Introduction
The Celtic Engineer is a weekly newsletter produced by Celtic Engineering Solutions. We hope you enjoy it. If you have any suggestions for topics, would like to give feedback or want your email added to the distribution list please send an email to [email protected].
Reach out and touch someone
Sensors are the things that allow us to interact with our environment. They are transducers that convert a parameter of interest into an electrical signal (sensor that convert a parameter into a chemical signal are also sensors, but we will not consider those). Our eyes, ears and skin are sensors. In engineering, we have sensors that measure temperature, pressure, position, force or light intensity.
These devices convert those parameters into a voltage or current that we can use to understand or to control. Sensors should not be confused with their inverse, the actuator, which takes an electrical signal and causes a change to one of these parameters. Examples of actuators include speakers, heaters, lights and motors.
Most sensors that you will run across are analog. That is, the output of the senor will change proportional to the input. If you are lucky, it will do so in a somewhat linear way. However, sensors can be digital as well. A good example is a button. Elevators make use a great many digital sensors to operate.
Just the facts ma’am
What makes a good sensor? To be good, you would want the device to be sensitive to the parameter you are interested. For example, if you wanted to measure pressure, you would want a transducer whose output (voltage?) changed with changes to the input – pressure. We call this sensitivity. It is the slope of the transfer function, and assumes a linear interaction, at least around an operating point. That is why you might see a specification that says 25mV/K at 273 K. Just a note about temperature. Fahrenheit and Celsius are measured in degrees but Kelvin is not. You can smack people who say “0 Celsius” or “273 degrees Kelvin” both of which are wrong.
In addition to being sensitive to the parameter that you are interested in, you want your transducer (sensor) to be insensitive to other parameters. Going back to the pressure sensor. If your sensor is very good at sensing changes in pressure, but is also sensitive to changes in temperature, that is not as good as a sensor that is only sensitive to pressure. You can fix such a device by accurately measuring temperature and compensating for it, but that is an extra step. Unless you are brand new to electronics (If you are welcome!), then you know that everything is sensitive to temperature.
He keeps looking at me
Something that we must clearly understand when we start snooping around in our environment is the ‘Observer effect.’ Whenever we observe the environment, we necessarily change it. When you measure the voltage across a resistor, you are in effect placing a resistor in parallel with the one you are measuring. That will lower the overall resistance seen by the circuit, and will likely raise the current slightly. When you stick a thermometer into a cup of coffee to see if it is just the way you like it, you will have decreased its temperature. When you measure the pressure in your car’s tire, a small amount of air escapes, decreasing the pressure in the tire.
So, the third thing that makes a sensor good is that it should not influence the property you are trying to measure any more than necessary. Incidentally, you will often hear people say this is an example of the Heisenberg uncertainty principle, which states that you can know the position or the velocity of an object with high precision, but not both. That is not strictly correct. The Heisenberg uncertainty principle is describing a principle of quantum mechanics, one that is inherent in wave-like systems and it is not making a statement about how well our technology can measure something. The Observer effect on the other hand expresses that to measure something you must interact with it and that interaction changes the parameter you are interested in knowing.
Accuracy, sort of
There are many things that will affect the accuracy of your sensor. Most sensors are not perfectly linear. Unless you are designing your own sensor, a manufacturer will go to great length to make a sensor that has an output that is linear to the input. Having said that, one only has to look at the transfer function of a thermistor to know it is a rule and not a law.
If the error between your sensor output and reality is a constant, you have a bias error. If the output changes with time when the input is constant, you have a drift error. If you get a different answer based on the previous state of your senor, you might be dealing with a hysteresis error.
These are all static errors. If your error changes with the speed of the measured property, you are dealing with a class of errors know as dynamic errors.
Except for noise, all of these errors are addressed by calibration. Noise is a separate animal and must be reduced by other means such as filtering, board layout, or shielding.
Is it accurate or is it precise
You might hear someone say that their device is very precise, they can measure something to within +/- one iota. The problem arises when the accuracy is poor. Precision refers to the closeness of two or more measurements while accuracy is a measure of how close a measurement is to the actual or true value. An example of this is when you are shooting at a target. If you place three shots so close that they are toughing you have high precision, but if those three shots are on your neighbor’s target when you were shooting at your target, then you have poor accuracy.
Final thoughts
This newsletter is sponsored by Celtic Engineering Solutions LLC, a design engineering firm based out of West Jordan, Utah, which can be found on the web at: www.celticengineeringsolutions.com. You can find the newsletter on the company blog, LinkedIn or by subscribing. Send your emails to The Celtic Engineer at: [email protected].
