I am still struggling with my temperature sensors. For some reason, they are just always off. I put a regular thermometer next to them, and they are off. I then adjust them by putting in the reference temperature, and then they seem to be fine, and then a day or two later, they are off by like 5 degrees or more. This seems to be the case with all my inexpensive small sensors. The only temperature that is every right is the one shown by my two Zen thermostats, which always seem to agree with my thermometers.
Am I doing something wrong, or are my sensors just crap? But could I have really gotten unlucky enough to have different types of crap sensors with the same problem? It seems very odd.
One thing I have notices is that the amount off is not linear. That may be why it seems to change over time... ie the amount off at 60 degrees is not the same as 32 etc.
Do you find the error is + one day then - the next?
When you test your sensors relative to your reference you should put both in a small cardboard box. Place the box in a location that is not subject to air currents etc. Let it stabilize for at least two hours.
On your preferences, what is the Temperature Reporting value? Default is .5C. I guess that setting could be off. Are you talking 5C or F difference? 5F is still a lot.
The temperature I get is often high and usually off by 5-8 degrees F. I have a few different sensors, but the ones a got a whole pack from is Smekitlly ZigBee Mini Temperature and Humidity Sensors.
Are these sensors on or near an exterior wall or heat generating device. I.e a computer, television, etc?
If so move it away from the exterior wall or heat generating device.
Do these sensors sit in an area that have large humidity swings? If so, it can throw off your temperature readings.
The device "calibration" may just be putting in an offset, which isn't a true calibration, in which your kinda stuck dealing with the problem.
Confirm that the temperature reporting frequency setting isnt different than your expected temperature measurement. I.e if your expecting deg f, confirm your reporting setting is for deg f.
If all else checks out, then the sensors your using aren't accurate enough for your use case.
Location can make a huge impact. I am playing around with some Iris v2 motion sensors. For an experiment I put ten of them outside. (I’m testing different setups to reduce false triggers from the neighborhood cats.)
When I set them up they were all pretty much showing the same temperature. Now they are all over the board. Here’s the current values.
You can see the range goes from 67 to 97. The two reporting 67 are currently in a box inside and pretty close to accurate. The other ones all are exposed to varying levels of sunlight. The weather app is reporting 68 right now.
While I’m testing motion I was curious to see what temps they were reporting.
I don’t really have a solution but thought the variations outside were interesting. Averaging two or three sensors might give you more accurate results. Also placement of the sensors might make a big difference. Try moving them around the room and see if it impacts the reporting.
If a sensor is in direct sunlight outside without a sun shade, you will get very different readings. If you put them all outside but in the shade, I'd expect them to be within 1-2 degrees of one another. If not, then id say they need to either be calibrated if possible, or the sensor is not accurate
Yep, my point exactly. Placement is going to make a big difference. I was just giving an example for the OP.
I’ve experimented with the Iris sensors and a calibrated thermometer in the past. This case was just out of curiosity while I was testing the motion. The interesting thing is that at night they all report approximately the same temp and it is six degrees below the locally reported temperature. I haven’t used the calibrated thermometer outside but I’m guessing that’s my normal divination from the local station.
