The gap between a Delrin datasheet and real parts is often framed as Delrin Datasheet vs Real Parts: Where Published Values Drift. On an injection moulding line equipped with a 25 mm diameter screw, an L/D 24 reciprocating screw, a 1,200 kN clamp force, and a water-heated mould held at 90 °C, the as-moulded tensile yield stress of an unreinforced POM-H can depart from the published ISO 527-2 value as holding pressure, gate geometry, and local cooling rate interact. A datasheet value is typically generated from dry-as-moulded 4 mm ISO 3167 multipurpose specimens, conditioned at 23 °C and 50 % RH according to ISO 291. The real part, particularly a 1.2 mm wall section, develops a skin layer of highly oriented crystallites and a slower-cooled core with different spherulitic dimensions. The resulting tensile stress at yield may be 6–10 MPa below the published value when the specimen is cut from a moulding rather than injection-moulded as a dedicated test bar. The drift is not a defect; it is a consequence of the standard specimen geometry, cooling rate, and gate-induced orientation.

How Does a Weld Line Reduce Published Charpy Notched Impact Energy?

POM homopolymer datasheets commonly list a notched Charpy impact strength of 7–10 kJ/m² at 23 °C when measured on a 4 mm machined notch according to ISO 179-1/1eA. In a two-gate housing, the meeting of melt fronts forms a V-notch zone where molecular diffusion across the interface is incomplete. Flow-induced orientation runs parallel to the weld line rather than across it. A production validation with a 32-cavity hot-runner mould and a melt temperature of 190–210 °C can show weld-line Charpy values of 3–5 kJ/m², corresponding to a 50–65 % loss. Increasing mould temperature to 95–100 °C and reducing packing pressure oscillation raises weld-line impact energy, but does not return it to the datasheet value. The published datasheet value therefore applies to an unwelded coupon; it cannot qualify a multi-gate load-bearing component without weld-line testing on specimens prepared from a purpose-built tool according to ISO 294-3.

In hot-runner manifold service, the melt residence time becomes an independent process variable that modifies the melt volume-flow rate. At 230 °C in a closed manifold, acetal homopolymer can undergo unzipping and formaldehyde evolution; the melt volume-flow rate at 190 °C and 2.16 kg measured to ISO 1133-1:2022 can increase by 10–25 % after 15 min of hold time. The shift is batch-dependent and occurs even when the screw barrel temperature is within the grade range. In production, a colour change or a hot-runner drool event may conceal a viscosity loss that does not appear on a standard certificate because the certificate sample is moulded from pellets, not from melt held in the manifold. When the same lot is processed on a cold-runner machine with a shot-to-shot residence time below 5 min, the drift is typically smaller. The practical limit is therefore not a fixed barrel temperature; it is the time–temperature history in the plastication unit and hot runner. Residual PVC compound in the screw or manifold is an additional incompatibility because chlorine-containing residues accelerate formaldehyde generation and surface splay.

When a Published 0.25 % Water-Absorption Value Is Applied to an Immersed Bearing

POM datasheets often list water absorption of 0.25 % after 24 h immersion in 23 °C water, following ISO 62. That short-immersion figure is repeatedly misapplied. The same standard also reports saturation water absorption near 0.8 % for POM-H, and prolonged immersion can produce dimensional change up to 0.9 % in an unrestrained moulded plate. A machined bearing with a 50 mm bore can therefore grow by 0.4–0.5 mm across the bore diameter when moving from dry-as-moulded to moisture-equilibrated service. If the datasheet 0.25 % value is used to calculate clearance, the calculation is incorrect for a wet-running application. The effect is not purely isotropic; flow direction and skin orientation produce greater linear change perpendicular to the flow path. This drift is most important in precision gears, valve bodies, and pump rotors that are machined to ±0.02 mm and then exposed to water or high humidity.

Creep Modulus, Isochronous Strength, and the Misuse of Short-Term Tensile Stiffness

When a constant load replaces a short-term tensile ramp, a datasheet tensile modulus of 3,100 MPa at 1 mm/min is no longer the governing input. Under a constant applied stress, the same POM-H part relaxes and creeps. Tested according to ISO 899-1:2017 at 60 °C with an applied stress of 20 MPa, an unreinforced POM-H can show a creep secant modulus of 1,500–1,900 MPa after 1,000 h. The ratio of creep modulus to short-term modulus is therefore around 0.50–0.60. A designer who uses the datasheet tensile modulus directly for a spring-loaded clamp will overpredict retained clamping force and underestimate long-term strain. The correct approach is to use isochronous stress–strain curves or a creep-allowable stress corresponding to 1 % total strain at the intended service temperature. Published values do not drift; the operating condition is outside the test definition, and the design calculation is what shifts.

The compiled drift directions for an unreinforced POM-H medium-flow grade are shown below. Absolute values vary by grade, but the direction and method are typical of commercial acetal homopolymer datasheets.

Surface tribology follows a similar pattern. A POM homopolymer datasheet may report a dynamic coefficient of friction against steel of 0.20–0.35, commonly measured with a thrust-washer configuration. In a real moulded gear pair, the counterface is often glass-filled POM, textured steel, or anodised aluminium, and the contact zone experiences start-stop cycles. When the countersurface roughness is below Ra 0.2 µm, stick-slip and noise can increase; when wear debris accumulates, the local coefficient can exceed 0.40. The pressure–velocity limit is not a single datasheet constant. It depends on ambient temperature, lubrication, and tolerances. Published data for a specific gear set is limited; the datasheet friction coefficient should be treated as a laboratory ranking value, not a final input for gear wear simulation.

Does Dilute Organic Acid Attack an Unannealed POM-H Part More Than the Chemical Resistance Chart Suggests?

In an injection-moulded latch, frozen-in orientation and residual tensile stress are concentrated at gates, sharp corners, and weld lines. Chemical resistance charts for POM are generated using unstressed or low-stress plaques. When such a part is exposed to dilute organic acids at 40–60 °C, stress cracking can occur even though the unstressed datasheet coupon shows no significant mass change. The relevant test is not simple immersion; it is environmental stress cracking under constant strain, such as a three-point-bend method based on ISO 22088-2. A moulded latch stressed below the datasheet yield point can fail in 24–72 h if the moulding has not been annealed and the acid concentration is outside the manufacturer’s stated resistance window. Because POM-H is moisture-insensitive but acid-sensitive through hydrolysis of acetal linkages, the process combination of high residual stress and low pH is an operational boundary. For applications involving food acids, cleaning agents, or metal chloride road salts, the real part must be tested at the as-moulded stress state, not as an unstressed ISO coupon.

Accelerated weathering of natural POM-H under ISO 4892-2 provides another case where the datasheet value does not transfer directly to field service. A natural, unfilled POM-H plaque can retain a high proportion of its initial tensile yield stress in a short laboratory exposure, yet lose surface gloss and develop microcracks at the moulded skin after prolonged xenon-arc ageing. The same grade with carbon black or a UV-stabiliser package can retain impact energy substantially longer, but the datasheet for a natural grade may not communicate this boundary. When a datasheet lists notched Charpy impact after only 500 h of accelerated ageing, it is not a substitute for multi-year outdoor data under a specific mounting stress. Components such as exterior door handles, fuel-flap hinges, and agricultural sensor housings require cut-out testing from the actual tool because the weathered surface layer is exactly the zone that the standard 4 mm ISO 3167 coupon does not reproduce.