Warpage Complaints on a PBT Housing: Nine Causes to Check in Order

Warpage in semicrystalline polybutylene terephthalate (PBT) housings is rarely a single material failure. It is typically a sum of frozen-in stress, anisotropic shrinkage, and non-uniform solidification history. The following sequence orders the checks from fastest drying verification to post-mold dimensional drift. It applies to injection-molded housings with flat sealing surfaces, snap-fit ledges, and wall thicknesses from 2.0 to 4.0 mm. Table 1 lists each check with the recommended measurement and a typical acceptance window.

Moisture Uptake Before Molding Is the First Variable to Eliminate

Hydrolysis is the fastest way to destabilize a PBT housing. Commercial PBT grades exposed to 50% relative humidity at 23°C approach an equilibrium moisture content near 0.2 wt%. Even 0.02 wt% residual moisture can reduce molecular weight during plastication, producing non-uniform viscosity, splay, and lower local shrinkage. Drying should be performed in a closed-loop desiccant dryer with a hopper-inlet process-air temperature of 120°C, a hopper-outlet dew point below -30°C, and a residence time of 3–4 h for virgin material. A moisture analyzer using ISO 15512 or ASTM D6980 should confirm residual moisture below 0.02 wt% before molding. If the dryer process-air temperature falls below 110°C, the desiccant bed cannot maintain the required dew point. At relative humidity above 60%, open containers should be avoided and the hopper should remain sealed. This check is placed first because it can be verified in the molding cell without removing the tool from the press.

How Differential Cooling Drives Out-of-Flatness in Thin Rib Sections

After moisture is eliminated, the second check is mold surface thermal balance. PBT solidifies rapidly, and local mold temperature differences of 5–10°C are sufficient to alter crystallization rate and final crystallinity. A hotter cavity face allows lamellar growth to continue longer, producing higher local density and greater shrinkage relative to a colder core face. The housing then bows toward the hotter side after demolding. Unfilled PBT is normally molded at 80–100°C mold temperature, while glass-reinforced PBT often requires 100–120°C. For flatness-sensitive housings, a thermocouple array or calibrated IR imaging should verify that the entire core and cavity surface stays within 5°C of the controller setpoint. If the cooling-channel layout is asymmetric or the tool has thick and thin sections on the same cavity, the result is a temperature map with a diagonal gradient. Correct the circuit balance before any material change; differential cooling symptoms cannot be compensated by increasing pack pressure. A two-plate mold with drilled cooling channels should also be checked for scale and blocked circuits because PBT tools frequently run at temperatures that promote mineral deposition in untreated water.

Gate Freeze-Off Pressure and Packing Imbalances

Once the thermal map has been verified, the third check is whether the gate remains open long enough to pass pressure into the cavity during the shrink phase. PBT housings with a large sealing face often fail when the gate freezes before the melt in the thick mounting bosses has fully packed. Cavity pressure transducers should be placed at the gate and at the last point to fill. On a closed-loop hydraulic machine, switch-over from velocity to pressure should occur when the cavity pressure reaches 4–6 MPa; the exact value is tool-specific and must be established by short-shot and response studies. The hold time is increased until the shot mass reaches a plateau; the point of mass stabilization is the gate seal time. For a 2.0–3.0 mm nominal wall, gate seal time commonly falls between 5 and 12 s depending on gate diameter, mold temperature, and hot-drop temperature. Holding pressure is typically set at 60–80% of the injection boost pressure. If the mass plateau is never reached before 15 s, the gate is too large or the melt temperature is too high; if the plateau occurs before 3 s, the gate is too restrictive and the part is being ejected with partially packed thick sections.

Because flow-induced orientation is the fourth check, the injection velocity is recorded from the machine linear transducer and compared with the fill pattern. On an electric injection-molding machine with clamp force between 800 and 2000 kN, screw advance is typically read directly from the linear transducer rather than calculated from flow rate. High injection speeds create a highly oriented frozen skin; low speeds allow molecular relaxation and reduce anisotropic shrink. A short-shot series in 5% increments will show whether the melt front is balanced. Mold shrinkage measured according to ISO 294-4:2018 on plaques provides a baseline: unfilled PBT commonly exhibits 1.5%–2.0% free shrinkage, while a 30% glass-reinforced grade can exhibit 0.2%–0.5% in the flow direction and 0.8%–1.2% transverse. When the measured shrinkage difference between two housing axes exceeds 0.5% and the short-shot study shows jetting or an unbalanced front, the gate location or the injection speed should be changed before additional tooling work is performed. Reducing injection velocity is not always correct; for thin housings, too low a velocity may create premature freeze and worsen warp.

Thermal history is checked next if orientation adjustments do not restore flatness. For unfilled PBT the recommended melt temperature range is 230–250°C; for glass-reinforced PBT it is 250–270°C. Barrel rear zones are generally held 10–20°C lower than the nozzle to prevent premature melting and feed-bridging. Residence time above 4–5 min at the upper limit causes molecular weight loss, yellowing, and a measurable increase in melt flow rate. A sample of dried virgin material and a purged sample from the barrel should be tested per ISO 1133-1:2022 at 250°C and 2.16 kg. An MFR shift greater than 15% from the certificate of analysis indicates degradation. Degraded PBT exhibits reduced skin crystallinity, lower melt viscosity, and non-uniform packing, which typically appears as a saddle-shaped bow on flat housings. The correction is not higher pack pressure but a purge, lower barrel setpoints, and a shorter cycle. If the same mold runs multiple PBT grades, lot-to-lot variation in initial MFR should also be checked because a low-viscosity lot will flow differently under the same pack profile.

When Fiber Orientation Gradients Dominate Shrinkage Anisotropy

For glass-fiber-reinforced PBT housings, the sixth check is fiber orientation. Glass fiber reduces total shrinkage but introduces strong anisotropic behavior. A 30% glass-reinforced PBT may show a tensile modulus near 9.5 GPa in the flow direction but significantly lower modulus transverse to flow per ISO 527-2:2012. The diagnostic procedure is to cut three sections from the warped housing, polish them, and evaluate fiber orientation by reflected-light microscopy or ashing per ISO 3451-1. A gradual skin-to-core orientation gradient is normal; an abrupt transition from highly aligned skin to nearly random core is associated with high shear at the gate and rapid cooling on the cavity wall. If the fiber orientation gradient dominates the warpage, changing mold temperature alone will not correct the flatness. The gate thickness, injection rate, and wall thickness must be modified to produce a more uniform orientation state through the cross-section. Published data for a universal numerical orientation-gradient threshold that applies to all housing geometries is limited; part-specific polished-section comparison against a conforming housing is the most reliable diagnostic.

Geometry-induced shrinkage mismatch is the seventh variable to inspect after the material and machine settings have been ruled out. PBT housings with ribs, screw bosses, and sealing ledges often contain local wall thickness variations that produce differential cooling and bending moments. A rib thickness should not exceed 60% of the adjacent nominal wall in semi-crystalline PBT; thicker ribs remain molten after the main wall has frozen and they shrink later, pulling the surface inward. A CMM or ultrasonic thickness gauge should record at least 9 points across the flat surface. A local deviation above 0.05 mm from the CAD nominal is enough to alter cooling history and produce visible waviness. If the design cannot be changed, a moldflow analysis with a PBT-specific pvT curve should be used to determine whether local mold temperature increases or added gate pressure can compensate. The same geometry can be accepted for polycarbonate but warp in PBT because of the higher crystalline shrinkage. This check is often delayed because it requires cutting or disassembling a housing, but it is less costly than changing the tool cooling layout without dimensional evidence.

The eighth check is tool hydraulic balance and gate placement. Short-shot studies will show whether two flow fronts meet in a thin section and create a weld line; in PBT a weld line is not only a tensile strength loss but also a frozen-in stress riser that can bend the housing after ejection. Gate placement should fill thick sections first and allow flow to continue into thin sections, not the reverse. On the cooling side, the mold should be fitted with flow meters on each circuit. Laminar flow or a Reynolds number below 10,000 reduces heat transfer and allows the coolant to exit more than 3°C hotter than the inlet. An inlet-outlet ΔT above 3°C across any circuit means the circuit is not removing heat uniformly. Correcting this may require a larger pump, larger cooling-channel diameter, or removal of flow restrictions such as undersized quick-disconnect fittings. Flow meters with 0.1 L/min resolution are sufficient to detect circuit-to-circuit imbalance. This problem is common when a mold originally designed for a lower-temperature material is converted to PBT without increasing coolant flow.

Post-mold dimensional drift is the final check and is frequently misclassified as a material lot problem. If a PBT housing is ejected too hot, stacked immediately, or measured before it has reached equilibrium, residual stress relaxation will continue for hours. Conditioned dimensions should be measured at 24 and 48 h after molding according to ASTM D618-21 at 23°C and 50% relative humidity. A flatness drift greater than 0.05 mm between those intervals indicates that stress relaxation or moisture uptake is still occurring. For severe stress relaxation, the housing can be annealed briefly at 130–150°C for 0.5–2.0 h in a fixture, but this must be qualified because annealing can change crystallinity and dimensions beyond the original tolerance. If post-mold annealing is used, the fixture must support the part without applying a bending load. A final measurement after annealing must be compared with the original drawing profile using the same datum surfaces as the CMM inspection. Failure to use the same datums often creates a false pass or fail because the part is simply measured in a different orientation.