A seven-layer membrane-electrode hot press usually arrives at your plant after weeks of assembly and dry running at the builder's site. The factory acceptance test is the last moment when a problem can be corrected with a screwdriver rather than a freight container. What decides whether that visit protects you is not the test itself but the acceptance plan written months earlier, while the contract is still open.
A plan that works does three things. It checks the machine against contract data rather than marketing copy. It runs the machine with your materials, under conditions close to your production floor. And it states, in advance, what counts as a pass, what counts as a hold, and what can be carried onto a punch list and fixed after delivery.
Write the plan against contract data, not brochure data
Every acceptance item should trace back to a document: the purchase order, the technical agreement, and the drawings annexed to both. A figure that appears only in a datasheet is difficult to enforce once the machine is on the loading dock.
Take the published specification for PowerDe's membrane electrode hot press bonding machine as an example: fitting error ≤0.2 mm, working temperature ≤150 °C, 2 t press pressure, and a production cycle of at least 1.5 pieces per minute. Each of those numbers needs a test method, a measurement position, and an acceptance tolerance written next to it. Without them, the figure is a claim rather than a requirement.
The plan should also state the conditions under which each measurement is taken: ambient temperature, compressed air dew point, power supply, and the material thickness being pressed. A result recorded under unspecified conditions cannot be repeated at your plant, which turns a passed test into an argument.
Check the frame and the platens before anything moves
Start with the static machine. Level and platen parallelism are recorded before heating, and the footprint, service clearances, and lifting points are compared with the shipping drawings. This is the stage at which a missing handrail, a non-standard power connector, or a machine that will not pass through your door is cheapest to fix.
Prove alignment and fitting accuracy across the working area
On a multi-layer press, fitting error is the specification most worth verifying first. The machine has to register several layers — membrane, electrode layers, gaskets and support frames — into one bonded stack, and a small offset at the feeding station becomes a rejected stack at the press.
Specify how registration between the feed station and the platens will be checked, how the fitting error will be measured on layered test coupons, and how many repeated measurements are required. One measurement in the centre of the platen proves very little. Ask for several positions across the working area, and for a second set of readings after the machine has run long enough to reach thermal equilibrium.
Test the thermal system where the product sits
Platen temperature uniformity should be mapped at the production setpoint, not at idle. State the number and placement of thermocouples, the soak time before recording, and the maximum deviation from setpoint that you will accept. Cover ramp rate and stability over a continuous run, and record how the control system responds when a cold batch of material is loaded — that recovery behaviour is what an operator actually feels during a shift.
If the same line includes a membrane electrode vacuum hot pressing machine, add vacuum pull-down time and a leak-rate hold to its own acceptance plan. Reaching the specified vacuum once and holding it for an hour are two different results.
Verify pressure accuracy and platen parallelism together
The force the platen applies to a layered stack is what determines bond quality, so the two cannot be tested separately. Calibrate the load cell against a reference, measure platen parallelism across the working area, and demonstrate how the machine handles thickness variation from one stack to the next. Whether the press controls by force or by stroke, and how it compensates during the dwell, should be observed running rather than assumed from a drawing.
Measure throughput with real part flow
A cycle time obtained by dry-cycling empty platens tells you almost nothing. Define one cycle precisely — load, align, press, unload, and any indexing between those steps — then require a continuous run long enough to expose handling faults and thermal drift. State how many interruptions are permitted and how stoppages are recorded, so that the reported rate can be compared with the rate you will schedule against.
Run the machine with your own materials
This is the part of a factory acceptance test that most often gets shortened for time, and it is the part that protects you most. Send your membrane, electrode and support layers to the builder in advance, and require the press to produce bonded stacks from them. Inspect the output for wrinkles, bubbles, edge lifting, and layer offset, using an inspection method agreed before the test week rather than negotiated during it.
Fix the sample quantity and the pass criteria in the plan. A handful of good samples from a machine that has been adjusted all morning is not evidence of a stable process; a documented run, with the parameters that produced it, is.
Check controls, recipes and data logging
Ask for a demonstration of a full production recipe, including recipe changeover and the interlocks that stop a cycle when a layer is missing or misaligned. Inject one fault deliberately and watch how the machine and the alarm list respond. Then export the log for that run. If traceability matters in your line, the export format and the fields recorded should be confirmed here, because retrofitting data logging after delivery is rarely straightforward.
Confirm safety, guarding and your supply conditions
Guards and interlocked access, emergency stop placement, and the surface temperature an operator can reach during a normal cycle all belong in the plan. So does a comparison between the machine's utility requirements and what your factory can actually supply — power, compressed air, floor loading, and the height of the door it has to leave through.
List the documentation you expect to receive
The acceptance plan should name the deliverables: as-built mechanical and electrical drawings, the spare parts and consumables list, calibration certificates for the instruments used during the test, and the signed acceptance report with raw data attached. A report that records only pass or fail, without the measurements behind it, is of little use when the same item has to be re-checked after delivery.
Define the rules of the test itself
State who attends and whether remote witness is accepted. Describe how a hold point works, and how findings are classified: ship-stopper, or correction after delivery with an agreed date. Set the retest rule for any item that fails, and link the outcome to the payment milestone and the delivery date. Name the people on both sides who are authorised to sign. These clauses cost nothing to write and settle most disputes before they start.
| Item | What to verify | Evidence to keep |
|---|---|---|
| Contract baseline | Every figure traced to the technical agreement and annexed drawings | Signed specification list |
| Frame and platens | Level, parallelism, footprint and lifting points before heating | Measurement sheet |
| Alignment and fitting | Registration across the working area, repeated hot and cold | Coupon readings with positions marked |
| Thermal mapping | Uniformity at setpoint, ramp rate, recovery after a cold load | Thermocouple log export |
| Pressure and parallelism | Load cell calibration, force distribution, dwell behaviour | Calibration record and force curve |
| Cycle run | Continuous run with defined cycle and permitted interruptions | Counted run log with stoppages |
| Buyer-material trial | Bonded stacks from your own layers, against agreed criteria | Inspection report and retained samples |
| Controls and logging | Recipe changeover, injected fault, exported traceability record | Alarm list and data export |
| Safety and utilities | Guarding, stops, touchable surface temperatures, supply match | Checklist signed by both parties |
| Documentation | As-built drawings, spares, instrument calibration certificates | Document handover list |
Agree the plan before the contract is signed
An acceptance plan is a negotiation document as much as a technical one, and the buyer who writes it first sets the terms. If you are specifying a hot press for a membrane-electrode line, ask the builder to review your plan at the same time you review their machine drawings — while both sides still have room to adjust.
PowerDe builds thin-film solar equipment for membrane-electrode, perovskite and related processes, and supports specification, testing and customisation through its engineering services. Send your acceptance requirements with your enquiry through the contact page.
