A roll-to-roll magnetron sputtering machine is rarely bought for its nameplate. It is bought to put a consistent conductive or transparent-conductive coating onto a moving web, repeatedly, with the next order already waiting behind it. The difference between a machine that ships on paper and one that runs on your floor is decided during commissioning — the stage where the supplier proves, in front of you, that the line can do what your application needs. The checks below are the ones worth insisting on before sign-off, written from the buyer's side of the table.
Before the machine is installed: agree on the web, not the vacuum
Most commissioning disputes start before the machine arrives, because the acceptance criteria were written around vacuum hardware instead of the material being coated. Fix the web definition first: substrate type and thickness, web width, the tension window the web can tolerate from unwind to rewind, the line speed range you intend to run, and the coating you expect on it. A roll-to-roll magnetron sputtering coating machine specified around the wrong web — too thin, too wide, or pulled too tight — will pass a vacuum test and still fail in production.
Write these numbers into the purchase documents as measurable acceptance conditions, with the test method attached. "Uniform coating" is not an acceptance condition; a sheet resistance measured at defined positions across the web, at a defined line speed, is.
Check 1: coating uniformity across the full web width
Sputtering uniformity is the first thing to verify, because it is set by source-to-web geometry, shutter and target conditioning, and the way the web passes the target. Ask the supplier to demonstrate it, not promise it: run a representative length of your own substrate, then measure — don't eyeball the colour.
- Measure across the web at several machine-direction positions, not one point at the centre of one sample.
- Repeat the measurement at low, nominal, and high line speed. Uniformity that holds at one speed may drift at another, and it is at speed changes that a line reveals its real behaviour.
- Check the edges. Edge thinning or edge build-up matters for downstream cutting and device areas, and it is often where a source layout shows its limits.
Agree in advance what variation you will accept across the width and along the length, and who measures with which instrument. Two parties holding two different meters will never agree on pass or fail.
Check 2: web handling and tension, at the speeds you will run
A coating defect that repeats every metre usually has a mechanical cause, not a plasma cause. During commissioning, watch the web itself: how it threads, how it behaves through the winding zones, and how it responds when the line accelerates and decelerates.
- Run the web through the full tension range and look for stretching, waviness, curl, and wrinkles — record which of these appear at which tension.
- Inspect the web for marks from rollers, guide edges, and vacuum-chamber hardware. Roller marks become permanent once coated.
- Verify splice behaviour and the procedure for a web break: how the line stops, how it re-threads, and what happens to the roll that breaks mid-coat.
- Confirm that the unwind and rewind units hold stable tension over the full roll length, not just at the start of a roll when the diameter is at its nominal value.
Tension faults are the cheapest thing to fix at commissioning and the most expensive thing to discover in production, when the reject roll is already coated and the next job is already scheduled.
Check 3: process repeatability from run to run
A single good roll proves the machine can work; a series of good rolls proves the process is controlled. Build repeatability into the commissioning test rather than leaving it to the production schedule.
- Run several consecutive rolls with the same recipe and compare results roll to roll.
- Repeat the run after a full vent and pump-down cycle, so the test includes the recovery you will actually live with between jobs.
- Ask for the recipe structure to be visible and documented: what is logged, what can be changed by an operator, and what is locked to the manufacturer.
Target conditioning, gas-flow stability, and pumping behaviour all shift between runs. Commissioning is the moment to establish, with the engineer present, how long the machine needs to stabilise and what its first-off-the-roll output will look like after a vent. If you learn this from your own operators months later, you have paid for the lesson twice.
Check 4: uptime, service access, and consumables
Sputtering sources, traps, and vacuum hardware need routine attention, and how much attention they need depends on the mechanical layout you agreed to buy. Walk the machine with the supplier's engineer and check what maintenance actually requires.
- Can targets be changed and the chamber re-prepared by your own staff, using tools you were given and procedures you were handed?
- Is there clear physical access to the parts that wear — no need to dismantle a housing to reach a feedthrough or a sensor?
- Are pumps, gauges, and power supplies standard parts you can source locally, or items that only come from the machine builder?
- Is the consumables and spare-parts list written down with the expected service interval, and priced in the same currency you budget in?
A line that looks excellent on a demonstration day and impossible to service on a Tuesday night in your own plant is a bad purchase. Service access is a commissioning item, not a post-sale surprise.
Check 5: control system, data, and training
The interface is where your team lives for the life of the machine, and it should be checked as carefully as the vacuum system. During commissioning, confirm that the machine tells you what your engineers need to know.
- Recipe management: can a proven process be recalled without re-tuning, and are changes recorded?
- Fault reporting: does an alarm name the subsystem and point to the next step, or does it only stop the line?
- Run data: is the history available for the roll you are investigating, days after it ran?
- Language and documentation: are the interface and the manuals in the language your operators actually work in?
Training should happen on the machine you bought, on the substrate you run, with your own engineers present. Ask them to run a full job end to end without supplier help before you sign — a supervised dry run is a far better acceptance test than a demonstration.
Check 6: the paperwork that closes the deal
Commissioning ends with documents, and the documents are what protect you afterwards. Before releasing the final payment, expect a package that records the results rather than describing the equipment: measured uniformity data from your substrate, the tension and speed ranges demonstrated, the repeatability runs, the as-built configuration list, the spare-parts and consumables list, the warranty terms with what they do and do not cover, and the acceptance sheet signed by both parties with any outstanding item named and dated.
If a check is left open, write down who closes it and by when. An unsigned "to be confirmed" is the start of next year's argument.
Where a build partner changes the picture
Roll-to-roll sputtering sits inside a wider set of thin-film processes, and a line rarely stands alone: coating, handling, and the process steps before and after it have to be planned together. Shenzhen PowerDe Automation Precision Equipment builds this class of thin-film solar equipment alongside related process tools, and its engineering-services pages cover integrated engineering team services, equipment manufacturing automation services, and steam cleaning equipment product development work — services that matter when the question is not only which machine to buy but how the process around it will be put together.
Send your web definition, target coating, and speed range to PowerDe's contact page and ask what commissioning tests the company is prepared to sign for in writing. A supplier who agrees to measurable acceptance conditions before the machine ships is making a statement about the checks above — and about how the line will behave in your plant, not in their demonstration room.
