Commissioning paperwork for an automotive display line usually arrives as a checklist copied from a nearby project: dimensional reports, an uptime figure, a cycle-time figure. Those numbers get signed off in a week and the line still misses its real target, because nothing on the sheet tested what an automotive display line actually fails at — mixed model changeover, alignment on curved and soft-to-hard bonding, handling of glass and film at large sizes, and the vision and recipe setup behind every bond. The checks below are the ones that expose those problems while the supplier's engineers are still on site and adjustments are cheap.
They apply to the bonding and lamination stations in the Semiconductor Process Equipment range as much as to the transfer, joining and test stations around them. Where a station is listed as customised for a specific customer's product, treat that customer's process window as your starting point and ask for the tolerance stack-up rather than the machine brochure.
1. Model mix and changeover, run before anything else
Automotive display production rarely runs one part number for a week at a time. Instrument clusters, central multimedia units and passenger displays share a line, and the line's real output is set by how it behaves at the changeover. Run the changeover test first, not last, because a station that needs a new fixture, a new program and a fresh alignment study every time models change will never hold a stated cycle time.
- Time the changeover from the last good unit of model A to the first good unit of model B, including the first-article check. Record it three times.
- Confirm that recipes are stored and recalled by model, with the parameters an operator can legitimately change separated from those locked to engineering.
- Run a mixed queue — A, B, C, A, B — and verify that no station silently keeps the previous model's recipe, film thickness or pressure profile.
- Check changeover on the smallest and largest model in the mix, not just the nominal one.
2. Alignment and registration, measured not asserted
Alignment is where automotive display bonding is judged, and it is where a "meets spec" statement means least. Insist on measurements from the machine's own alignment system plus independent measurement of the same parts.
- Record fitting error against the figure stated for that station. Some bonding stations in this class of equipment are listed with a fitting error of 0.2 mm or better; if the station on your line carries a stated figure, ask for it in writing and test against it. If the specification for your configuration has no published figure, ask for it before signing rather than accepting a verbal claim.
- Measure across the whole working area, not the centre. Corners and edges reveal plate flatness, roller pressure variation and vacuum uniformity problems that a centred test part hides.
- Test soft-to-hard and hard-to-hard combinations if your product mix includes both, and keep the results separate. A station tuned for one pairing is not automatically good at the other.
- Run repeated bonding on the same part and measure the scatter, not just the best result. Scatter, not a single good reading, predicts what your first-pass yield will look like in month six.
3. Bond quality on the defect classes that matter in vehicles
Automotive displays sit in a hot, vibrating, UV-exposed cabin, so the defect bar is different from a handheld device. Agree the defect classes and their acceptance limits with the customer's quality team before commissioning, then test for them deliberately.
- Bubbles and inclusions after bonding — count and size them, and confirm that the debubbling station brings parts inside the agreed limit rather than just reducing the count.
- Delamination and lifting at edges and curved transitions, checked after thermal cycling, not only at room temperature.
- Film wrinkle and bridge on curved surfaces, including on parts with the tightest bend radius in the mix.
- Adhesive coverage and void at the bond line, on a sampled part sectioned or otherwise destructively examined.
For the debubbling step, check that the process parameters — pressure, temperature, ramp and hold — are recorded per recipe and repeatable between runs. A machine that reaches the right pressure is only half the check; the ramp profile is what decides whether a small bubble shrinks or a large one becomes a permanent void.
4. Cycle time measured on the real queue
A cycle time quoted on a single test part under ideal conditions is not a line capability. Some stations in this equipment class publish a cycle time derived from a specific test size and pairing; use the published number as a comparison point, and confirm the number that applies to your part and product mix in writing.
- Measure station cycle time in a full mixed queue over at least one full shift, and report the distribution rather than an average.
- Identify which station sets the line pace, then test whether the line still meets the target when that station runs the slowest product variant.
- Include load, alignment, bonding, unload and transfer in the timing, and check that the transfer between stations does not add waiting time that no single station's timing shows.
5. Availability and the stop-recovery path
Uptime claims are only useful with a definition attached. Agree what counts as a stop, what counts as a micro-stop, and who is called, then test recovery rather than trusting a percentage.
- Log every stop during commissioning with cause and duration. A station that recovers in seconds is a different problem from one that needs a technician on site.
- Test the machine after an interruption: emergency stop, air or power interruption mid-cycle, and a jam at a transfer. Verify that the part in process is either recoverable or reliably scrapped, and that the station resumes with the correct recipe.
- Check that alarms name the failed function clearly enough for a shift technician to act without a manual open. Ambiguous alarms are a common cause of long stoppages on lines that otherwise meet their cycle time.
6. Process control, recipes and traceability
Automotive customers audit process data, not opinions. Confirm before sign-off that each station records and exports what your customer will ask for.
- Per-unit traceability: which station, which recipe, which operator, which time, and which film or adhesive lot, tied to the part identifier.
- Parameter logging of pressure, temperature, vacuum and cycle events at a sampling rate fine enough to show a deviation after the fact.
- Recipe version control, so a changed parameter is visible rather than absorbed silently into a station's program.
- Data export in a format your customer's quality team can actually ingest, tested once during commissioning instead of negotiated later.
7. Consumables, maintenance and the line's own upkeep
Commissioning is the moment to establish what the line costs to keep running, while the supplier's engineers can still answer.
- Wear items on bonding stations — vacuum seals, release and protective films, roller surfaces, heater elements, nozzles — with their expected intervals and a spare list priced before the first year starts.
- Calibration intervals for the vision and alignment systems, and who performs them. An alignment system out of calibration produces good-looking parts that fail registration in the field.
- A written fault tree for the likely stoppages on each station, kept at the station rather than in the manual.
- Utility loads — electrical, compressed air, vacuum, exhaust, and nitrogen where a station uses inert atmosphere — verified against what the building actually delivers, at the socket and manifold, not at the nameplate.
8. Operator training and the first-article routine
Run at least one shift with your own operators, with the supplier observing and correcting. Their ease with the machine tells you little; your crew's first attempt tells you a lot.
- Require your operators to perform a changeover, a recipe recall, a jam recovery and a first-article check unassisted, and record what they could not do.
- Fix the first-article routine at commissioning: which dimensions and which defect classes are checked after every changeover and after every maintenance intervention, and who releases the line afterwards.
- Write the work instructions at the machine during commissioning, using the numbers measured in these tests, not generic text.
9. Documentation set that has to be complete before sign-off
- As-built specification for each station, including the figures that apply to your configuration — fitting error, working temperature and vacuum ranges, press or roller pressure, cycle time — with the test conditions stated alongside each number.
- Dimensional and foundation drawings, utility connection points, and the machine footprint including handles and access openings.
- Risk assessment and safety function descriptions for guards and interlocks, and for any heated, pressurised or vacuum chamber, plus the procedure for entering a chamber safely.
- Spare parts list with part numbers, a recommended initial stock, and lead times.
- Acceptance report signed against the tests above, with the residual open items listed and dated rather than omitted.
Where the supplier's own engineering services change the picture
If your line includes a station that has not been built for your product before, commissioning alone will not remove the risk — process development beforehand will. Shenzhen PowerDe Automation Precision Equipment lists automated production line solutions alongside engineering services including product development and hardware testing, which is the route when a bonding or joining process still needs proving against your glass, film and stack-up. Send the part drawings, the model mix and your target takt to the contact page, and ask for the station specification with the commissioning tests above written into the acceptance terms rather than negotiated afterwards.
