In pre-production automotive wiring harness connector trials, Arnite T AM1210 is dried at 120 °C for 4 h in a desiccant-bed dryer to reduce residual moisture below 0.02 % by ISO 15512:2019 before it enters the powder bed. The powder is processed on a 30 W CO₂ laser powder bed fusion system with a 120 µm layer thickness, a recoat speed of 120 mm/s, and a laser beam offset of 0.15 mm; these are starting values that must be corrected for the individual machine’s thermal gradient and laser spot size. The bed setpoint is positioned below the DSC melt peak measured to ISO 11357-1:2016 so that fused layers resist curl without forming a hard cake that cannot be broken by the recoater. Connector housings printed under these conditions are used for wiring harness assembly trials, terminal insertion force measurements, and lock-tab cycling tests. The limiting dimension is not the pin pitch but the lock-tab root radius, where porosity and semicrystalline spherulite size control fracture behavior. Build orientation is rotated so that the Z-axis runs parallel to the tab’s flexural plane, not perpendicular to the root, because the interlayer boundary acts as a crack path under repeated deflection. Failure modes observed on production lines include recoater blade dragging from electrostatically charged fines, short feed caused by powder aging, edge curl on flat mating-face side walls, and brittle breakage at the tab root when reclaim powder is blended without virgin stock.

Why Knife-Edge Contact Retention in Printed Connector Bodies Does Not Track Injection-Moulded PBT Values?

Injection-moulded unfilled PBT components subjected to repeated insertion of folded sheet-metal terminals commonly retain holding force through the high surface shear strength of a homogenous matrix. When the same geometry is produced by laser sintering, the applied load crosses interlayer boundaries that contain a higher free volume fraction than the bulk PBT. The result is not uniform creep but localized microfracture at the contact edge, which shifts the terminal retention force downward after 50 to 100 insertions. In printed parts, the notch-like discontinuity at each layer line behaves as a stress concentrator under the line load of the metal tang. The effect is measurable under IEC 60512-15-1 for connector contact insertion/withdrawal and under IEC 60512-16-20 for tensile holding of crimped contacts. To maintain holding force above 4 N per contact, designers increase the contact boss thickness by 0.4 mm to 0.6 mm relative to an injection-moulded baseline. This compensation is not needed in X-Y oriented parts that carry the insertion load in the build plane, but most connector housings are built with the mating face upward, so the tab root is a Z-plane feature.

The comparative tracking index of unfilled PBT is tested to IEC 60112:2020 and is sensitive to surface finish; printed surfaces with Ra above 15 µm can create additional tracking paths, so creepage distances for PCB connectors should be based on printed-surface data rather than injection-moulded plaques.

Fuel Vapour Sensor Housings and Long-Term Exposure to Aromatic Hydrocarbons

The polybutylene terephthalate backbone resists swelling in aliphatic hydrocarbons, but aromatic fractions in commercial gasoline can plasticize the amorphous regions at elevated temperature. Printed housings for fuel vapor pressure sensors therefore demand a low-porosity shell with sealed top and bottom skins. Open porosity in the as-sintered part provides capillary paths for fuel condensate, and the pressure gradient across a sensor wall drives liquid into lateral layer seams. This is not a bulk chemical compatibility failure; it is a fluid ingress failure caused by interlayer voids. When the seam is sealed by a post-process treatment and passes a differential pressure leak test at 50 kPa, the housing can survive exposure to gasoline vapor, diesel vapor, and ethanol blends at temperatures below 60 °C. Above this threshold, aromatic diffusion accelerates and hydrolysis risk increases if water is also present. The component is not recommended for direct immersion in hot methanol or for coolant-containing environments because PBT hydrolyses at ester linkages under hot water and high pH. Long-term evaporative emission cycling data for printed PBT housings is limited; OEM validation should follow SAE J2044 for quick-connect fuel line assemblies or the chemical resistance portions of ISO 16750-5 for road vehicle environmental conditions.

For under-hood wire harness clip consolidation, a two-piece PA66 clip and steel nut clip is replaced by a printed PBT one-piece design with a snap arm that retains the harness bundle without a secondary fastener. The part is built flat at 0.15 mm layer thickness so the flexural arm lies in the X-Y plane, preserving maximum tensile strength under cyclic loading. Vibration exposure is assessed to ISO 16750-3, with a resonance dwell at 15 min per axis and sinusoidal sweep from 10 Hz to 1,000 Hz. Printed PBT exhibits a higher storage modulus than PA12 when measured by DMA to ISO 6721-11:2019, but a lower elongation at break, so the clip arm is thickened at the hinge and the snap angle is reduced from 45 ° to 35 ° to avoid overstrain during installation. The primary process risk is not layer adhesion but warpage at the base, where the printed part is flat and long; edge lift of 0.8 mm is enough to trap the recoat blade and shift the build. Therefore, the part is nested with a border matrix and the bed temperature is allowed to float within a narrow band tied to the crystallization onset.

If Bed Temperature Falls More Than 6 K Below the Non-Isothermal Crystallization Onset, Recoater Impact Drives Delamination

Differential scanning calorimetry to ISO 11357-1:2016 defines the non-isothermal crystallization onset, which is the critical process control input for this powder. The SLS bed setpoint is not chosen from the melt peak alone; it is selected so that fused material remains above the crystallization onset for at least the time required to deposit the next 2 to 3 layers. When the bed temperature drifts below the onset by more than 6 K, the top of a fused layer begins to solidify and contract before the next powder layer is applied. The contraction produces upward edge curl on large flat parts, and the recoater blade then strikes the raised edge on its return stroke. This impact generates a shear pulse that propagates into the part along the Z-axis interlayer boundaries, causing delamination that is visible only after breakout as a smooth cleavage plane. The failure is not detected by the machine’s powder feed sensors because the part remains attached to the build plate. The process correction is to raise bed temperature in 2 K increments while monitoring the layer surface for gloss transitions and the part edge for curl shadow. Nitrogen flow below 5 % oxygen extends the permissible dwell time, but it does not eliminate the need for temperature control, because oxidative degradation is slower than crystallization under these conditions.

Operators monitor the layer surface after each recoater pass because the earliest indicator is a change in gloss: a semi-solid surface remains glossy, while a crystallized region scatters light and appears matte. A hard cake forms when the bed is too hot and the unsintered powder starts to sinter outside the laser scan path; this condition is detected by increased recoat torque or chatter. The safe processing corridor for this material is therefore bounded on the low side by curl and on the high side by cake formation, not by laser energy alone. Laser power is then trimmed to maintain melt pool depth at 1.5 to 2.0 times the layer thickness without increasing the scan width beyond the test coupon dimensions.

Powder Refresh Ratios, Thermo-Oxidative Aging and the Ductile-to-Brittle Shift in Thin-Walled Relay Sockets

Thin-walled relay sockets with snap-fit covers are sensitive to changes in molecular weight distribution that accumulate in recycled PBT powder. In long builds, the unsintered powder is held near the crystallization temperature in an oxygen-poor atmosphere, which still permits slow chain scission and crosslinking. The melt flow rate measured to ISO 1133-1:2022 shifts relative to virgin powder after each build cycle, and the shift is larger when the refresh ratio drops below 30 %. Below this threshold, the socket’s living hinge becomes brittle, and the snap cover fails by tensile fracture rather than bending. A practical control is to blend reclaimed powder with at least 40 % virgin material when the socket wall is below 1.0 mm, and to track MFR shift as a batch release property. Batch-to-batch variance in particle size distribution also affects the powder bed density, but the dominant failure signal is not particle size; it is a reduction in elongation at break measured on X-Y tensile specimens to ISO 527-2:2012. When elongation at break falls below 3 %, the socket is no longer acceptable for snap-fit assembly.