The cheapest way to reduce the risk in a UV printer order is not a longer warranty or a bigger discount. It is a sample test on your own material, run under recorded settings, with acceptance criteria you wrote before the print was made.

This guide explains how a UV printer sample test should be structured before you commit to a configuration: what the test is meant to prove, how to choose a sample that represents production, what to send with your artwork, which acceptance criteria to fix in advance, and what a test report must contain so the result can actually be reproduced on the machine you buy.

What a UV printer sample test proves

A sample test proves that one combination of substrate, preparation, ink and settings can meet criteria you defined. It does not prove throughput, repeatability across a shift, or behaviour on a different batch of the same material - those come later, with a pilot and a lot-change check.

Understanding that boundary stops the test from being asked to do too much. The test answers a specific question: can this material be printed to an acceptable standard with this process, under conditions that can be recorded and repeated? If the answer is yes, you have a process starting point and an approved sample to compare against. If the answer is conditional, you have learned which variable is unresolved.

Structure the test around the decision you are facing. If you are choosing between two machine classes, run the same artwork and substrate on both and compare adhesion, edge definition and handling time. If you are validating a substrate you have never printed, run a small matrix of preparation variants instead of a single perfect sample. The test is only as useful as the question it was built to answer.

How do you choose the sample that represents production?

Use production material, cut to the size you print.

Use material from the batch you will actually run; a sample from a different supplier, coating or storage condition answers a question you are not asking.

Send enough pieces to allow repeat prints: three to five of the same item is a reasonable starting point, because one acceptable print may be a coincidence. Where your product is cylindrical or tapered, match the diameter, taper and length of production parts rather than sending a generic offcut; fixture behaviour changes with geometry. Where the surface matters - textured MDF, coated metal, film-protected acrylic - send the material as it arrives, film and all, so preparation can be tested as it will occur on your line.

Say where the part will be used, too. A promotional product that must survive handling, a panel exposed to sunlight, and an industrial control faceplate that must resist wiping are three different acceptance problems on the same substrate, and the ink and layer strategy changes with the requirement.

Which artwork and layer requirements should you send?

Send the production file and the layer order you need.

State white ink for opacity, varnish for protection or effect, and any raised detail. Without it the test answers a colour question and misses the structure question.

Files should be provided in the format your production workflow will use, with the same resolution and spot channels you intend to print. If the job includes fine text, gradients or tight registration around cut-outs, say so; those features decide the print mode and the pass count, and they are often what fails first when a substrate is difficult.

State the colour reference as well. A Pantone reference, a previously approved physical sample or a measured value gives the operator a target; "the blue from our logo file" does not, because the printed result depends on the substrate and the ink. Where the intended market requires specific limits - for example substances restricted under REACH in the European Union - say so before the test, not after, because it may change the ink selected.

Acceptance criteria should be defined before the test

Write the criteria before the sample is printed, and make each one measurable. Criteria written afterwards tend to match whatever came out of the machine, which is not evidence of anything.

A practical set for decorated rigid parts covers adhesion, appearance, durability and handling. For adhesion, specify a method rather than an impression: the cross-cut tape test described in ASTM D3359 is a widely used reference for coating adhesion on test panels. For appearance, specify the reference sample, the acceptable colour difference and the edge registration tolerance. For durability, specify the cleaning chemical, the wipe count and the abrasion test you will apply. For handling, specify when the part can be stacked and packed.

Where the parts will be inspected in batches, define the sampling rule in advance. The acceptance-sampling approach set out in the NIST engineering handbook gives a defensible basis for deciding how many pieces to examine and how to interpret the count, so the decision does not depend on who is standing at the bench.

What the test report should record

The report is the real deliverable, because it carries the settings that make the result repeatable. Photographs alone cannot be reproduced on a machine in another country.

Require the material and batch identification, the surface preparation actually applied, the ink set and layer order, the print mode, pass count and resolution, the curing arrangement, the tests performed and their results, and the settings that produced the approved output. Ask for images of the finished part and of the adhesion test area, and for a note of anything that was marginal. Where appearance has to be agreed with a customer, published print research from bodies such as Fogra is a useful reference for how viewing and measurement conditions affect the result being approved.

Component-level detail helps too. Where the machine uses standard industrial piezo printheads, the head model belongs in the record, because the technical documentation published by the component maker - Ricoh's industrial inkjet head documentation is one example - describes how head geometry affects drop placement, which is part of why a setting that passes on one head generation needs review on another.

How do you read the result: pass, conditional pass or fail?

Against the criteria you wrote, not the best-looking area.

A pass means every criterion was met; a conditional pass means a named variable is unresolved and needs a defined next step.

Push back on a conditional pass that arrives without a next action. "Adhesion was acceptable but the substrate is difficult" is not a result. A usable conditional pass names the variable - surface energy, coating type, cure settings, layer order - and defines the test that will resolve it, who runs it and what the outcome has to be for the process to be approved.

Keep the approved sample. It becomes the physical reference for the first-article approval on the delivered machine, and later for incoming material checks, because a standard held in a drawer is what lets a different operator in a different month judge the same way.

From sample to pilot batch

The sample proves a setting; the pilot proves the setting survives a shift. After an approved sample, run a pilot batch with the real jig, the real file handling and the real inspection routine before you commit to volume.

Define what the pilot has to demonstrate: accepted parts per hour, the rejection reason that occurs most, the time taken for setup and changeover between products, and whether the operator can hold the approved standard without supervision. Record the settings at the start and end of the shift, because a process that drifts is more expensive than a process that is merely slower.

Practical markers to confirm at this stage are the sample policy, the training included with the machine and the spare-parts availability period for the model. AndresJet prints the first three samples free of charge and includes three days of on-site training; the same questions are worth asking of any supplier, because they determine how quickly a pilot becomes routine production. Industry bodies such as Printing United Alliance publish wider guidance on print production practice, which is useful when a pilot result has to be explained to someone outside the shop.

Mistakes that invalidate a sample test

Most invalid tests fail for administrative reasons rather than technical ones. The print may be excellent and the evidence useless, which is the worst combination because it feels like progress.

• Testing a substrate that is not the production material, or not from the batch you will use.
• Sending artwork without layer or durability instructions, then judging a result against unstated expectations.
• Judging by eye with no acceptance criteria and no recorded settings.
• Skipping surface preparation because the sample "looks clean", so the test measures a process you will not run.
• Accepting an adhesion claim without a defined method; a tape test performed differently on each side is not a comparison.
• Treating a single acceptable print as proof of repeatability.
• Leaving the approved sample unlabelled, so no one can tell which settings produced it three months later.

Avoiding those seven mistakes costs almost nothing and is the difference between a test that supports a purchase decision and a print that looked good once.

The checklist to send with your sample

Send the checklist with the samples, not after the results arrive. It takes ten minutes to complete and it removes the most common cause of a delayed or ambiguous test: a supplier trying to guess what "acceptable" means for your product.

Include the substrate and batch, the size and quantity sent, the artwork with layer instructions, the colour reference, the durability requirement, the acceptance criteria with methods, the destination market, and the date you need the result by. Add who will judge the test on your side, and whether you want the printed parts returned. With that page attached, the test becomes a documented comparison rather than a favour - and the report you get back is one you can attach to the purchase file.