The extrusion of polyamide 6,6 into monofilament with a target diameter of 0.20 mm for surgical suture and high-tenacity fishing line is carried out on a 60 mm single-screw extruder with an L/D 30:1 barrel, a 10 cm³/rev melt gear pump, and a three-zone filtration head operating at a melt temperature of 290 °C and a head pressure setpoint of 180 bar. The first-stage filtration medium is a sintered metal fibre element rated at 15 µm absolute according to ISO 16889:2008; the normal differential pressure across this element is 45 bar, leaving a margin of 25 bar before the safety bypass valve opens at 70 bar differential. When the bypass opens, unfiltered melt containing gel bodies, degraded stabilizer residues, and fine metal oxides transfers directly to the spinneret pack. The resulting diameter excursions are recorded by a three-axis laser gauge at 200 Hz sampling and reach 0.24 mm from a 0.20 mm setpoint. The diameter standard deviation increases from ±0.005 mm to ±0.018 mm within 12 s of the bypass event, and the coefficient of variation shifts from 2.5% to 9.0%.

A second bypass pathway is created by the annular seal of the screen pack itself. If the screen pack is seated with a dimensional mismatch of 0.3 mm in the filtering cavity, a gap of approximately 0.1 mm remains at the periphery. Under a head pressure of 180 bar, the leakage flow through this gap follows the annular leakage model and can reach 0.4% of total throughput. A leakage fraction of this magnitude delivers enough unfiltered particulate to change the spinneret hole pressure balance. Field observations on polyamide 6,6 lines using melt pressure transducers with repeatability of ±0.25% of full scale show a correlation between seal leakage and a 0.2 bar root-mean-square pressure ripple at the spinneret. The diameter control loop compensates for long-period drift but cannot remove short-period oscillations below 2 s; therefore leaks that produce intermittent plugging of individual holes create non-compensable diameter variance.

Comparative filtration media performance under continuous operation:

The influence of bypass on diameter variance also operates through the residence time distribution of the melt stream. A bypass path that opens only at high differential pressure creates a periodic high-shear lane in the spinneret manifold. The polymer passing through this lane experiences less filtration and a lower pressure drop; the local residence time decreases by 8%, which shifts the temperature of the melt reaching the spinneret holes by 1.5 °C. Because the shear viscosity of polyamide 6,6 changes by approximately 10% per 10 °C in the 280 °C to 300 °C range as determined by capillary rheometry under ASTM D3835-16, this small thermal shift produces a measurable change in die swell and extrudate diameter. The interaction is nonlinear, and the diameter control loop must be detuned when a bypass event is detected to avoid amplification of the oscillation.

What Mechanisms Permit Filter Bypass During Continuous Screen Changer Operation?

Continuous screen changers eliminate the pressure transients of discontinuous slide-plate units, but they introduce bypass mechanisms through the scraper seal and the solidified polymer heel that separates the clean and dirty melt chambers. In a rotary continuous screen changer with a 4.5 MPa operating pressure and a 0.8 m² filtration area, the seal gap between the rotating screen disc and the housing is maintained by a wear ring. The manufacturer specifies a gap of 0.02 mm to 0.04 mm; after 5,000 h of service the gap can increase to 0.08 mm due to abrasive wear from titanium dioxide pigment. At 4.5 MPa melt pressure, the leakage rate across a 0.08 mm gap is estimated from the parallel-plate leakage equation as 0.25% of throughput. This leakage bypasses the filter element and carries titanium dioxide agglomerates with an effective particle diameter of 22 µm directly to the spinneret. The resulting spinneret hole plugging produces measurable diameter loss in the affected filament. Published data for this specific leakage configuration is limited, but the leakage fraction is within the range reported by filter manufacturers for worn rotary seals.

Thermal degradation in stagnant melt pools functions as a second mechanism. In screen changers, the breaker plate and the filter element chamber contain dead zones where polymer can remain at 310 °C for 20 min to 45 min. Polyamide 6,6 is susceptible to oxidative branching and gel formation under these conditions. Oxidation induction time measured by differential scanning calorimetry under ASTM D3895-19 shows that the oxidation induction time of unstabilized polyamide 6,6 at 300 °C is approximately 4 min; stabilizer packages extend this to 20 min. When the screen changer indexes, a portion of this degraded gel-rich material is displaced from the dead zone into the main melt stream, bypassing the filter element because it originates downstream of the filter. The gels are not removed by downstream mesh and reach the spinneret as viscous inclusions. The diameter gauge records a negative spike of 0.015 mm to 0.03 mm because the gel temporarily reduces the effective hole flow area.

Pressure-induced deformation of the screen pack support forms a third pathway. Multilayer wire mesh packs assembled without a heavy breaker plate can flex under 60 bar differential. The flexure opens a peripheral channel where unfiltered melt travels along the housing wall. If the pack is replaced at intervals longer than 72 h, polymer oxidation products accumulate in this channel and form carbonaceous deposits that further increase the channel depth. The diameter variance symptom is an intermittent high-frequency flutter of ±0.008 mm at the same frequency as the gear pump teeth, approximately 300 Hz. The flutter is not present when a new pack is installed and disappears after replacement. Maintenance logs from production-scale extrusion lines indicate that 30% of diameter variance alarms are resolved by replacing the screen pack and inspecting the breaker plate for flatness deviations greater than 0.05 mm. Amine-based additives should not be introduced upstream of a sintered metal fibre filter rated at 15 µm because amine end groups accelerate oxidative gel formation and reduce the filter service life by 50%.

When Melt Pressure Oscillation Exceeds 0.5 MPa, Diameter Variance Becomes Nonlinear

The positive displacement melt pump is the primary source of short-period pressure oscillation. A gear pump with 10 cm³/rev displacement running at 30 min⁻¹ produces a meshing frequency of 300 Hz and a theoretical flow ripple of 0.9% to 1.4% at a discharge pressure of 120 bar. The ripple magnitude depends on the compressibility of the melt and the inlet pressure stability. If the inlet pressure falls below 20 bar, the pump cannot fully fill the gear pockets, and the discharge pressure ripple increases to 0.8 MPa. At this level, the spinneret hole exit velocity fluctuates in proportion to the square root of the pressure drop across the hole, according to the orifice flow equation. For a 0.60 mm diameter spinneret hole and a land length of 1.2 mm, the pressure drop at 180 bar head pressure is 120 bar. A 0.8 MPa ripple corresponds to a 0.67% change in pressure, which produces a diameter change of 0.34% for Newtonian flow. For polyamide 6,6 at 290 °C and 1,000 s⁻¹, the shear-thinning index measured by capillary rheometry per ASTM D3835-16 is 0.75, which amplifies the diameter response to 0.5% for the same pressure ripple. The response becomes nonlinear above 0.5 MPa ripple because the elongational viscosity in the spinneret entry region changes the effective hole length.

Filtration bypass lowers the damping capacity of the system. When the bypass valve is closed, the filter element acts as a flow resistor and damps pressure fluctuations entering the manifold. When the bypass valve opens, the hydraulic capacitance of the manifold increases because the filter element no longer contributes to the pressure-drop budget. The transfer function from pump ripple to spinneret inlet pressure changes, and the phase lag shifts by 18°. A control loop tuned for normal operation with a proportional gain of 0.4 and an integral time of 12 s then exhibits a resonance at 0.8 Hz. The diameter gauge records a sinusoidal variance of ±0.012 mm at this frequency. Operators can identify this state by a simultaneous reduction in the pressure drop across the filter to below 5 bar and an increase in the diameter variance amplitude at 0.8 Hz. The condition resolves after the bypass valve is reseated or replaced.

Diameter variance is also generated by pressure oscillation at the melt pump inlet. If the extruder screw is operated with a compression ratio of 3:1 and a screw speed of 120 min⁻¹, the melt temperature at the screw tip is not uniform. A temperature gradient of 5 °C across the melt causes viscosity differences of 18%, which produce pressure pulsations with a period of 0.7 s. A melt pump can correct volume flow but cannot fully compensate for viscosity fluctuations entering the pump. The residual volumetric ripple is 0.4% and translates to a diameter variance of ±0.006 mm in a 0.20 mm monofilament. In this operating window, pre-drying to 0.08% moisture and maintaining a stable feed hopper throat temperature of 70 °C are required to keep the ripple below the 0.5 MPa threshold. Above 60% relative humidity, the resin absorbs surface moisture within 15 min, and the resulting hydrolytic degradation changes both the melt viscosity and the filtration pressure drop. Resin moisture measured by Karl Fischer titration per ASTM D6869-03 must remain below 0.08%.

Cross-flow Filtration Media and Backflush Screen Changer Considerations

Cross-flow filtration is not common in monofilament lines, but backflush screen changers use a secondary clean melt stream to flush contaminants from the filter surface. In a twin-arm continuous screen changer with sintered metal powder discs rated at 5 µm, the backflush sequence opens a valve for 0.8 s every 15 min to reverse flow through one filter cavity. The reverse pressure is 6 bar, which is 10% of the normal forward differential of 60 bar. If the backflush pressure exceeds 12 bar, the filter cake is not dislodged but forced into the support mesh, and the filtration efficiency falls from β₁₀₀₀ = 12 µm to β₁₀₀₀ = 20 µm within 50 cycles. The resulting diameter variance is characterized by a slow upward drift of ±0.002 mm/h as the filter cake builds. Published data for this specific configuration is limited, but filter manufacturers' technical bulletins recommend a reverse-to-forward flow ratio of 0.05 to 0.10 to prevent cake breakthrough.

The filtration medium itself can contribute to diameter variance if it sheds fibres. Sintered metal fibre elements rated at 15 µm are composed of 6 µm diameter stainless steel fibres. Repeated backflush cycles create fatigue cracks at fibre junctions after 50,000 cycles. A fibre fragment with a length of 200 µm and a diameter of 6 µm can pass through the downstream support and lodge in the spinneret hole entry. This blocks approximately 20% of the hole area and causes a diameter reduction of 0.04 mm in the affected filament. The event is detected by the laser gauge as a single-channel dropout. Manual removal of the spinneret pack reveals the fibre fragment under magnification. To avoid fibre shedding, the filter element is replaced after 10,000 backflush cycles or 18 months, whichever occurs first. Compliance with the filter element integrity test is verified according to ISO 16889:2008 at installation and after each 5,000 cycles.

Backflush screen changers also create a bypass condition during the switching transient if the clean melt flush reservoir is not preheated. The reservoir volume is 0.25 L; if the melt in this reservoir cools to 260 °C during a 30 s flush interval, the viscosity rises by 35%. The next backflush pulse delivers this cold, high-viscosity polymer to the exit stream, and the melt pump discharge pressure increases by 15 bar. The spinneret sees a temporary pressure surge that increases diameter by 0.005 mm for 2 s. The temperature control system must maintain the reservoir at 300 °C with a tolerance of ±2 °C to prevent this transient. This temperature tolerance is also required to limit the diameter standard deviation to ±0.003 mm in high-tenacity monofilament. Temperature sensors used for this loop are platinum resistance thermometers conforming to IEC 60751, class A.

The interaction between filtration bypass and diameter variance is often separated from the quench and draw stage, but the two are coupled by melt pump control. In a two-stage water quench system for 0.40 mm PET monofilament, the first quench bath is maintained at 65 °C and the second at 25 °C; the draw ratio is 4.2:1 over a hot shoe at 180 °C. When the filtration bypass valve opens for 3 s, the melt pump speed is increased by the diameter control loop to compensate for the apparent diameter loss. The resulting flow increase of 3% changes the draw-down ratio at the spinneret exit from 1.6:1 to 1.7:1, which increases the extrudate strain rate from 80 s⁻¹ to 90 s⁻¹. The molecular orientation in the as-spun filament changes, and after drawing the final diameter variance is shifted permanently by ±0.007 mm until the control loop returns to setpoint. This transient cannot be removed by later annealing because the orientation-induced crystalline structure is already set.

A static component of diameter variance is generated by spinneret hole tolerance. A spinneret plate with 0.60 mm holes drilled by EDM has a hole diameter tolerance of ±0.005 mm and a land length tolerance of ±0.05 mm. If one hole has a diameter of 0.605 mm and another 0.595 mm, the volumetric flow difference at constant pressure is 1.7%; the resulting filament diameter difference is 0.85%. For a 0.20 mm final monofilament, this corresponds to 0.002 mm inherent variation. Filtration bypass adds a time-varying component on top of this static offset. The standard deviation of the online diameter gauge increases from ±0.004 mm to ±0.014 mm when bypass is active. The dynamic contribution from bypass is superimposed on the static hole-to-hole offset, which is why a process capability study must separate the two sources using a gauge repeatability and reproducibility protocol based on ISO 2060:1995 for linear density correlation.

Diameter variance sources and associated measurement standards for a 0.20 mm polyamide 6,6 monofilament:

Thermal Degradation By-products in PET Monofilament Extrusion

Polyethylene terephthalate monofilament for paper machine clothing is extruded at 280 °C to 295 °C with an intrinsic viscosity of 0.85 dL/g. At these temperatures, thermal degradation produces acetaldehyde, cyclic oligomers, and gel bodies. The gel bodies have a mean diameter of 10 µm to 30 µm and are removed by a 15 µm sintered metal fibre filter. When the filter is bypassed, the gels reach the spinneret and create intermittent hole blockage. The diameter variance of the filament measured by a dual-plane laser gauge increases from ±0.010 mm to ±0.022 mm over 8 h. The molecular weight distribution measured by size exclusion chromatography shifts, and the tensile strength at break, tested according to ASTM D2256/D2256M-21, decreases from 4.2 cN/dtex to 3.6 cN/dtex. This loss of tenacity corresponds to a reduction in intrinsic viscosity from 0.85 dL/g to 0.78 dL/g. Intrinsic viscosity is measured according to ISO 1628-1:2009.

The operating boundary for PET monofilament is defined by residual moisture and residence time. Pre-drying at 170 °C for 4 h to a final moisture content below 0.005% by weight is required to prevent hydrolytic degradation. The residence time in the extruder and melt pump should not exceed 8 min. If the filtration system is operated with a safety bypass valve set at 80 bar differential and the normal pressure drop is 55 bar, a viscosity shift from 0.85 dL/g to 0.75 dL/g lowers the pressure drop to 45 bar, but the gel content increases by 50%. The filter may appear clean, but the bypass valve remains closed. The diameter variance in this situation is caused by the degradation products themselves, not by filter bypass. Distinguishing between the two requires in-line rheometry or melt pressure fluctuation analysis. Residual moisture is measured by ASTM D6869-03.

Acetaldehyde is a volatile degradation product that does not affect diameter directly but indicates thermal history. In PET monofilament designed for food-contact applications, acetaldehyde is limited to 10 ppm under FDA 21 CFR 177.1630. High acetaldehyde is accompanied by lower melt elasticity and reduced die swell. The diameter of the as-spun filament at constant pump speed decreases by 0.008 mm when acetaldehyde increases from 5 ppm to 15 ppm. This shift is easily mistaken for a melt pump calibration drift. Process operators should maintain a weekly acetaldehyde trend and compare the diameter offset with the melt pressure drop across the filter. If the pressure drop is unchanged but the acetaldehyde has doubled, the diameter variance source is thermal degradation, not filtration bypass. Acetaldehyde is measured by ASTM F2013-10.

Polyamide 6,6 and PET monofilament lines share a common limitation: the filtration bypass valve is a safety device, not a process control element. When it opens, the melt pressure downstream of the filter becomes more uniform, but the particulate concentration becomes heterogeneous. A diameter gauge with 0.2 µm repeatability can detect the transition because the diameter variance increases before the pressure gauge alarm. The diagnostic sequence is to log the variance at 1 s intervals, compare it to the filter differential pressure, and inspect the screen pack for evidence of peripheral leakage. If the variance remains high after replacing the filter pack, the spinneret hole plate must be removed and each hole measured with a calibrated stereomicroscope at 50× magnification to identify gel or fibre fragments. Only after these checks should the melt pump calibration and the quench water temperature be adjusted.