[RE-wrenches] Unusual problem with SMA TL inverter on metal roof

[boB at midnitesolar.com]

I could see how they might get confused between hi-pot 
something-or-other and capacitance.
But for hi-pot arc-over, you don't really have to have much 
capacitance.  Just one spot with
too little spacing for it to arc across is enough to fail.   i.e. Hi pot 
problems unfortunately don't
creep up usually, they just happen when you pretty much least expect it.

oOOOPS.  Forgot...  In addition, if you somehow do try to measure the 
capacitance of the array,
there will probably want to be at least some PV current flowing from 
plus to minus when they
are shorted so early morning or evening would work as well as full short 
circuit current during
the day I would think.

Some current because the measurement might not get the full array of 
capacitance to ground
if all those PV diodes aren't at least forward biased which would happen 
when light out.
Don't want any current through the capacitance meter itself of course.

boB



On 3/13/2014 11:29 PM, Jerry Shafer wrote:
>
> I had this issue way back with pvl's unisolar and it was always 
> written off as hi pot leakage in the conductors and its good to see 
> that it was put to paper
>
> On Mar 13, 2014 11:20 PM, "boB at midnitesolar.com 
> <mailto:boB at midnitesolar.com>" <boB at midnitesolar.com 
> <mailto:boB at midnitesolar.com>> wrote:
>
>
>     Interesting app note.  Interesting problem.
>
>     How far from this metal roof (grounded I presume) is the bottom of
>     the modules ?
>
>     This app note is mainly talking about the area of the module from
>     the mounting bracket which should
>     be quite a bit less than the modules area from a distance above a
>     grounded metal plate like this roof.
>
>     How far above the metal roof are the modules ?  If they are real
>     close, raising them slightly might
>     be enough to lower that capacitance enough to stop the TL from
>     tripping.
>
>     Another thing you could do, but would of course cost money would
>     be to split the array up for
>     more inverters.
>
>     This app says that 1400 nano-Farads is an approximate point for
>     problems.   That's for 50 Hz.
>     It will be around  6/5 worse at 60 Hz.
>
>     If you have a capacitance meter (impedance bridge or meter), you
>     could measure the capacitance
>     of the array.  Disconnect the + and - from the inverter first of
>     course.   Then, tie the plus and
>     minus of the combined array together and connect that to one
>     terminal of the capacitance meter.
>     Connect the other lead of the capacitance meter to the roof and/or
>     ground and it ~should~ show
>     the capacitance you are interested in.
>
>     This is one of those newer compromises associated with the higher
>     efficiency TL inverters (anybody's,
>     not just SMA's of course), and also the newer code requiring TL
>     (non-isolated) inverters above 300 volts
>     and everything else new we have to all learn about.
>
>
>       Welcome to the future !      ( from Firesign Theater's "I think
>     we're all [solar] bozos on this bus" )
>
>
>     boB
>
>
>
>     On 3/13/2014 7:31 PM, Marco Mangelsdorf wrote:
>>
>>     Friends,
>>
>>     We have run across an odd technical problem at our commercial
>>     install at a laundromat on our island.   Attached is the
>>     description of the issue from SMA.  We don't know much except
>>     that we are switch out the TL inverters to the old style US
>>     inverters because SMA has no fix for the phantom ground tripping.
>>      The issue is beyond our feeble understanding.
>>
>>     Any observations to share?
>>
>>     Thanks,
>>
>>     marco
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20140314/39f3b85f/attachment-0004.html>