Designing in PBT: wall thickness, draft and the features that must not be ignored requires simultaneous attention to the semicrystalline solidification behaviour of polybutylene terephthalate. A nominal wall section between 1.5 mm and 4.0 mm is standard for unfilled injection-moulding grades, while 30% glass-fibre-reinforced grades can be processed at 1.0 mm to 3.0 mm for small connector bodies. Draft is not a post-mould correction but a design variable that determines ejection force, surface drag and the onset of stress whitening. Features such as abrupt wall transitions, undercuts, bosses and rib intersections create local heat-transfer differentials; these zones govern cycle time and distortion more than the nominal wall does. Production audits and mould-filling simulations indicate that wall-thickness imbalance commonly produces non-uniform crystallinity across the part, creating warpage that cannot be compensated by tooling adjustments alone.

When a 1.0 mm Wall Section Replaces a 2.5 mm Nominal Thickness

Reducing the nominal wall from 2.5 mm to 1.0 mm increases the pressure drop required to fill a given flow length. In semicrystalline PBT, viscosity at commercial processing temperatures is shear-thinning; a thin wall increases volumetric shear rate and locally reduces melt viscosity, but this shear heating cannot fully offset the narrower channel. For a constant volumetric flow rate, the pressure requirement scales approximately with the inverse cube of wall thickness in simple slit-flow approximations. A 1.0 mm section may therefore demand more than 5 times the filling pressure of a 2.5 mm section at the same flow length. Injection moulding machines with clamp capacities below 1,500 kN often reach the machine pressure limit before the cavity is packed when flow-length-to-wall-thickness ratio exceeds roughly 120:1 in unmodified PBT. Glass-reinforced grades reduce post-filling shrinkage but increase melt viscosity; a 30% glass-fibre grade can have a melt volume-flow rate of 10–20 cm³/10 min at 250 °C/2.16 kg per ISO 1133-1:2022, whereas an unfilled extrusion grade may exceed 30 cm³/10 min. Thin-wall designs also cool faster, which reduces cycle time but produces a high orientation gradient. The highly oriented skin restricts later contraction, while the more slowly cooled core crystallizes and shrinks, producing anisotropic residual stress. Tools for thin-wall PBT should be designed with intensified cooling circuits and a mould-temperature unit capable of maintaining 60 °C to 80 °C for unfilled grades and 80 °C to 100 °C for reinforced grades.

Where wall thickness must change, a step transition exceeding 25% of the thinner wall creates a local hot spot and a shear gradient that can trigger jetting, gas entrapment and sink marks. A nominal 3.0 mm section adjacent to a 1.5 mm section produces a ratio of 2.0:1; the melt preferentially fills the thicker channel, then returns to pack the thinner region, creating an unstable flow front and a weak weld line. For semicrystalline PBT, the resulting differential cooling rate shifts the onset of crystallization. Thick regions continue to crystallize after ejection, whereas thin regions are already frozen; dimensionally, the thick region shrinks more, bending the boundary inward. To limit this effect, thickness transitions should be tapered over a length of at least 3 times the thickness difference, or the thicker section should be cored using ribs and bosses with controlled wall thickness. Sharp internal corners further concentrate stress in PBT because the notch sensitivity of glass-reinforced grades is higher than that of unfilled grades, particularly under rapid loading. A minimum internal radius of 0.5 mm is standard for unfilled PBT; for 30% glass-reinforced PBT the internal radius should be at least 0.75 mm to 1.0 mm, or equal to 0.6 times the adjacent wall thickness, whichever is larger. Radii below these values do not necessarily produce short shots, but they create high local shear rates that degrade molecular weight and lower elongation at break when tested according to ISO 527-2.

What Limits Packing Efficiency in Rib-to-Nominal-Wall Junctions?

Rib thickness in PBT parts should be limited to 50% to 70% of the nominal wall when high-gloss, sink-free surfaces are required. A rib base at 100% of wall thickness creates a local mass accumulation that remains molten after the adjacent wall has frozen. During the packing phase, the still-molten core feeds further shrinkage, but the frozen wall cannot deform to accommodate the volume change; the resulting vacuum void or sink mark appears above the rib base. This effect is more severe in semicrystalline PBT than in amorphous resins because crystallization produces an additional volumetric shrinkage of approximately 1.5% to 2.0% beyond thermal contraction. Pack pressure must be sufficient to maintain melt flow into the junction until the gate freezes; if the gate freezes before the rib core solidifies, no amount of pressure dwell can compensate for the remaining shrinkage. The rib height should not exceed 3 times the rib base thickness for unfilled PBT, and 2 times for glass-reinforced grades, because long unsupported ribs can warp from fibre orientation anisotropy and differential cooling. Draft on ribs should be at least 0.5° per side for smooth tooling; shallow ribs with no draft accelerate sticking and may pull the part from the core during mould opening. Gas-counterpressure or coining can reduce visible sink, but these methods change tooling cost and should be evaluated against the mechanical requirement of the rib.

For bosses intended for self-tapping screws or heat-stake inserts, the same mass-control rule applies. The outside diameter of a boss in PBT should be approximately 2.0 to 2.5 times the screw diameter, but the boss wall should be cored to 60% of the nominal wall to avoid a thick mass at the intersection with the base wall. If the boss is attached to a side wall, gussets should be placed at 50% wall thickness rather than as a solid junction. In flow terms, a boss on the part perimeter acts as an obstacle that divides the melt stream; the two fronts recombine downstream and generate a weld line. The remaining strength at the weld line in unreinforced PBT is generally 60% to 80% of the base resin tensile strength, but in 30% glass-fibre PBT the weld-line strength often falls to 30% to 50% because the glass fibres align parallel to the weld plane rather than across it. When weld lines cannot be repositioned, local heating of the mould insert or a sequential valve-gate system can shift the weld line to a less loaded region. Venting at the weld location is critical; a vent depth of 0.01 mm to 0.02 mm and a land of 1.0 mm are typical for PBT, while deeper vents may allow flash because of the low melt viscosity at 250 °C to 270 °C.

Draft Angle Requirements Are Derived From Ejection Forces

From an ejection-force perspective, draft angle design begins with the coefficient of thermal expansion and the elastic modulus at the ejection temperature. As the part cools from the mould temperature to ambient, it contracts onto the core. For unfilled PBT with a linear mould shrinkage of 1.5% to 2.0% in the flow direction and 1.8% to 2.2% in the transverse direction, a deep core without draft can generate an ejection force high enough to puncture the part with the ejector pins or cause stress whitening. A minimum draft of 0.5° per side is required for smooth mould surfaces, but this value is insufficient for textured surfaces. Texture depth is commonly specified according to VDI 3400; for a VDI 24 equivalent, draft should be at least 1.0° per side, and for an equivalent of VDI 36 or heavier, the draft should increase to 2.0° to 3.0° per side. Draft on bosses and ribs should be measured from the root, not the tip, because the variation in thickness along a drafted rib changes the local cooling rate. A rib drafted at 0.5° per side with a root thickness of 1.5 mm may have a tip thickness near 1.0 mm depending on height; this produces differential shrinkage along the rib that can bow the part after ejection. Mould release coatings reduce ejection force but do not replace geometric draft, because PBT can abrade coatings over production runs exceeding 100,000 cycles, leading to inconsistent release and increased cycle variability.

In glass-reinforced PBT, gate location determines fibre orientation, weld-line placement and out-of-plane warpage. A gate placed at one end of a rectangular connector creates a predominantly unidirectional flow; the fibres align along the flow axis, giving high tensile strength in the flow direction but lower strength in the cross-flow direction. According to ISO 527-2 tensile data for 30% glass-fibre PBT, the flow-direction tensile strength is typically 130 MPa to 145 MPa, while the cross-flow value may fall to 70 MPa to 85 MPa. This anisotropy is not a defect but a design condition; stress-critical features such as snap-fit arms should be oriented to load fibres in tension along their length. A centre gate or multiple gates can reduce flow-length differences and produce more balanced orientation, but each additional gate creates a weld line. For unfilled PBT, film or tab gates with a land length of 0.5 mm to 1.0 mm are preferred because they reduce jetting and permit a smooth transition into the cavity. When a direct gate is used, the high shear at the gate creates a local reduction in viscosity and a visible gate blush; the gate diameter should be limited to 0.8 mm to 1.5 mm for small parts. Shear rates above roughly 100,000 s⁻¹ in PBT have been associated with surface melt fracture and local molecular weight reduction in manufacturer processing guides, although published data for specific grades is limited.

Moisture Control Boundaries and Hydrolysis Kinetics in PBT Melts

At melt-processing temperatures, PBT is hydrolytically sensitive. Resin exposed to ambient humidity above 60% can absorb moisture above 0.1% within hours, and melt processing at 250 °C can reduce intrinsic viscosity through ester hydrolysis. The practical limit is usually set at 0.02% moisture before moulding. Drying in a desiccant-air dryer at 120 °C for 4 hours is typical for cold, moist pellets; oven drying without desiccated air is less reliable and should not be used when the ambient dew point exceeds −20 °C. Hopper residence time should be limited to avoid moisture regain in high-humidity plants. At processing temperatures of 240 °C to 270 °C, the maximum recommended melt residence time is generally 10–15 minutes for unfilled PBT and shorter for flame-retardant grades, because halogenated or phosphorus-based additives can accelerate degradation and release acidic by-products. Shorter barrels with an L/D of 20:1 to 22:1 and a compression ratio of 2.5:1 to 3.0:1 reduce total residence time, but excessive backpressure above 1.5 MPa can raise melt temperature and trigger degradation. The processed melt can be checked by measuring melt volume-flow rate after moulding; a shift greater than 20% from the virgin value indicates hydrolysis or thermal breakdown and correlates with brittle failure in weld-line or snap-fit regions.

For dimensional control, tolerances for PBT parts are commonly established using DIN 16742 or the equivalent ISO 20457. A general tolerance of ±0.1 mm for nominal dimensions up to 40 mm is often used for unfilled PBT, but glass-reinforced grades can hold tighter tolerances in the flow direction because the fibres reduce mould shrinkage. The same part may exhibit a tolerance band wider than 0.4% across the flow direction due to differential shrinkage. Post-ejection moisture absorption of unfilled PBT is low—typically 0.1% after 24 hours in water at 23 °C per ISO 62—but dimensional changes can still occur in hot, humid service. Electrical connector housings designed in PBT are typically specified with a comparative tracking index of 250 V to 600 V per IEC 60112 for unfilled grades and are rated as HB or V-0 under UL 94 depending on the flame-retardant package. Wall thickness, draft and feature transitions should not be considered independently in tolerance reviews, because warpage from non-uniform crystallization and fibre orientation frequently exceeds the linear mould-shrinkage allowance supplied on the material data sheet.