There are three degrees of freedom in this equation (offset and span vs. temperature and sensitivity vs.
pressure), each of them is described with a quadratic equation. If we don’t know much about the sensor
(besides that it can be approximated with the Equation1), we need to measure three points on each degree of
freedom for curve-fitting, i.e. 9 measurements: zero- and full-scale output at 3 temperatures each for offset and
span and output at 3 pressure level for sensitivity. Since the min and max pressure levels for the sensitivity
can be combined with zero and span measurements, this reduces the total number to 7 measurements: zero-,
half-, full-scale at one of the temperatures; zero- and full-scale at each of two other temperatures. This is
called the
3 temperature + 3 pressure calibration
.
Since over-temperature calibration in an oven is often a lengthy and expensive process, it is often preferred to
characterize the span and offset drift non-linearities on some number of sensor samples and avoid using the
3
rd
temperature test in production. In this case, if we know that the sensor exhibits, for example 9% span drift
non-linearity over the full temperature range and we measure only two points on this curve, we can find the
equation coefficients. The same is true for offset drift and pressure non-linearity. By using this pre-
characterization method we can reduce the number of required measurements to
2 temperature
points with
2
pressure
levels at each (
2 temperature + 2 pressure calibration
)
In cases when the pressure non-linearity is not known due to bridge resistor value mismatch or other reasons,
three pressure levels need to be measured at one of these temperatures (
2 temperature + 3 pressure
calibration
).
The PGA309 Calibration Spreadsheet includes templates for each of the three methods described above. The
exact equations for curve-fitting the sensor are not included in this document, but can be extracted from the
spreadsheet cell formulas.
9.2 Calibration of Signal Path Errors
Since we are interested in calibrating the output of a full module, “Sensor+PGA309”, we can measure the
output of the PGA309. Then we curve-fit the result to the same model. This will take into account all offsets,
gain errors and drifts of the amplification signal path.
PGA309 and sensor must be pre-configured with desired settings: reference, excitation level, coarse gain,
offset, A/D configuration, etc. The coarse settings should be based on preliminary characterization of the
sensor and be such that the PGA309 produces a reasonable valid output signal range (say within 20% of the
target) at both zero and full scale pressure level without saturating the internal signal path. Since the circuit
offsets can be different at different gains and offset levels, the values of the all coarse settings must remain
constant during the calibration processes.
The Lookup Table containing the values of the fine Zero and Gain DACs vs. Temp ADC output is created
during the calibration to produce the desired output signal at all temperatures. We recommend to set both Zero
and Gain DACs at mid-scale each during calibration as this allows most adjustment range for calibration.
NOTE:
Set LinDAC=0 (linearization disabled) during calibration or a repeated re-calibration. Its value is also
calculated during calibration.
Page 38 of 53