[RE-wrenches] Trouble Shooting Bad Diode

Greg Seelhorst greg_s at sonic.net
Wed Nov 12 21:16:56 PST 2008


Kirpal -

I have had relatively bad experiences with the BP3160 line of modules, as
stated previously in earlier wrenches contributions.  We have had all sorts
of problems with them, from the j-box actually blowing up (while the
customer was working in the yard) to burn marks on the busbar of the front
of the module, to burn marks on individual cells of the panels.  I am
attaching a product advisory from 2006 which sounds like it falls under your
SX170 line of modules possibly.

What BP has told us is that individual module failure is likely due to close
range shading of the array from a nearby tree, chimney, or some other sharp
silhouette that may cause more stress of the current through the cells
without a voltage (the shaded modules).

In my experience the fastest way to find a panel in these strings of 12
modules or less:
Step 1: Get individual string Voc [for your example, 362 and 312]

Step 2: Cut string in 8 in half  [312 in your case, yielding 170 and 132]

Step 3: Cut string of 4 in half and do the same [yielding say 78 Voc and 54]

Step 4: the two worst panels [say 78 Voc and 54] are one of the 54 Volt
panels

With four steps you have gotten your two worst modules in an array of 16
panels.  Believe me, the Enphase inverters would utimately cut all of this
troubleshooting wasted time out because at the end of the day BP only gives
you like $100 per panel.  That is to cover your time apologizing to the
client, drive time, and troubleshooting time, not to mention time dealing
with the warranty claim report, e-mailing BP, and packaging the module in a
carton to send it back.  If this is your first time dealing with such an
issue, you have also just wasted your time driving to the site without a
replacement panel.  I would suggest that you demand BP to provide you with
backup panels now for dealing with the issue at a financial loss to your
company.  For example, we now have seven panels which we rotate in and out
of our 3,000+ BP3160s installed.  This last quarter, we have had at least 15
more bad panels since I wrote the e-mail below; mostly now with only burn
marks, but they were ready to go.  Believe me, I would suggest spending more
time on a class action lawsuit than trying to figure out bad diode problems;
[I¹m half kidding here, contact me offline if you are ³inspired²]

Here is an e-mail that I forwarded to a colleague of mine who wanted to
bring this to UL¹s attention due to his position in the solar industry and
safety concern for module failure in the industry.

³These photos are typical of about 20 of our 41 failed panels for which I
have shown two of the more seriously failed panels at the j-box.   I assume
that the summer this is more likely to happen or in cold weather with a lot
of irradiance.  

We are at almost a 1.5% failure rate [41 out of 3057] BP 3160 series panels
installed since 2001.  Also (and more importantly) inconveniencing 13% of
our customers (13 out of 95 BP installations) with failed BP panels.  We
don't know the panels have failed unless they tell us either 1) they heard a
noise and are smelling burnt rubber or electric wires, 2) their production
has been cut in half, or 3) their anniversary card shows that they are not
getting as much total output during the year.

In essence, I was wondering what you thought our possibilities for pursuing
entire array change-outs or litigation.  At what point do manufacturers seem
willing to change out an array, (2 out of 20 panels [10%], 5 out of 20
[25%]) etc.?  In the mean time, we are trying to be proactive and the
manufacturers are approving change-outs of panels with burn marks around the
buss bar on the cell side of the junction box.²

I have attached below the content of an e-mail confirming similar issues
with this line of modules from our distributor, at the time.

 mes 2 centimes quoi,


Greg Seelhorst
Installation Manager
NABCEP Certified Solar PV Installer®
Solar Works
400 Morris Street, Suite C
Sebastopol, CA 95472
707-829-2518 (work)
707-849-6819 (cell)
707-829-8283 (fax)
CA Solar License (C-46) #878231




BP Solar Product Advisory
Dear Conergy Customer:

BP Solar has identified a potential problem with certain BP Solar modules
operating in specific installation environments and applications.  Under the
circumstances outlined in the attached BP Product Advisory (Advisory), there
is a limited risk of overheating of the junction box, which can lead to roof
damage and the potential for fire.  Damage to UL790 Class A fire resistant
roofs ­ such as typical asphalt shingle, metal, and tile roofs ­ is
unlikely, and the current Advisory only concerns installations on non-fire
resistant roofs.

The problem has been traced to failed solder joints within the low profile
junction box on MC connected modules, and failures tend to occur within the
first few years of operation.  Approximately 0.02% of US market modules are
affected, including BP Solar 3, 4, and 5 Series modules and the SX line
manufactured between 2002 and 2004.  The Advisory is further limited to high
voltage, grid-tie installations with open circuit voltages above 100 VDC,
and where flammable materials are located within approximately 2 inches of
the module.  Please see the attached Advisory for the exact criteria used to
determine affected installations.

If any of your installations meet all of the criteria listed in the
Advisory, please contact BP Solar directly at bpsolarwarranty at bp.com or call
1-800-891-2163.  BP Solar will immediately arrange for system disconnection
and inspection at their cost, and will also compensate for power lost.

Distributor X [name purposefully removed to protect the innocent] stands
ready to provide support to you, our customer, for any BP Solar product
issues associated with products we distributed.  Distributor X is committed
to minimizing any hardships the Advisory may incur upon you and your
clients, and to working with you and BP Solar to resolve any issues that may
arise.



On 11/12/08 8:03 PM, "Kirpal Khalsa" <solarworks at gmail.com> wrote:

> Chris......Less than 2 years.......
> 
> 
> On Wed, Nov 12, 2008 at 7:59 PM, Chris @ the Oasis <info at oasismontana.com>
> wrote:
>> Hello Kirpal:
>>  
>> What is the relative age of these modules?  For starters....  Just curious.
>>  
>> Chris Daum
>> Oasis Montana Inc.
>> 406-777-4309 local line
>> 406-777-0830 fax
>>  
>> 
>> 
>> From: re-wrenches-bounces at lists.re-wrenches.org
>> [mailto:re-wrenches-bounces at lists.re-wrenches.org] On Behalf Of Kirpal Khalsa
>> Sent: Wednesday, November 12, 2008 8:33 PM
>> To: RE-wrenches
>> Subject: Re: [RE-wrenches] Trouble Shooting Bad Diode
>> 
>> Bill and all.......So we were at the site today and these were our findings:
>> We covered each solar panel with cardboard 1 at a time.....we continued to
>> see the same lower voltage on string #2.  String #1 had an expected Voc.
>> String 1 had a Voc of 368, String 2 had Voc of 312.
>> We decided to disconnect the entire string.  Luckily the interconnects were
>> accessible without physically removing the modules.......I love array's that
>> are only 2 rows deep with service aisles all around.....
>> The system was 2 strings of 9 BP SX170's connected to a PVP2800XV.
>> Module nameplate Voc was 44V.  Cell temp of the modules at time of
>> measurement was roughly 85-90 degrees F.
>> all modules were in the 40.1-8 range....except module # 3 from the positive
>> homerun side was 38.8V, module 6 was 36.6V and module 8 was 10V.
>> We had no trouble determining that module # 8 was defective......In these
>> modules I believe that the entire 72 cells are in a series.string...(am i
>> right?roughly .5+V per cell.).
>> Covering 1 module at a time did not give us any indication as to which panel
>> was defective.  The Voc with 1 module covered ranged from a low of 299VV to a
>> high of 304V. This did not make sense to me as to reducing the string Voc
>> from 312 to 304 by covering 1 module.........
>> 
>> Anyhow........the 2 modules that had reduced voltages......1 with ~38V and 1
>> with ~36V......would you consider those defective?  Worthy of trying to get
>> BP to warranty those?
>> We will be filing a warranty claim on the single panel that is obviously
>> defective.......
>> Thanks for your advice and input.....
>> Cheers,
>> 
>> Kirpal Khalsa
>> Renewable Energy Systems
>> 
>> 
>> On Fri, Oct 31, 2008 at 10:33 AM, Bill Brooks <billbrooks7 at yahoo.com> wrote:
>>>  
>>>  
>>>  
>>> 
>>> Kirpal,
>>>  
>>> 
>>>  
>>>  
>>> 
>>> When you parallel,  your voltages have to be the same, so you are correct in
>>> that  assumption.
>>>  
>>> 
>>>  
>>>  
>>> 
>>> Your assumption about  the diodes is not correct. Diodes don't work like
>>> that. You could have a  shorted module (all three diodes shorted) which
>>> would lower the Voc by one  module. It looks like you have nearly two
>>> modules worth of problems. It could  be a total of 5 bad diodes, but that
>>> would be weird. It could be you have a  ground fault, and haven't tripped
>>> the GFP yet because of lower sunlight  levels‹(it has to be very low sun to
>>> stay under 1 amp). You also could have a  wiring mistake, either because the
>>> factory reversed wired a module (rare, but  it happens)‹that would be 2
>>> modules low; or, you could have skipped two  modules somehow.
>>>  
>>> 
>>>  
>>>  
>>> 
>>> It does seem like a  quarter of the bad string is out. Your assumption about
>>> bad diodes could be  correct, but it only affects the groups of cells its
>>> connected to. Each diode  protects 24 cells‹it looks like you have 5 diodes
>>> worth of loss. It could be  in 2 modules, or up to 5 modules.
>>>  
>>> 
>>>  
>>>  
>>> 
>>> The shading test is  the best way to find bad diodes. Disconnect the bad
>>> string from the good one  and take a piece of cardboard the dimension of the
>>> module. With your meter  testing Voc on the back string, completely cover
>>> each module, one at a time.  When the Voc is not affected by complete
>>> shading of a module, that module is  completely bad. If the shaded module
>>> only loses a portion of one module's Voc,  then you have one or two, of the
>>> three diodes, bad.
>>>  
>>> 
>>>  
>>>  
>>> 
>>> This is a great  opportunity to learn how to troubleshoot a PV array. Take
>>> the time, and the  effort, to run through it carefully, learn the problems,
>>> and report back with  your findings. It's a beautiful thing to learn
>>> something new.
>>>  
>>> 
>>>  
>>>  
>>> 
>>> Bill.
>>>  
>>> 
>>>  
>>>  
>>>  
>>> 
>>> From: re-wrenches-bounces at lists.re-wrenches.org
>>> [mailto:re-wrenches-bounces at lists.re-wrenches.org] On Behalf Of  Kirpal
>>> Khalsa
>>> Sent: Friday, October 31, 2008 10:43  AM
>>> To: RE-wrenches
>>> Subject: [RE-wrenches] Trouble Shooting  Bad Diode
>>>  
>>>  
>>>  
>>>  
>>> 
>>>  
>>>  
>>> 
>>> Greetings all......I was hoping i could have someone help me confirm a
>>> troubled series string on a system we have installed.....
>>> System consists  of 16 BP 170 W 24V nominal modules......2 strings of 8
>>> modules.....Vmp of each  module is roughly 34V and Voc of each module is
>>> roughly 41V.  One string  is under-producing......String 1 Voc is 346V and
>>> String 2 Voc is 272V.   String 1 Vmp under load is 270V and String 2 Vmp
>>> underload is 270V.  I am  hypothesizing that String 2 has a bad diode on the
>>> 3rd panel in and is  reducing the Voc of that string by th 2 panels ahead of
>>> it in the  string.  I am hoping to figure this out without removing the
>>> entire  string and checking each panel if possible.....
>>> I think the Vmp is equal  due to the fact that both strings are paralleled
>>> in the inverter and the 1st  string is being pulled to the same level V as
>>> the 2nd string under  load......Is my logic good?
>>> Thanks for the help and  advice.....
>>> Cheers,
>>> 
>>> 
>> 
>> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20081112/e9ca2aeb/attachment-0004.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: BPS Product Advisory.pdf
Type: application/octet-stream
Size: 59584 bytes
Desc: not available
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20081112/e9ca2aeb/attachment-0014.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fargo Bad BP Panel Photos.jpg
Type: application/octet-stream
Size: 234093 bytes
Desc: not available
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20081112/e9ca2aeb/attachment-0015.obj>


More information about the RE-wrenches mailing list