I'm building a small, low-power device that will be used inside a standard consumer refrigerator (approximately 2°C). I've had trouble with Alkaline batteries at low temperatures, and was wondering if anyone has any recommendations for cells to use under these conditions. How do Lithium coin cells hold up, for instance? The datasheets I've found for them don't seem to take temperature into account.

Update: it looks like I'll have to source ~50mA for short bursts, so the low-end lithium manganese dioxide cells I've been using aren't going to cut it. Any other suggestions?

  • \$\begingroup\$ Has this worked out for you? Let me know how things go! \$\endgroup\$
    – Kortuk
    Jun 9 '10 at 14:41
  • \$\begingroup\$ I'm close to finding out-- most of the rest of the device is running, so I'll be testing this in a fridge soon. I'll let you know! \$\endgroup\$
    – phooky
    Jun 12 '10 at 1:55
  • \$\begingroup\$ Just don't forget, if you have problems I may be able to give advice. \$\endgroup\$
    – Kortuk
    Jun 25 '10 at 19:55
  • \$\begingroup\$ Thanks; I'm pretty close. The xbees I'm using are having some severe reliability issues, but it's not temperature-related. \$\endgroup\$
    – phooky
    Jul 5 '10 at 19:55
  • \$\begingroup\$ 50 mA is not low power. \$\endgroup\$ Jul 15 '10 at 2:58

I would suggest a lithium CR2032 battery. feel free to pick a large package or place a few in parallel.

Lithium Polymer will have lower capacity at lower temperatures.

However, to quote battery university,

"Lithium-ion works within the discharge temperature limits of -20°C to 60°C (-4°F to 140°F)."

As a side note, they often list 1mA as maximum current on one of these, but they can handle 10mA without much of a battery life loss(this will be worse at lower temps), pass 10mA and you will be killing batteries right and left.

Here is an example of a discharge curve based on temperature. It is the second page of the data-sheet. There is a clear decrease in life(~25%), but you should still be very able to use the battery. It looks like -20 degrees C is where you start to have large reductions.

Please let me know if there is anything I can add to make this more clear or more valuable to you.

  • \$\begingroup\$ Sounds ideal! I have a few CR2032s hanging about; I'll give them a whirl. \$\endgroup\$
    – phooky
    May 28 '10 at 21:13
  • \$\begingroup\$ Just watch their voltage, and remember to always measure a batteries voltage under load to check how much life is left. \$\endgroup\$
    – Kortuk
    May 28 '10 at 21:25
  • \$\begingroup\$ Well, it turns out the lithium cell is fine for my instrumentation amplifier and microcontroller, but the xbee I'm using draws far too much current. I'll see what else is out there. \$\endgroup\$
    – phooky
    Jul 15 '10 at 1:34
  • \$\begingroup\$ Haha. I am sorry, i had no idea you were using an X-Bee, I could have told you that would be too much. If you place about 10 of the coin cells in parallel it will be fine for a while... \$\endgroup\$
    – Kortuk
    Jul 15 '10 at 13:49
  • \$\begingroup\$ You can get much more powerful lithium batteries, such as the CR123A which used to be popular in cameras. \$\endgroup\$
    – Simon B
    Jan 20 '16 at 11:19

I would not recommend the lithium batteries for low temperature applications. They suffer badly from low temperature, more than either the alkaline or NiMH chemistries.

I'd recommend Nickel-Cadmium (Ni-Cad) batteries, as they have have the best temperature range of any of the typical batteries, They also have very low internal resistance (less than half that of a NiMH) and therefore have very good maximum currents (think of an electric drill).

Some things to watch out for are their memory characteristic (Don't recharge it before it's almost dead, or you'll lose the capacity you don't use) and the fact that they contain toxic cadmium. Use a decent charger or charge IC, and you'll be fine. You might have some trouble trying to do a dT(emp)/dt(ime) cuttoff, and they don't have a great -dV/dt characteristic either.

Another idea (if you own the fridge) - Can you just open up the temperature control panel and light at the back of the fridge, get at the power source, say, for the light (before the switch, if you want it to be running when you close the fridge), and stick in an outlet or a couple (INSULATED! WELL INSULATED!) spade terminals and plug in a wall wart? That's what I'd do, rather than messing with batteries.


If your device works continuously, don't worry about the performance of lithium batteries at charging and discharging. they just have a lower capacity which you can check in their data sheets. But if you using battery some in a while and you will have a voltage drop in the start because of passivation. The only way to avoid this problem is to take a higher current in start, for example, with a resistor, and then continue with regular condition.


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.