AERO61021 Chap.2 Sensor Characteristics and Measurement Terms
Sensor Characteristics and Measurement Terms
Sensor Characteristics and Measurement Terms builds the vocabulary used on every sensor data sheet and turns it into numbers you can calculate. Range sets the minimum and maximum measurable values and the full scale against which most errors are expressed.
Sensitivity is the change in output per change in input, and for ratiometric transducers it also scales with excitation voltage, so ratings such as mV/V/N must be multiplied by the supply. Linearity is tested by sweeping the input, fitting a straight line and finding the largest residual as a percentage of full scale, while cameras follow an averaged convention over part of their saturation range.
Offset is the reading with no measurand applied and often drifts with temperature. Hysteresis is the difference between rising and falling sweeps at the mid-range input, divided by the output span.
The chapter works both percentages from one calibration run, shows which error should drive the correction you apply, and practises the two calculations that most often trip students: excitation changes in ratiometric devices and offset removal before applying a sensitivity. Every later chapter on calibration and acquisition reuses these terms.
What this chapter covers
- 01
Range and full-scale range of sensors and converters
- 02
Sensitivity as the gradient of the response
- 03
Ratiometric ratings and excitation voltage
- 04
Percentage non-linearity from calibration residuals
- 05
Offset and its dependence on temperature
- 06
Hysteresis error from rising and falling sweeps
- 07
Choosing which error to correct first
Worked example · free
Calculate hysteresis error from a load sweep
- 1The mid-range input is 0 + (100 - 0)/2 = 50 N, where both outputs are compared.
- 1The vertical gap between the sweeps at 50 N is |1.46 - 1.58| = 0.12 V.
- 1The output span is 3.10 - 0.10 = 3.00 V.
- 1%HE = 0.12/3.00 x 100 = 4.0%, so a single rising-load calibration would misread unloading values by up to about 4% of span.
Key terms
- Full-scale range
- The total span between the minimum and maximum values a device can measure.
- Sensitivity
- The change in a sensor's output divided by the change in the input that caused it.
- Ratiometric output
- An output that scales with the excitation voltage as well as with the measured quantity.
- Hysteresis error
- The difference between rising and falling responses at one input, expressed as a percentage of output span.
- Offset
- The value a sensor system reports when no measurand is applied to it.
Sensor Characteristics and Measurement Terms FAQ
What does a rating such as 10 mV/V/N mean?
It is a ratiometric sensitivity. For every volt of excitation the sensor produces 10 mV per newton, so at 5 V excitation it gives 50 mV per newton. Because the output depends on the supply, keep the excitation steady and record it alongside the signal so that any drift can be corrected later.
Is resolution the same as linearity?
No. Linearity describes how closely the response follows a straight line across the range, measured from residuals about a fit. Resolution is the smallest change the system can detect. A sensor can resolve tiny changes yet still curve away from a straight line, so data sheets quote the two separately.
Why calibrate in both directions?
A sweep that only rises cannot reveal hysteresis, where the output depends on whether the input is increasing or decreasing. Moving parts and friction commonly cause it. Sweeping up and back down shows the loop, lets you quantify the gap at mid-range and tells you whether falling readings need separate treatment.
Assessment move
Learn these terms by calculating with them, because coursework and quizzes reward the number and its interpretation together. Make a one-page sheet with the definitions of range, full scale, sensitivity, ratiometric output, linearity, offset and hysteresis, and next to each write the unit it is quoted in and the question it answers about your measurement.
Then practise the two percentage formulas with invented calibration data until you can set them up without looking: largest absolute residual over full-scale range for non-linearity, and the mid-range gap over output span for hysteresis. Each time, write the range first; most slips come from dividing by the reading instead of the range, or from using the bit count rather than two raised to that power for a digital device.
Pair every ratiometric rating with an excitation voltage and work out what happens when the supply sags, so the habit of logging excitation becomes automatic. For offsets, solve problems in both orders, removing the zero before applying sensitivity and then deliberately ignoring it, to see how large the resulting error is.
When you meet a real data sheet in the laboratory, for example the potentiometer or pressure sensor used in the hands-on sessions, locate each term on it and note anything that looks inconsistent, since suppliers sometimes quote figures that do not hold on your rig.
Use your own laboratory data where possible: fit a line, plot the residuals, and decide whether linearity or hysteresis is the bigger problem before choosing a calibration model. In your group report, state which errors you measured, how you measured them and what correction you applied, instead of quoting manufacturer numbers alone.
Finally, keep raw voltages in every saved file so that improved offsets or sensitivities can be applied after the event. Revisit the sheet after each laboratory session and add any new term you met, together with its unit.