MarkT:
The general rule-of-thumb is use PV panels with a voltage of 1.5 x battery voltage.In low light conditions the output voltage drops considerably, and the open circuit output voltage
spec of a panel is only for full tropical sunlight conditions.So for 5V that means 7.5V panel (or more to allow for regulation losses).
The other approach is use a boost-buck converter with MPPT so that the voltages can be different,
and this is what a good solar charger can give you - the MPPT is going to optimize the load the
panel sees.You could stack your panels in series groups of two, and then use a buck converter to drop down.
(remember the extra diodes to prevent reverse driving of shaded panels if you stack in series).
Your last suggestion did finally occur to me. It's so obvious I'm surprised I didn't think of it earlier. Just double the voltage so that voltage never falls below the buck converter's threshold minimum input voltage. I'm sure that there's a tradeoff in this approach, though. Is there?
My inclination though is to go with the MPPT approach even if it's overkill for this application, just because, ideally, it would put out the max output for any given 24 hour period. Or so I'm led to believe. Is the $5 board I linked to above going to do this? It claims to be MPPT but I'm wondering if that's possible at this price point.
This is the board I meant:
MPPT Solar Panel Controller 5A DC-DC Step-down CC/CV Buck Charging Module
Btw power banks have 5V USB inputs, but generally use 3.7v LiPo batteries. So, we're dealing with two stages of power conversion and the unavoidable extra energy loss that comes from that. So it probably makes more sense to forgo a power bank and connect a battery or battery bank to the solar charging circuit directly. Which might itself call for a different charging circuit than the one used to charge a power bank.
The other person was correct in that you have to start with the intended application and design the circuit around that, and choose components accordingly. That's the engineering (i.e. correct or ideal) approach.
But, right now I'm just playing around with solar to get a "feel" for it, and the various challenges it presents and solutions that exist to deal with them. This might seem academic to old hands here, but when you're learning something for the first time, you basically have to reinvent the wheel by way of "experiments" whose outcomes is well-known, for it to all sink in.