In the world of tactile pressure sensors, you’re often only as good as what you’re able to demonstrate. As such, calibrating pressure sensors is a crucially important process.
In a perfect universe, all sensors would have the exact same reaction to a given input. When the input is, say, 1 psi, every sensor element would produce the same number of counts in response. Unfortunately, manufacturing variations and various control issues mean nothing is ever really identical. For example, 1 psi might produce an output of 1,000 counts in one element and 990 in another. To account for these disparities and attain a truly accurate pressure reading, it’s thus often necessary to calibrate.
Sample calibrated output from PPS sensor
(Click to download sample report)
To begin this process, a known, uniform pressure is applied to a sensor, and then a calibration coefficient is derived. If 1 psi results in an output of 1,000 counts from an element, the calibration coefficient for that element is 1 psi per 1,000 counts. This figure is the key to converting raw sensor output into pressure values; when the sensor output for a given pressure is multiplied by the calibration coefficient, the result is the true pressure.
If a given pressure produces an output of 500 counts in a sensor with the above calibration coefficient, for example, then we know the pressure measurement is 0.5 psi. When the calibration is verified and the same load is applied to the sensor, the ratio should be 1:1. On a graph, this would translate to a perfectly straight line, which is a sign that the calibration was successful and proof that the sensor is accurate.
Of course, ensuring that a uniform, known pressure is applied in the first place can be tricky. Consider a 1 kg weight. One might reasonably assume that placing the weight on a sensor would apply the pressure associated with 1 kg. But it’s not that simple; the pressure is often not actually uniform. Nothing is perfectly flat, and the bottom surface of a weight is no exception. The small surface variations on the weight means that it’s only placing peak pressure on the sensor in some locations, while in others it’s applying far less pressure. If the peak pressure is high, this can lead to saturation in the sensor, which can result in errors.
One possible solution might be to place a piece of foam between the weight and the sensor to ensure that the pressure is distributed more evenly. This brings its own challenges, of course. If the foam is too hard, it will act too much like the rigid object you’re trying to buffer in the first place; if the foam is too soft, it may bottom out, negating the benefit of having the foam there at all. Rubber might compress more effectively than foam, but it, too, can cause complications, namely by gripping the sensor as it squishes, sometimes to the point that the sensor is ripped apart.
Pressure Profile Systems (PPS) has had success using leather as a calibration material, especially when working with high pressures. The leather is strong enough to remain intact, but also compliant enough to evenly distribute the pressure. An even more consistent method, however, involves the use of a thin bladder inside a chamber. The sensor is placed inside the chamber and the bladder is inflated to the desired pressure, thus applying uniform pressure to the sensor. This method can be tailored to the pressure ranges being tested: the lighter the pressure, the thinner the bladder.
For other types of sensors that aren’t suitable for the bladder and chamber, PPS will use a calibration pod. In this method, the sensor is covered with a thin membrane and then placed into a pod, which is subsequently pressurized. This approach, in turn, results in the application of uniform pressure.
The calibration process can become even more complex yet, often defying the relatively simple notion that pressure equals force over area. Consider a wiper blade. When it’s in use, the shape of the blade and the point of contact between the blade and the windshield are not constant. As the blade moves across the glass, it can flatten and straighten. A sensor designed to verify the pressure applied by the blade must be able to accommodate the variable nature of that pressure; calibrating it with a uniform, static pressure, then, might not make sense. In such cases, devising a calibration method that mimics the actual use conditions—such as a robotic calibration fixture that presses down on a piece of the blade, making it slide as if it were in use—is preferable.
Though it can be tough to grasp and easy to oversimplify, the calibration process is a key part of developing capacitive tactile pressure sensor technology, and one that must not be overlooked.