Posted by Doug Baillie on July 14, 2008, 11:07 am
Hello community !
Looking for help here on a couple of simple questions, having just
taken responsibility for debugging an small experimental concentrated
There appear to be dead or damaged cells within the module. Shading
tests show areas of the module which are insensitive to illumination.
the cells are currently hardwired in a mixed series/parallel
configuration and not easily accessible, so before starting to
dismantle the module to cell level, I want to go a little further in
understanding the likely failures. All interconnections appear to be
unbroken, and each leg of the system produces an output voltage, but
several or many cells do not seem to contribute photocurrent. The
whole system has been left unattended for some months, may have been
under illlumination some of the time, and under unknown load
So the questions are -
What are the most likely failure modes of such a system ?
What kind of failures are likely if the output has been short-
circuited ? Any problems if left illuminated without any load ?
What kind of damage or degradation to the silicon could be expected
from being left intermittently for a few minutes or tens of minutes
under ten suns concentration without adequate cooling ?
Any recommendations on debugging tests at either cell or module
level ? Any good resources on this ?
Is this a good forum for technical questions and ideas ? Any other
Thanks for any help. Looking forward to making contact with the
Posted by bealiba on July 14, 2008, 12:50 pm
Temperatures over 70C will destroy cells.
Posted by Duane C. Johnson on July 14, 2008, 1:29 pm
> Looking for help here on a couple of simple questions,
Maybe, but the answers may not be simple.
> having just taken responsibility for debugging an
> small experimental concentrated photovoltaic system.
> There appear to be dead or damaged cells within the
> module. Shading tests show areas of the module which
> are insensitive to illumination.
Big mistake if bypass protection diodes were not
present. I wrote about this here:
OK, this was based on whole PV panels but the concept
is the same.
The problem is that at 10X concentration the currents
are much higher than at 1X. The shaded cells in the
presence of a string of illuminated cells will have high
voltage across the shaded cell. This can breakdown the
cell and cause a short. This damage is permanent
especially when the currents are high.
The fix is to install "Shunt Protection Diodes" every few
cells to or so. Say every 5 or 6 cells. This limits the
breakdown voltage to 3 or 4 volts. The current rating of
the Shunt Protection Diodes must be capable of conducting
the full cell currents of the series string multiplied
by the number of strings. Maybe the diodes should be
heat sinked also.
These shunt currents can be reduces to that of a single
string if "Series Protection Diodes" are employed.
Note, the cell breakdown voltage tends to be lowered as
the temperature is increased as happens in concentrator
Some of the Spectrolabs triple junction cells have these
diodes built in every 2 cells.
> the cells are currently hardwired in a mixed
> series/parallel configuration and not easily accessible,
> so before starting to dismantle the module to cell level,
> I want to go a little further in understanding the likely
> failures. All interconnections appear to be unbroken,
> and each leg of the system produces an output voltage,
> but several or many cells do not seem to contribute
> photocurrent. The whole system has been left unattended
> for some months, may have been under illumination some
> of the time, and under unknown load conditions.
The second thing that can damage these cells is high
temperatures. This can also cause shorted cells.
Make sure that you have heatsinks on the cells.
I recommend using passive air cooling on the heatsinks
for low concentrations like 10X.
> So the questions are -
> What are the most likely failure modes of such a system?
1. No shunt protection diodes.
2. Over temperature.
> What kind of failures are likely if the output has
> been short- circuited?
Generally, short circuit operation does not cause
> Any problems if left illuminated without any load?
Yes, overheating if active water cooling is employed.
I would choose a solar tracker that has a Remote Shutdown
feature. A thermostat on the heatsink tells the tracker
to move off the sun when it gets hot.
I can add Remote Shutdown to my LED3X trackers. See:
> What kind of damage or degradation to the silicon
> could be expected from being left intermittently
> for a few minutes or tens of minutes under ten suns
> concentration without adequate cooling?
Over heating doesn't take long at 10,000W/m^2.
> Any recommendations on debugging tests at either
> cell or module level?
Do testing under a bright light without concentration
and measure cell voltages.
> Any good resources on this?
Some of the NREL papers.
Search on "+concentrating +photovoltaic".
> Is this a good forum for technical questions and ideas?
Yes, if you ignore the crazies, but they are everywhere
> Any other recommendations ?
1. Use Shunt Protection diodes.
2. Use Series Protection diodes.
3. Use good passive cooled heatsinks.
> Thanks for any help.
> Looking forward to making contact with the community.
Home of the $5 Solar Tracker Receiver
Powered by \ \ \ //|
Thermonuclear Solar Energy from the Sun / |
Energy (the SUN) \ \ \ / / |
Red Rock Energy \ \ / / |
Duane C. Johnson Designer \ \ / \ / |
1825 Florence St Heliostat,Control,& Mounts |
White Bear Lake, Minnesota === \ / \ |
USA 55110-3364 === \ |
(651)426-4766 use Courier New Font \ |
email@example.com (my email: address) \ |
http://www.redrok.com (Web site) ===
Posted by mundt on July 15, 2008, 4:55 pm
Posted by mundt on July 15, 2008, 5:42 pm
I think that Duane provided some pretty good information (importance
of bypass diodes etc.) but did not necessarily answer your questions.
You indicate that each leg (series string) of the system produces a
voltage and that shading some cells indicates that they are not
contributing (insensitive). You also indicate that the
interconnections appear to be unbroken.
This would suggest to me that the "insensitive" cells are shorted.
Under the combination of temperture and current the have seen, the
semiconductor junctions have "failed". This could be caused either by
the metalization having diffused into the junction or a localized
defect in the junction resulting in a "run away" heating .
As Duane indirectly indicted, testing a cell by shading it is not a
real good idea. This can stress a good cell and maybe cause it to
fail. It would be better to use a bright light (halogen spot) and
illuminate the cells one at a time to identify bad ones.
Since you have already identified the bad cells (the insensitive
ones), you should simply replace them. I would suggest that you might
want to run the system as is (with proper cooling and no shaded cells
of course) for a while prior to replacing the cells. If some of them
have failed, then it is possible (even likely) that others have been
damaged and are near failure. It would be wise to weed out the
cripples early - you don't want to chase failures forever.
Either short or open circuiting the system should cause no problems
with "good" cells. Under short circuit conditions the junctions are
conducting a fairly high current but at almost no voltage drop (i.e.
no power produced). Under open circuit conditions the junctions
develop their full PV voltage but with no current flow (i.e. again no
power). The short circuit condition is the one one that could put the
most stress on a "bad" cell, if there is defect within the junction
resulting in a localized high current then this can lead to localized
high temperture - resuling in localized heating - resulting in in even
more current flow, etc. etc. until the defect becomes a short.
As the others have indicated, if the cooling on a concentrator system
fails, it can be destroyed very quickly. Cells that don't fail
immediately are still possibly damaged such that their reliabilty
(lifetime) will be substantually shortened.
I hate to say it, but I think that your "responsibility" for this
concentrator system is likley to evolve into a continous
troubleshooting and cell replacement task.