Hey all,

I just completed code to read the LM35dz tempature sensor. I tested it while the sensor was on a breadboard.

Unfortunately, the sensor does not appear to provide accurate or consistent readings when I place it on 25ft of cat5 cable. The tempatures comming back to the arduino may have a degree difference of 100 degrees. In fact there not 5 consecutive reading s that similar thoughout 4 hours of testing (generating a reading every 5 seconds).

Has anyone used the LM35dz with 25ft or more of cat5?

Is there anything I need to do as I connected the catfive directly to the sensor and the other end directly to the arduino?

I use the same ones, to aleviate variation I always average 10-20 readings, and at the moment average from two adjacent sensors to get a more stable reading. Overall I get the same reading as my regular temperature guage! there may be a slight drop over 25meters of cable, if you have a spare sensor you could compare both readings if they are next to each other (one with the 25m of cable attached!)

With such kind of sensors “averaging” is generally done by a hardware low-pass. This is not strictly necessary when you understand what your software should do instead …
The resolution of the LM is 10mV per K, 10mV is 2 points on your ADC readings. This is something you could expect to get on your breadboard (i.e. +/- 1/2K) With such a wire inbetween this accuracy can no longer be upheld. There is unavoidable noise and systematic bias.

The bias can be re-calibrated.