3
\$\begingroup\$

I intend to use an Arduino in an environment with highly variable weather. It will be an outdoor project staying outside during the entire year with temperatures ranging from a maximum high of approximately 37C (~100F) to a maximum low of approximately ~35C (-31F). It is unlikely that these temperatures will actually be reached, but it isn't unheard of. The temperature will most likely fluctuate between -15C (5F) and 30C (86F).

I have read that the ATmega32u4 and XBee 1mW U.FL (which I also intend to use) datasheets that both of these should operate properly between -40 and 85 degrees Celsius. Does this mean that the Arduino itself should be able to carry on normal operation? (if it runs a bit slow, that is fine, as long as it functions normally)

Also, I am concerned about condensation. I intend to use the following enclosure for the project http://www.adafruit.com/products/905 and will need to drill a few holes in it to get wires in and out. Will I get condensation in the enclosure? How big a problem is this? Is there an easy way to prevent it? Also, I'm curious whether the holes I drill for the cables will be problematic? Does anybody have any tips on sealing them?

\$\endgroup\$
7
  • \$\begingroup\$ With reference to comments about incorporating a heater: If the only challenge was cold temperature operation, your circuit itself is a heater. Wrap it up well with several tight layers of glass wool, foam polystyrene or some other very good insulator, then put it in the enclosure, and the insulation will keep the heat inside and ensure that the board stays warm as long as some power consumption occurs. If you insulate it too well, you might actually end up with the opposite problem, too much heat! Now, at the high temperature end of the operating range, you have a different problem :-) \$\endgroup\$
    – Anindo Ghosh
    Commented Nov 12, 2013 at 11:15
  • \$\begingroup\$ Thanks for the advice! I wonder if it might be possible to do bi-annual maintenance and just pop the wool in and take it out... Perhaps. \$\endgroup\$
    – raphaeltm
    Commented Nov 12, 2013 at 12:46
  • \$\begingroup\$ I have been involved with one project where the entire circuit board is wrapped in glass wool, epoxy sealed, placed inside a steel canister pressure-sealed with a brass (I think) cap. The deployment is cryogenic, and even an annual maintenance cycle would be unthinkable. The thickness of the control cable's conductors had to be empirically determined, such that the heat extracted via the conductors would just about balance out the heat generated within the ultra-low-power circuit inside. \$\endgroup\$
    – Anindo Ghosh
    Commented Nov 12, 2013 at 12:59
  • \$\begingroup\$ Wow! That sounds like a pretty intense project! Fortunately, what I'm working on does not require much precision at all. It would likely simply be a question of popping open a fairly standard enclosure and closing it again. \$\endgroup\$
    – raphaeltm
    Commented Nov 12, 2013 at 13:20
  • 1
    \$\begingroup\$ @TimH those are big packets! They might not even fit in a RaspberryPi housing. Smaller packets would be much cheaper. \$\endgroup\$
    – Jeanne Pindar
    Commented Jul 14, 2018 at 16:26

4 Answers 4

Reset to default
2
\$\begingroup\$

Take a look at this crazy Arduino experiment with liquid nitrogen http://3.14.by/en/read/arduino-liquid-nitrogen-overclocking Although this is not exactly what you asked for, nevertheless you could take some ideas from the conclusions at the end of the article.

\$\endgroup\$
4
  • 2
    \$\begingroup\$ The most relevant part being the caps dramatic changing of capacitance under 0°C. \$\endgroup\$
    – Passerby
    Commented Nov 12, 2013 at 23:38
  • \$\begingroup\$ Do not forget about BOD failure, this means analog reference is temperature dependent too. I would not rely on ADC functionality either if using internal reference in given temperature range. \$\endgroup\$
    – x4mer
    Commented Nov 13, 2013 at 6:18
  • \$\begingroup\$ I think what I'm getting from all of this is that my best bet is to keep my enclosure heated. Would you agree? Any ideas on preventing or limiting condensation? \$\endgroup\$
    – raphaeltm
    Commented Nov 14, 2013 at 16:50
  • \$\begingroup\$ Definitely. Most outdoor video surveillance cameras have internal heating of the camera box to prevent condensation on the front glass. And they are quite ok without sealed connectors, usual cable glands suffice. I suggest also to cover the board with insulation varnish. Although cameras inside these boxes have normal PCBs without have any kind of varnish covering. IHMO you will not need underwater class sealing. Rain and snow box cover, cable glands with rubber inlets, heating and varnish will do. \$\endgroup\$
    – x4mer
    Commented Nov 15, 2013 at 15:17
0
\$\begingroup\$

Feeding cables

For your holes, there are these push-out-holes in your selected box that fit a cable gland to feed a cable that have a rubber seal. You can only feed a single cable per gland, because the seal is round. These glands are usually available at any DIY or hardware store.

Temperature and humidity

I don't know about the rated temperature for Arduino, but for a Dutch rail road project we used to mount a heating element and thermostat under the equipment that was rated for 5 degrees and higher. That would keep the device just warm enough and dry. I think the biggest issue will be condensation.

Are you sure overheating will not be an issue?

\$\endgroup\$
3
  • \$\begingroup\$ Thanks for the info on feeding cables. I hadn't realised that. I hadn't thought of overheating. I assumed that since both parts seem to be rated for 85 C, that wouldn't be an issue. Do you think it's something I should be more weary of? \$\endgroup\$
    – raphaeltm
    Commented Nov 12, 2013 at 7:30
  • \$\begingroup\$ I'm also curious what sort of element you used to keep your equipment warm. Did you use something small? \$\endgroup\$
    – raphaeltm
    Commented Nov 12, 2013 at 8:05
  • \$\begingroup\$ Power resistors (the large wire-wound kind) are good for heating. Don't forget some sort of thermostatic control to turn it on and off, could be from the Arduino but better to do it failsafe in the electronics. If it's battery powered heaters may not be an option. \$\endgroup\$
    – pjc50
    Commented Nov 12, 2013 at 15:08
0
\$\begingroup\$

I don't know on what kind of budget you are running, but I also think that condensation will be the biggest issue. I worked on a submarine project where we had to seal everything. In order to acheive that we used sealed connector, like these ones. This way you could seal the interface with your box and have a closed environement inside with close to no humidity in and none entering. You could then isolate a bit the board to make sure the temperature stays within spec.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks. That looks like it could be very convenient. When I got the response from @jippie about cable glands, that seemed like it would be a good solution, but this looks even better. \$\endgroup\$
    – raphaeltm
    Commented Nov 14, 2013 at 5:49
0
\$\begingroup\$

Assuming a Arduino Leonardo which has the 32u4, you have to keep in mind that every part of the board, from the pcb and SOLDER, to the passives, to the connectors, to the ICs have operating temperature ranges. Most are not designed for subzero operation (-18°C/0°F)

A cursory look at datasheets of the Leonardo parts shows that the only IC that won't reach past -45°C is the NCP1117ST50T3G, the 5v Regulator. It has a minimum operating temperature of 0°C to a max of 125°C. If you bypass the 5v regulator by using a dedicated 5v charger, you have to worry about that 5v charger's circuit working at such low temperatures.

The PCB itself, at -30°C, should avoid any unneeded shocks. And to quote a comment from What is the minimum temperature for FR-4 PCBs?

It's not what you're asking, but...In addition to looking at the operating temperature range of each part, remember to look at how they interact. Different materials have different coefficients of thermal expansion (CTE), and parts with CTE much different from that of FR4 tend to be the ones that fail (either break or pop off the board) due to temperature swings. Large ceramic parts (like some crystal oscillators) are the most common source of problems in my experience.

In fact, the crystal used on the Leonardo, a SMD KX-7 16mhz crystal, not knowing the specific manufacturer, seems to show multiple manufacturers citing it at -20°C minimum temperature. And if the parts heat up from use while the board stays colder, it can cause the parts to snap themselves off. I recommend reading that other question for some more details.

Common passive capacitors will experience a negative capacitance shift that will make them useless in most circuits.

As for the enclosure, you may want to look into ones designed for sub-zero temperatures, with heating.

\$\endgroup\$
3
  • \$\begingroup\$ Thanks for the informative response. Do you think some of these problems could be avoided by enclosing some sort of heating element inside the enclosure? Do you have any recommendations for enclosures? Here is one I have found which the site suggests is "sub zero," but I can only find a maximum temperature in the datasheet. l-com.com/… \$\endgroup\$
    – raphaeltm
    Commented Nov 12, 2013 at 7:57
  • \$\begingroup\$ Also, since you seem to have good insights, I'm curious what your opinion is on the question of condensation? Is there any way I can prevent it? \$\endgroup\$
    – raphaeltm
    Commented Nov 12, 2013 at 8:03
  • 2
    \$\begingroup\$ Passive components also have temperature limits. Especially electrolytic caps and Y5V ceramic capacitors. Crystal will have significant frequency shift at negative temperatures. Heating it to positive temperature is definitely the simplest way. \$\endgroup\$
    – BarsMonster
    Commented Nov 12, 2013 at 23:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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