A test plan for an automotive-display production line reads very differently from a test plan for a consumer panel line. The units are larger, they are bonded to curved or irregular surfaces, and once a display module is installed in a vehicle it is expensive to reach. That is why the plan cannot be written as one inspection step at the end of the line. It is written as a chain of checks, each placed where a defect can still be corrected. Below is the structure we use when we discuss semiconductor process equipment and bonding lines for instrument clusters and central multimedia displays.
1. Define the product and the process window before defining the tests
The first page of the plan is not a test list. It is a description of what is being bonded. An automotive display may combine cover glass, a touch sensor, an adhesive layer such as OCA, an LCD or OLED module, and a decorative or functional film on the back side. Each interface has its own failure mode. The plan should state, for every interface:
- the material combination and the supplier lot traceability you will record;
- the alignment tolerance the bond must hold, and how it will be measured;
- the process conditions that must be logged per unit: temperature, pressure, vacuum level and cycle time.
Without that baseline, later test results cannot be traced back to a cause, and the plan turns into a sorting exercise.
2. Put an alignment check before lamination, not only after it
Misalignment is the cheapest defect to find and the most expensive one to leave. In our bonding equipment for 5–15 inch touch glass and LCD or LCM assemblies, alignment is handled by CCD vision together with an XYθ (UVW) correction stage, so the correction happens before the film contacts the substrate. The test plan should mirror that logic. Include a pre-bond alignment measurement with a recorded value, not a pass/fail judgment, so that drift can be seen while it is still small.
For larger formats the same principle applies at a different scale. Automatic laminating machine lines for large LCDs are planned around registration across the whole area of the panel, where the corner of a large glass sheet can be out of tolerance even when the centre is correct. A test plan that measures only one point will pass parts that fail in the field.
3. Plan the bubble and void checks as two different tests
"No bubbles" is not a test specification. Two different defects are usually lumped together under that phrase:
- alignment bubbles, created when the film lands at the wrong position and traps air along an edge;
- process voids, left inside the adhesive when the vacuum level or the pressing sequence is not holding.
They are corrected by different controls, so the plan should separate them and should say who owns each correction. Where a post-bond step is used to remove residual voids, as with a high-pressure debubbling machine whose tank is built in SUS304 stainless steel, record the treatment parameters with the unit serial number. Otherwise a part that passes at final inspection cannot be linked to the cycle that made it pass.
4. Add the environmental and durability blocks that an in-vehicle display has to survive
Automotive displays sit in a cabin that swings between very hot and very cold, is exposed to sunlight, and is cleaned with ordinary agents. The test plan should therefore contain a durability block that is run on samples rather than on every unit, and that block should be defined before the line starts, not after a complaint. Typical items to define are thermal cycling, humidity exposure, and a cleaning and chemical-resistance check on the outer surface. The plan needs to state the sample quantity per lot, the frequency, and what happens to the in-process material between the qualification sample and the next production run.
A practical point customers often raise with us is the surface finish. When we explain why an equipment housing is stainless rather than painted steel, we describe the difference as ease of cleaning and resistance to repeated wipe-down, not as a guaranteed service life. The same discipline belongs in your plan: describe an acceptance criterion in terms you can actually measure with the instruments you own.
5. Define how the line itself is tested, not only the parts
A test plan for a production line has to test the line. Three items are usually enough to make the equipment performance visible:
- availability, measured as the share of scheduled time the station is actually producing, with stop reasons logged;
- cycle time, measured over a full shift rather than from one good cycle;
- changeover, timed every time a different display size or film stack runs through the same station.
For reference, our 3D surface vacuum bonding machine listing cites availability of at least 98% and a 20-second cycle, both based on a 5.5-inch soft-to-hard test. Quote such figures as the result of a defined test, and require your supplier to state the test conditions behind any number in your acceptance document. Targets carried over from a different panel size are the most common source of dispute at acceptance.
6. Plan for curved and irregular surfaces explicitly
Automotive displays increasingly follow the shape of the dashboard. A flat-panel plan does not transfer directly. For curved work the plan must add checks that only matter in that geometry: film bridging at the edge of an arc, uneven pressure across a curved surface, and registration on a large-size curved-surface lamination process where the alignment reference itself may sit on a non-flat face. If a recipe per product shape is available on the equipment, the plan should say which recipe is expected to be loaded for which part number, and how that is verified.
7. Close the loop: records, rework rules and the acceptance document
Two pages of the plan decide whether the rest is usable. The first is the record sheet: unit serial number, station, recipe, operator, measured values, and any rework the unit went through. The second is the rework rule: what may be re-bonded, how many times, and what must be scrapped. A plan without an explicit rework limit quietly accepts invisible repair history inside automotive parts.
Finish with the acceptance document that ties the plan to the equipment purchase: which test is run, by whom, on which sample set, and what result closes the project. If you want the line specification reviewed together with the test plan, our engineering team can look at both in the same pass through our automated production line solutions service, which avoids the situation where a line is accepted on criteria nobody measured.
