Interpreting transducer specifications can always present a challenge. Manufacturers can obscure performance using misleading terminology and selective data presentation. A transducer's performance is usually determined by examining the various errors that affect its behavior.
Error components are:
Zero or Null Offset/Shift
This is the output (electrical) of a transducer when subjected to zero differential pressure across its measurement membrane.
Temperature effects are composed of two components
- Zero or Null Temperature Shift: this is the shift in the transducer zero (i.e. with zero differential pressure applied) resulting from changes in temperature. This effect causes the entire transducer curve to shift.
- Sensitivity or Span Shift: Span is typically defined as the difference between the output under full load and that at zero load. Thus the temperature span shift is the change in sensitivity caused by a change in temperature.
Typically the largest component of transducer error. This is a measure of how much the transducers output varies from that of a straight line. Linearity (or alternatively, non-linearity) is usually reported using two methods:
- Best-Fit-Straight-Line: This method uses a least squares regression to fit the transducer's output curve. (most commonly used by manufacturers)
- Terminal Based Linearity: This method involves drawing a line connecting the two terminal data points. A line is drawn perpendicular to this line so as to give the largest deviation from the actual experimental data. The length of this line is then a representation of the error. The error of this method is approximately twice that given by BFSL.
This is a measure of the difference in an output reading when repeatedly approached from the same direction (e.g. applying increasing pressure repeatedly).
This is a measure of the change in an output reading when a test point is approached from below and above. Thus, pressure is increased until the measurement point is reached, continuing pressure is applied until full scale. Pressure is then reduced until the measurement point is reached again.
This is a measurement of the change in transducer's output over a specified time period.
Summing the Errors:
Assume a Transducer's Specifications are given as (of Full Scale):
- Drift = ± 0.1%
- Null temperature shift = ± 0.15%
- Repeatability and Hysteresis = ± 0.25%
- Linearity = ± 0.2%
Using this data, the transducers performance is well represented by summing the error components using the Root Sum Square Method. Here, the total error is calculated using
RSS Error = (0.12 + 0.152 + 0.252 + 0.22)½
RSS Error = 0.37% of F.S.