Short answer
Before asking for a laser cutting machine, do not begin with a preferred power rating, brand or table size. First describe the real part mix: which parts are made, their material and thickness, production volumes, edge requirements, and what happens to the part after cutting. Those data make it possible to assess not merely whether a machine cuts metal, but whether the future cell fits the actual production route.
A part mix is not a list of names such as “bracket” or “housing”. It is a set of production facts. A 3 mm structural-steel bracket made in lots of 10,000 and a 1.5 mm stainless enclosure made in lots of 20 create different needs for loading, automation, assist gas, file preparation, quality control and economics. Select the machine around material, geometry, task and flow rather than one power figure.
Why “steel from 1 to 20 mm” is not a technical brief
That statement hides the decisive facts. It does not say whether 20 mm occurs daily or once a quarter, whether the edge goes straight to welding, whether parts contain many small holes, whether non-standard sheets are needed, whether the shop works one shift, or whether cutting is the first step before bending, welding and painting. Without that context, a supplier either needs many follow-up questions or produces a formally correct but weakly comparable offer.
Do not collect only the maximum thickness. The maximum matters, but seldom describes the load. Show the distribution by material and thickness, identify regular series and exceptions, then discuss the base configuration, infrequent work and alternative routes separately.
Build a representative sample, not a perfect archive
There is no need to digitize every historical job. Start with a period that represents normal work, such as the last 6–12 months or a typical production cycle. Include serial parts, common materials, large parts, small parts with many contours, parts with demanding edge requirements and several problematic positions. State seasonality explicitly where it exists.
The most useful input is a package of drawings or DXF files for representative parts, material specifications and a short volume table. Files show geometry, but not whether a part is one-off or contract-critical, nor whether it will be bent, welded, machined or shipped after cutting. Add that operating context.
Data worth collecting
| Data block | Record | Why it matters for the decision |
|---|---|---|
| Material | Grade or at least group: carbon steel, stainless steel, aluminium, galvanized sheet; surface condition | Affects the process window, assist gas, edge requirements and productivity assessment |
| Thickness | Distribution of positions and area by thickness, not only minimum and maximum | Shows what the machine will do most of the time |
| Geometry | Overall size, small holes, narrow bridges, long contours, part density | Indicates format, quality and cycle-time constraints |
| Volume | Parts, sheets or total area per month/year; typical lot size | Separates occasional tasks from the primary load |
| Quality | Critical edge appearance, repeatability, spatter, weld preparation | Prevents unlike requirements being merged into an abstract “high quality” |
| Next operation | Bending, welding, painting, tapping, assembly, shipment | Shows which edge properties and deviations really matter |
| Current route | In-house equipment, subcontracting, manual steps, bottleneck | Provides a baseline for comparing a new route |
| Work organization | Shifts, staffing, loading, floor space, sheet and part logistics | Shows whether the cell can accept the expected flow |
Material and thickness: do not reduce everything to “metal”
For a first questionnaire, material groups may be sufficient if an exact grade is not yet fixed. But “steel” is too broad: structural, stainless and galvanized steel, as well as aluminium, behave differently and create different surface and edge expectations. State if sheets are film-coated, painted, contaminated or stored outdoors. This is not a request to prescribe a cutting recipe; it prevents a demonstration on an ideal sample from being treated as proof for real stock.
Record thickness in two forms: range and frequency. “1 to 16 mm” does not reveal whether most work is 1–3 mm and 16 mm is exceptional. Show the share by thickness group and identify critical positions separately. Do not mix sheet, profiles and tube; they are different equipment classes and require separate analysis.
Part geometry and sheet use
Part area is not the same as complexity. A small item with many holes, slots and thin bridges may require more machine time and control than one simple large contour. Describe typical blank sizes, the largest working part, frequency of small elements, and any genuinely planned marking, engraving or identification operations.
Record the sheet formats actually purchased. A nominal table size does not automatically make every sheet convenient to load, nest and remove. Continue the format analysis in ART-032: 1500×3000 or 2000×6000 working area. Also collect information about costly materials, scrap, non-standard remnants and any requirement to track them; nesting calculations are only as realistic as their geometry and production rules.
Volumes, lots and the real production rhythm
Record annual volume separately from lot behaviour. Annual area is a useful base, but it does not show whether orders arrive evenly. Ask how often material changes during a week, how many rush orders occur, whether nesting programs repeat, and how quickly small lots must be switched. The same annual sheet area can require very different automation and work organization.
Do not turn a forecast into a guarantee. Future contracts, new products and expected growth belong in a distinct “development scenario” section. The base assessment then rests on facts, while investment scenarios retain their stated assumptions.
Edge quality and downstream operations
“Quality cut” and “clean edge” are too vague. One part may accept a process edge before welding or painting; another may require a visible surface; a third must fit a precise assembly. State what is checked after cutting: size, burrs, distortion, protective-film condition, fit in an assembly, or fitness for the next operation.
Where reference parts exist, send them for a test with acceptance criteria. Do not ask for an “ideal quality” promise without a specified material, geometry and assessment rule. Treat the test as data for discussing configuration and process, not as a universal guarantee for the whole part mix.
Site data and the production route
Even a well-selected machine does not work in isolation. Early assessment needs floor area, delivery access, the sheet feed path, storage and part-removal zones, electrical infrastructure, compressed air, extraction, gases and responsible staff. This is not an instruction to design utilities independently: final requirements come from the specific manufacturer documentation and the site project.
Describe the current route as well: what is cut internally, what is subcontracted, where queues form, and which parts delay assembly. The purpose is not to blame current suppliers. It is to identify whether the new cell should provide reserve capacity, bring regular work in-house, reduce internal movement or enable a new product.
A practical request format
A useful first request contains a short letter, one part-mix table and representative files. Describe the production, the assessment goal, the data period and attachments. Ask the supplier not to name “the best machine”, but to explain the assumptions behind the recommendation, the decisive positions and missing data.
This also disciplines the comparison of offers. When suppliers see the same data package and answer the same criteria, it is easier to separate actual configuration from general wording. See ART-038: how to compare two commercial laser-machine offers.
Separate production fact from assumption
Add a “data source” column to the part-mix table. A fact can be supported by ERP data, material invoices, production plans, part files or shop-floor measurement. A future contract, expected range expansion or “there will probably be more stainless work” is an assumption. Both are useful, but they must not be merged in one total.
For an investment decision, an accountable offer should show two scenarios: what it covers for the current part mix and which growth assumption it uses. That makes the cost and logic of each scenario visible and makes a later review possible when the order mix changes.
Working with confidential drawings
Confidential drawings are not a reason to replace technical data with abstractions. You can share a limited sample under an agreed process, remove non-essential identifiers, or first agree the data required for assessment. Geometry, material, thickness and acceptance criteria still cannot be fully replaced by words. Anyone receiving files must operate under the agreed confidentiality terms.
For a first discussion, part types and the structure of the part mix can be enough. A final decision or test needs a more precise package. That is a normal sequence: form a hypothesis first, then test it on real samples. Sending every archive without structure makes analysis harder, not more reliable.
Who should gather the data
Do not assign the questionnaire only to purchasing or only to a technologist. Purchasing knows material formats, prices and supply; production knows actual bottlenecks and rhythm; a designer or technologist knows part requirements; finance knows scenario limits; and the site owner knows infrastructure. They do not need a multi-volume project. One agreed file with clear columns is usually more valuable than several incompatible presentations.
Agree in advance who accepts test results. If a welder or quality controller decides whether an edge is suitable, that role must see the samples and criteria. Otherwise the technical discussion returns after purchase, when changing the configuration is much harder.
Common mistakes in the source table
Do not total parts by unit count when their area and contour complexity differ radically. Do not show maximum values without each group’s share. Do not place current volumes and future assumptions in one column. Mark parts critical to delivery, quality or contract. State the next operation; otherwise cutting can be assessed without understanding what the edge is for.
Copying another company’s request is equally risky. Shops with the same materials may have different lots, queues, staffing and logistics. A good request is short but specific to the real production, not a collection of catalogue phrases.
A short validation before sending the package
Before sending, bring together the request owner, a production representative and a technologist or designer for a short review. Take the five highest-volume parts, two geometrically difficult parts and one unusual part. Check the table against each file or drawing: material, thickness, overall size, annual volume, next operation, and whether the entry is fact or forecast. Put disagreements in a note instead of hiding them in an average.
Then compare the totals with purchasing, ensure serial parts were not lost under identical names, and make sure one part was not counted twice in different revisions. This does not make the table perfect; it removes obvious errors before it is used to compare equipment, format or configuration.
Checklist before sending a request
- [ ] A list of 20–50 representative parts, or another agreed sample, is available.
- [ ] Each main group has material, thickness, overall dimensions and annual/monthly volume.
- [ ] Regular, critical and rare positions are identified.
- [ ] DXF files or drawings are available where possible and have a clear revision.
- [ ] The downstream route after cutting is described.
- [ ] Acceptance criteria exist for critical parts.
- [ ] Actual data are separated from growth forecasts.
- [ ] Sheet formats, shift organization and basic site constraints are described.
- [ ] The request contains specific supplier questions, not only a request to “calculate a machine”.
What cannot be determined without data
Without a real part mix, it is not possible to determine reliably the needed power, table format, automation level, shift throughput, gas consumption, payback period or final configuration. One demonstration sample cannot honestly promise quality on every material. Site readiness cannot be assessed from room photographs alone. Those decisions need part, material, flow, infrastructure and manufacturer-specific data.
Select a laser-machine configuration