A gel battery would probably be an even better sollution The one I use(d) is an old one (label has come off, dunno part number), which is similar to this one: https://www.amazon.co.uk/YTX16-BS-Exide ... B00AO2S36C I do not believe that gasses are an issue with this type of closed (maintenance free) battery.
That is NOT a battery, that is a USB power bank which contains a 2000mAH LiPo battery, charge control circuitry and an output regulator circuit to provide a steady 5VDC that meets USB voltage specs. The confusion is created because manufacturers want to claim a high mAH rating but advertising rules in many countries say you can't lie in specifications. With this particular device the output that you can use is fixed at 5V. If they called it a 2000mAH USB power bank it would have to provide 10WH (5V * 2000mAH). It actually has an average voltage at the battery during useful life of 4V which = 8WH (4V * 2000mAH). Converting the batteries average voltage up to 5V also causes some loss of power so even the full 8WH is not available at the output. The amount actually available is dependent on the boost converter design and is AFAIK never specified by power bank manufacturers. If you assume 90% efficiency, not unreasonable, there will only be 7.2WH available.
Raspbian writes a time record periodically and *any* writes have the potential to trigger wear leveling activity on the card itself. So while you will *probably* get away with sudden power drops, over a long enough term, it carries some real risk.kneekoo wrote: ↑Sat Nov 04, 2017 7:43 amThanks for your replies!
@broe23: As this Pi0W rarely has to write on the microSD card, I'm less worried about having it shutdown properly before it runs out of power completely. So when the power comes back, I assume the battery will start the Pi back up. I'd have to test this anyway, but my first thought was if the battery can at least supply the required power for normal operation.
Tool...cheap: https://www.adafruit.com/product/1852 There was a relatively recent post that showed a version of this that displays voltage and current simultaneously, so there are variations that will work.@W. H. Heydt: Besides, I have no tools to measure the power consumption and I wouldn't know how to use them even if I had them.
Paul Hutch wrote: ...
I'm currently experimenting with a generic power bank someone gave me that does pass through power. The second hurdle I've also passed is it responds quickly enough when switching over from pass through mode to battery power so that the RasPi0W doesn't crash. Still have to test out length of back up time and figure out if it recharges while passing through power. Assuming it turns out to be usable I'm thinking I'll put a transistor circuit on the power input side that will signal a GPIO pin when power in is lost. Then the RasPi can time how long power has been off and automatically shut down before the power bank runs out of power.
It works great!