Plastics contract manufacturing: grade, regrind and the process window

Specify the polymer, not the plastic

Naming a family such as polypropylene or polycarbonate leaves enormous latitude. Grades within a family differ in flow, impact resistance, thermal behaviour, additive package, colour stability and how they respond to processing, and a converter will select whatever runs well and prices well unless instructed otherwise. A usable specification names the grade and its maker, or defines the properties that matter with test methods attached and an obligation to seek approval before substituting. Parts approved on one grade and produced on another look identical and fail differently, generally in the field and long after anyone would connect the two events.

Regrind policy has to be written down

Converters reprocess sprues, runners and rejected parts, which makes economic and environmental sense and changes the material. Reprocessed polymer has shorter chains, different flow and reduced impact performance, and the effect compounds with each pass. The question is not whether regrind is acceptable but how much, from which source, how many times, and whether any is permitted in parts carrying structural or contact requirements. Silence in the contract means the converter decides, and their decision drifts with the price of virgin material. Ask what current practice is rather than what the policy document says.

Colour is matched in the part, not in the pellet

Colour arrives either as pre-coloured compound or as masterbatch dosed at the machine, and the two differ in consistency and cost. What a customer sees depends on wall thickness, texture, gate position and the moulding conditions on the day, so matching has to be judged against a physical part produced on the actual tool. Agreeing colour from a swatch or a screen guarantees an argument later. Keep signed reference parts with a stated tolerance, assess them under a defined light source, and treat any change of masterbatch supplier as a re-approval rather than a purchasing decision.

Validation proves the process, not the sample

First articles produced under an engineer's eye on a slow cycle prove very little about what happens on a night shift at production rate. What matters is that the process window has been established and recorded: which settings produce acceptable parts, how wide the acceptable range is, and what gets monitored during running. Buyers should ask for parameters to be documented and locked, for changes to require approval, and for a defined check at the start of every run. Without that, quality is a property of whoever last set the machine, which is a fragile place to keep it.

Resin price moves, and the agreement should say how

Polymer prices track feedstocks and move independently of anything either party controls. A fixed piece price simply means the converter built in a buffer, and when a move exceeds it you get a renegotiation at an inconvenient moment. The alternative is an explicit mechanism: material cost shown separately from conversion cost, adjusted against a published reference at agreed intervals, with the conversion element held. It takes more trust and more arithmetic, removes the recurring argument about whether an increase is real, and lets you see what you are actually paying for processing.

Frequently asked questions

Can the moulder substitute an equivalent resin?

Only with your approval, and equivalence is not something the converter can establish alone. Grades sharing a datasheet headline can differ in additive package, colour stability and behaviour under processing, and the consequences show up in service rather than at inspection. Write a clause requiring prior written approval for material changes, ask periodically what is actually being run against what was approved, and treat an unnotified substitution as the serious matter it is.

Should we allow regrind in our parts?

Often yes, within limits. Reprocessing internal scrap is standard practice and an outright refusal raises cost for little benefit in undemanding applications. What matters is setting a maximum proportion, restricting the source to your own parts rather than mixed material, excluding it where structural or contact requirements apply, and requiring the limit to be recorded and auditable. A blanket prohibition tends to be ignored quietly, which is worse than an agreed limit somebody actually checks.

Why do the first parts look right and later ones warp?

Usually because the early parts were made under careful conditions and the later ones were not: a shortened cycle, insufficient cooling, a different material lot, a mould running hotter after hours of production, or parts packed while still warm. Warpage is a process and cooling problem more often than a design fault. Ask for the process records from both runs and compare them before concluding the tool is wrong, because modifying a correct tool solves nothing.