Flexible Panels, Mounting Methods and Heat

[Barboots]

Did you do a amp test on the Gamy panel to see if it was putting out much power .
Dibs

I can get anywhere between 0A and a bit over 7A depending on where I tap them.

[dapope]

Hi Martin how how’s goes it we’re in terry’s in front of the shed for Xmas . Anyways I have always wondered how these Flexi panels stand up to wind flap if there not fixe to a solid backing as no one has ever commented it mustn’t be an issue . I just thought it might cause connection failer . That’s a weard breakdown a .

Dibs

Heya
Even my wiring is not terribly well anchored, however there is no evidence of it going anywhere in particular. I was toying with the idea of putting holes in the alloy frames of the old panels to let the heat out /encourage airflow, but have not. I do know the panels cool and become more efficient when driving, and thats not always taken into account. Definitely no panel flap though.

[T1 Terry]

After rather costly field testing I think we have an answer to the panel failures, and it isn't what we thought. I discovered it heading from one direction and Peter (Oldtrack) found it from the tech blurb http://thegreynomads.activeboard.com/t6 ... t-64104633 . The problem is also what makes these panels perform so well when they are actually working, the have very low "reverse bias breakdown voltage" no idea if this is in the diode block mounted under the junction block or actually internally wired under the cells themselves. The fix has been to install 10 amp blocking diodes in each positive solar cable preventing the reverse voltage punching through the internal diode and turning the weakest cell into a heater and destroying it.
These fractured cells can be seen clearly while wired up on the roof, take them off and they are near impossible to find. Some of them even work again once cooled, but I'm certainly not going to risk fitting them to anything else but a test bed.
After way more practice at this than I'd hope I'd even have, here it my method for replacing failed flexible panels.
This replacement method involves leaving the van/panels in the sun till it get real hot, then start at one edge of the panel and peel the panel off the Lexan. If you get one where the Lexan will let go of the roof then peel the whole thing off and put them in the sun, the panel will peel off once it gets hot.
Next, place the new panel over the Lexan and mark out where the eyelet holes are, then cut away enough Lexan to fit a 80mm x 80mm x 3mm aluminium foot you later glue to the roof. These aluminium feet need a 4mm x 40mm countersunk screw fitted roughly centre and 25mm in from one edge for the centre ones and 25mm from the corner for the 4 corner ones.
Glue the piece of Lexan to the top of the foot with the threaded part through it, probably best to add a lock nut to the counter sunk screw to hold it firmly in the plate and slide the piece of Lexan over it.
Now mount these to the new panel, bits of scrap 3mm aluminium plate cut in a triangle with a 4mm thread tapped in it work will as securing washers and a lock nut added on top reduces the risk of them coming off when you don't want them too.
Now you have the feet attached to the panel, fit the panel in place so the aluminium feet contact the roof, a good layer of Selleys Armourflex on the base of each foot and set it into position. 2 x 2ltr milk bottles full of water each side of each foot will hold it down so the glue spreads out and make a good bond, 12 hrs or more later and the bottles can be removed, job completed except for the wiring .... and don't forget the diodes.

T1 Terry

[Barboots]

Everyone loves a photo. Lots of bubbles under the panels... some corresponding to the surface defects pictured previously, but some not.

[Barboots]

I cut the back of the panel open where one of the rear bubbles was, but where there was no corresponding top surface defect. It sort of looks burnt...

[Barboots]

As I can't be sure if it was mechanical or electrical stress which has caused the failures, I'm definitely going to heed warning and address both. As far as a blocking diode option, the specs on this are appealing... just 28mV to drive it:
http://www.ti.com/product/SM74611

Of course I managed to dredge up some information to concern myself with:
https://e2e.ti.com/support/power_manage ... 6/t/264615

Whaddayathink?

[Barboots]

I observed a considerable amount of perishing of the polycarbonate Thermoclear where it was in contact with the Armorflex. I thought that it was a chemical incompatibility. Then as I removed the last panel I found this darkened band under the Armorflex. The adhesive itself appeared normal, however the polycarbonate was discoloured and almost sticky.

Hmmm... interesting.

[Barboots]

I don't know if I'm just chatting with myself now? Anyway I finished a crash Google research binge regarding blocking diodes for paralleled strings... and ended up quite confused by the differing opinions.

Obviously trying to use a meter to measure the presence of a diode across a functional panel isn't going to work. However two of the panels I have are zero voltage... and both measured ~0.12V with my meter on diode testing mode. I figured that was indicating that blocking diodes were fitted... but to be fully sure I started dissecting the junction box. There was too much sealant inside so I ripped it off the panel to guarantee that everything else was out of circuit.

Confirmed. Blocking diodes fitted and still functional.

[Barboots]

Carving the panels off with the bread knife was hard work. I'm going to have a beer and wait to see if any mates turn up.

[dapope]

Thanks Barboots. I am following it, and I think theres a few others too..the two panels on the swagman do not have diodes that I can see. Not sure about the other one.