Nozzle melt temperature reveals molding quality.

A novel approach to injection molding control has been built around subtle but informative thermal changes in the melt stream as it leaves the nozzle. Measured by a new infrared sensor, this once-hidden temperature information promises answers to fundamental questions about the many variables affecting final part quality.

The first commercial machine monitoring system employing the new sensing technology grew out of experiments conducted over the past two years by Dynisco Inc. of Sharon, Mass., and Mannesmann Demag Hamilton (MDH), the British arm of the German machinery supplier. Last month, the two companies unveiled the hardware and software developments needed to sense small-scale melt temperature movements and then display them graphically as unique curves.

MELT-STREAM MONITORING

Each of the melt-temperature curves or "thermal portraits" recorded by the Dynisco/MDH system corresponds not only to the process and material variables from an individual shot but ultimately to final part characteristics as well, according to Dynisco marketing manager Gene Yazbak.

Under the new sensing system, temperature and pressure control for machine and mold remain as crucial as ever, but they don't paint the full picture of a shot. "Most people only focus on temperature control of the mold barrel and nozzle" Yazbak explains. Yet there's often a false assumption that this machine metal temperature is the same as the melt temperature." What's more, melt temperature may provide a more sensitive characterization of the molding process than pressure because the former changes with both compression and shear, Yazbak says.

For now, the system works as a real-time monitoring tool with MDH's IBED machine controller processing and displaying the melt-temperature signal from Dynisco's MTX probe. At the same time, the IBED stores a "golden curve" representing the optimal processing parameters. By overlaying the current and optimal curves on the same screen, the system provides a basis for comparison. Users can then harmonize current operating conditions with the optimized ones by manipulating processing variables such as injection speed, temperature settings and nozzle pressure. The system also yields information on another important variable as well: material variations. "The real wild card here is the material itself," Yazbak says.

In developing the new system, the two companies began with the assumption that thermocouples alone don't truly measure the melt-stream temperature. Yazbak explains that information provided by flush-mounted thermocouples near equipment walls doesn't really represent the temperature of the rest of the melt steam due to boundary-layer flow effects and the fact that polymers are poor thermal conductors. The same goes for the measurements from intrusive thermocouples, which may actually create localized temperature discrepancies by disrupting flow.

Before introducing an IR-based method, Dynisco and MDH tried numerous flush-mount and intrusive thermocouples from several suppliers, including Dynisco. None could detect the short-term temperature variations essential to the new SPC method. IR pyrometers, however, have faster response times, typically around 10 millisec, according to Yazbak. The fastest thermocouple response times are around 300 millisec, he says.

Even with the MTX IR sensor, the transient temperature variations initially showed up only as a small blips--in other words, the variations were both brief and small in magnitude. "At first we though it was just noise," says Yazbak, noting that these temperature variations for an amorphous resin often amount to as little as 6|degrees~ C and do not last even a second.

About a year ago, however, MDH process technologist John Goff and Dynisco UK's Bill Davies theorized that these blips might be something more than noise and might even reveal something about the molding process. By amplifying the IR pyrometer's signal, the fleeting temperature phenomenon became easily observable. MDH then developed a software option for its IBED machine controller, allowing it to display the curves graphically. "The portraits are measurements of these small temperature nonuniformities as a function of time during an individual shot," Yazbak explains. "These thermal portraits contain the essence of the quality of a shot."

DECODING THE CURVES

Moving beyond empirical observations from monitoring individual processes, the next R&D step for Dynisco and MDH will be to break the curves down into regions that correspond to particular processing and quality parameters. For example, the shape of curve's leading edge now appears to relate mostly to the polymer's rheological characteristics, Yazbak says. "We believe there will be universal families of curves generated for related materials," Yazbak says.

Already, various segments of thermal portraits have been correlated with final quality parameters such as surface quality, part weight, mechanical properties and temperature resistance. One unnamed university has come up with a proprietary method for interpreting the curves, which should be commercially available within the year. Similar work also proceeds at the Polymer Development Center in Athlone, Ireland.

CURRENTLY AN OPTION

As an option, all MTX pyrometers can now handle the unique signal-processing requirements of the melt-stream monitoring system. So far, though, the only commercial machine controller ready-equipped to process the enhanced MTX signal is the IBED controller from Mannesmann Demag Hamilton. Six molders in Europe have been using a forerunner of the current system over the past year.

However, the Dynisco MTX system also can work with any machine controller that has the ability to graphically display a curve, such as one depicting mold cavity pressure. And within two months, Dynisco plans to release a stand alone system that will include the modified MTX and a graphic display.

Likely applications for the new system will include high-value molding such as electronic, medical and optical parts, Yazbak says. The price will be around $3000.