Solyndra, the missing story September 16, 2011Posted by Maury Markowitz in solar.
There’s been a lot of press on the Solyndra meltdown, but it’s entirely focussed on politics. I guess that’s not entirely surprising, given the way the press works.
But one would hope that someone would cover what caused the company to fail in the first place. From what I’ve seen, no one has – they’re just happily parroting Solyndra’s claim that China forced them out of business.
Poppycock! The product stank. Here’s why…
Rolling the dough
Normal solar panels have a list of problems that would fill this blog, but for this article there’s two issues that need to be considered.
The first is easy. Think about the difference in the heat from the sun in the summer and the winter. All of that is due to the angle of the sun – directly overhead in the summer and way down on the horizon in the winter. The same is true as the sun moves during the day; in the morning and sundown the sun simply doesn’t have the power that it does at noon.
A conventional panel can only face one direction at a time, so when you mount them to something like a roof you’re accepting the losses when the sun isn’t above the panel. You can address this by using a “tracker” which keeps the panels pointed at the sun, and generally in the Ontario area that’s worth a whopping 40% more power. But the trackers are pretty expensive, so the economics are always an issue.
Solyndra’s “whole idea” was a solution to he tracking problem. Instead of fixing the cells to a sheet of flat glass, like a normal panel, they wrapped theirs around a glass cylinder. When you laid the cylinder along a north-south line, some portion of the panel was pointed directly at the sun all of the time. It gained you some of the benefits of a tracker, but with no moving parts.
Aside from the obvious benefits, there’s a more subtle one to consider. Since the “panels” don’t have to be pointed directly at the sun, it’s OK to leave them lying flat on the roof. This means there’s nothing above the roofline for the wind to catch, which seriously reduces the loading. With normal panels you have to weigh it down or screw it into the roof, but with the Solyndra you just sat it right on the surface.
Basking in the summer sun
If this is such a great idea, why didn’t anyone do it earlier? Here’s where it gets a little complex.
Solar cells work by having sunlight boost the energy of an electron, technically from the “Fermi level” to the “conduction band”. The amount of energy you get is a function of the difference between these two energy levels. As the cell heats up, the Fermi level moves, and the energy that’s captured in this process drops. So that means that solar panels are at their least efficient on hot sunny days!
The 240 Watt Sharp panels we’re installing these days have a ” Temperature Coefficient” of -0.485%/°C . That means it will lose almost 1/2 of a percent of output for every degree the temperature goes up — or about 20% over Toronto-standard range of -10 to +30 degrees. That’s pretty typical for conventional panels.
Changing the cell material can improve things dramatically. In the case of one alternate material, CIGS, this is fairly dramatic. First Solar rates their Temperature Coefficient at 0.25%, just about 1/2 that of silicon.
Given the Toronto conditions, if I had these panels on my roof the variance in output would be only 10%, meaning I’d get 10% more power than I do now when it’s hot. When is it hot? In the summer, when I’m getting 15 hours of sunlight. Since I make about 2/3rds of my power during May to September, these panels would gain me about 15% ro 20% more power over the year.
That’s nothing to sneeze at.
The Solyndra difference
Solyndra’s product was a CIGS thin film wrapped around a glass tube, which was then inserted into a slightly larger glass tube. It looked exactly like a black fluorescent bulb. Solyndra can get away with this manufacturing, which is basically a miniature greenhouse, by relying on that low temperature coefficient of the CIGS cells. So it’s not so much that the CIGS drove the design, but that the CIGS allowed for their design to be built.
Each tube is inserted into a rack with electrical connectors. The racking leaves considerable room between the tubes, which lowers the wind, snow and rain loads on the tubes. The tubes themselves also have a limited area inside that’s covered with the cell because of the gap between the inner and outer surfaces. Maybe 40% of the overall area is covered with a cell.
Ok, so here’s the value proposition: using Solyndra panels you can just lay the “panels” right on the roof, no tools, no ballast, no worries. It also means you don’t have to space them out to avoid shading. And don’t get me wrong, as someone that’s been through repeated rounds of engineering due to wind loading, this is a major bonus.
So here’s the problem
So this all sounds great until you consider the fact that less than 40% of the roof is actually covered with cells. And since those cells are CIGS, not silicon, they’re also producing less power per square meter anyway. The result is that a Solyndra array produces way less power than a conventional one on the same roof.
Consider the case where the power being produced is a fixed price, say 25 cents a kWh. In this case the income from the array is a simple multiple of the “peak power” the array will generate on the perfect day. In the case of Toronto, a conventional panel at 30 degrees tilt will produce about 1150 kWh per 1000 Watts of peak power you install.
Now what do you do with that money? In the market Solyndra was targeting, large commercial roofs, you use the income to lease the roof from the property owner. You keep the difference between the income and the lease payments.
Since the Solyndra panel has so much less power per square meter, that means the system owner is making less money. They means they can afford to pay less to the property owner… maybe to the point where it doesn’t make sense to even bother. The only principle when designing for these sorts of systems it to get every single Watt you can onto that roof.
A conventional system on the same roof faces problems of its own. Since the panels have to be tilted up to face the sun, you need to space them out to avoid the panels shading each other when the sun is low in the sky in the winter. But that spacing, even up here in Toronto at high latitudes, takes up about 50% in the worst case, and you can reduce that if you’re willing to accept a little loss in performance. 30 to 40% loss is not unreasonable.
So in one case I have conventional technology that’s being driven down in price through supply and demand and gets about 60% “area efficiency”, and in the other I have a custom system that’s made by a single vendor and gets maybe 30% area efficiency — oh, did I mention CIGS is less efficient than silicon?
It just didn’t make sense
I ran the numbers six ways to Sunday and it just never worked out. There were definitely some cases where I might consider it, where the structural limits of the roof were questionable for instance, but all of the other benefits, including much lower install costs, were wiped out by being able to use off the shelf equipment when you use standard panels. I can’t imagine that other installers came to radically different conclusions.
Solyndra, and the administration, can blame this on China all they want, but in my opinion the problem was the product, not the country.