If this has already been ask I apologize as I cannot find it using the search option, please send link.

For my first project I'm thinking of building a monitor, in stages, for my outdoor pond (about 30 feet from the house) with wireless data transmission for various conditions like water level, waterfall pump failure, icing, and maybe even the occasional Heron that likes to feed on my Koi. I'm also leaning heavily towards the Netduino board as I would like to leverage my C# skills.

Anyhow, when I look at the specification for the Netduino board, it specifies an operating temperature of 0 to 70C. Right now it's -8C outside and will probably go to -13C tonight. The coldest I ever remember it being here (Rochester NY) is maybe about -28C (-18F maybe -20F) but that is very very rare.

The Arduino site lists no specific temperature specification, but links to the Atmel specification.

The Atmel processors (ATmega328 and AT91SAM7X512) for both both boards are rated from -40C to +85C; well withing my requirements.

So the question is, can I also assume the Arduino board is also limited to some range less than the microcontroller? What board factors might limit the temperature range for such a project?


  • \$\begingroup\$ At a range of 30 feet... you could run an extention chord out there and run a heater in an enclosure with the 'duino j/k :) \$\endgroup\$
    – vicatcu
    Dec 15, 2010 at 20:52
  • \$\begingroup\$ Actually my second question would have been if it's worth it to have the board monitor and maintain it's own temperature. I do have 120 at the pond on the pole I would mount the monitor to. Thanks - good suggestion. \$\endgroup\$
    – ddm
    Dec 15, 2010 at 21:12
  • \$\begingroup\$ I have a Davis Instruments (davisnet.com) soil temp sensor with one probe at the bottom of my pond and another in shallower water. I have a USB interface on a computer in the basement that logs the data and updates a web page. I've been thinking about setting up a water "too low" sensor that would manipulate a 3rd input on the soil temp station. I like the idea of detecting pump failure as well. \$\endgroup\$
    – TomG
    Dec 16, 2010 at 3:06
  • \$\begingroup\$ Very cool. I could run cables to the pond from a basement Microcontroller, then make available via webpage for monitoring. The big problem in winter is detecting icing as I need to plug in the horse trough heater (that if left too long heats the pond too much and activates the Koi, goldfish and maybe the frogs in February at -5C). I've run 120, why not underground cables as well. Thanks. \$\endgroup\$
    – ddm
    Dec 16, 2010 at 22:48

3 Answers 3


Well, my first guess is that temperature operation would be limited by the oscillator. Now, I'm flying completely by the seat of my pants here, as I haven't looked at the parts on the Arduino. If you pull the specifications on the parts, my guess is that the oscillator is the part with the most limited range. Typically, the bigger the temperature range, the more expensive the oscillator. The other factor that may come into play is the frequency offset.

As the temperature shifts, the oscillator frequency will shift as well. For a wireless system, this may mean that the transmitter/receiver frequency will shift as well, possibly enough to make the range decrease.



If you can't identify the exact parts used on the Arduino board, you aren't going to be able to verify that the Arduino/Netduino board is rated for those temperatures. This would apply to everything in your system, from the water level sensor to the wireless transmitter. Some parts are harder to find based on markings. I wouldn't worry too much about resistors and capacitors.

Most manufacturers won't tell you what happens when you exceed the temperature ratings. I'd recommend you take your system and put it in a cooler full with dry ice, and see what happens.

  • \$\begingroup\$ Most electrolytic caps suck near and below 0C. \$\endgroup\$
    – tyblu
    Dec 15, 2010 at 22:09
  • \$\begingroup\$ From all the responses I think I need to move the controller inside, the sensor cabled outside. Good point - thanks. \$\endgroup\$
    – ddm
    Dec 16, 2010 at 22:49

Most electronics will still operate at temperatures lower then their rating, it's just that they may drift outside their specifications. Wireless communications would be the only thing that would be fairly sensitive to this; the tolerance on most wired embedded communications is fairly loose.

What I would be more concerned of besides temperature is condensation on the board. If you put the Arduino into a relatively enclosed box in your warm house then take it into the freezing cold, all the excess moisture will come out, possibly onto your board. Despite me being "more concerned" about this, my level of concern would still be very low unless you're doing high precision analog readings or the like.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.