Accurate Temperature Sensor

[Soda Ant]

I was planning to use an LM75 for a temperature sensor for a weather station I'm building until I read the datasheet and found that its accuracy is only +/- 2C.

Is there another sensor I can use that's accurate to +/- 0.5C in the temperature range -10C to 50C?

 

[Mike_2545]

Dallas semiconductor makes several good choices for accurate temperature measurement.

 

[dknguyen]

Why do you need that accuracy?

Maxim (aka Dallas Semiconductor) carries single chip silicon-based temperature sensors accurate to 0.5C:
The DS600 is the only analog output device they have with that accuracy:
http://www.maxim-ic.com/quick_view2.cfm/qv_pk/4253

THe rest are digital interface such as the DS620 or DS7505:
http://para.maxim-ic.com/en/search....M|Period µ Temp.|Pulse Delay&tree=master

Needless to say, you could do the same thing with an RTD (with a much greater stability and accuracy over a much larger temperature range) at greatly increased cost, complexity, and hard to find parts.

 

[Dean Huster]

Why do you need that accuracy?

If I were building a weather station that I wanted to be taken seriously, I'd want that kind of accuracy. Same for wind speed and rainfall.

Dean

 

[Russ Hensel]

if you have a computer involved you can add a calibrate function to that program. this avoids the additional expense of more acurate sensors

 

[Diver300]

http://www.electro-tech-online.com/custompdfs/2009/05/LM92.pdf is +/- 0.33 °C accurate

 

[bigfarmerdave]

Texas Instruments TMP112 is rated from -40 to +125 and I believe is rated at .5deg C over the entire range.

 

[KMoffett]

http://www.electro-tech-online.com/custompdfs/2009/05/LM92-1.pdf is +/- 0.33 °C accurate

Not exactly. Check the chart in Note 9 on page 6.

Ken

 

[Mr RB]

The dallas 1wire DS18b20 has 1/16 'C resolution, and you can calibrate it in software.