Spherical Solar Cells Soak Up Scattered Sunlight | Hacker News Reader

Spherical Solar Cells Soak Up Scattered Sunlight

spectrum.ieee.org

44 pointsheadalgorithm33 comments

My home's solar installation exhibits the curious behavior of producing its maximum overall single day output when there is summer morning fog which burns off around 9AM. This effect is clearly visible in the production graphs, which show a steeper energy production increase earlier in the day when these conditions are present.

My theory is that the fog moves the onset of generation earlier by reflecting low angle morning sunlight from the east down to my solar panels. The 9AM fog burnoff is critical to maximizing production because after that time, the fog begins to inhibit the production of electricity rather than enhance it.

It is also true that heat will add resistance to the electrical installation. So most of the time energy producten is way better when there is sun but it's cool.

Energy producten can be much better in spring than in summer for this reason.

Reduced heat might be the most likely explanation – https://www.renvu.com/Learn/How-Temperature-Affects-Solar-Pa...

I thought so initially too, but during the early hours of the day when the phenomenon occurs, the temperature is nearly always very cool regardless of whether there is fog cover or not - well below the 25°C (77°F) described in that article.

Ambient air temperature is not the same as what a dark surface exposed directly to the sun will be at. The solar panels will be much hotter than the ambient air temperature if they're in direct sunlight.

In full sun panels can get up to ~150f, but they are not going to be that much above 77f ambient at 9AM.

The panels could have a small amount of condensation on them?

Other people are implying it, but to be more specific, the boost in cooling is the evaporative cooling impact of dew.

Long ago the World Solar Challenge competitors noticed that their cars had peak power output not on clear days but on overcast days. They explained it as having something to do with what frequencies get bounced back rather than radiating out into space.

Later on while discussing a solar power result that sounded like Maxwell's Demon, someone pointed out to me that solar cell performance is affected by the temperature of the cell. Either or both could be in play here.

Did you plot temperature versus output by any chance?

I didn't, but this is just after sunrise in the Bay Area, so the temperature is pretty low - not in the zone where solar panel efficiency starts to decline.

Right, wouldn't lower morning temperature be better for solar cell efficiency?

It's been a long time since I took semiconductors, but I vaguely recall a decrease in voltage at higher temperature.

You'd need to measure the temp of the actual cell, because it gets wonky. You can't just look at air temp.

It would be surprising to see such huge difference in cell temperature in sun vs fog when the ambient temperature itself is about only about 60F, which is typical, or maybe even high for 7-9AM on the type of days I'm describing.

Even if the cells were 80F on the sunny mornings vs 60F on the foggy ones (or a difference of roughly 10 Kelvin) that would result in about .5% to 1% drop [1] in efficiency, so I should see only slightly lower output on sunny vs foggy mornings.

Instead what I see is that initial production on foggy days is several times higher than on sunny days. The only explanation I have for this is some kind of local effect where the fog reflects light that would have otherwise gone straight west in the morning down to my solar panels.

1. https://i2.wp.com/synergyfiles.com/wp-content/uploads/2017/0...

Morning fog means the dew point is being crossed. If you control for temperature, as you have here, then the humidity is the deciding factor.

Can a hot solar panel shed heat faster into humid air than dry air? I'll bet if we look it up that the answer is 'yes'.

Enough to cause a 2-3x increase in panel efficiency? Seems unlikely. Instead the panels must be getting more light somehow.

I think you'd have to measure the panel temperatures to know when it gets too hot, rather than just tracking air temperature, true.

All objects lag behind changes in ambient temperature. I don't expect solar panels have a lot of thermal mass, but whatever they sit in front of might...

Yes, the panel temperature can be hugely different due to absorption of radiation (I don't think it's the thermal mass, which as you say is going to be negligible).

Ambient air temperature is not a good way to approximate the panel's temperature. There's a huge difference between ambient air temperature and an object's surface temperature due to (among other things) the latter's absorption/emittance of radiation.

As I mentioned in another comment, I think what's happening is the dew (fog) on the surface of the panel is evaporating and the panel is cooling due to the evaporative cooling impact. Specifically, the surface gets cooler as latent heat is used to phase change liquid to vapour, and degrades the air film on the surface (which also contributes to thermal resistance of panel). This would also explain why you're getting such a huge peak, since the latent heat impact is nonlinear. So the temperature of the solar panel is increasing linearly as the air/sun rises, then the dew starts evaporating and the temperature increase will stop.

That's surprising because what gets bounced back from overcast skies is long-wave radiation (infrared) - which I didn't think had an impact on solar panel production. This occurs because the absorptivity of a clear sky is higher then it's emissivity, so clear sky will in net absorb the long-wave radiation, which is why you get dew on grass and car windshields overnight.

Solyndra tried something similar (cylindrical cells) and it just wasn't worth the effort. Flat cells are so, so cheap now.

I think it's pretty much a hallmark of engineering that the cheapest physical solution will overcome 99% of all clever solutions that cost more than a fraction more. This also goes for conventional solutions vs exotic ones.

It's like "better, faster, cheaper - pick cheaper"

Alliterative articles are agreeable and alleviate anguish

AFAIK large part of price of classical solar cells is attaching the leads (by manual soldering). Tesla solar roof is one ingenious solution to this problem.

I suspect it could be a problem with these spheres.

Genuinely ignorant, how does the solar roof reduce the need for manual soldering?

Connectors.... thousands of them. Small tiles On flexible structures like roofs have connectivity issues too.

Tesla's only party trick is tiles that looks a little like roof shingles. They look less intrusive than flat panels. Imho it is about as useful as painting turbines green so they look like trees.

Man, if that’s really the case, the only screwup in the first principle thinking is relying on downstream habits to adapt to solar vs traditional shingles. Working with traditional materials is a completely different building flow than the solar panel, obviously. Smart overall, might be difficult to execute.

The roof part of Tesla solar roof should presumably last a long time. It also avoids drilling holes in your shingles, combined that could represent a significant competitive advantage.

The solar part on the hand is overpriced and may limit how long the roofing lasts. Which means it’s only really competitive if you’re already constructing/replacing the roof and don’t yet have solar.

Aka best case is after ~25 years the roof reverts to a normal roof and then goes on to last another 25+ years. The worst case on the other hand is if the roof needs to be replace as soon as the solar panels reach end of life.

>> The roof part of Tesla solar roof should presumably last a long time.

Ya, but people aren't buying a tesla roof for the roof part. It is the solar part that matters. Roofs move, they swell and shift with the seasons. Over 25 years at least some of those thousands of connectors between solar tiles will have degraded and/or disconnected, leaving entire sections/rows of tiles disconnected. That is why, when you need to make thousands of connections in an electrical project meant to last many years, manual soldering is still the preferred tactic.

I worked in the Tesla factory making solar panels and I can attest that they are more snazzy and lighter then normal panels I've bought for my home (the wait list was to long for Tesla ones) I'm not sure they're much more durable.

They can take a hit straight on but if anything comes at the wrong angle the top glass layer shatters. It was so common that we had a designated spot to go dump the ones that would frequently break during the manufacturing process.

I have ~8yr old panels manufactured in China, one of them was shattered by what I suspect was a falling bullet (must've been a wild party...). Since the panels are wired in series, that reduced the output of the system by about half (the shattered panel was in the middle). Ironically I couldn't easily replace it, as new panels are higher output, so I would have to replace my inverter. I opted to just bypass the dead panel and have 19 working instead of 20.

> They can take a hit straight on but if anything comes at the wrong angle the top glass layer shatters.

hopefully the cybertruck and the panels will get better in future revs. :)

I want to know if they tested a hemisphere, and how well it did. Intuitively, it seems like the lower half can't be as productive per unit of solar material as the upper half.

Yeah, or even a quarter sphere since the sun only occupies the southern part of the hemisphere (north of equator).

Presumably you could have the spheres mounted on a reflective surface so you'd still power the upper- and lower- hemi's at the same time (or oblique spheroids or whatever they are when the sun's at an angle)

Why a sphere and not some other shape?

Reply on news.ycombinator.com