re: string sizing for your 2500 watt inverter
22 oct 2002
bill kaszeta / photovoltaic resources wrote:
>do it your way and let us know if you have any insurance coverage
>if you have a fault.
it'll be about 50' from the house. my largest worry is theft.
if it is not "in the house," this nec section need not apply:
690.7(c) photovoltaic source and output circuits. in one and two-family
dwellings, photovoltaic source circuits and photovoltaic output circuits
that do not include lampholders, fixtures, or receptacles shall be
permitted to have a maximum system voltage of up to 600 volts.
>good thing you do not have any building inspectors to deal with.
i've heard the code inspectors in phoenix are murder. our township has
a building inspector, but various independent code inspectors check
the electrical work. one just paid the $4k cost of rewiring my neighbor's
barn, which they inspected 4 years ago, because they failed to notice
that there was no ground wire inside the metal conduit.
>once a fire or arc gets started in a pv system, it is difficult to put out.
in this case, there won't be much to burn... 12 or 15 sharp 165 w panels,
some steel unistrut, some wire, and an inverter. most likely a sun tie,
at this point, with rainwater trickling over the panels, which might contain
some copper sulphate (the navy had algae problems in an early experiment)
and a diffuse parabolic overhead reflector to raise the peak output by 25%.
>the french almost lost the nice airport terminal many years ago to a fire
>in a pv system. this lead to some of the requirements in the nec. there
>are reasons for most of the requirements.
s'plain why sunny boy's need dc disconnects in the us and panels need to be
grounded, s'il te plait. germany doesn't require dc disconnects, with their
great emphasis on worker safety ("berufsgenossenschaft") and according to sma,
panel grounding prevents using some very efficient inverter topologies. other
hvdc nec sections _prohibit_ this kind of grounding, for safety reasons...
>the only unreasoned restriction i know about in the nec is a restriction
>on a wire type in battery rooms that no one now knows the history of...
poke around, and you may find more. having served on other standards
committees, i realize now that commercial interests can have a great
influence on code requirements.
nonetheless, the intent was to prevent fires and injuries, and i have
a nice shiny new copy of the 2002 nec code and handbook and sandia report
sand2001-0674 "photovoltaic power systems and the national electrical code:
suggested practices," by john wiles (a 101-page bargain, at $0.00.)
i haven't read them entirely yet, but it seems to me that a series string
with an external clamp circuit to limit voc to 600 in the rare event that
a 15 panel sharp string is unloaded would be entirely nec-legal, and ul-able,
if any single failure would leave it safe. sharp might build something like
this into their panels, to limit voc to 40 v max...nec section 690.7(a) "max
pv system voltage" says "...where other than crystalline or multicrystalline
silicon photovoltaic modules are used, the system voltage adjustment shall
be made in accordance with the manufacturer's instructions." i would argue
that a panel containing a clamp circuit is "other" above.
or we might use a voltage divider and a comparator to turn on a relay to
short the lowest 4 panels in a string of 15, if the string output would
otherwise exceed 600 v. it's my understanding that sunny boy movs can limit
the input to 600 v, but not often. the mov voltage thresholds decrease with
use until they prohibit normal operation.