I’m hoping to build a simple outdoor thermometer which will communicate via an xbee. I was thinking I would power the xbee using a solar cell, LiPo battery, and charger. But I recently discovered that one should not charge a LiPo battery at temperatures under 0C/32F. Where I live, temperatures can be below 0C/32F for 2-3 weeks at a time.
Does this mean that LiPo batteries are not a good idea for year-round outdoor applications in a temperate climate?
Or perhaps the xbee can run for weeks (hibernating between transmissions) from a decent-sized LiPo battery without recharging?
More generally, I’m wondering whether it is practical to operate an xbee year-round with solar power, or whether I should instead just run a buried power cable from my house to the thermometer.
LiPo battery charging requires special circuitry and is inconvenient at any temperature. If you don’t care about size and weight, lead acid batteries are extremely forgiving, easy to use and charge and are the way to go for rechargeable DIY projects, especially for a semipermanent installation.
Consider a completely different, extremely low power design that uses a supercapacitor instead of a battery. It is very thoroughly documented here: http://gammon.com.au/forum/?id=12821
Alternatively, use alkaline AA batteries and an extremely low power design. I have remote temperature sensors that run year around on a single AA battery. I replace it about every two years. Gammon has a thorough review of low power approaches with ATMegas on the same web site http://gammon.com.au/power.
Many of the commercial remote weather sensors have been decoded, so you can save yourself a lot of trouble by just using them and read the data with your own equipment.
If you’re recharging the battery, NiCd/NiMH aren’t a particularly good choice, as they are pretty unhappy about low temps; when I had a solar-charged weather station in the Reno, NV area (low temps into the single-digit sub-zero F range), the OEM NiCds did not provide enough power once the temp dropped below around 5 degrees F. I ultimately replaced them with NiMH, which were better but not great.
I’m currently working on (remote) solar-charged sensor nodes, and I’m liking LiFePo4 batteries (in addition to lead-acid gel cells, which I’ve used in the past); much safer than Li-Ion, since they tolerate overcharging better, they handle cold better (typically rated to -20C, some studies have indicated they actually hold charge better at lower temps than at room temps after 50 or so charge cycles), and they are increasingly more available, since they are getting used in solar landscaping lights in the 14500 form factor (AA). Hacking a landscape light to get the solar cell, charge controller and LiFePO4 battery is pretty easy (although not necessarily wildly efficient).
There is LiFePO4 support in several of the common (expensive!) MPPT solar charge controller chips out there (for example, with a little hacking, the Sparkfun Solar Buddy can support LiFePO4). The chips tend to be expensive (6-7$ each in quantity 1 from US suppliers), so I haven’t pursued building my own controller yet.
Biggest problem with them is that they are nominally 3.2V, so to be safe I think you really need to use two to supply a 3.3v project.
I can, but they are changing on an almost daily basis, so you’re probably better off finding them yourself when you’re ready to do the project. Go to Amazon and search for LiFePO4 solar and you’ve got a set of results that are pretty reasonable. The hacked lights may not have enough power for your project; only you can tell that. You can generally find these one banggood.com or aliexpress.com for less than Amazon, but it’s more of a hit-or-miss game there - they frequently don’t list the battery chemistry for the lights you find there.
The suggestion to use sealed Lead Acid is a good one as well; 6v lead acid cells are fairly cheap in the smaller sizes (http://www.amazon.com/gp/product/B00BHM … bw_g263_i6) and would probably be adequate for powering the system you’re talking about, and lead-acid solar chargers are reasonably cheap.
I know that for me, running any kind of wired power to my sensors would be at least $100, even if I did all the work myself. There’s probably no money savings in using the solar/battery approach, but there’s a significant savings in hassles…