The measurements reported in this article were obtained using the three coils of one Articulograph, on one day. The variability of the Articulograph fields from day to day, or from device to device, has not been addressed in this study, but some speculation is possible.
The positions of the receiver coils are calculated by comparing the transduced magnetic field to a standard level, which is defined when the device is calibrated during system installation. Any significant change over time in the magnetic fields would require the device to be recalibrated. Since recalibration is a time consuming and undesirable task, the device has presumably been designed to minimize any change in the magnetic field strengths over time.
Electrical failures are extremely unlikely to cause transient increases in magnetic field. Each transmitter coil is presumably the inductor in a simple resonant circuit; a short circuit anywhere in this circuit will cause the current in the coil, and the magnetic field it produces, to decrease monotonically toward zero. Although an electrical failure seems to present no danger of transient magnetic fields, it does seem to present significant danger of electrical shock, and safety precautions to prevent an electrical failure and protect subjects from shock are therefore discussed at some length in the user's manual (1993).
The transmitter coils in the Articulograph are apparently hand-wound, and although they must meet relatively strict specifications in order to be useful in the device, some variation from device to device is to be expected. In the measured device, for example, the blue coil was somewhat stronger than the red and green coils, with field strength constants of A=1910 and A=1750 uT cm², respectively. If these three measurements are presumed to be taken from a Gaussian distribution, standard statistical analysis yields a 95% confidence interval between 1620 and 1980uT cm², or about +/-10% of the mean value. Since the three coils in the measured device were presumably created in sequence, it is possible that the variation among devices may be somewhat larger than the variation among these three coils.