So I have this crazy idea of taking 3 AAs and connecting them directly to the Photon (in series), using something like this 3xAA holder. I have some rechargeable NiMh AA batteries that I’m not using at the moment.
Thing is, I’m totally new at this. Is it a bad idea? In theory 3x1.5 gives me 4.5V, right in the range for the Photon. Everyone seems to be using fancy LiPo battery packs though, are they better in some way? Am I crazy for trying to use what I already have handy?
You can but they probably wont last more than a day under continuous Wifi connection 
Thanks for the quick reply! I’m glad it can work, at least to a point.
I’ll have to experiment: for the application I’m thinking about I don’t need to be connected all the time so it may still be workable. I can try to put the chip to sleep for a bit in between sensing, and only turning on wifi when there’s something to report.
Also there is new power shield coming up, which should have more regular shape. It allows to charge the battery and power the device at the same moment. There is even battery fuel gauge. It doesn’t last any longer of course, less with included 400 mAh battery.
3 standard batteries will drop below the 3.6V minimum VIN pretty soon, you need to boost it or use 4 to get full lifetime out of them, but 4 is over the max for VIN so need a buck converter.
However they likely wont last as long as you want, I did a test with a UDP packet being sent every 2 minutes and deep sleeping the rest, I got a week out of 3xAAA.
The power usage of photon is just so much larger than an 8bit mcu, sitting idle with wifi on is 80ma+, sleep currents are ~90ua.
At ~1000mA you get 12.5hours of running (45000seconds) not counting sleep losses.
Say you use 5 seconds every 2 minutes, that gives 90000wakeups @ 2minutes = 18000 minutes = 12.5days.
Outside with a small solar cell maybe, but otherwise it does not seem like a good match for battery power.
Let’s dig deeper and plug in the numbers for my scenario.
I’m thinking about 1m of operation per 5min, without wifi. Send a summary message once a day.
The data sheet lists:
deep sleep: 80ua
operating, no wifi: 30mA.
operating with wifi: 80mA.
So the amount of energy used in a day for my scenario would be roughly
(243600 * 4/5 * 80e-6)+ (243600 *1/5 * 30e-3)+ (60 * 80e-3)
= 529As
each AA battery claims 2000mAh. Let’s say, conservatively, that in reality the whole pack together only deliver 2Ah before the voltage drops too low.
2Ah = 7200As
at 529As per day of usage, that means that (if my math is correct) the gizmo would last 13 days. Just like in your scenario!
Indeed, that’s not very good. We’ve been conservative, so we might get as much as 4x as much. Still, that’s more battery changing than I might have hoped.
Oh right rechargeable batteries can give a bit more than the 1000-1200mah alkaline.
But the power usage of photon is based on vin, so 3.6V at least, so in rough math you only get the mah rating of one battery since you need 3 in series to get the voltage.
We may be able to cut a few seconds from each wakeup by using wifi.on wifi.connect, instead of connecting to cloud, and once the queued firmware is here, every x wakeup could connect to the cloud.
-That only works if you dont use the cloud though.
It’s better than an Arduino with ESP8266 Wifi chip attached though, so with similar connectivity options it’s still pretty good. I would love a lower power no WiFi mode though and I bet that Coretex M+ still has some untapped sleep modes that won’t lose SRAM data either. I have measured running from USB at 5V and see about 360 mW with WiFi and about 220 mW without.
Right, @MORA. The number I got is assuming I only connect to the cloud once per day (trying to stretch the battery life). So using Spark.disconnect() and wifi.off() most of the time.
@spacenick88, noted. Plus the cloud platform here is really nice so I don’t think I’d even consider going with an Arduino. Still, changing batteries every month is no fun.
The STM32F has several sleep modes which can wake after a time or a pin event. Furthermore, it seems the broadcom wifi chip may have a low power wifi mode which has not been tapped yet though this may be very dependent on the (not yet open) WICED library. If you want REALLY low power usage and Particle cloud compatibility, you may want to consider Bluz.
@peekay123, wow, that Bluz looks like it’s going to be very cool! I didn’t know BLE reached that far (60ft). I’m not entirely clear on the Gateway, though. Is that a self-contained BLE<->Wifi bridge, or does it need to plug into a computer running specific software? If so, do you know if it’ll work on Linux?
@cyan, the gateway has a BLE module and a Photon or Core on it and is a full Internet and/or Particle Cloud gateway. A cell phone can also be used. The gateway can have up to 8 Bluz connected to it and the Core/Photon can do more than just run the gateway code. 
You can also get something like this http://www.exp-tech.de/seeed-studio-lipo-rider-pro
(these can also be found on Aliexpress) or http://www.adafruit.com/products/390 which are both available now and
should work with the Photon. I have tested the Chinese version of the first one and it doesn’t drop power when unplugging the battery but keeping USB or pulling USB and keeping the battery and the same should be true for solar as well. I will
be testing this one with a USB solar module in the coming weeks too.
Looking forward to bluz too, although the limit of 8 per gateway is a bit odd, not sure what the actual reason for the limit is.
Maybe it wont be an issue since BLE is somewhat short range anyway.
I used NRF28 on “breadboard arduino” before, 3xAAA gave me 6months of lifetime, but debugging radio issues becomes annoying fast.
@peekay123
I dont think the WiFi powersave will help much in these setups, AFAIK wifi powersave is a way to generally use less power at the cost of latency so useful for devices that need to run continously and can live with 250ms latency, but still need to preserve power.
@spacenick88, most of those chargers only work on single-cell batteries, no? That means about 2500mAh, not much more than the rechargeable AA batteries I have. Solar isn’t an option in my case, but it’s nice to see this technology is progressing!