Which battery technology for low temperatures is the best right now?

[Johannsen]

Hi,
I need for a small battery powered outdoor device a power source that works also reliable at -15°C for at least >1month. The usual current is around 1mA@3.3V.
I'm using a step-down converter for getting the 3.3V.
So far I considered eneloop batteries, Energizer Ultimate Lithium, Li-Ion, LiFePO4.
I read that LiFePO4 has a good discharge curve but struggles wit low temperatures.
What technology would you use these days?

[beanflying]

Eneloops SUCK in spite of their claims of -20C along with most NiMh batteries their impedance goes way up. Sounds like you are only running very low power so this may not be an issue. In R/C use at just on or below zero C Eneloops nearly cost me a $2k glider into the dirt as the digital servos loaded up during a launch, the glider lost control and luckily flicked off the winch high enough for it to get control back. Since then I have gone back to lower impedance Sanyos with zero issues (other than the higher self discharge).

Typically Lithium batteries faired even worse but a quick Googling got this https://www.grepow.com/page/low-temperature-battery.html There should be plenty of Avionics/space type of data out there as it is now a common use case for them.

The general rule of batteries is they like to be charged and used warm in the 10-40C range. Above that has benefits for high discharge use as the impedance goes down. For lower temperatures there is a reason battery warmers are a thing.

Random other thoughts are watch out for condensation with your Battery and Electronics package. There is some threads around here on that so a search might pay.

[Phoenix]

Lithium thionyl chloride is an extreme temperature and low self-drain chemistry, but it is only primary cell. For a single cell the voltage might be a bit low at low temps though.

https://au.element14.com/eve/er14250w/battery-lithium-wired-1-2aa/dp/1365942?st=lithium%20thionyl%20chloride

[Siwastaja]

Li-ion is fine. Any chemistry, modern NCA cells in 18650 form factor, or LiFePO4.

Primary lithium is also fine, so is alkaline.

Internal resistance is just higher at cold temperatures. This means voltage (therefore efficiency!) drops under heavy load.

But 1mA is nothing. Efficiency is like 99.9% at room temp and it's no big deal if it drops to 99.0% at -15degC. You can't even easily find small enough cell not to be able to produce 1mA at -15degC.

Capacity does not drop (except by the mechanism that you reach low-voltage cutoff earlier due to I^2 * R drop, leaving unused capacity). But at I=1mA, this isn't a problem so you can expect near full capacity.

Also you can expect great calendar life as the cell aging mechanicms basically stop.

The only limitation is that charging at -15degC might be damaging to the cell, even at low currents (or, at least, officially unsupported), but I'm assuming you are taking the empty cells with you to be charged elsewhere.

[Johannsen]

Thanks a lot for your answers! Yes I'm taking the cells with me for charging. In future solar charging could be handy..but first step by step. I'm glad to hear that this low current use is obviously not that critical for various battery technologies.
I checked that grepow-page but it seems they only offer custom designs or bulkier packages. I just need a couple if them for a reasonable price.
In general if more current is needed, let's say around 500mA. Which technology would be superior? Or is this already a case where  battery warmer are needed?

Thanks for the info with condensation, I'll search here for threads. 

[Siwastaja]

Well, -15degC isn't that cold so I bet the best general idea is to use whatever cell type is available at best cost / energy density / etc., and then just increase the capacity of the pack so that the current rating goes up too, so that it can then take derating needed for cold temperatures.

For example, a single $2 3Ah 18650 cell can normally supply some 1.5A with good efficiency and without dropping too much voltage; or up to 3A with still acceptable efficiency in most cases. This is at room temperature.

At -15degC, that could drop to like one fifth or so.

If you expect both cell aging and low temperatures and want to add some margin against uncertainty, it's not a bad idea to derate the current rating to one tenth. So maybe you would use the 18650 cell up to 150-300mA. Then if you need 500mA, just use 2 or 3 in parallel! You obviously get more capacity; if you don't need it, then it's just an unnecessary side effect but likely not a bad one.

Maybe if you go to -50degC then you might want to look at special unicorn products. Maybe. But they also have special unicorn prices.

[Peabody]

I looked into this a while back for a mailbox notifier, and concluded that the Energizer Ultimate Lithium batteries had the best low-temperature performance of any widely available type.   But my research wasn't really all that thorough.  Anyway, I ended up settling on three AAs in series directly powering a Wemos/Lolin D1 Mini Pro.  I later gave away that project, but the guy who ended up with it said it survived the big freeze of last winter, and he hasn't changed the batteries.  As I remember, those batteries would normally last a number of years.  Everything was powered down via a mosfet all the time except for twice a day for a few seconds each, once when the mailman delivered the mail, and once when I retrieved it.

I think the big problem with cold weather is trying to recharge batteries.  If you look into Cubesat projects, you'll see they provide for battery heaters.  Actually, Cubesat solar power systems are really interesting.  You're in bright sunlight for 45 minutes, then in total darkness for 45 minutes.  So during the 45 minutes of sun, the panels have to provide enough power to reheat and recharge the batteries, and power the load, and there's some rule of thumb that you can't discharge the batteries at "night" more than 20% or so.

Anyway, if you aren't recharging, it probably doesn't matter all that much which type you use as long as you over-provision a bit on current.