The Optimizer Advantage?

Updated 01/25/25 with additional detail about the voltage/amperage of the optimizers and panels.

Can someone explain to me what’s going on here? This is not how I’d expect an optimizer system to work, at least based on how it’s advertised.

A customer pointed out an interesting situation as his snow is melting off the roof this week. He has one string that is completely unshaded from snow, and one string that is probably about 1/3rd covered with snow. The production on the unshaded modules from the partially covered array are producing at about 1/4 of any module from the unshaded side. See the pictures below.

Before you start your comment, I took the screencap of the array from the playback view (notice watts instead of watt-hours), so it’s not representative of any snow that may have been present earlier in the day.

I drew an example of the approximate shading so it’s easier to see on the screencap. Of course I’d expect partially shaded modules to produce way less. But it seems like the difference between module voltage and where the string voltage needs to be is so stark that it’s actively dragging down the ability to produce for most of the string.

The inverter doesn’t have MPPTs (that I know of), so it’s entirely reliant on the optimizers to create the proper voltage. I’m guessing what’s happening is that the optimizers are boosting the voltage so high that they can no longer run the module at the best voltage for production.

What’s important here is that if you ask literally anyone about what would happen if half of an optimizer string got covered by snow, most people would say that the rest of the modules/optimizers would produce just fine, no problem. Apparently that’s not the case.

UPDATE: I got a request to add more details about the optimizers themselves, so below is the opti wattage and voltage, and panel voltage and amperage.

After some spirited speculation, I think the answer comes down to this comment from Gabriel Chong of Sunspear Energy:

These are S1201 SolarEdge optimizers, whose minimum string length is 15 optimizers/29 panels. So it turns out, this installation is at the minimum allowed string length. What seems to be happening is that once another optimizer or two gets completely covered and can’t produce energy, the rest of the optimizers can no longer “optimize” for best production and are focusing entirely on meeting the bus voltage, compensating for the MIA optis.

Such an interesting edge case, and would seem to suggest that aiming for the middle spot between minimum and maximum optimizers is the best design practice.