Interesting Lesson

[linesrg]

Good Afternoon,

I've learnt a useful lesson this last 36hrs as we've been enjoying some lovely sunny weather up here in Aberdeenshire and having fitted an Elster TGM to my 2.56kW array yesterday lunchtime. At the end of the day I was a little alarmed at the fact that only 0.6kW had been generated by the array at the end of the day. The 1.28kW array had an output more consistent with what I was expecting. I did all the usual things like check all connections etc. etc.

My suspicions fell on the inverter, a second-hand unit and I contacted the seller. In the meantime I changed over the inverters during the night and prepared for some close monitoring in the morning.

Well there was nothing obviously wrong with the suspect inverter but the other did seem to be doing better but it caused me to look more closely at how my 32 ground mounted panels are distributed. They are split in to two lots of 16. I have 10 leaning against one part of the outbuildings with the other 22 further around in a 'sheltered' but SSW facing recess. Stupidly the ones that make up the other 6 of the first 10 mentioned are the furthest 'sheltered' so I have 10 panels in full sun and 6 shaded and then the sun strikes the start of the other 16 and completely illuminates them before it gets to the final 6 of the first string (hopefully you're still following this). Thus the inverter is not seeing the trigger voltage as early as it might.

It is also making me wonder if I shouldn't connect the tracker mounted 16 panels to this second (hopefully FIT'able) inverter as they are generating power way earlier than the other 32 will.

As an example the 1.28kW array has propduced 4.5kW today so far whereas the 2.56kW array has only made 1.8kW.

I'll need to see what the MMP voltage is set at on the SMA's and see if it is down at 250V or the manufacturer default of 300V to see if that will help.

Regards

Richard

[linesrg]

Good Evening,

I'm answering my own query or maybe more accurately I'm looking for confirmation.

At the end of the day the 1.28kW array produced 7kw and the 2.56kW array produced a mere 2.7kW.

I'm also thinking that having 16off BP380's and 16off Chinese panels in parallel may not be a good thing?

Both inverters seem to be tracking at around 250V DC on all three sets of 16 panels.

The 2.56kW array was in full sun for at least 3 hrs.

I'd welcome suggestions as to what I must be missing here.

Regards

Richard

[Bill H]

You are on the right lines  - sounds like you are not getting value as part of the string is shaded.  In this case you will much less or zero power from the string in question.   Is there no way you can move the panels (even temporarily) so that they are all illumuminated/shaded at approximately the same times ?  You should then be convinced !

regards

Bill

[rogeriko]

All panels have to be in full sun all day even if you move them some distance away the voltage loss at these high voltages is negligable. A house near me unoccupied for the winter has 2.4 kw system installed but because 4 panels out of 11 are half shaded all day by a wall they only generate 0.7 kw all day, in the summer with higher sun they get 11 kw a day. Move them all 50 meters or so to a clear spot.

[billi]

i guess thats why some GTI have more then one MPP trackers inbuilt to get better results of each string

[linesrg]

Good Morning,

I agree shade will be an issue but from about 1200-1500 all 48 panels are effectively fully exposed and thus I'd expect to see the larger array produce alot more power in that 3 hrs as a minimum i.e. double. Thus if say the 1.28kW produced 800W an hour I'd expect the 2.56kW to produce 1600W an hour or 4.8kW over that period (conservative/ realistic figures).

My Efergy meter was showing a maximum of about 1.65kW at one point - I'll need to do a quick check on average 'fixed load' so I know what to add to this for a better indication of total output. OK so the base load, according to efergy is circa 0.25kW - I don't know how much this device is affected by temperature. Thus I was outputting circa 1.9kW post lunchtime.

Could it be the new Elster? I should have thought to check the DNO import/ export or my own import/ export export meter I guess?

Regards

Richard  

[Paulh_Boats]

Partial shading is a old problem with well known solutions. Do you have bypass diodes installed?

http://pvcdrom.pveducation.org/MODULE/Bypass.htm

http://www.daviddarling.info/encyclopedia/B/AE_bypass_diode.html

http://www.ibselectronics.com/pdf/ac/Diotec/applications/solardiodes.pdf

And if you have a HomePower subscription some real world tests:

http://homepower.com/article/?file=HP107_pg86_Byington

cheers
Paul

[RichardKB]

You all seem to missing the point, he is saying all 48 panels are in full sun from 1200 to 1500 but the larger setup is not providing the amount it should compared to the smaller setup during this time.

Hi Richard

Do you have a DC current clamp meter, if so it makes it easier check the Amps in all strings against the voltage at different times in the day, say every 15mins or 30mins assuming you are about and this should give you a good idea of the power input. If you assume inverter efficiency of approx 90% you should be able to check both setups.

Rich

[billi]

ou all seem to missing the point,

   so we are all happy that you managed to be foccused

[Alan]

The Efergy meter gets a bit confused with some of the pulse width modulation that comes out of grid inverters. Mine has registered  17 K.W. when the actual generated was 6 K.W.

Amp meter on D.C. as suggested is a better bet.

Was going to Ponder one day how confused a modern day electronic generation meter gets, when I get some time.

Regards

Alan

[linesrg]

Guys,

I was starting to come to the same conclusion wrt to bypass diodes.

The BP380 panels I have fitted to my tracker, as I recall, quite clearly have a diode visible in the junction box.

The newer BP panels I have, manufactured in China, don't have this diode yet the accompanying literature clearly states that all panels over 40W are fitted with a diode.

The (other) Chinese panels come ready sealed with leads and MC4 connectors. I'm assuming they do have diodes.

Thus there is a question mark over the BP380 panels. I think I'll break one of the the BP380 string MC4 connections and run with the Chinese panels.

Yes I do have a clamp meter. I'll try this in due course.

Regards

Richard

[Outtasight]

In the morning, the shaded array will produce zip as the MPPT will lock on to the Vmp of the good string at a higher voltage and this will be way above the Vmp of the shaded string so it will put out nothing.  If the MPPT software isn't very clever it might lock on to some random voltage between the two peaks and so be optimised for neither of them  .

More cunning than that, the shaded array will contribute to losses by back feeding in parallel with the good string.  The bypass diodes in the modules won't help here as they are for bypassing forward current in the string when blocks of cells are shaded while others are in the sun.  The lower voltage string will absorb power from the higher one as the PV cells look like diodes pointing from the + terminal to the - terminal of the module (not the other way round as you might have thought).  When a module is reverse fed, the internal cell "diodes" are actually forward biased!

All module strings in parallel (especially ones prone to asymmetric shading) should have blocking diodes (one per string) in addition to the built in bypass diodes.  Working at ~280VDC and about 4.5A you should use common silicon rectifiers rated maybe 400V 8A minimum.  You'll never notice the 0.9V diode drop but you will stop one string back feeding the other when it's partially shaded.

I'd check inside the 380Us as well...  The two I bought recently had Schottky blocking diodes in the junction boxes (in addition to two bypass diodes) - fine if you're just connecting single modules in parallel or straight on to a 12V battery but no good if you're wiring modules in series and then parallel.  The diodes were 40V rated so could not be relied to block the reverse Voc of two modules connected in series (as the Voc would be 45V or more).  If yours have the same diodes in them then you'll be losing 16x 0.7V when the diodes are forward biased in sun (about 50W).

BTW... for the record, you're madder than me.  I've only got 1.4kW propped up against a garden wall 

[linesrg]

outtasight,

If it ever stops snowing in Aberdeenshire I'll get the panels up on the steading roof.....................

Thanks for the comments about the diodes - I'll get some ordered. I have 1N5408's readily to hand but also a range of 600-800V 5A+ TO-220 diodes. I do need to order some Schottky's for the panels themselves though I think or maybe some of theses here new fangled silicon carbide jobbies.

Regards

Richard

[Outtasight]

If it keeps snowing there, you won't need scaffold to get 'em up to the roof...

[linesrg]

Good Evening,

Well I'm back onshore and I fitted blocking diodes (FES16J for the technical as I have them in stock) in the +ive paths from both strings to the one SB2500 and at the end of the day the 1.28kW tracker produced some 3.5kW and the 2.56kW array a mere 0.3kW.

I'm reading the PV Shading thread with interest and am giving serious thought to mounting my 'spare' Windy Boy SB2500 such that each of my 3off 1.28kW strings will have its own inverter.

There will be an element of shading on both the strings constituting the 2.56kW set-up.

I'm also going to carry out a test on these BP panels (and come to that the Chineses ones as well) to see if any bypass diodes are fitted. Anybody have a suggestion as to the simplest way to do this?

I'm also determined to get at least one string mounted on the roof such that they will effectively all see the same light and not be subject to any shade.

Regards

Richard