How do you compare and contrast different types of temperature sensors?

Some applications may have challenges maintaining calibration as contact sensors cover a very small area and may be influenced by fouling, loss of circulation, changes in contact media, and poor placement. In one case, the sensor required a secondary water bath which significantly delayed control loops. Once the delay was accounted for the process came back within specifications.

I have designated an integrated a plethora of sensors into variety of applications This article is NOT even mediocre. All sensors have limitations . Just saying that “xyz technology sometimes does not quite work in some conditions” immediately qualifies this post as a junk If you want to be serious , than setup struck rules. Best to follow approach of classical Analog ICs Cookbooks Bob Pease once said: “don’t just reuse your friends circuit, it might not work for you “

Worked with a variety of non-contact sensors which are area sensors which can be influenced by environmental factors. In one case, ambient lighting would force changes in readings during sunlight exposure. Sensor alignment combined with vignetting resolved the issue.

Ran into some issues with digital sensors drifting in and out of calibration. Turned out that power supply generated noise was the culprit. De-coupled the digital sensors with selected R-C networks which removed the changing offset voltage.

Agree that analog sensors can be finicky. Handling and shipping can have a significant impact on usability. In one case, contamination led to surprising results. Once the source was identified, the process was controllable.

You've missed an entire class of temperature sensors: those based on Raman spectroscopy. You could start by dividing sensors into those which measure temperature at one point vs those which simultaneously measure temperature at multiple points. You divide sensors into contact vs non-contact, which is fine. But then you don't make clear that you could have digital contact, digital non-contact, analog contact etc. The text reads like you get four different kinds. Well, then you throw in smart and hybrid which feels non-sequitur. Each description of sensor type is repetitive. Reads like it was written by a machine, to be honest. Maybe more generally, it is unclear where LinkedIn is going with this project. Wikipedia?

Ran into a situation where temperature sensors were noisy at certain temperatures. Turned out to be batch control issues. If the date codes were near a holiday, shut down, or significant event, the unit would not stay in specification. Directed purchasing to avoid accepting units within certain date codes.