Tooling Is Just a Deposit: What Really Costs on LED Optical Extrusion Is the Supplier Switch

1. Why optical parts can’t be compared by “price per meter”

Two PC extruded covers may differ from tens of grams to over a kilogram per meter — behind that are four completely different accounts: material, tooling, tolerance, and yield. A “quote without drawing” essentially means the supplier is guessing for you, and the guess will always favor them.

A quote usually hides five variables:

1. Weight: the main engine of material cost, and the source of rigidity and assembly feel;
2. Formulation: flame rating, UV resistance, diffuser system — often more price-defining than the base resin;
3. Tooling: the more complex the structure (multi-cavity, micro-structure, asymmetric snap), the heavier the amortization;
4. Cut-to-length, packaging, shipping: extruded parts are bulky and light, so freight is often an invisible cost tier;
5. Yield and scrap allocation: almost no supplier writes the scrap rate explicitly, but it is always in the unit price.

So the correct first-round posture is not price comparison, but aligning assumptions: have all suppliers quote against the same drawing, the same weight target, and the same compliance set. Only then is the quote comparable.

2. Four hidden bills of switching suppliers

1. Time. Re-development usually needs two to three sampling rounds, each plus logistics is two to four weeks. This bill is most expensive when a project window is tight, yet invisible in Excel.

2. Optical drift. The same cross-section, made by different factories with different extrusion process and diffuser dispersion, will drift in transmittance and haze. The luminaire side must re-do photometry and customer approvals — money spent far exceeds the unit-price saved.

3. Assembly. Snap dimensions, straightness, twist usually appear only as tolerance bands on the spec, but on the line they convert directly into cycle time and rework rate. Two minutes extra per assembler per day becomes significant cost over a year.

4. Tooling ownership. If tooling ownership isn’t clarified at kickoff, moving volume production elsewhere later gets blocked. The cost of asking one question before tooling is lower than litigation after.

3. Front-load uncertainty: a qualification drawing package

Almost all later disputes trace back to unclear requirements at inquiry stage. Make it a standard template sent to every candidate supplier:

1. Profile cross-section (DWG / PDF), marking key dimensions and tolerances;
2. Functional-face notes: defect tolerance for output face, mating face, decorative face;
3. Material and color / haze target values;
4. Target weight range (noted as target, not ceiling);
5. Length, cutting method, packaging (film required? stackable?);
6. Volume ladder: annual estimate, MOQ, first trial batch;
7. Use environment and compliance (flame, RoHS, REACH, outdoor?);
8. Whether batch traceability and shipment inspection report are needed;
9. Sampling requirement: sample length, whether photometry validation on your own luminaire is required;
10. Project milestones: sampling, approval, mass-production dates.

With this package, quotes become truly comparable and later communication has a common baseline.

4. How to judge a worthy extrusion partner

Skip the brochure; look at these five actions:

• Asks for the drawing before quoting. One that quotes full-range prices without a drawing can be dropped.
• Gives manufacturability advice. A real extrusion engineer proactively raises wall-thickness equalization, fillet transitions, snap rebound margin — not just takes the order as given.
• States its process boundaries clearly. The weight range it covers, the width it can form, production scale — if these numbers match sampling results, the interface is trustworthy.
• Commits to batch consistency. At least a clear sampling method and non-conformance handling; ISO 9001 is just the entry ticket.
• Ships enough sample length. One that only sends a ten-centimeter stub and won’t give enough to build your light likely lacks confidence in its own batch consistency.

Conversely, three signals suggest cutting losses: promising “all standards met”; giving only photoelectric data and no physical sample; a quote with zero stated assumptions.

5. Division of labor for engineers and procurement

The only way to avoid “both sides assumed the other was managing” is to write the standard into one shared document:

• Engineer owns: the acceptance boundaries of optics and dimensions — transmittance–haze range, allowed color and black-dot density, tolerance budget of key assembly dimensions.
• Procurement owns: the commercial boundaries — tiered pricing, standard lead time, tooling ownership, annual reduction mechanism, minimum economic batch.
• One shared “spec freeze sheet”: any change goes through change evaluation, not verbal adjustment. Most “cost overruns” start from one undocumented verbal change.