Tim Wescott wrote:

John Popelish wrote:

Just for efficiency reasons, I think you would want ot have enough
capacitance across the regulator input that the cell resistance drops
voltage only with respect ot the average output current, not the
switcher peak value. This can be a pretty big factor in the overall
efficiency. Using a switcher that has little ripple current on its
input (two phase boost, for instance) makes this much easier.


That's not the point. Because a switcher tends to draw a constant power
from a load it's input impedance has a negative resistive component. If
you match this with a source that has a too-high impedance it'll be
_unstable_; a big capacitor would just slow it down in this case.

Presumably what you need is a controller that detects when the supply
voltage gets down to some threshold, then regulates the supply-side
current rather than the load-side voltage.

Come to think of it that'd be a fun thing to design...

Very few switchers draw an instantaneously constant power from the
unregulated source. Almost all can draw an average constant power
(over the switching period). The difference means a lot when you
consider what the variations do to the total losses in the solar
cells. You missed my point, completely.

--
John Popelish