How to distinguish a showcase laser-cutting demonstration from a test that reproduces your production ART-044

Why an attractive cut is not yet a production result

A supplier will usually have a well-adjusted demonstration machine, prepared sheet, familiar settings, and an operator who has performed the same operation many times. This is a normal part of presenting a technology. The problem arises when one clean contour is used to conclude something about the entire future process.

In serial or small-batch production, the result is determined by more than the beam. It is affected by the grade and condition of the metal, thickness, protective film, contour geometry, the number of lead-ins, small holes and narrow tabs, cutting sequence, gas, changeovers, unloading, and stability throughout a shift. A demonstration cut may honestly show the capability of the technology while failing to check those variables.

It is therefore useful to separate three levels of evidence. The first is a capability review: the machine cuts typical samples. The second is an applied test: it cuts your parts or agreed equivalents within a specified range. The third is an acceptance test with written criteria, samples, conditions, and a result. An investment decision normally needs at least the second level; delivery and start-up need the third.

What task the test must solve

Do not set the task of “checking the machine in general.” Formulate the decision to be made after the test. For example: whether the proposed configuration gives an acceptable edge on a steel part of a specified thickness; whether holes and slots from your product mix pass; whether the cycle becomes unacceptable because of a large number of piercings; whether it is possible to move frequently between thin and thick material without risk.

The test should not turn into concealed process development at the supplier’s expense. Its purpose is to collect data for a decision, not to prove that every part must be cut unconditionally on every laser. If a part is technologically difficult, an appropriate outcome may be a recommendation to change the design, the operation sequence, or the configuration.

| What is compared | Showcase demonstration cut | Production test | |---|---|---| | Input samples | Convenient for demonstration | Your parts or agreed equivalents | | Purpose | Show the technology | Check a specific solution | | Conditions | Defined by the demonstration site | Recorded before the test starts | | Geometry | Often one or two striking contours | Typical and risk-bearing product features | | Time | One successful run | Cycle, preparation, changeover, repeatability | | Conclusion | Indicative impression | Decision with limitations and next actions |

How to choose representative parts

The most common mistake is to submit the largest or most attractive part for testing. It may be important, but it does not always account for most hours, defects, or cost. For a test, it is better to select a small “basket of parts” containing three to five positions. One should be typical by volume, one critical by quality, one complex by geometry, and, where necessary, one more with another thickness or material.

Along with the files, provide an explanation: material and thickness, critical surfaces, unacceptable defects, the required subsequent operation, approximate volume, and what counts as an acceptable result. There is no need to state exact tolerances immediately if the drawing does not define them. It is better to mark separately: “this edge will be painted,” “this hole must receive a fastener,” “this part will be bent,” or “this slot is important for assembly.”

To compare cycle time, it is not enough to record the time shown by the controller during cutting. Machine time must be separated from sheet preparation, positioning, material changes, unloading, and manual rework. On one demonstration sheet these operations can seem insignificant, but in the real route they are what change throughput.

What to record before starting

An agreed list of checks makes the test fair to both parties. Before it begins, record the file version, sample material, thickness, surface condition, machine configuration, gas type, quality-assessment principle, and those responsible for acceptance. It is not necessary to require the supplier to disclose every process parameter. More important is that after the test it is clear under which conditions the result was obtained and whether they can be reproduced.

It is useful to agree which indicators are observations and which are decision criteria. These may include whether the contour is completed, burr visibility, stability of narrow features, need for manual deburring, repeatability across two or three runs, cycle time, and extent of manual intervention. If measurement is needed, determine its method and instrument in advance. No categorical conclusion should be made from a measurement if the inspection method was not agreed.

When a production test is not yet FAT

The term FAT is often used for any test before delivery, but it does not itself turn a demonstration into an acceptance procedure. FAT makes sense only when the project documents have defined in advance the subject of the check, specification version, sample, acceptance criteria, participants, method for recording the result, and the procedure if a criterion is not met. Without this, even a careful cutting of several parts remains an applied test for selecting a configuration.

This distinction benefits both parties. The buyer does not mistakenly transfer a demonstration result to installation, training, and daily operation. The supplier does not receive an undescribed obligation to ensure a result for parts that were not provided or had no agreed criterion. In a production-test protocol, state its status directly: “input data for a configuration decision” or “preliminary product-mix check.” If the test later becomes the basis for FAT, agree its expanded plan separately rather than merely attaching old photographs to the contract.

ISO 9001 is relevant here only as a general framework for managing documented checks and nonconformities. It does not establish cutting parameters, allowable geometry, or the number of repeats for a specific laser machine. Such criteria come from the drawing, specification, documentation of the particular system, and the parties’ agreements.

Verification algorithm: from file to protocol

1. Select parts by frequency, quality risk, and complexity, not only by size. 2. Check that the file, drawing revision, and material are identified identically by all participants. 3. Record the critical features: holes, slots, edge, flatness, and subsequent assembly. 4. Agree the configuration scope and test limits: what is checked and what is not. 5. Perform not one “successful take,” but the agreed number of repeats for risk-bearing parts. 6. Inspect parts before cleaning and after required standard rework; do not hide additional labour. 7. Record observations, deviations, limitations, and the next check in a short protocol.

Which mistakes distort the conclusion

Confusing maximum thickness with the working product mix. The fact that a machine physically cuts a thick coupon does not mean that it provides the required rate, edge, and cost on your parts.

Using only ideal material. If production includes sheet with film, different batches, or surface variations, note it. There is no need to deliberately prepare poor material, but neither should conclusions suggest that the demonstration sample represents all deliveries.

Not checking the subsequent operation. An edge that looks clean may behave differently in bending, welding, painting, or assembly. If that operation is critical, add at least a simplified check of it to the test.

Comparing only cutting seconds. For batches with frequent job changes, preparation, file entry, setup, and unloading logic also matter.

Turning the test into a promotional argument. It is better to record a limitation than to force a technology to appear universal. An honest “this needs separate checking” is more valuable than an unconfirmed promise.

Working scenario: how to run a test in one day

Imagine a manufacturer of enclosure products. It has thin panels with holes, several parts made of thicker sheet, and one item that goes directly to bending after cutting. If the test contains only a large panel, the result says almost nothing about bottlenecks. It is much more useful to provide a set of four files: a standard panel, a perforated panel, a bracket with critical holes, and a part for bending.

Before the test, the manufacturer defines what matters for each position. For the panel, it is the absence of excessive manual deburring; for perforation, the stability of small features; for the bracket, holes passing and accuracy of subsequent assembly; for the bending part, edge condition after the normal operation. The supplier is not obliged to guarantee a process result without a drawing and agreed conditions. It can, however, show which constraints it sees and cut the parts so that both parties assess the same things.

After cutting, do not rush to a conclusion. Number the parts, inspect them, photograph critical areas, and, where needed, pass them to the next operation. The protocol needs only short, verifiable records: “contour completed,” “the edge has an area requiring separate assessment,” “additional checking is needed after bending,” or “time is machine time and does not include sorting.” This document is not a contract and does not replace acceptance, but it prevents facts from later being substituted by recollection.

Questions to ask the supplier

The questions should concern reproducibility, not secret parameters. What exact system configuration was used for the test? Does it differ from the offered configuration? What material was used, and is its manufacturer or certificate known? Which file features can affect the result? What shown work is typical for the system, and what requires separate process verification? Which conditions at the customer’s site can change the result?

Also ask to show the delivery boundaries. Does the proposal include commissioning, training, consultation on site preparation, gas and extraction requirements, and test acceptance? If any part remains the company’s responsibility, name it before the contract is signed. This is not distrust of the supplier; it is a normal division of responsibility.

How to interpret a negative result

An unsuccessful or partly successful test does not always mean the equipment is poor. It may show that the part needs another technology, the material has particular characteristics, the configuration does not suit the task, or the quality criterion was not formulated. The worst response is to hide the result or repeat attempts until one attractive sample appears. The better response is to record where the deviation occurred, which causes are possible, and which next action is agreed: clarifying the drawing, another material, another configuration, a separate test, or an alternative route.

How to prepare the conclusion after the test

The conclusion should be short but separate fact from assumption. The “confirmed” block records only what was seen or measured: which parts were cut, under which conditions, what was observed at critical features, and whether the subsequent operation passed. The “requires verification” block keeps questions for which there was no sample, time, instrument, or agreed criterion. The “next decision” block names one responsible action: change the configuration, conduct FAT, collect additional data, or proceed to agree the delivery.

This form protects both parties. The buyer does not receive a vague “everything is fine,” and the supplier is not held responsible for properties of a part that were not provided or checked. If a test is conducted before the contract, it is especially important not to present the document as a guarantee. If it is part of acceptance, the criteria, measuring means, sample, and procedure on deviation must be separately defined in the project documents.

What cannot be determined without data

Without specific drawings, materials, their origin, batch volume, subsequent-operation requirements, and the actual configuration, it is impossible to state honestly the productivity, cost, consistent edge quality, or suitability for every part. A test at one site also does not replace verification of installation, electrical supply, extraction, gas systems, staff training, and acceptance after delivery.

Checklist before a laser-machine test

• [ ] Three to five parts representing the real product mix have been selected.
• [ ] Material, thickness, file revision, and critical features are stated for every part.
• [ ] It is agreed what counts as the criterion: geometry, edge, repeatability, cycle, or subsequent operation.
• [ ] The configuration and conditions under which the result is obtained are recorded.
• [ ] Machine time is not mixed with preparation, unloading, and manual rework.
• [ ] Deviations and limitations are entered in the protocol rather than left “verbal.”
• [ ] The next decision is defined after the test: additional check, configuration clarification, or acceptance scenario.

A production test need not be long or bureaucratic. Its strength is that it asks real questions of real parts and leaves a verifiable trace. Then the demonstration becomes not a spectacle, but part of a well-founded choice.

Minimum protocol format

The protocol does not require a complex form. Six fields are enough: part name and revision; sample material and thickness; system composition used for the test; criteria checked; actual observations; unresolved questions and next action. Where possible, add photographs of the same areas that were assessed, not only the best overall view of a part. This lets another specialist understand the context without oral explanation.

It is particularly useful to state every manual intervention separately. If an operator cleaned a part, changed a file, repeated an operation, or used a special placement method, this does not invalidate the test. Without that record, however, the buyer cannot see the actual process boundary. Similarly, if a sample passed only after the input data were changed, record both versions. The test then becomes material for a joint decision rather than evidence of one side’s position.

For a serious purchase, return to the protocol when agreeing acceptance. There is no need to copy demonstration conditions unchanged: the buyer’s site has different people, infrastructure, and organisation. The initial test nevertheless helps formulate the questions: which parts to repeat, what is acceptable, who assesses them, and when the result requires corrective action. This link between testing and start-up reduces the risk that important expectations remain only in correspondence.

Role of the technical reviewer

If the company has a process engineer or designer, involve them not to search for faults in the demonstration but to translate product requirements into clear criteria. They can explain which dimensions, holes, surfaces, or subsequent operations really matter. At the same time, a technical reviewer should not replace the entire decision process: an investment conclusion also depends on work flow, infrastructure, service route, and financial assumptions. Separating the roles helps make the test useful rather than unnecessarily complicated.

Need help selecting equipment?