From Sample to Production in 6 Weeks: What to Check at Every Gate

Extrusion line operator checking a running profile during production
An operator checking a running line. On a six-week schedule this is what Gate 2 looks like in practice — dimensions measured on the line, not after the batch.

Abstract: Many projects treat “sample to production” as “get a sample fast, then go mass production”, and then stall for two or three months reworking at the T1 first-article, small-batch trial and mass-production SPC stages. The 6-week schedule below compresses that path into gates with defined deliverables and acceptance actions, and flags the three hidden risks that most often blow the timeline: setup scrap, batch colour drift and bow. Timelines are given as experience ranges; confirm them against your supplier’s takt time and tooling complexity.

Why a schedule beats “get a sample fast”

Extrusion failures rarely happen at the first article itself. They happen in the gap where the sample passed and the batch failed. Without a schedule, the supplier pushes at its own pace: tooling first, then production, then trial, then back to rework the tool when the trial fails. Every loop adds two or three weeks.

A 6-week schedule puts acceptance actions at each gate so that “not good enough” is caught at the earliest, cheapest point.

Week 1: Drawing and tolerance review (Gate 0 · tooling review)

Deliverables: signed drawing, tolerance classification table, tooling-concept sign-off.

What to check:

  • Separate functional tolerances (affect assembly and photometry) from cosmetic tolerances (appearance only) and relax the latter, because an unnecessary tight tolerance directly raises tooling cost and scrap rate;
  • Confirm wall thickness is uniform and the prism can be filled stably, so the tool does not end up theoretically extrudable but practically unfillable;
  • Agree the retained-sample and re-inspection scope here; if you skip it this week, batch consistency becomes unenforceable later.

Week 2: T1 first article + full dimensions + optical measurement (Gate 1)

Deliverables: several T1 parts, full-dimensional inspection report, transmittance / haze / chromaticity data.

What to check:

  • Measure transmittance window, haze and chromaticity, not just dimensions. Many “dimensions OK but light wrong” problems should surface here;
  • Check prism pitch and apex angle; these decide UGR and distribution and cannot be judged by eye;
  • Write the first-article pass criteria on the review sheet, so “looks about right” does not become the acceptance standard.

Weeks 3–4: Small-batch trial run, 200–500 parts (Gate 2)

Deliverables: trial parts, measured scrap rate on the line, per-part assembly-minute count, light-leak / bow log.

What to check:

  • Run the real assembly line, not supplier-hand-picked “showpieces”;
  • Count scrap rate and per-part assembly minutes; accept on data, not feel;
  • This gate specifically catches the “invisible at first article, shows in batch” issues: colour drift, light leak, clip misfit, bow.

Weeks 5–6: Mass-production SPC on first three batches + retained-sample re-inspection (Gate 3)

Deliverables: SPC control charts for the first three mass batches, retained-sample ledger, inspection-cadence table.

What to check:

  • Run statistical process control on wall thickness, transmittance and colour ΔE, and read the spread rather than the mean;
  • Establish a retained-sample re-inspection cadence to make batch consistency enforceable;
  • Only after the first three batches are stable is it truly mass production; later variation is covered by inspection.

Who signs off at each gate

The schedule only holds if each gate has a named owner on both sides. A gate without a signatory turns into a conversation, and conversations do not close.

Gate Your side signs Our side signs Deliverable that closes it
Gate 0 · drawing and tolerance Design engineer Toolmaker Signed drawing with marked criticals and a tooling concept
Gate 1 · first article Design engineer + optics Toolmaker + QC Full-dimensional report plus transmission and haze data
Gate 2 · small batch Quality or incoming inspection Production + QC Sample from the batch, measured against the Gate 1 report
Gate 3 · production Purchasing Production SPC record for the first three batches, retained samples logged

What actually adds a week

  • Waiting for the drawing to be signed. The single largest cause. Nothing runs while a drawing sits in review.
  • Approving a sample by eye. If the acceptance criteria were not written at Gate 0, the first article is debated instead of measured, and the die comes out of the press for a change nobody defined.
  • Re-specifying colour after the first article. A colour change is a material change. Done late, it restarts the trial sequence.
  • Discovering the end cut matters. Cutting to length is a secondary operation with its own tolerance. If it appears at week five, it appears as a delay.

Where the six-week figure comes from, and where it does not apply

Six weeks is an experience range, not a commitment, and it is worth being explicit about what sits inside it. The figure assumes a section inside our normal band — a linear profile up to 400 mm across, walls that can be balanced, one material, one colour — and it starts from a drawing that is already signed. What it covers is die design, die making, first article, correction passes and a production-ready line.

It does not cover four situations. A section that needs a new die concept rather than a variation of an existing one. A project where the optical target has to be measured on a goniometer before the section can be accepted, because photometric rounds add calendar time. A colour that is matched to a physical sample and then has to be re-matched. And any project where the drawing changes after Gate 0 — a revision restarts the sequence from wherever it lands.

The practical advice is to treat the schedule as a floor and plan procurement against it, rather than treating it as a promise and discovering the revision cost later. We would rather tell you in week one that a section will run long than tell you in week six.

Three points where schedules slip

  1. Setup scrap: every batch change or tool change throws off the first dozens of parts as scrap. In small batches this is the largest single cost item, so budget for it in both price and lead time.
  2. Batch colour drift: three consecutive batches from the same tool, five parts each, and measure ΔE to read the batch-to-batch drift rather than a single good batch.
  3. Bow: thin-wall, asymmetric long profiles deform in storage and transit; age-stabilising and partitioned packaging are mandatory, or parts look flat on arrival and twist after mounting.

A reusable 6-week Gantt, in one table

Week Gate Core deliverable Show-stopper
W1 Drawing / tooling review Tolerance class, tooling sign-off Functional tolerances not classified
W2 T1 first-article measurement Full dimensions + optical data Transmittance / distribution fail
W3–W4 Small-batch trial Scrap rate, assembly minutes Scrap rate over threshold
W5–W6 Mass SPC Control charts, sample ledger Batch-to-batch ΔE drift

What the schedule changes

A 6-week plan costs nothing to write and moves the acceptance criteria to the point where a change is still cheap: the drawing review in Week 1. Tooling speed rarely decides lead time. Whether the acceptance standard was fixed early does.

Related guides and products

Scroll to Top