Dimensional variation is usually analyzed by measuring the part and correlating the dimensions with the process parameters. The thermocouple's data is the process parameter that is most directly linked to the part's thermal history, and its inclusion in the analysis narrows the root cause more efficiently than the part measurements alone.
The first use is the cavity-to-cavity comparison. A part dimension that varies between cavities is correlated with the cavities' probe readings, and a dimension that correlates with a probe's deviation is a dimension that the probe is causing. The correlation is established by plotting the dimension against the probe's calibration deviation, and a systematic pattern-one cavity's parts are consistently different, and its probe consistently deviates-is the evidence.
The second use is the time-series comparison. A dimension that drifts over a production run is correlated with the probes' readings over the same period, and a drift that tracks a probe's drift is a drift that the probe is causing. The correlation is established by plotting the dimension against the probe's reading over time, and a parallel trend is the evidence.
The third use is the cycle-to-cycle comparison. A dimension that varies from cycle to cycle is correlated with the probes' readings at the same cycle, and a variation that tracks the probe's variation is a variation that the control is causing. The correlation requires the probes' readings to be logged at the cycle rate, and the logging is the enabler.
The fourth use is the gate freeze comparison. A dimension that is sensitive to the packing history is correlated with the near-gate probe's freeze time, and a dimension that tracks the freeze time is a dimension that the gate's thermal behavior is causing. The correlation is established by plotting the dimension against the freeze time across a range of cooling times, and the relationship is the evidence.
The fifth use is the zone isolation. A dimension that is sensitive to one zone's temperature is isolated by varying that zone's setpoint and observing the dimension, and the probe's reading confirms the setpoint's actual effect. The isolation is a designed experiment, and the probe is the instrument that records the independent variable.
The practical discipline is that the probes' data is logged at the same time as the part measurements, and the two are analyzed together. A dimensional analysis that uses the machine's setpoints without the probes' readings is an analysis that assumes the setpoints are the actual temperatures, and the assumption is the variable that the analysis should test. The thermocouple is the evidence that the setpoint and the actual temperature are the same, and when they are not, the difference is the root cause.
