Panel output dropped off and burned but it doesn't make sense

[dudeinthewoods]

So I've currently got 13 250w panels on top of a shipping container.

I noticed the last few days the power they put out is pretty terrible. I used to get 2kw~ or so and now I'm getting like 350w combined.

I've got some pics I'll attach. They look burned on all the wires around the cells. The bus up top etc too.

Now a little bit ago I had a relay fail and it wrecked my lifepo4 setup (really pissed). I reversed polarity and it burned out a charge controller (maybe internal fuse, haven't pulled it).

So I was thinking that may have wrecked the panels. Here's the part that I don't get. 3 SCC and all 3 sets of panels are burned. So I'm thinking they're partially intact. I'll test each panel individually to see how they're doing soon.

That sort of made sense even though the surge was all through one SCC but maybe somehow it happened, I don't know.

Then I realized that 2 panels I had just sitting next to the others are burned. They're not connected at all to anything. They're sitting on the container (it's actually 3 containers one set of panels per).

I'm just scratching my head now.

Also there dusty as hell. They cut hay a few days ago so now that I saw how bad they are I'm going to clean them and see if production goes up too.

Any thoughts?

 

[snoobler]

If a SCC failed and allowed the batteries to discharge through the panels, and you don't have suitable fuses/breakers, you could have cooked the panels.

Dirty panels can have substantially reduced output.

 

[Hedges]

No reason for discharge from batteries though a failed SCC to damage panels.
Almost all SCC require PV Voc and Vmp to be higher than battery voltage. If you connected the PV panels directly to battery, they would eventually overcharge and cook the battery. But battery voltage, being lower than Voc of the panel or string of panels, will not back feed more than Isc. It will be much less.
(Look at the IV curve for a PV panel. A "12V" panel will have Vmp = 17V, Voc = 21V, or something like that. Look at the current for 14V. Compare that to the current for 0V. The difference is how much current a 14V battery would backfeed into a "12V" panel. Very small, maybe 10% of Voc)

I'm not clear what you mean by panels "burned", don't see obvious damage.

If is possible for diodes in panels to be damaged. That can happen (to some brands) if part of the panel is shaded and the rest has direct sun.

Test a panel by itself - what Voc? What Isc?

 

[snoobler]

What's panel Voc at night?

 

[Hedges]

At night the panel doesn't produce any current or voltage, so Voc at night is zero.
Think of its Voc and Isc readings in the sun as a test to determine parameters.

A PV panel is a string of diodes. They are forward biased in operation and the photocurrent they produce leaks back through them. Voc is the voltage they reach when no current is taken from the wires, and all leaks back through the forward biased diode (voltage about 0.7V per cell is just part way up the diode's exponential curve.)

A "12V" 100W panel has about 21 Voc and about 6A Isc.
If you apply 21V to the panel in a dark room, about 6.5 A will pass through it.
Apply 17V (Vmp) and about 6.0 A will pass through it.
Apply 14V and about 0.25A will pass through it.

You could put a 12V panel on a 12V battery, leave it in the sun all day and all night. The panel won't be damaged at night.

 

[snoobler]

Thanks. I was under the impression there was a mechanism for batteries to be able to discharge through the SCC in the event of a blocking diode failure in the SCC.

https://www.electronics-tutorials.ws/diode/bypass-diodes.html

 

[Hedges]

Thanks. I was under the impression there was a mechanism for batteries to be able to discharge through the SCC in the event of a blocking diode failure in the SCC.

https://www.electronics-tutorials.ws/diode/bypass-diodes.html

The "PV Equivalent Circuit" shown in that document is incorrect.
What they showed was an 0.58V source in series with a resistor. That is known as a Thevenin model.
It produces a straight line through (0.58V, 0.0A) and (0.0V, 0.58/R A)

A better model for a photocell is a non-linear one as shown here:

Fig. 1: Equivalent circuit of a photocell.

Download scientific diagram | Equivalent circuit of a photocell. from publication: Modelling and Experimental Analysis of a PEM Electrolyser Powered by a Solar Photovoltaic Panel | In the present paper a proton exchange membrane (PEM) electrolyser for hydrogen production powered by solar...

www.researchgate.net

That produces an exponential curve.

It would take a short in the panel (Voc < Vbattery) and failed SCC to cause high circuits that could burn something up.
But some systems could have a smaller drain through the panels, discharging the battery. It was for this the article suggested blocking diodes.

 

[dudeinthewoods]

Hmm all that makes sense. On tbe second pic it doesn't look grey on the leads that connect all the cells? Those were white. I will wipe it off and take better pics. It looks really obvious in person. The entire panel looks different.

I'll take some individual readings today too when the sun is out.

The weird part is the 2 (or 3 I forget) that aren't connected to anything that have the same issue.

I'll get back to you later with an update.

 

[DThames]

I have some panels with similar looking traces but they appear to still work normally. Notice the one on the left of the image. I bought these used. I don't recall seeing this before but it could have been there all along. They do look worrisome.

 

[Hedges]

I have some panels with similar looking traces but they appear to still work normally. Notice the one on the left of the image. I bought these used. I don't recall seeing this before but it could have been there all along. They do look worrisome.

I don't think that is breaks in the metal, just epoxy bond between wafer face and glass affecting how light reflects.
My 15+ year old AstroPower AP120 panels have some differences but where the active area bonds to glass, so would affect light transmission.

 

[DThames]

I don't think that is breaks in the metal, just epoxy bond between wafer face and glass affecting how light reflects.
My 15+ year old AstroPower AP120 panels have some differences but where the active area bonds to glass, so would affect light transmission.

View attachment 33093

I was thinking it looked like it was separated from the glass and tarnished looking. It does look bad.

 

[Hedges]

Might look "bad", but isn't that just cosmetic? If performance isn't affected, may not be significant.
Some PV panels have cells with broken traces, things that affect current output.

I posted a link to a site that has reports showing various failures, and accelerated stress test results. Corrosion, cracks, etc.

Is there a reliable resource to compare panel makes/models

Well that sounds like a heck of a deal! Glad you placed your order and can't wait to see how it all works. That is a very good price, and I should not be surprised but that confirms they get a great deal when they upgrade. I wonder if they went with Trina on the replacement units? Any idea?

diysolarforum.com