There are 2 theories you can use.
1) Any of the PPS+ pools, due to the nature of what you are trying to do in terms of not always having them running due to power then you don't want them on a PPLNS with ramp up and ramp down times. Going to be way to variable. You are better off with one with close stratum nodes but in the end that probably matters very little.
2) Point them towards a solo pool and hope for the winning ticket. But the odds really are that you will never find a block and never get any payouts. So although good as a project it's never going to get you any money.
-Dave
I agree that PPLNS most probably would be a waste of time and energy
Braiins wouldn't be too bad, however, I haven't figured out how to get detailed information about individual miner's activities. One solution would be to produce logs for cgminer and send them to me all couple of months, but this seems to be too much work unless I invest a lot of time for writing an analyzer tool.
Solo pool is not an option. If someone has one of the solar minerchargers, he should get an incentive. A lottery ticket is imho not the right way to go...
Never worked with Luxor pool before, testing it out and they have a nice API.
At the end of the day: Yes, there will be some Sats per user. But the idea is that a minercharger can further help to decentralise. Using a gekko-pod, for example, produces about 1.7 Th/s, which is some cents per day. And more people will get into Bitcoin - which would be the nicest side-effect ever.
Just an example: My tractor's alternator broke down. Instead of trying to organize a new one, I charge the battery now through a solar panel which I use as a roof. Once the battery is charged, the "tractor" starts mining. In a typical summer, this produces about 100'000 Sats. Not a lot, I know. But it is like an alternative form of staking sats