2
\$\begingroup\$

I am exploring a project which would require stationing several sensor packages outdoors in a semi-protected environment. I am wondering wether adding a simple, low-power heater circuit to the package might improve its robustness in winter time. Basically, the idea would be to keep the package warmed within its spec'd operating range, and also possibly to evaporate any dew or other condensation that might otherwise form inside the package.

Is this a reasonable idea? What kind of heating element would be suitable for this application?

\$\endgroup\$
  • \$\begingroup\$ what about dipping the complete and entire end product in high weather resistant epoxy, and let the temp sensor be exposed. It would stop water damage, protect from rapid and harsh temp changes.. but ultimately would not stop it from becoming the same as the environment temp. If its low power device it should be OK like this. \$\endgroup\$ – Piotr Kula Dec 1 '11 at 16:52
  • \$\begingroup\$ The component I am most concerned about is a Parallax PING))) sensor. This is an acoustic ranging device and thus it cannot simply be sealed in epoxy to protect it from the elements. It is also not calibrated for temperature variations, although that's probably not critical for my application. I'm most concerned about moisture from condensation ruining the device. \$\endgroup\$ – Kaelin Colclasure Dec 1 '11 at 18:31
  • \$\begingroup\$ I saw once a dip that sealed electrical equip from water damage, eg notebooks,mobile phones; dunked in water for hours- removed and worked for days.. i cant remember the name but it could be worth a search.. \$\endgroup\$ – Piotr Kula Dec 1 '11 at 23:08
4
\$\begingroup\$

A heater is a pretty drastic solution, although there are appropriate applications for heaters.

What minimum ambient temperature must your device operate at? Just about all electronic devices are rated down to 0C. -40C versions are usually available. If that's good enough, I would stop there and not bother with a heater.

How is this device powered? A heater will consume significant power. This is where thermal insulation helps. The better insulated it is, the less power will be required to maintain temperature in the cold. Of course, this works the reverse in hot conditions. How much does your device dissipate on its own? What maximum ambient temperature must it operate in? You don't want to add so much insulation that it gets too hot in the summer.

If you just want to keep condensation away, then perhaps the normal power used by the device is enough. Again, we need specs on your device.

If you do go with active heating, plain old resistors are the simplest. Place a few resistors around the board, and turn them on when the temperature gets below some setpoint. A thermistor on the board would be a good idea so that the heater is not turned on needlessly. Simple on/off control based on a thremistor threshold will oscillate a bit around the set point, but as long as the heaters aren't too overpowered that will be fine.

\$\endgroup\$
  • \$\begingroup\$ Each package will be powered by PoE, so battery life is not a factor. I do still have a relatively limited budget of current to work with, though… And of course I do not want to build a needlessly inefficient device. \$\endgroup\$ – Kaelin Colclasure Dec 1 '11 at 18:34
  • \$\begingroup\$ The prototype currently consists of a Freetronics Eleven controller board, ethernet shield (w/ 802.3af regulator), and a Parallax PING))) sensor. \$\endgroup\$ – Kaelin Colclasure Dec 1 '11 at 18:42
  • \$\begingroup\$ @Kaelin: If I remember right, POE is limited to around 13W. There will be some inefficiency and of course your circuit will use some. How much is left over? With reasonable insolation, even 5W of heat should be able to raise the temperature significantly, assuming a physically small device. We need more parameters on your device. \$\endgroup\$ – Olin Lathrop Dec 1 '11 at 19:31
  • \$\begingroup\$ The Freetronics shield / regulator is spec'd for a maximum of 12W, but I am hoping to remain substantially below that. The prototype device uses less than a watt in standby and can momentarily draw ~3W when active. \$\endgroup\$ – Kaelin Colclasure Dec 1 '11 at 20:19
  • \$\begingroup\$ @Kealin: So dig out the specs and tell us the answer. This is your problem, so that's your job. \$\endgroup\$ – Olin Lathrop Dec 1 '11 at 20:21
1
\$\begingroup\$

Low-power is difficult to do with any heater but your best bet would be a power resistor or a large gauge piece of wire with about 4 ohms resistance and enough wattage rating to handle the load you'll place on it. The load will be dependent on the voltage you're putting through the resistor - Power (W) = I^2 * R of course.

If you want to save power you can avoid one high power resistor to heat the entire enclosure and instead focus on placing lower-power resistors next to sensitive components (if there are specific sensitive components). Otherwise, low power options include a controller that keeps it off except for specific times, or a temperature feedback controlled current source that varies the current through the resistor and therefore the heat it produces to maintain a set point.

If your application is running off of batteries it's probably not a reasonable idea at all. You'll be pulling (conservatively) an amp or so from your supply. You'll be lucky to find a reasonably sized battery with more than single-digit of amp-hour capacity. So if you're lucky you'd find a 5V 5AH battery which would heat your enclosure for 5 hours and then die. Probably less. Heating is expensive - period.

\$\endgroup\$
1
\$\begingroup\$

Most electronics can be made temperature happy down to say30 degrees celsius below freezing, and say -50 is doable. The item which gets very very affected by temperature is the battery. The best batteries readily available will usually operate to say -30C but capacity as a portion of nominal is very low - maybe 25%, maybe even less, at -30C.

As noted, heating is expensive energy wise. If the item is battery powered then you can "afford" heating which increases energy availability in excess of the amount lost to heating. That's often not a lot of available energy to attempt to achieve magic heating with. You can work out how feasible this is by working out how much wattage a battery gains by heating and the working out what sort of thermal insulation is needed to achieve this with the available energy.

Example. 4 x AA NimH 2000 mAh =~ 9 Watt hours at full capacity.
Say capacity falls to 25% at -30C but is 100% at 10C.
Gain is 9 Wh x 75% = ~ 7 Wh gained.
Say we want to achieve a 1 week battery life and that the 2 Wh is adequate for driving the sensors etc. Energy available for heater is 7 Watt-hour / 168 hrs/week =~ 40 milliWatt. You'd be lucky if 40 mW would heat the skin on a rice pudding. You can work out how much insulation you'd need to reduce heatlosses to 40 mW for a temperature differential of 40 C (-30 to +10) BUT it's an immense amount of insulation. Assume instead 1 day holdup and you get 280 mW - enough to noticeably heat a small resistor or small IC package. component. Still a massive amount of insulation needed!

\$\endgroup\$
0
\$\begingroup\$

Resistance wire is what you want. This is used in (at least older) space headers. But make sure you have a fail-safe.

\$\endgroup\$
0
\$\begingroup\$

I worked on a system that had LCD display and a PC on top of vehicles. We had heaters/resistors to get use above 0deg C before booting the PC. Once the PC and backlights were on, heat was plentiful and the summer problem.

\$\endgroup\$

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.