Properties that can be measured for testing purposes
Some testing is accomplished through digital interfaces, but a lot of testing is accomplished through measuring various properties of the UUT by attaching sensors to the device. Here are some of the properties that can be useful to measure for product testing purposes and a short list of common associated sensor types.
- Temperature – Thermocouple, RTD, infrared, thermistor
- Pressure – Capacitance, piezoelectric, piezoresistive
- Vibration – Accelerometer, prox, strain gauge
- Current – Current transformer/probe
- Voltage – Voltage transformer/probe
- Load – Load cells, piezoelectric, strain gauge
- Force – Strain gauge, accelerometer
- Flow – Volume flow, mass flow
- Light – Photodiode, photomultiplier, CCD, bolometer
- Acoustics – Ultrasonic, hydrophone, piezoelectric
- Location/orientation – GPS, accelerometer
- Distance, proximity – Laser, electro-optical sensor, camera, hall effect, infrared, capacitance
- Vision – Optical, infrared (see Light)
- Humidity/moisture – Capacitive, resistive
- RF Emissions – RF receiver
- Magnetic Field – Hall effect
The selection of each particular sensor is dependent on the exact type of test being performed as well as the characteristics of the component being tested.
Measured Signal Bandwidth
For example, testing for current draw by a motor could call for a current transformer (CT) with an RMS output, if you intend to know the quasi-steady state current draw of a motor that had been brought to a steady operating speed under a specified load. Or you may be interested in pull-in current when the electrical power is first applied to the motor, an important characteristic to know in order to properly size the power supply that will be supplying the current for the motor. Then, the test will need a fast response CT. And, if your motor is an AC motor, perhaps that sensor should be fast enough to measure the individual cycles of the AC driving the motor.
Similar considerations of responsiveness will apply to other sensors as well. Beware that some sensors require signal conditioning due to electrical isolation needs, to prevent ground loops, and may also be active (rather than passive), sensors such that excitation power needs to be supplied to them, often in the form of a DC current source. These additional electronics will affect the bandwidth of the signal acquired from the sensor.
For example, a typical accelerometer for vibration measurement that requires IEPE excitation will have a high-pass filter (HPF) to block the output DC voltage offset caused by feeding the IEPE current to the sensor. That HPF will attenuate very low frequencies of motor vibrations. Typical sensors will have a cutoff frequency of 0.5 to 2.0 Hz. As another example of bandwidth effects, consider that electronic signal conditioning modules used for electrical isolation often apply a low-pass filter (LPF).
Another example occurs with encoders. Many encoders convert light intensity to a voltage level and note the angular motion by chopping a light source with a mask that alternates between passing and blocking the light source, creating an on/off output digital signal. Of course, the electronics that convert the light to voltage have a maximum bandwidth due to the responsiveness of the switch from off to on or on to off, and not necessarily with the same time response. That time response will limit the usefulness of the encoder when the motor speed is too high. A similar argument applies to the measurement equipment’s ability to respond to changing input voltage levels.
Inherent Sensor Bandwidth
The sensor itself can limit the bandwidth of your measurements. For example, if you want to test the temperature at the input electrical power connections during start-up, so you can watch the rate of heating, you should consider a temperature sensor that has a small mass compared to the connection material. A temperature sensor heats inversely proportionally to its thermal mass, and larger mass will act as a lower passband on the bandwidth of the sensor.
In general, be sure to choose your sensor responsiveness to match your needs.
Sensor Sensitivity and Accuracy
Sensors are designed to cover a finite range of the input physical parameter. For example, a flow sensor designed to measure the output of a kitchen water faucet will be overwhelmed by the flow from an fire engine pump. The first sensor covers a much smaller range than the second. You might wonder why a sensor with the large range shouldn’t be used for a small-range application. The reason is that the sensor sensitivity is related to its noise output. The electronics that convert the physical parameter to a voltage (or current) have inherent non-zero random noise. The amount of this noise generally increases with the range, so that as the range increases so does the noise level. A small change to the input of large range sensor would be overwhelmed by the output noise, making it difficult to detect that small input change: the sensor is not sensitive enough. More expensive sensors generally have lower noise and are thus more sensitive.
Accuracy accesses the correctness of the sensor output. Suppose you purchase a flow sensor of model M and that unit outputs Y volts when the flow is X cc/s. Will another unit of model M flow sensor also output Y when the flow is X? Generally, each unit of model M will output a slightly different Y value. The deviations across all units of model M is a measure of that model’s accuracy. More expensive sensors have better accuracy. Accuracy usually roughly scales with range, for the same reason that sensitivity decreases with range, i.e., the electronic components can add both systematic errors and random noise proportional to the range. Calibrating each individual sensor will improve that sensor’s accuracy by correcting for the systematic errors.
When choosing a sensor, make sure that you choose one with an appropriate range and accuracy.