(article) I'm not sure what the quality of journal is. You have to scroll down to 4th author to get to the top school in the country.
I'm pretty sure Solar Edge require central coordination unless you buy the unlocked ones. From what I gather the unlocked ones output a fixed voltage, current, or passthrough. I'm pretty sure you loose all the smarts with that "upgrade"
Yes, SolarEdge has some kind of "constant global voltage" control. There's no way an individual optimizer can do that without feedback.
Not sure exactly what you mean by losing all the smarts.
Some tea leaf reading for guessing the IndOp vs Standard algorithm. It teases some hints via describing the different specs/behavior you get from each mode. Instead of just writing up a clean whitepaper on what it is... (maybe my Google-Fu needs to be better).
Interesting tidbits are that the output current (10A) and VoC are lower. The VOC is capped at the PV module VOC, which is good because it means you won't boost the voltage up and kill the string inverter. It does not say that module level MPPT is removed. I don't know if the 3rd parties reprogramming SE are able to force them into boost mode.
For some reason it says minimum 4 optimizers per string. It is probably related to 1V safe mode when off (also on this datasheet)
Tigo works fine with distributed control and buck mode. I think adding their CCA/TAP primarily gives debugging and firmware upgrades.
One very simple way to do distributed control on the SolarEdge would be to always output 10A. If the buck ratio exceeds the range of the converter, bypass. This would be able to scale from 80W to 400W. However, the problem with this approach is that it might easily drop below the start voltage of the connected inverter. A patch to deal with that is for the optimizers to detect this and do a coordinated drop in current to reduce the buck ratio. EG, go to 9A, 8A, ... until the string starts again.
(On Tigos the maximum drop is somewhere in the 20%-33% range, so you can design the string against that to ensure that the string can still start)
As I posed a few messages up, parallel wiring would avoid current matching... just make them all run at 350VDC and peace out. I can sort of imagine that this will require switching to more expensive transistors that can handle 350VDC, or there could be tradeoffs in using these vs <100V ones. Maybe working voltage will jump up a lot when everything moves to higher breakdown transistors (MLPE, FET BMSes, ...)