Polybutylene terephthalate (PBT) is a fast-crystallising semi-crystalline polyester with a melt temperature near 223 °C and a freezing behaviour that leaves little time for correction after the gate has sealed. The material’s warpage response is the geometric sum of local shrinkage differences, crystallinity gradients, fibre orientation and residual stress. Shrinkage measured on standard plaques according to ISO 294-4:2018 cannot be transferred directly to a complex connector or housing because wall thickness transitions, gate position and cooling asymmetry change the local solidification sequence. A PBT part that is flat at ejection may continue to move during conditioning at 23 °C/50 % RH or during service near 80–120 °C as low-crystallinity skin regions reorganise. The processing parameters that dominate warpage are packing pressure integration across the gate-freeze interval, mould temperature symmetry, and cooling-channel heat extraction balance.

Why Packing Pressure Does Not Behave as a Linear Correction

The packing phase is a pressure-limited compensation window. During injection, the screw transfers velocity to melt. At transfer from velocity control to pressure control, the cavity is normally 95–98 % full. If transfer occurs too late, the cavity pressure spike can exceed 100 MPa and create flash or mould deflection. If transfer occurs too early, the pressure-controlled phase must push the flow front the remaining distance, usually at lower velocity and with higher orientation near the end of fill. A two-stage packing profile is sometimes used to separate gate compaction from long-range shrinkage compensation. A first stage of 70–80 MPa specific cavity pressure held for 0.3–0.5 s can compact the gate region; a second stage of 40–60 MPa held until gate freeze can feed the core without excessive residual stress.

Machine hydraulic pressure is not a useful comparator across presses because the intensification ratio changes with screw diameter. On a press with a 10:1 intensification ratio, a hydraulic hold pressure of 4.0 MPa corresponds to approximately 40 MPa melt pressure, while 7.0 MPa corresponds to approximately 70 MPa. For unreinforced PBT, the useful specific cavity pressure range in thin-wall parts is commonly 40–80 MPa; for 30 wt% glass-reinforced PBT, the range shifts upward to 60–100 MPa because higher viscosity and faster solidification reduce pressure transmission. The relationship between pack pressure and warpage is non-linear. Too little pressure produces shrinkage-driven warpage, sink marks and low part weight. Too much pressure produces residual stress, overpacked gate areas and stress-relaxation warpage after annealing. Published processing guides for PBT often recommend starting near 60–70 MPa for unreinforced grades and adjusting in 5–10 MPa increments while monitoring cavity pressure traces.

On a 2.0 mm wall section with a 1.0 mm gate, the gate freeze time for PBT is typically under 1.0 s; hold time beyond gate freeze has no effect on cavity mass. Cavity pressure traces show gate freeze as an inflection from a stable hold plateau to a pressure decay curve. When the gate freezes before the core is adequately packed, the far-end cavity pressure at that moment is often below 25–30 MPa and longitudinal shrinkage variation increases. PBT melt compressibility means that pressure release at the end of hold also causes a small volume recovery. If the pressure is released too suddenly, the gate region may register a short reverse-flow pulse that induces sink near the gate. A stepped reduction is therefore used for thicker sections.

Mould Temperature Gradients and the PBT Crystallinity Skin-Core Structure

PBT crystallises quickly, but the cooling rate imposed by the mould determines the skin-core crystallinity distribution. With a cavity surface at 60 °C, the frozen skin is thick and the core may remain hotter for longer, leading to a post-ejection crystallinity rise and dimensional change. Raising the mould temperature to 80–100 °C reduces the temperature gradient across the part wall and produces a more uniform morphology. The practical upper limit is set by ejection stiffness and cycle time. Unreinforced PBT may be moulded with mould temperatures of 60–90 °C; glass-reinforced PBT typically requires 80–120 °C to reduce surface defects and improve dimensional stability. The mould temperature controller setpoint is not the cavity temperature. The actual cavity face temperature can be 5–15 °C lower than the supply setpoint if the cooling circuits are long, scaled, or operated at low flow. Surface thermocouples or infrared pyrometry are used to map the cavity face before adjusting the controller.

The warpage-relevant variable is not average mould temperature but temperature symmetry. A fixed-half and moving-half difference as small as 5 °C can cause a flat PBT part to bow because the hotter side shrinks later and leaves a measurable curvature after ejection. Coolant flow should be turbulent. The onset of turbulent flow is near 4,000 Reynolds number, but injection mould designs usually target 10,000 or higher for consistent heat transfer. Laminar boundary layers in drilled channels produce convective asymmetry and change the effective cavity temperature from one side of the mould to the other.

Cavity-to-cavity temperature balance can be a larger source of warpage than the packing-pressure setpoint. On multi-cavity PBT moulds with a common return header, the coolant path length can differ by 1–2 m, producing a 5–10 °C temperature spread between the first and last cavities. This spread causes between-cavity shrinkage differences that cannot be corrected by a uniform packing pressure. Individual flow controls, pressure gauges and isolation valves are used to bring each circuit within 2–3 °C of the target. The temperature rise across any single circuit should be kept below 2–3 °C at steady state; larger rises indicate low flow or restricted circuits. For deep cores, bubblers or baffles are required to reach the hot tip of the core. A bubbler tube with an annular gap of 3–5 mm and coolant flow in the turbulent regime can reduce the core tip temperature by 5–10 °C compared with a blind drilled channel. Without this correction, the core side retains heat, and the difference between the core-facing and cavity-facing surfaces becomes a bowing force.

When Glass-Fibre Orientation Turns Packing Optimisation into a Warpage Trade-Off

In glass-fibre-reinforced PBT, shrinkage is anisotropic. The longitudinal shrinkage in the flow direction is usually lower than transverse shrinkage because glass fibres restrain matrix contraction along their axis. Packing pressure after fill can amplify anisotropy if it causes additional flow while fibre orientation remains mobile. For this reason, increasing packing pressure can reduce average shrinkage but increase differential shrinkage in glass-filled PBT. When flatness is the critical requirement, moulders may reduce packing pressure slightly and compensate with a higher mould temperature to improve symmetry. The optimum is compound-specific and should not be transferred between fibre loadings, coupling agents or regrind percentages. Published data for mineral-filled and impact-modified PBT grades is limited, and processors should verify shrinkage anisotropy on a cavity-specific basis using plaques moulded to ISO 294-4:2018. Fibre length, loading and coupling-agent selection all shift the flow-direction to transverse-direction shrinkage ratio, and a narrow pack-pressure window may exist for high fibre contents.

Gate location often determines the direction of PBT shrinkage gradients. A single edge gate on a long flat part may create a large flow-direction and transverse-direction difference, especially in glass-filled grades. Moving the gate to a centre point or splitting into multiple gates reduces flow-length differences but introduces weld lines and requires more uniform packing. Tunnel or submarine gates with diameters of 0.8–1.5 mm may freeze before the cavity is fully packed. Hot-tip gates maintain melt temperature near the gate but require precise thermal separation. Moulders evaluating warpage should not change gate geometry before confirming the cavity pressure decay at the existing gate is acceptable.

Gate Freeze Must Be Measured, Not Inferred from Hydraulic Pressure

Indirect machine pressure monitoring cannot determine when the gate freezes because the hydraulic system continues to apply force to the screw and runner after the cavity has sealed. Cavity pressure sensors are used to detect the gate-freeze moment. A sensor located near the gate shows a stable pressure plateau during effective packing and then a decay beginning at gate freeze. A sensor at the last-filled region shows the pressure loss across the cavity and whether the far end was packed before freeze. The pressure difference between the gate sensor and the last-filled sensor should typically be less than 15–20 MPa for a well-packed PBT part. If the last-filled pressure is below 25 MPa at gate freeze, the far end has not received enough volumetric compensation, and differential shrinkage across the part is likely.

Sensor placement must avoid creating a witness mark on functional surfaces; behind-pin sensors are preferred for cosmetic parts but introduce a slight signal lag due to pin clearance and thermal expansion. Direct cavity sensors provide faster response but leave a small ring or button mark. Both piezoresistive and piezoelectric cavity pressure sensors are used; quartz piezoelectric sensors are preferred for fast decay signals because of high response and temperature stability. The switchover point from velocity to pressure should be set so that the screw reaches the transfer position at the same time in every shot. Shot-to-shot switchover variation of 0.05 s can change the amount of fill at transfer by several percent in a thin-wall PBT connector. The resulting packing variation produces cavity pressure scatter and warpage variation that is not visible as a shot-weight change.

Injection velocity influences warpage through shear heating and the thickness of the frozen layer. A high injection velocity reduces the frozen-layer thickness during filling, which can improve pressure transmission and reduce differential shrinkage. However, excessive velocity in glass-filled PBT causes high shear and local heating at the gate, shifting the gate-freeze time and creating gate-brush marks. Published process data for PBT are often given with injection times of 0.5–1.5 s for thin-wall parts; faster times require venting adequate to avoid burn marks. The velocity profile should be set so that the flow front remains continuous through the last-filled region. A hesitation at a thin rib or boss creates a local cold front that freezes before packing and forms a shrink point. The result is a local warpage hinge. Rib-to-wall thickness ratios should be kept below 0.6–0.7 to reduce differential shrinkage at the junction.

Standardised shrinkage testing does not validate warpage directly, but it provides the material-specific shrinkage anisotropy inputs used to set packing and cooling adjustments. The standards shown below are the minimum set for PBT process validation.

Operational boundaries for PBT warpage control include the material’s moisture and thermal stability limits. PBT should be dried to 0.02 wt% residual moisture or less. Drying at 120 °C for 4 h in a desiccant dryer with a dew point below -30 °C is a common starting condition. Melt temperatures above 270 °C for unreinforced PBT accelerate thermal decomposition and increase the risk of gas burns, splay and molecular-weight loss. The combination of PBT with alkaline or amine-based additives is undesirable because it can promote hydrolysis and shift crystallisation behaviour. Post-mould annealing at 120–150 °C for 1–2 h is sometimes used to relax moulded-in stress, but it also changes dimensions and must be applied only after shrinkage allowance has been evaluated. If long-term flatness at service temperatures near 120–150 °C is required, dimensional inspection should be repeated after 24–48 h because ejection-time warpage data understates relaxation movement.