Accuracy TMP 36?

[TheRarePinkSheep]

Recently, we've used a TMP 36 sensor with an arduino to build our sensor for regionals already. While our line of best fit wasn't the best, I noticed on the data sheet that the accuracy could be off as much as 2.

Would it be worth it to look into another sensor? I'm most likely right now going to fix the line of best fit, but my main concern is getting different readings when the sensor is measured at different times at the same temperature. Or does anyone know who's also worked with the TMP 36 know it can be worked around and is feasable for states?

[pepperonipi]

Recently, we've used a TMP 36 sensor with an arduino to build our sensor for regionals already. While our line of best fit wasn't the best, I noticed on the data sheet that the accuracy could be off as much as 2.

Would it be worth it to look into another sensor? I'm most likely right now going to fix the line of best fit, but my main concern is getting different readings when the sensor is measured at different times at the same temperature. Or does anyone know who's also worked with the TMP 36 know it can be worked around and is feasable for states?

Yeah, I would look into another sensor. You may no longer have the linear relationship you are able to get with the TMP 36 with your new sensor, but 2 degrees of error is a lottttt for a sensor for this event.

[knightmoves]

Recently, we've used a TMP 36 sensor with an arduino to build our sensor for regionals already. While our line of best fit wasn't the best, I noticed on the data sheet that the accuracy could be off as much as 2.

Would it be worth it to look into another sensor? I'm most likely right now going to fix the line of best fit, but my main concern is getting different readings when the sensor is measured at different times at the same temperature. Or does anyone know who's also worked with the TMP 36 know it can be worked around and is feasable for states?

How does your sensor behave? If you measure the same temperature of water on successive days, how close do you get?

"2% accuracy" could cover a lot of sins. It could describe a sensor that was stable over time, but had some random factory variability - in that case, you can calibrate it, and then you have a good sensor. Or it could describe something that drifts all over the place from day to day, and what you have is "it's a bit warmish".

I rather think you need to take a lot more data and characterize your device. Then you'll know.

[LIPX3]

Another possibility is water damage - water does strange things to sensors.

[NTangerine]

Another possibility is water damage - water does strange things to sensors.

I had the exact issue with 3 of the same sensors. Pretty much scrapped my design for a NTC thermistor.