On a gas train or a pressure instrument, the injection-moulded parts are rarely the items anyone specifies carefully. They are also, quite often, the items that fail first. Here is what to put in the enquiry.
Walk along almost any gas or pressure assembly in a building — a regulator on a cylinder manifold, a pressure gauge on a plant-room header, a flow indicator on a distribution line — and count the plastic. The handwheel that sets the flow. The valve stem guide. The seat inside the check valve. The graduated tube. The protective boot around the gauge lens. Almost none of these get more than a line on the schedule, and almost all are injection-moulded.
That asymmetry is the problem. A brass body gets a material certificate, a pressure rating, and a witnessed test. The moulded handwheel next to it gets specified as “nylon, black” — and nylon, black, is not a specification. It is a category containing materials whose stiffness, moisture behaviour and temperature performance differ by more than a factor of two.
My background is not building services. I run a second-generation injection moulding factory in Ningbo, China, and for the last seventeen years most of what we have moulded has gone to a European medical-gas equipment OEM: regulator handwheels, check valve seats, valve stem guides, complete float-type flow meters. The environment is different from a plant room, but the physics of a small moulded part sitting in a pressurised gas assembly is not. The five questions below are the ones that, in our experience, separate a part that lasts fifteen years from one that comes back.
1. Which grade, not which family?
“Nylon” covers PA6 and PA66, glass-filled and unfilled, and the gaps between them are not cosmetic. On the handwheel family we build — around forty variants from 26 to 66 mm across — the material is chosen by torque, not by preference. Small flow wheels run in unfilled PA6. Higher-load wheels move up to PA66. The largest cylinder-valve wheels use 30% glass-filled nylon, which roughly doubles stiffness.
That stiffness has a price. The 33% glass-filled PA66 on our shelf has a tensile strength of 186 MPa but an elongation at break of about 3%: strong, and brittle. Put a sharp internal corner on a part made from it and you have designed in a crack initiation site. Ask for the grade designation and the datasheet, not the family name.
2. What moves after the part leaves the mould?
Two mechanisms make a moulded part a different size in service than at final inspection, and both are routinely missed at specification stage.
The first is post-mould shrinkage. Semi-crystalline materials keep crystallising after ejection. On the acetal homopolymer we run for check valve seats, mould shrinkage is roughly 2.8–2.9% along the flow direction and 2.1–2.4% across it, and the part is still settling for about twenty-four hours after it comes off the machine. We hold parts a day before dimensional inspection for exactly that reason. A supplier who measures at the machine and ships is measuring a part that no longer exists. Shrinkage is measured to ASTM D955; ask which value the tool was cut to.
The second is moisture. Unfilled PA6 takes up somewhere around 1.5–1.8% water in twenty-four hours’ immersion and reaches roughly 2.8–3.5% at equilibrium in 50% relative humidity, and it grows and softens as it does so. In a dry, heated plant room, a nylon component behaves differently in February than the one you tested in August. If a dimension is critical to a seal or a fit, either the moisture uptake belongs in the calculation or the material is wrong. Water absorption is tested to ASTM D570.
3. What is the real temperature, and what does the datasheet actually promise?
Heat deflection temperature is the number most often quoted and most often misread. Our acetal grade lists an HDT of 115 °C at 1.8 MPa, but its maximum continuous service temperature is 100 °C — and continuous service is what a plant room delivers. Meanwhile the glass-filled PA66 that measures 186 MPa in tension at 23 °C measures about 110 MPa at 77 °C. That is a 40% loss inside the normal working range of a boiler house.
None of this is exotic; it is on the datasheet. It just has to be read against the service condition, not the test condition. HDT is determined to ASTM D648, at a stated load, on a bar — not on your part, under your load, for fifteen years.
4. How many cavities, and what does that cost in tolerance?
This one never appears in enquiries and it should. A mould with more cavities is cheaper per part and less consistent between parts. Our rule of thumb, developed over roughly thirty years of family-factory production, is that each additional cavity costs about 5% of dimensional precision, and that where precision genuinely matters you should not exceed four cavities. We hold ±0.05 mm as standard and ±0.02 mm where the drawing demands it, and cavity count is part of how that gets achieved — we run tools from two to eight cavities depending on what the part has to hold. If your part has one feature that decides whether it works, say so in the enquiry, and expect the tooling proposal to reflect it. If you want to see how the tooling and machine side of that decision is set up in practice, you can learn more here.
Related: ask where the parting line falls. On a barbed hose connector we mould, the whole function lives in ridges a fraction of a millimetre oversize relative to the hose bore. A parting line across a sealing surface, or a trace of flash on a barb edge, scores the hose on assembly and produces a joint that looks perfect and weeps from day one — a tooling decision made before the first shot, and invisible on the finished-part drawing.
5. Is the acceptance test dimensional or functional?
A calliper cannot tell you whether a flow indicator reads correctly. The float in a tapered tube sits at a height set by the tube’s internal geometry; if the bore drifts slightly, the part passes every dimensional check and the reading is still wrong. On the flow meters we build — 0 to 50 L/min, in 2.5, 3.5 and 4.5 bar versions — every single unit is flow-calibrated against a standard gas source before it ships, to about ±4% of full scale against the printed scale. Not a sample. Every unit.
The same logic applies to elastic parts. A protective boot for a gauge grips because its opening is moulded deliberately smaller than the gauge; a calliper reading cannot describe that grip, so batches get fitted to a reference gauge instead.
So the last question is the simplest: what does the supplier’s inspection actually prove? If the part’s job is to measure, seal, grip or hold pressure, dimensional inspection alone proves that it is the right size. It does not prove that it works.
Writing it into the enquiry
None of this requires a bigger budget. It requires four extra lines in the enquiry: the grade designation rather than the material family; the service temperature and humidity rather than ambient; the one feature that actually decides function; and the functional test you expect on the certificate. Moulders who cannot answer those four will quote lower. That is usually the tell.
Frank Lai is the second-generation owner of Weilin Plastic, an injection moulding factory in Ningbo, China, running sixteen injection moulding machines from 100 to 1,000 tons. The factory has supplied a European medical-gas equipment OEM for seventeen years and a North American dental consumables brand for twenty-nine, and holds ISO 9001 certified production.