Automated Harmonic Phase Compensation for Ultrasonic Sensors under Non Uniform Multi Cavity Tool Cooling

Transit

Piezoelectric transducers mounted behind ejector pins or seated against cavity insert walls deliver high-frequency mechanical pulses into the steel substrate at centre frequencies between 2.0 MHz and 10.0 MHz. A pulser-receiver card excites the lead zirconate titanate disc with a negative 200 V square pulse. The resulting compression wave propagates through the mould plate, crosses the steel-polymer interface, traverses the molten polymer core, reflects off the opposing cavity boundary, and returns to the transducer face.

Longitudinal sound speed in hot tool steel hovers near 5,850 metres per second at 20 °C. Sound travels through steel fast. That velocity drops linearly by roughly 0.75 metres per second for every degree Celsius rise in tool temperature. When molten polyoxymethylene or polyamide enters the impression at 220 °C, the instantaneous acoustic impedance mismatch between hardened steel and liquid polymer rejects over ninety percent of the incident acoustic energy at the boundary.

The minute transmitted fraction samples the cooling melt directly, returning a back-wall echo that shifts in both arrival time and amplitude as the polymer densifies. Acoustic velocity in molten semi-crystalline resins runs between 1,000 and 1,200 metres per second in the melt state. As crystallisation initiates and the polymer contracts against the cavity walls, acoustic velocity climbs rapidly toward 2,200 metres per second.

Monitoring this acoustic time-of-flight allows press-side operators to detect the precise moment of volumetric gate freeze without relying on proxy cavity pressure transducers. Cavity pressure pins frequently seize from resin flash or register false plateaus caused by mechanical pin friction in worn guide bushings.

A polished transducer interface transfers acoustic energy cleanly without grease couplants under thirty megapascals of clamping preload.

Thermal variations across multi-cavity tooling distort this acoustic measurement. Tool cooling circuits never extract thermal energy with perfect symmetry. In an eight-cavity or sixteen-cavity tool, coolant flowing through serpentine drillings gathers heat as it moves from the supply manifold to the discharge port.

Cavities positioned near the return manifold operate against water that has already absorbed thermal loads from upstream impressions. Steel expands under load. The physical path length through the tool insert lengthens while the steel acoustic velocity drops simultaneously due to the elevated temperature.

These coupled thermal phenomena generate an artificial phase lag in the ultrasonic arrival signal that mimics delayed polymer cooling. A fixed-window time-domain peak detector registers this thermal drift as incomplete resin solidification, instructing the moulding press controller to extend hold times unnecessarily. Automated harmonic phase compensation separates the stationary steel thermal drift from the dynamic phase shift driven by polymer crystallisation through multi-frequency spectral tracking.

Chill

Water channels drilled through tool steel plates exhibit continuous temperature climbs along their circuit length. In a standard sixteen-cavity medical connector tool running polybutylene terephthalate on a twenty-four-second cycle, the coolant loop experiences a temperature rise of 6.5 °C between the supply port and the discharge line at a flow rate of twelve litres per minute. Laminar flow regimes inside poorly sized cooling channels amplify this disparity.

When the Reynolds number drops below 4,000, convective heat transfer rates collapse from roughly 4,500 Watts per square metre Kelvin down to less than 1,200 Watts per square metre Kelvin. Coolant lines build mineral scale. Downstream cavity blocks consequently run hotter than inlet cavity blocks by 9 °C to 14 °C across a single moulding frame.

This temperature spread alters both the mould steel acoustic velocity and the polymer solidification rate across separate impressions. In an insert fabricated from 1.2343 tool steel through-hardened to 52 HRC, an 11 °C cavity-to-cavity temperature differential shifts the ultrasonic round-trip transit time through a 40 mm steel wall by approximately 24 nanoseconds. Because the entire ultrasonic transit variation during polymer gate freeze spans roughly 80 to 140 nanoseconds in thin-wall geometries, a 24-nanosecond baseline shift represents an error margin exceeding twenty percent of the active measurement window.

A temperature spread of twelve degrees across cavity inserts shifts round-trip steel acoustic transit by twenty-six nanoseconds in standard tool steel.

Tool builders confront immediate physical failure modes when non-uniform cooling distorts sensor data without algorithmic phase compensation.

• Asymmetric Gate Freeze terminates pressure transmission prematurely in colder cavities while hotter cavities continue drawing packing volume, yielding cavity-to-cavity part weight deviations exceeding three percent across the shot.
• Localized Sink Depressions develop on show surfaces opposite heavy structural ribs in thermally retarded impressions when machine holding pressure cuts off based on false sensor signals from colder reference cavities.
• Differential Radial Warp twists cylindrical part geometry out of roundness tolerances specified under DIN 16742 TG4 because divergent cooling rates lock in asymmetric residual stress distributions between distinct cavity rows.
• Pin Seizure Failures jam ejector mechanisms when mismatched thermal expansion rates between hot cavity inserts and cooler core backing plates gall hardened ejector pin surfaces.

Cavitary cooling balances govern part roundness. When machine controllers terminate holding pressure based on an uncompensated sensor mounted in Cavity 1, Cavities 6, 7, and 8 remain molten at their gate cores. The melt in those downstream cavities depressurises and back-flows into the runner system, causing out-of-spec shrinkage and part rejects at the customer receiving dock.

Waveform

Broadband piezoelectric pulses excited by short unipolar spikes contain energy spread across multiple harmonic frequencies. When an ultrasonic pulse enters the mould plate, the returning echo train carries discrete reflections: the initial interface echo between the sensor face and steel pocket, the primary tool-polymer interface reflection, and subsequent reverberations from the opposing cavity steel boundary. As cooling fluid temperatures fluctuate, the steel insert acts as a dispersive acoustic waveguide.

Temperature alters acoustic velocity directly. The phase angle of each spectral component rotates in direct proportion to frequency and local thermal state. Traditional envelope detection or threshold zero-crossing algorithms register gross signal shifts, mistaking thermal velocity changes in the steel for density shifts in the solidifying resin.

Which Acoustic Modes Drift under Coolant Starvation?

Longitudinal modes dominate ultrasonic measurement because shear waves fail to propagate across liquid polymer melt pools. As thermal gradients intensify across the mould, acoustic attenuation climbs alongside phase delay. High-frequency harmonic components above 7.0 MHz suffer severe damping in molten polymers due to viscoelastic absorption mechanisms described by dynamic mechanical analysis.

Phase delay tracks bulk density. Lower frequency components near 2.5 MHz pass through the melt pool with minimal attenuation, maintaining clear reflection profiles off the back cavity wall throughout the entire injection and hold sequence.

Acoustic phase velocity shifts across harmonics reveal structural transition points before volumetric shrinkage pulls the polymer clear of the steel insert.

Tracking the cross-spectral phase difference between the fundamental excitation frequency and its second harmonic extracts the real thermal state of the steel. The acoustic path through the mould insert involves two distinct domains: the invariant steel thickness and the contracting polymer wall. Because tool steel exhibits negligible acoustic dispersion across the 2 MHz to 10 MHz spectrum, phase shifts induced by steel temperature changes scale strictly linearly with frequency.

The solidifying polymer displays strong frequency-dependent phase velocity dispersion during its amorphous-to-crystalline phase transition.

1. Signal Digitisation captures the raw radio-frequency echo train at a sampling rate of 100 megasamples per second with twelve-bit vertical resolution to resolve sub-nanosecond temporal shifts.
2. Windowed Gating isolates the interface echo from the mould cavity reflection, eliminating secondary mechanical ringing originating within ejector pin shoulder steps.
3. Fourier Transformation converts both windowed time-domain signals into complex frequency spectra, isolating amplitude and phase responses across fundamental and overtone bands.
4. Dispersive Phase Separation calculates the linear phase slope attributable to steel insert temperature drift and subtracts that vector from the total phase angle.
5. Crystallisation Detection flags the sudden departure in the compensated second-harmonic phase angle, establishing the true physical gate freeze timestamp for that specific cavity.

Fourier analysis isolates fundamental frequencies. Tooling suppliers often assert that simple thermocouple offset tables stored in the press controller adequately correct for thermal gradients across the tool frame.

Compensator

Digital signal processors embedded directly in press-side acquisition hardware execute automated phase correction at cycle update intervals under two milliseconds. Field-programmable gate arrays receive multichannel digitised acoustic streams from up to thirty-two cavities simultaneously. The processor calculates the cross-power spectral density between the stored reference echo of a cold tool and the live shot echo.

By determining the instantaneous phase slope over the frequency interval from 3.0 MHz to 6.5 MHz, the compensation engine isolates steel thermal expansion from polymer cooling dynamics.

Will Discrete Fourier Tracking Eliminate Cavity Skew?

Discrete Fourier tracking calculates the exact phase difference between adjacent frequency bins, isolating the invariant linear delay caused by the steel block from the nonlinear dispersion produced by the cooling polymer. When cooling line flow rates drop in downstream channels, the algorithm updates the steel baseline matrix without requiring human intervention or machine recalibration cycles. Processors pay for scrapped parts.

Dynamic matrix adjustment ensures that an 8 °C thermal divergence between Cavity 1 and Cavity 16 does not shift the perceived crystallisation point.

Integrating ultrasonic transducers into multi-cavity injection moulds demands strict hardware discipline during tool design. Transducers mounted without rigid mechanical isolation suffer signal degradation from mechanical machine vibrations and mould clamping strain.

• Transducer Preload Mechanisms maintain constant mechanical contact pressure exceeding thirty megapascals against insert pockets using disc spring washers to eliminate thermal coupling loss.
• Coaxial Cable Shielding isolates low-voltage microvolt transducer signals from electrical noise generated by machine servo drives, hydraulic proportional valves, and internal band heaters.
• High-Temperature Piezoelectrics fabricated from modified bismuth titanate or lithium niobate prevent sensor depolarisation when cavity steel inserts reach operating temperatures above 180 °C.
• Reference Insert Probes positioned in dead steel zones outside the melt impression capture pure steel thermal drift, providing an independent baseline validation channel for the harmonic engine.

Dimensional drift compounds across cycles. Whether automated phase compensation can completely decouple three-dimensional thermal gradients across thick-walled optical parts remains a subject of active tooling debate.

Dispute

Procuring a multi-cavity injection mould equipped with in-cavity ultrasonic monitoring creates contractual friction between brand owners, tooling shops, and contract moulders. Toolmakers routinely guarantee mould steel workmanship and basic water line flow rates under dry shop conditions. Contract moulders assume responsibility for cycle time and process capability inside production environments.

When a brand owner discovers that parts from Cavities 12 through 16 fail dimensional verification under DIN 16742 tolerance group TG3, the contract moulder points to uneven cooling water distribution inside the factory manifold. The toolmaker asserts that the cooling drillings match the signed engineering drawing. Gate freeze terminates pressure transmission.

The buyer stands between both parties, facing production delays and unrecoverable tooling capital.

Ultrasonic phase tracking provides objective physical evidence that arbitrates this conflict. By capturing real-time solidification curves across all cavities simultaneously, the automated harmonic compensation engine records the exact microsecond when each gate seals. If Cavities 1 through 4 seal at 5.2 seconds while Cavities 13 through 16 remain unsealed until 6.8 seconds, the thermal imbalance of the tool stands mathematically demonstrated.

Moulders can no longer attribute dimensional scrap to polymer lot variation under ISO 1133 melt mass-flow rate spreads when cavity-specific acoustic profiles reveal an eight-degree cooling imbalance across the mould plates.

Commercial contracts referencing DIN 16742 tolerance class TG3 enforce cavity-level process capability indices exceeding 1.67 before formal tooling sign-off.

Tooling capital budgets increase when incorporating multi-channel ultrasonic compensation systems. An eight-cavity diagnostic setup adds between 18,000 and 32,000 euros to initial tooling invoices for embedded transducers, high-speed digitiser boards, and hardened wiring channels. Amortising that capital expense across high-volume production programs shifts the financial calculation.

In a program running eight million medical syringe barrels annually, shaving 1.4 seconds of unneeded hold time off a sixteen-second cycle recovers the sensor equipment cost within seventy operating days. Cavity four freezes first. Tool trials prove sensor calibration.

Supply agreements incorporating precise verification clauses stipulate that mold acceptance testing mandates a process capability index of at least 1.67 across all cavities under production cycle rates, compelling toolmakers to rectify cooling circuit imbalances before steel arrives at the buyer manufacturing plant.