how should we tape the sensors?

saucepan
I’ve been testing whether it matters how our DS18B20 temperature sensors are affixed to the radiators – it’s a classic round peg in a square hole problem.  We want to know what temperature the water is inside the radiators, and all we can do is tape a sensor to the surface metal.  The question is really, how fiddly should we be about that process, and what do the readings actually mean?  I had the thought that shielding the side of the sensor that’s away from the radiator might make the reading approach the truth better, and so I set up that test. A saucepan stands in the for radiator; a sensor in the water gives the actual temperature, and two sensors strapped to the side give the insulated and uninsulated results. The insulation is just 6 layers blackout curtain lining I had lying around; it’s suppose to have reasonably good thermal properties.

insulation-vs-none

I thought we were getting a nice clear picture here, but we aren’t.  It’s not that the insulated sensor reads higher; it’s just that there’s a lot of variability in the readings.  In the next test, I made sure the water level was higher compared to the sensors, added some more with bare tape, and gave it some more ergs.  The insulated sensor reads about the same as some uninsulated ones; it’s just a matter of how much contact I managed to get, I think.

many-sensors

So what’s the answer to the question? I think:  (a) don’t bother with the insulation, and (b) the readings are going to be well short of actual temperature.  I suspect saucepans and radiators are made of pretty similar stuff (apart from radiators have an extra level of paint).   We can still learn about timings from sensors on the radiators, and  I think we can still learn something about temperatures across a system this way, but we will have to take this variability into account.  We probably can’t know much about the temperature across a radiator; if that’s meant to be 11C, and these vary by 7-8C, that doesn’t really leave room for interpretation.

Of course, Iain’s results might be a bit neater.  I never could wrap Christmas presents neatly, and this is kind of similar!

Leave a Reply

Your email address will not be published. Required fields are marked *