper, up to 20 mm in plate, with the iron worker covering 30-40 mm for heavy plate work. Downstream, the 200 T five-axis brake with a 4,450 mm bed, three CLOOS robotic MIG cells and the seven-tank pretreatment and powder line let a long part be formed, welded and coated in one release. The plants hold ISO 9001:2015 and ISO 14001:2015.

The laser cutting capability page lists the materials and envelopes, and bending long panels without distortion covers what happens to that panel on the brake.

What should a DXF for a long part contain?

1. The flat pattern at 1:1, with the bend lines and the bend direction on a layer the nesting software ignores.
2. Material grade and thickness, with the Indian equivalent named if the drawing uses a European grade.
3. Overall length and width, plus the tolerance basis: general tolerances or named dimensions.
4. Every hole with its functional size and position tolerance, and a note on holes that can be cut undersize and reamed.
5. Edge quality call, at least for the edges that will be handled, welded or coated.
6. Splice detail if one exists, with the fastener pattern and the joint preparation.
7. Grain or rolling direction if the part is visible or formed, because a long part cut across the grain behaves differently.
8. Quantity, release pattern and whether a remnant should be returned.

The DXF preparation checklist covers layer naming and units, and the laser cutting cost estimator gives a first-order price before the drawing is final.

When is laser cutting the wrong process for a long part?

Laser is the wrong route when the part has to be bent beyond 4,450 mm, when the thickness exceeds 20 mm, or when the annual volume is high enough that a press tool would blank the part faster and cheaper. It is also a poor fit for a long part with a single large opening and no other features, where a punch and shear route on a coil line may be quicker than a nest.

Ganesh does not cut structural sections, tube or pipe on the flat-bed laser, and does not offer plasma or waterjet for thicker plate. If a 30 mm base plate is part of the assembly, it comes off the iron worker and is drilled or punched, not laser cut, and the drawing should say so before the RFQ is priced.