Project background
Veltra Meccanica S.p.A. is a Modena-based tier-one supplier of structural battery pack enclosures and chassis cross-members for European electric vehicle programmes. Around 400 people work across two buildings, with one laser cell and one manual load area serving all four customer programmes.
Battery pack work changed the shape of the plant. The rails are 3 mm S700MC and 4 mm 5754 aluminium, cut to a ±0.5 mm positional tolerance because they are later seam welded and then leak tested. A visually perfect cut that sat 0.8 mm out of position still failed at the weld.
The existing cell was a plasma table working 3 mm mild steel, with 5–6 mm high-strength work sent to a laser job shop in Emilia. Cycle time on the job shop parts was 41 minutes per kit, and the queue in front of that supplier had twice caused a line stoppage at Veltra.
Loading was entirely manual — one operator per machine — and the plant ran three shifts, which meant three trained laser operators for a machine that only needed attention at nest change.
The challenge
Positional accuracy was the binding constraint. S700MC springback at 3 mm added up to 0.6 mm of variation at the rail ends, and the weld cell rejected one kit in six.
The four customer programmes each ran their own programme library on two different machines, so an operator moving between the plasma table and the job shop kit was effectively changing jobs.
Manual loading meant the machine idled during every nest change, and the rate of a 3 mm mild steel blank was limited by the operator's ability to place the next sheet, not by the cutting head.
And there was no realistic path to taking on the 12 mm structural parts that were also outsourced — the plasma table could not hold the flatness, and there was no floor space for another machine with a manual load area.
Our solution
Veltra installed an LP-F3015-12kW in the existing laser building in early 2025, replacing the plasma table for all 2–6 mm work, including the pack rails and the 3 mm mild steel blanks that were the plasma cell's core workload.
A second machine, the LP-F4020-10kW, was positioned on the other side of the plant with an LP-A2000 Automated Loader serving the pallet magazine. The loader was the key decision: it allowed the 10 kW machine to sit on the floor space freed by the plasma table rather than requiring a longer building.
Nesting for all four programmes is now driven from one order book, with machine assignment by thickness. A single operator covers both cells during the day shift, and the night shift runs with the LP-A on the 10 kW machine and a cross-trained operator on the 12 kW.
The fixture and support strategy on the rail nests was rebuilt around the springback behaviour — support points were moved off the cut edge, and the rails are now cut with a two-stage strategy on 3 mm S700MC that holds the corner within 0.3 mm.
The 12 mm structural cross-members were brought in-house on the 10 kW machine in the first quarter, ending the second off-site supplier relationship on the pack programme.
Delivered outcome
Mixed-model takt on the pack rail programme improved by 26%, and the improvement is visible on the loading side as much as on the cutting side, because the LP-A loader removes the nest-change idle time entirely.
Weld rejections attributed to positional error fell from 16.2% of kits to 3.1%. Seam tracking on the pack rails now holds ±0.4 mm against a ±0.5 mm requirement, with the leak-test reject rate down from 2.8% to 0.9%.
Material yield across the four programmes rose by 6 percentage points once nesting was allowed to run across the shared order book rather than per programme, which alone is worth more than the machines' electricity cost by an order of magnitude.
Direct laser operating hours per part fell 34% through the combined effect of the loader and the single-operator arrangement, and the plant has since redeployed one of its three laser operators into the pack sub-assembly area.
The 12 mm cross-member outsourcing ended in 2025 and both off-site laser relationships on the pack programme were closed. On the original business case, the two machines and the loader paid back in 26 months.