Solar Electric Panels vs Lightning EMP

[TerraHertz]

Here's an update on those 'lighting struck' solar panels I mentioned here: https://www.eevblog.com/forum/testgear/list-your-test-equipment-score-here!/msg908060/#msg908060

The story became quite interesting, quite apart from the good news that all the panels work (after some repairs.)
Writeup here: http://everist.org/NobLog/20160504_solar_panels_vs_emp.htm

I still have some sections to add to that, discussing the nature of EMP and how it produced the visible cell-surface-tracking effects.
A couple of indicative pics below, more in the article. They apparently were not struck directly, but suffered some EMP-like near-field breakdown effects from a very close lightning strike.

A question: has anyone broken open a large (190W) solar panel, and examined the internal construction? With these panels, looking through the glass it _looks_ like the cells are simply sitting in a cavity between the front glass and rear backing sheet. But the arc-tracing photos look like there's actually a very clear filler material. Which would make sense I guess, both for protecting the cells mechanically, and from moisture penetration.

The most amazing thing is that the panels with lots of visible arc tracing as in the pics, seem to still work fine, as well as the ones with no visible damage. Even perhaps a bit better. Which is really weird.

[coppice]

I think all these panels are filled. They would have to be filled or have air tight seals around the edge. Otherwise moisture would get in quickly, and eventually dirt. Expansion and contraction of the air in the cavity would be sufficient to pump pollutants from the outside to the inside.

[SeanB]

I would guess that the lightning damage is self limiting, in that it tends to blow out the aluminium contact on top of the cell, thus isolating any shorted silicon areas. The rest of the cell is still able to generate power, and the multiple tracks of the cell means that there are still power paths to carry the current.

As to the connectors, I think they are simply aluminium strips, as that is both cheap, compatible with silicon and also easy to spot weld ultrasonically onto the cells, and as they have not corroded with moisture ( as silver is want to do, going black very quickly in air exposure without a protective film) and the connection method of using a tool to press the strips down folded into a broad bladed socket is guaranteed to scrape the oxide off and make a gas tight seal with the stainless steel terminal strip.

[Red Squirrel]

Is the ground of a solar panel install typically bonded to earth ground?  I would imagine a lot of the energy just got shunted straight to ground, although that's still way more voltage that the cells are rated for so it is quite impressive they survived.  Wonder what is the best defence against this, a ground tower on the property perhaps?  Also want it in a spot that it won't cast a shadow on the panels.

[TerraHertz]

Is the ground of a solar panel install typically bonded to earth ground?  I would imagine a lot of the energy just got shunted straight to ground, although that's still way more voltage that the cells are rated for so it is quite impressive they survived.  Wonder what is the best defence against this, a ground tower on the property perhaps?  Also want it in a spot that it won't cast a shadow on the panels.

Individual panels don't have a 'ground'. They have plus and neg leads, which on roof installations are typically wired in series over however many panels there are. What happens in typical inverters with the leads coming down from the roof panels I don't know.

But it's important to realize that these panels were not actually struck by lightning. All that visible tracking on the panels was caused by a very fast risetime E-field pulse radiating from the nearby lightning strike. The field orientation was vertical (as you'd expect) hence all the tracings run pretty close to vertical on the panels.
So what was happening was charge carriers (electrons most likely) that were present on the cell surface were being pulled vertically along the surface. So fast that the metal (silver) conductor strips right beside the arc traces had almost no effect on the trace paths.

I still have to do some drawings of the panel series layout, relative to the field orientation. The diodes and all the inter-cell wiring would have played no part in the cell tracing, but would see sharp HV current pulses 'a while' (nanoseconds) later. Still enough to destroy most of the diodes.

It's a pity I have no way of knowing what order the panels were wired. I'd guess the ones with no damage were furthest from the strike.

[SeanB]

Not necessarily Guy, they just would have had to have been at the node of the EMP wave, which would have concentrated the energy nicely.  Initial pulse creates a sharp wave, which reflects from nearby things like the metal edges, roof struts, nearby poles and terrrain, and this reflection arrives as the next return pulse is travelling up the leader, and this mixing then makes areas of high potential and areas of low potential. Thus the lines curving alond, they follow a path of intersection of 2 or more curved reflections to make a standing wave pattern on the panels.

BTW how is your postal test coming along?

[System Error Message]

so how much damage does lightning cause to solar panels? Did you test whether or not it still produces the same amount of electricity? In tropical countries you get sun all year round but with bad storms.

[Red Squirrel]

I wonder if some kind of grid/mesh over the panels could help, or if the energy is simply too great.  Something like what they'd use for a dog cage.  Of course you'd want it to be grounded.     You would lose a bit of performance from the panels though.     As for panels being grounded I was talking about the string.  Ex: you have a string of 4 panels, would the negative side be earth grounded with the rest of the electrical system?  Or are they normally floating?  Wonder what is better?

[bitslice]

A question: has anyone broken open a large (190W) solar panel, and examined the internal construction? With these panels, looking through the glass it _looks_ like the cells are simply sitting in a cavity between the front glass and rear backing sheet. But the arc-tracing photos look like there's actually a very clear filler material.

Optically clear silicone is used
http://www.dowcorning.com/content/discover/discovershowcase/solar.aspx

It's really quite expensive too.

[jnissen]

Is the ground of a solar panel install typically bonded to earth ground?  I would imagine a lot of the energy just got shunted straight to ground, although that's still way more voltage that the cells are rated for so it is quite impressive they survived.  Wonder what is the best defence against this, a ground tower on the property perhaps?  Also want it in a spot that it won't cast a shadow on the panels.

Individual panels don't have a 'ground'. They have plus and neg leads, which on roof installations are typically wired in series over however many panels there are. What happens in typical inverters with the leads coming down from the roof panels I don't know.

But it's important to realize that these panels were not actually struck by lightning. All that visible tracking on the panels was caused by a very fast risetime E-field pulse radiating from the nearby lightning strike. The field orientation was vertical (as you'd expect) hence all the tracings run pretty close to vertical on the panels.
So what was happening was charge carriers (electrons most likely) that were present on the cell surface were being pulled vertically along the surface. So fast that the metal (silver) conductor strips right beside the arc traces had almost no effect on the trace paths.

I still have to do some drawings of the panel series layout, relative to the field orientation. The diodes and all the inter-cell wiring would have played no part in the cell tracing, but would see sharp HV current pulses 'a while' (nanoseconds) later. Still enough to destroy most of the diodes.

It's a pity I have no way of knowing what order the panels were wired. I'd guess the ones with no damage were furthest from the strike.

Panel frames and such should all be electrically grounded to earth ground. This is a requirement here in central Texas. Each panel must be wired into a solid copper grounding wire. In my case a single large gauge bare copper wire is used for multiple panels. They sell special clamp screws that cut through the frame anodizing and accept the ground wire.