Hi all, I have a Celsius sensor connected to an Arduino Uno via a BR logic level convertor. When running the standard example sketch I am getting air temperature readings about 1 to 1.5 deg higher than expected (checked with a K type sensor). I am careful to ensure that the sensor is away from the micro controller and any possible heat source.
Interesting when I handle the sensor there is an obvious thermal storage/release effect where the senor readings increase for few minutes after I isolate the sensor (i.e stopped holding it) and then the readings slowly fall at a rate of 0.01 deg every 20 s. The base readings always tend to be 1.5 deg above what is expected.
So has anyone had similar experiences with the Celsius sensor or do I have a dud?
I’d question your ground truth measurement of 22.6 degrees from the K-type thermocouple. Omega lists typical specs for these thermcouples based on calibration class, and K-type seems to have an error of 2.2degC, or .75% of measurement range. The chip used on the Celsius sensor has a manufacturer accuracy spec of 0.1 degC - so may be closer to the true temperature! It’s interesting that it is accurate enough to measure the affect of holding it - it is designed to have minimal thermal mass around the measurement chip, and so the “cooling off” post holding it would happen much faster in water.
Thanks for your thoughts and yes you are correct the K type sensor has a relatively range error band. However, the readings are all consistent with other temperature sensors I have and all it appears to be about elevated by 1-1.5 deg. I think this must be due to the alloy cage and agree that the effects should be reduced in water.
I plotted the temperature and have shown the period I was holding the sensor in red (held for 30s) and the dots are 1s readings. You will note that the temperatures decrease as soon as I isolate sensor and then it heats up again over a 1 minute period and then it slowly decreases to background temperature over 20 minutes. I note the strange increase in heat when there is no direct heat source when isolated happens every time I do this test…
So my only conclusion is that the alloy cage must have some thermal effect that affects the measurements and I do wonder if this can account for the elevated temps I observe. The simple answer might be that these sensors should only be used in water and you have to be careful if you want to capture any rapid temperature change.
I would appreciate any feedback on how users have found the performance of this sensor as I am keen to see if this is just an one off sensor issue (I dont have a second one) or it is just a function of the design.
