Originally Posted by socaldmax
Am I missing something?
What you are saying is fundamentaly right but the differences you are imagining in Vmp are much greater than are seen in practice.
Vmp does not vary much with light intensity.
Have a look at this graph for a Kocera panel:
The difference in Vmp between the top graph at at very bright 1000Wm2 and the bottom graph at a dull 200wm2 is less than 0.8v. In practice it will be slightly less than this due to temperature differences (at high illuminates the panel will heat up more reducing the Vmp).
The large sort of differences in Vmp you are quoting are not seen in practice.
If you draw a line down from the Vmp of top graph and operate the the panel exposed to 200Wm2 at this voltage, as would happen with a single controller, you can see the loss of power is very tiny less than 1W.
The Vmp in the panel specifications is for a very unrealistic cell temperature of 25c in practice the Vmp will be around 16v. You will not see 18v+ apart from very briefly if the sun comes out from heavy overcast, or in very cold temperatures.
If we take a typical realistic scenario where panel has a Vmp of 16v and the other a Vmp of 16.5v. To calculate the loss is quite complicated but we can make certain assumptions that make the maths easy. If we assume the solar controller uses the lower Vmp (instead of a compromise) and we assume the current does not rise as the voltage is lowered. These assumptions exaggerate the difference considerably, but even so, the loss in output of single controller verses one for each panel is only 1.5%. Given the higher self consumption and standby current of two units it is likely that that a single controller would be superior in these conditions.
Greater differences in Vmp between panels than 0.5v can, and do occur mainly with isolated shadows. I think we are both in agreement that a single high quality controller for each panel will generally give a higher output, but the magnitude of the difference is not great. In practice the difference is often between a single high quality controller, or multiple lower performance, cheaper units, when this is taken into account the difference becomes very small, or can swing in favour of the single unit.
Multiple controllers, including one for each panel are certainly worth considering. Two controllers one controlling the panels on the port side and one on the starboard side can work especially well, but often a single unit is the most practical option. With all these options it is worth remembering that we are only making very small differences that effect the last few percent of output.