Choosing designs of calibration transducer electronic data sheets

The IEEE 1451.2 standard introduced the notion of a calibration Transducer Electronic Data Sheet (TEDS), which contains the information needed by software for mapping between transducer-side values and physical unit values. When a calibration TEDS entry is created for a transducer, a calibration model (the degree of the multinomials and the placement of segment boundaries) must be chosen. Calibration TEDS must fit into small memories, and hardware floating point support is typically unavailable for correction function evaluation, but uncertainty due to error fitting the correction function must be minimized. This paper considers the computing of TEDS memory consumption, the number of floating point computations required for each correction, and the uncertainty when the correction function is a spline. It also describes a prototype tool that makes these computations and graphs them in forms intended to be of use in making calibration model design tradeoffs. The tool also contains a facility for automatic selection of "best" designs.