389 comments

[ 0.27 ms ] story [ 359 ms ] thread
Seems weird. How do you service it? How do conductors work? How do you keep random junk from blowing on top of it? How do you clean it? The photo in the story just looks like a giant square of PV material. Is that really what this is?
Maybe if it reduces installation cost enough they will be able to tolerate some inefficiencies.
Re cleaning, they have a cute little robot that you can see on this page: https://www.erthos.com/energyservices It's also visible in the photo in the article.
They should fix the wheel spacing so it rolls on the metal frame and doesn't scuff up the edge of the panels
> They should fix the wheel spacing

Sounds like they optimized it for their use case:

"The load of the robot is distributed almost entirely to the module frames rather than the glass module"

The robot is cute!

I wonder what happens after a major rain though. I suppose the panels are weatherproof. But they lie directly on the ground, and I did not notice any mention of a drainage system. The panels will eventually sag under load from rainwater, preventing it from flowing off them.

They mention that their installation can withstand a hurricane. I understand how it works for the wind load, but every hurricane I witnessed brought a lot of rain.

EDIT: Apparently they embrace flooding, and say that their panels and connectors can withstand being submerged in water. That's the spirit.

Presumably the plan is for them to be high enough relative to the surroundings that the water runs off rather than pools.
I wonder how they manage drainage and surface flooding?

Image of robot on panels: https://static.wixstatic.com/media/3b0818_1667facfc56e475e8f...

The image also seems to show water damage in the corner of the closest panel.

At a guess, they target areas without heavy rainfall, and fast draining soils. I didn’t see any drainage works in the video https://vimeo.com/556421759, nor did my google-fu help me find anything where they address the issue.

Edit: from https://www.canarymedia.com/articles/solar/erthos-rakes-in-1...

  How do the installations perform in the rain and snow?  “Our hydrology report proves that an Erthos plant is almost the same as native soils with respect to accumulated water depth and velocity in rainstorms. All of our designs include professional civil design that includes water runoff management and containment basins as per the jurisdiction’s application of building code and other local requirements.”

  What about flooding?  “The glass/ glass modules and the connectors we specify are all rated for submersion, so flooding is not a catastrophic event in case it does occur.”
A site graded to slightly angle towards the sun would also keep the panels drained.
Blemishes on solar panels aren't necessarily bad or have a huge effect on generation or durability. These could also be previously used panels in the image, which are often sold at enough of a discount to offset the loss in power generation.
Wish I'd had that at Burning Man this year. Getting dust off a 200 panel array was no fun job...
I doubt this robot would last long with gypsum dust.
It better. And rain plus other kinds of dust too. Reality is harsh.
Thanks for posting that! Made me realize that when all the panels are laid over a huge area in essentially the same plane that it must be so much easier and cheaper to clean. Just put this robot on it and let it go, roomba-style (OK, not exactly roomba-style of going over the same spot 50 times, but you get the idea). Seems like it would be a lot cheaper than what would be needed to keep rack-mounted systems clean.
> Re cleaning, they have a cute little robot

Ever cleaned dried bird droppings off a windshield? I'd be pretty surprised if solar-farm Roomba was up to it.

Any solar installation will face that problem though
My point only concerned the robot. Other solar installations won't have a tiny self-driving Zamboni that's ineffective at cleaning bird poo and is bound to be stolen. They use renewable and cost-effective elbow grease instead.
I can't see how bird poo could be a problem here: there's no food, nesting site, nor high spot to hang out anywhere near it. Theft could be a problem, though I'd assume they'd have some sort of protection hidden on it.
> won't have a tiny self-driving Zamboni that's ineffective at cleaning bird poo and is bound to be stolen

Why do you think it's ineffective, they could have a pressure washer inside that thing

> Why do you think it's ineffective,

It appears too small to have enough mass to use gravity to be able to do a thing with dried guano. Also, because it has wheels, it gives me the impression that it moves along and does whatever cleaning its capable of as it goes rather than lingering anywhere to thoroughly clean one spot.

> they could have a pressure washer inside that thing

How big of a water tank do you think?

The solution here seems to create a zone that is completely inhospitable to the ecosystem that should exist in that spot. In the picture in the article, the bird poop will be highly concentrated in the forest area that is nearby. While there will be some birds that fly over this dead zone, I bet the droppings will be minimal.

Edit: The article mentions a 100 MW installation. At 2.5 acres per MW, that is 40 acres or 0.25 miles x 0.25 miles. While there will certainly be some service roads, no matter what the ecosystem was before, it will be covered with something that doesn't support plant material, insects, etc. that may be consumed directly by birds or the small animals upon which birds prey. Birds will find a more hospitable place to poop.

maybe use a little potassium carbonate to convert the uric acid in the bird shit to dipotassium urate, thus increasing its water-solubility 300×, and follow up with a strong buffer that's mildly acidic like monopotassium phosphate to prevent any stray residues from the solution from causing alkali corrosion of the glass when it dries

probably there are a lot of possibilities you haven't tried on your windshield yet

Why is that not sold as a specific bird droppings cleaning product yet?
i don't think i've ever seen a specific bird shit cleaning product of any kind, much less one specifically for glass
Shouldn’t matter in the grand scheme and it will probably be clean after 5-6 passes
Thanks, I was wondering how they were going to keep it clean, and the linked article doesn't have the word "clean" in it at all, so it could be improved by discussing more about that. I also see now that the little robot is in the pic in the linked article too.
Cool. I can see myself using a Scooba. I’m serious.
(comment deleted)
First, they are banking on the fact that solar doesn't need a lot of repair and maintenance in general, and their design decreases some of the stresses that racked solar panels encounter. I imagine they are also over-sizing the system, and adding remote disconnects so they can disable a certain number of panels and still meet the contract.

And then when repair is needed, they just walk on it[1]. Seriously. I'm very curious as to what these pads they mention are like - big foam snow shoes, or walkways they rollout along a seam?

[1] https://www.erthos.com/reducing-degradation-rates-with-earth...

Likely something akin to deminers weight-spreading shoe pads. They have to meet panel manufacturer's certified pressure spec.
I assume they need to worry about scratches. An XXXXXL clean room bootie wouldn't work because it would pick up sand and grit as you walk across the panels.
A few scratches aren't going to be a big deal. Not like they not going to get dirty and dinged up from the wind anyway.
The real world just isn’t that clean. There’s no way that a weight spreading shoe will do more than some random frequent dust cleanings
Oh, that's interesting. And I guess if they do manage to break a few panels, they can be replaced cheaply. Its probably still cheaper than dealing with racking.
They are ridiculously strong. I've had a set of 8 on a tracker be blown over by the wind (in spite of ample foundation according to the people that sold me the gear), land on the edge and not get damaged at all. That was a pretty heavy impact too, the whole thing was 30 cm into the ground.
(comment deleted)
I would have expected it to be like agricultural products -- rows of corn with spaces between rows, so you could access the interior without stepping over the outer panels.
yeah good point, why not walking/drain space between rows? optimizing for minimal area?
That's a different approach, called agrivoltaics.

3 meter high frames with the panels on top, wide row spacing (about 10x), and crops in the ground. Minimal harvest yield loss (sometimes even improvement, as the shading reduces stress on the plants), and selling power from the same land. It could be marketed as "zero land needed" PV power.

With Erthos's on-ground panels, a robot with fat soft tires rolls over the panels. No walking needed.

That comment seemed to only refer to spacing panels like rows of plants are spaced in agriculture to allow easier access.
> How do conductors work?

I assume the conductors will be direct burial cable or single conductors run inside PVC conduit, just like any other outdoor electrical installation. The connectors and junction boxes are probably IP68 rated to handle flooding.

Agreed. The dust is likely to stay there and not blow off and also type of installation does not allow of undergrowth much less dual use of land (parking lot or other potential land use).
They have a little robot for the dust.

Land costs are still generally irrelevant for solar as in well under 5% the cost of this install and you can recover that after at the end of the panel lifespan.

I have a similar setup and rain takes care of the dust. A bigger problem is the leaves and bird droppings that get stuck, even that gets dealt with by the rain (eventually) but it has to be a pretty good downpour for that to wash off. Fortunately here in NL we have no shortage of those.
Install cost is ~50-75% the total cost of the system. If you can bring install cost down your tolerance for panel failure can be quite high while still having a better roi
It is correct that mounting costs and labor can be a large portion of the total BOM.

Even for a large off grid whole home PV system that can operate through December/January at high latitudes.

Let's say for an example you wanted to DIY a PV system that would be much too large to fit on the roof of a normal sized house.

Go calculate the cost of buying 30000 kW of good quality 72-cell PV panels rated at 380W STC each. It'll be something like 80 pieces at about $130 per piece.

Usually would ship as 20 panels per pallet, so call it four fully loaded pallets of 72-cell panels.

At 34 cents/W STC rating, PV panel cost from distributor something like $10,400 to $12000 USD.

The foundation work and poles/racking to do a basic ground mount will be a huge cost on top of that. Labor is a big part of it. If you're hiring people to build it the labor and ground mount gear and things like basic foundation work/screw piles/steel tubes set into concrete could easily cost you another 10 grand from a local contractor.

Something generally along these lines or an industry competitor of it:

https://www.ironridge.com/ground-based/

(Not discussing inverters/charge controllers/batteries/disconnect boxes and wiring here).

As a probably silly idea, I wonder if the panels could each sit on an inflated/foam and buoyant pad, be chained by flexible cables and then be pinned to the ground at intervals. That might give it some flexibility in the event of flooding.
I bought 20kw of solar from China that arrived this year, the quote I got for installation was about what I paid for the whole system. I'm thinking about just installing it myself.
What was the cost? And which panel did you buy?
I think including shipping and import fees it was about $26k. I looked at getting panels directly from a supplier, but it's very difficult to figure out which companies actually manufacture things. I ended up buying from one of the many companies that buys components and packages them together into complete systems.
I hope the fields never flood and the solar farm is under water
> The glass/ glass modules and the connectors we specify are all rated for submersion, so flooding is not a catastrophic event in case it does occur.

From: https://www.canarymedia.com/articles/solar/erthos-rakes-in-1...

Maybe I'm just being pessimistic, but the failure mechanism is the erosion of the underlying denuded ground in a large storm to the point that the underlying structure moves.
I don’t think that’s being too pessimistic. I didn’t think about that. The solution would be some sort of concrete footing or something—but I’m sure that is more expensive than just mounting these panels in the air.
Would it not contribute to flooding as well, unless drainage is somehow improved around each panel.
> less than 2.5 acres per megawatt

This is actually quite amazing. I wonder what the lifetime is for the panels.

That sounds like if I have a decently-sized suburban yard, I should be able to throw enough of these out there to run my life.
My co-worker did just that, but at a much higher latitude(49°) than Texas - he couldn't be arsed with building racks for the panels himself and having people do it for him would cost too much in his view.

While the panels are indeed cheap this, along with the DIY wood gas combined heat and power generator, are measures taken for energy security, not profit.

Yeah, nah, these are for utility scale PV plants, sorry.

In a few years, you might be able to buy a roll of PV film in a box and unroll it in your yard. The box it comes in would have all the gubbins to connect up to the house.

Not quite here yet, though.

You don't even really need to tile your yard with solar panels. Depending on location, you could be self-sufficient(ish) with 10 kW of solar, 25kWh of batteries, and a 10 kVA generator for cloudy weeks.

Total cost is around $40k.

However, unless you're rural or somewhere with poor grid reliability it's probably not worth the expense of being off-grid. Generally speaking, due to generous feed-in tariffs you would be (financially) better off staying connected to the grid than spending the additional money required to handle periods of cloudy weather.

Usually the expected life is ~30 years. They may last longer. I do wonder if pests and moisture will be early failure causes in this configuration though. Then again, if it saves enough money, maybe that doesn't matter.

TBH, the part that lasts the least amount of time is often the inverter.

> "Install in half the cost, half the time on 33% of the land"

If the panels don't point directly at the sun, then you lose much of the efficiency.

I wonder how the robot cleaner handles bird poop.

Horizontal panels also collect dust, so they still need to be propping them on the north side.
They don't look propped at all in the photos. Also I think propping would be hard for the cleaning robot as it looks like is a super simple little robot with small wheels and just rolls straightened over them.
Yah its probably a case that designing a vacmop style device (possibly purchased off the open market?) and having it drive around on a flat surface is both cheaper and keeps the panels cleaner than a custom bot that doesn't clean as frequently because it costs to much.

My question is whether sitting on the ground itself causes problems with efficiency due to the panels getting hotter than they would with some airflow under them.

Erthos spend quite a bit of time designing their own autonomous cleaning and repair robot.
All panels collect dust, but this setup is much more challenging to clean (and to generally service, i.e to replace a faulty panel)
This seems much easier to clean? A gigantic flat slab is pretty easy to automate something vs disconnected racks.
(Modern utility scale PV installations are already using automated cleaning solutions).

It is much better from POV of the cleaning robot maneuvering requirements. It's also much better in terms of single robot can access the entire installation.

But it's worse in terms of how much distance the dust should be pushed before it's off panel (as I don't see any gaps there)

Alternatively if the robot is a vacuum or something of the like, it could just store the excess dust and dump it at a station on the edge.
Service trade-off seems pretty marginal. They say the panels aren't tied down to the ground in any way, so you might be able to just pick one up and disconnect it
They have a robot that cleans the panels. It's a big, flat surface so they don't really need anything much more advanced than a Roomba.

Apparently they also have special shoes that maintenance people can wear that distribute the weight properly so they can walk on them if they have to.

Sounds like they're pretty confident this isn't an issue:

>Our fees are based on the plant producing at its optimal performance. If the plant underperforms for any reason, we curtail our fees – creating strong incentive and perfect alignment with the long-term asset owner.

It all comes down to the economic tradeoffs. Panels are cheap and everything else is expensive.

If you can get 75% efficiency for half the cost, your return on investment is 50% higher

Panels may be cheap, maintenance / replacement is less so, because humans at work.
They use a robot for regular cleaning; looks like the design will make that cheaper too as it's basically a fancy Roomba.
Replacing might be cheaper as well without any mounting racks or tracks.
Everyone is talking about maintenance ITT. Is it that big of a concern? My parents have had solar on their roof for almost 10 years now with zero issues. I imagine newer panels last longer.

So what's the actual maintenance cost? I could imagine it's going to be cheaper to just install new panels every few decades than constantly maintaining the installation to be 100% capacity.

For a grid scale installation, let’s say 5% of panels are dead due to whatever failure. I am not sure the rational response would be to do any maintenance vs installing more capacity. Twenty years later rip out all of the old panels, replace with new. Repeat forever.
Solar in Texas seems like an obvious win in just about any configuration. But I wonder how this would work at higher northern latitudes in the winter. I suspect efficiency would be pretty bad with a 20 degree incidence.
Doesn't have to be effective everywhere for it to be useful.
the unintuitive part is that land is even cheaper than panels, so you'd think that spending more panels to use less land would be a losing proposition

the interesting question is whether the costs of racking, grading, cleaning, repairing, etc., go up or down, and if they go down, whether it's enough of a reduction to make up for the larger amount of solar panels per average watt, as they say it will be

Based on my calculations, at my latitude (40 degrees North), you would need about 16% more panels to generate an equivalent amount of energy per year. This isn't taking into account potential issues with snow buildup (which theoretically would be worse with flat panels) or the effects of cooling (which theoretically could be better due to contact with a thermal sink, the ground), but it's probably pretty close. Even if 20% more panels are required, that means capital costs are superior as long as panel costs are less than 5x racking material and labor costs. Currently panel costs are more like 3x racking costs and will probably continue to decline. Racking costs will probably not go down unless steel prices go down pretty significantly. The only thing that surprises me is that there are not more companies doing this. Perhaps there are factors that neither Erthos nor I am properly considering, but I think this is how most utility solar projects will be done in 5-10 years.
We put in 40 KW and in the racking was a nightmare. The government required soil analysis, reinforced racks, and cement pilings 4 ft deep. Probably could have put in double the panels if we didn't have to deal with the f*** racks

I think this approach has interesting applications for small-scale solar in rural environments if the permitting can be streamlined

I mean with erosion, flooding, and the like, I'm not sure this sort of mounting on the ground would be any less of a hassle...
Most of those things seem nonissues in desert climates with flat Terrain.
steel prices are probably not a significant component of racking prices; scrap steel costs 50¢ a kg

the numbers they gave make the pv module prices seem slightly higher than the cited racking prices (15¢ per watt) but it's a little hard to be sure because of the numerous kinds of watts involved

What you lose with Erthos' approach is any possibility of using bifacial panels.

In regions that get snow, bifacial panels (that use light reflected onto the back of the panels as well as light from the front) get a lot better output in the winter, increasing the annual capacity factor and therefor return on investment. (Winter electricity can be more valuable too, in those regions.)

Horses for courses.

you can have a back mirror on this, they probably do
I don't understand why they cant dig some kind of trench or mound to put them on at least, to angle them more towards the sun. There's plenty of agricultural equipment designed to cultivate soil (first example I found[0]). That could take care of the drainage as well, although I don't know how much it rains in Texas.

[0] https://www.asa-lift.com/asalift/producttypes/bodenbearbeitu...

> why they cant dig some kind of trench or mound to put them on at least, to angle them more towards the sun

Trench will flood. Mound needs to be maintained. Cheaper to put more static panels flat than trying to get every electron out of them.

> why they cant dig some kind of trench or mound to put them on at least

Just use the rack then, it's exactly the same result.

I think using the dirt that's already there and building the structure with some kind of machine would be cheaper than having someone mounting them on a steel rack. See for instance large potato fields [0]. The point is instead of putting them on flat ground, make the ground have a better shape first.

[0] http://www.thompsonpotatofarm.com/common-tater/banking-the-p...

The racks are really easy, well known technology. That was my point. If you're going to give up the entire reason for laying them flat on the ground, you're going back to the way existing solar farms work, so just use the same proven technology.
(comment deleted)
> If the panels don't point directly at the sun, then you lose much of the efficiency.

This is true for a single panel. But the amount of sunlight which hits an acre of land is constant. If the land is 100% covered with panels, the panels will collect 100% of the available sunlight.

Installations that tilt the panels have lots of space between the panels.

> This is true for a single panel. But the amount of sunlight which hits an acre of land is constant. If the land is 100% covered with panels, the panels will collect 100% of the available sunlight.

That is not true from my understanding. The increased solar angle of incidence effects how much of the energy reflects back into the sky. Having more panels next to it without a gap won't change that. Yes you can fit more panels in the same area without putting them on an angle but they will be quite a bit less efficient because more light will be reflecting back up per sqft of panel which is what matters cost wise.

The anti reflective layers and texture on the panels are usually pretty good, such that then reflection is less than the area loss due to the angle.
Yes, one of the many areas of improvement of PV modules in the last couple of years has been reducing reflections from the surface of the glass and from the glass-cell boundary in panels. Also changed cell technology (n-TOPCon) helps with reflections and recombination within the cells.
it's true that you get 100% of the sunlight that falls on the land, or near enough (albedo is not literally zero at any angle) but you need more solar panels per output watt

it'll be interesting to know if they're derating the nameplate capacity by cos(latitude) as they should be

Their agreements with their local grid operator will specify power delivery. Marketing numbers like an individual panel's nameplate output are not important at utility scale, which this is designed for.
probably the local transmission companies are interested in not only average ouptut but peak output, but also, there were a number of numbers in the article cited as "per watt" which look like they might be talking about peak watts
(comment deleted)
You get the majority of the power when the sun is directly overhead. You can make up the lost difference by the space saving and just adding more panels.
The sun is never directly overhead except on the summer solstice at noon, assuming you're in roughly the part of your timezone.
And then only at the equator. Here in Iowa at 42' North, you'd want a panel tilted 42' South to start with.
Most utility scale installations don’t track the sun- too much cost. This takes cost-saving to a new level though.
The reason they are tilted is to maximize irradiance hitting the panel. At a 0 degree angle (flat on the ground) you get a a lot around noon and then very little.

This approach surely reduces land usage but what is the output per acre?

I’d be really surprised if it’s higher than with tilted modules.

They don't claim to outperform fixed-tile or SAT on that KPI. They claim to reduce upfront cost of installation, construction time, and general project risk.
I expect that they are getting lower output per acre, but in places where land is cheap and as solar panels continue to get cheaper, the money saved on building the support structures could be worth those losses.
They're getting quite a lot more output per acre, by being able to use more panels.
The article claims it's much higher output per acre:

> conventional solar technologies, which typically require five to 10 acres of land per megawatt of capacity. Erthos claims that its mounting scheme requires less than 2.5 acres per megawatt.

They claim the power per acre is 4x higher than tilted panels. Seems like a stretch, but I don't know how bad the density is in tilted installations. I guess I have seen some where you can drive between rows
Density in tilted installations is quite bad. If you want to capture morning and evening sun at an optimal angle you have to space the panels out a lot, like 5-10 panel heights. You can have them closer, but then you get shading, which defeats the purpose of tilting the panels.
The amount of power landing on an acre is fixed, what you can achieve by tilting is having less solar panel surface area per ground cover area. If solar panels are cheaper than the mounting hardware (wow) then there is no reason not to let them lie flat on the ground (it's not as if the racks were holding them above tree shadows, or anything).
> If solar panels are cheaper than the mounting hardware (wow)

I'm surprised this surprises people... Every electronics hobbyist knows that electronics are cheap as dirt while any kind of box, mount, rail or whatever is BY REALLY FAR the most expensive part of a project, even when buying massivly mass produced cheap Chinese junk.

The electronics you're talking about are rice grain sized little things, while the solar panels are big sheets of silicon inside of glass, etc.
Economies of scale make that irrelevant.
Those grain sized things have pretty high packaging costs in relation to what you are actually buying.
You are surprised that this surprised people because electronic hobbyists know this? Most people are not electronic hobbyists so this should probably not surprise you
I just find it interesting, the difference different perspectives can make, especially on a website where people are often bikeshedding things they have no experience with.
It's amazing how so few people understand this.
I think people are imagining that solar panels have directionality like most photodetectors do.
This is a great way of thinking about it, but don't you lose a bit more due to increased reflection from the glass surface at low incident angles? Probably not enough to make a difference a low latitudes in the summer, but at high latitudes in the winter I think it might be a significant difference: https://en.wikipedia.org/wiki/Fresnel_equations#/media/File:...
Partially answering myself, 'sacred_numbers' posted a link elsewhere in this thread that suggests this effect might be quite small: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6611928/. The paper concludes that for a cell with a good Anti-Reflective Coating (ARC), the reduction in efficiency due to reflection at a 60 deg incidence is only 2%.

On the other hand, the incidence for a flat mounted panel beyond 37 degrees of latitude on the winter solstice is greater than 60 degrees, and it's not clear (to me at least) how well the ARC on an average panel will continue working after years of outdoor usage. My guess is reflection is probably a real issue, but not a stopper unless one is already in a marginal situation.

The biggest issue going north is probably snow.

I'm at 54N and my output has been 0 for the past month as we've had a lot of snowfall. My panels are at 20 degrees, so even in the few sunny days we've had it's not been enough to melt the snow. A steeper angle probably would have cleared it a few times.

Usually the beginning of the year has more sun but it's colder, so I'll see what happens then.

The solar radiation wattage per unit of surface area is dependent on the angle that surface is to the sun. The angle is dependent on the season and time of day, so the amount of power is not fixed.
Per unit of panel surface area, not per unit of land area. If the sun is coming in at an angle, you'd be able to collect all that's available with less panel area than total land area by angling them (or equivalently, in this new configuration you need more panel area than you otherwise would), but in their estimation, it's cheaper to just get more panels than it is to buy and install racks.
What the comment wanted to say is that the power per unit of ground surface area is fixed for a given location and time, i.e. it does not matter [1] whether you cover a given area with panels angled towards the sun or lying flat on the ground, at least if one only looks at the available power. There is of course a difference in the solar panel area required to cover a given area of ground surface - solar panels lying flat on the ground will obviously have to have the same area as the ground surface area while panels angled towards the sun will only require a fraction of the ground surface area equal to the sine of the angle of the sun above the horizon.

[1] For a sufficiently large area so that effects on the edge are negligible.

You could write a few pages of all the things that the power available depends on, but you don't need to because it's fixed relative to the variables under consideration.
>The solar radiation wattage per unit of surface area is dependent on the angle that surface is to the sun.

A tilted solar panel casts a shadow that is bigger than its actual area. Mounting the panel flush to the ground means it casts a shadow exactly equal to its area.

The shadow represents the captured sunlight so the first panel covers more surface area than the second panel, which allows you to reduce the number of panels to cover the same amount of surface area. The entire point of this article is that you can just put the saved costs into buying more solar panels.

Used solar panels are very cheap but usually only the solar panels are replaced and the mounts are kept and fitted with new panels. So for companies that want to use used panels their primary cost is actually in the mounting hardware and not the panels.

the article claims 15¢ per watt of mounting hardware; last i checked solarserver says pv modules are about 20¢ per peak watt

presumably that was 15¢ per peak watt but the article doesn't actually say

> you get a a lot around noon and then very little.

That's a little harsher than reality. You get a very pretty bell curve. I have a flat panel on the roof of my RV and I track the output over time. I'm not 100% how much of the loss in output is because the incidence to the panel is changing, or because the light from the sun is going through more atmosphere. Probably a little of both, but in any case the panel is still plenty useful even when not pointed directly at the sun.

Seems like they would have to do some grading to clear anyway, why not grade with some tilt.
With tilted modules, you'd normally space them out quite a bit so the shadows of one aren't falling on the module next to it. If they're all flat, that's not a problem so you can space them closer. So, it makes sense that they'd get more power per acre than the conventional approach -- the panels are individually less efficient, but there's a lot more total solar panel area per acre.

That might not always be a good tradeoff, but maybe at least some of the time it is.

Texas is pretty far south. If you use https://pvwatts.nrel.gov/pvwatts.php there is about a 9% increase in total output over the year for optimal tilt(27 vs 0) but then you also need to space modules.

There is hourly data if you are interested but even Jan 1 the panels produce for ~7-8 hours. The 3 hour around noon it's about 1/2 the output for the day (for Jan 1).

Seems lying them flat also makes their cleaning robot able to easily maneuver, meaning they don't need to leave any space in between panels for humans to perform maintenance. Pros: reduces land usage as you mention, but also less humans needed for maintenance.
Going to the other direction, I wonder what would be the maximum archieveable output per acre without height limitations.

Would you have a forest of tall towers, or one really tall panel covered building, or something in between?

Per day? Per acre? Per dollar?

It's all about the metric you choose. That's the true issue.

We’ll the cost isn’t just the racks

You’ve got to deal with permits for the structures. Installation. And then you have to do lawn trimming around all the racks. This can save on all of that.

(comment deleted)
aussies are doing a similar thing although they install by pulling via tractor and the panels concertina out see https://suncable.energy/

they are doing an underwater cable to supply energy to singapore

Dead ground is not so good for biodiversity, with the standard way you have wild grass and flowers beneath.
Covering the ground with impermeable surfaces isn't great either. (Maybe they have drain holes at regular intervals so that's not as much of a problem?)

I'm guessing they will probably need a tall fence around the outside to keep deer away.

The optics are bad-- like this advanced civilization here on Earth decided to make a Dyson sphere of our own planet.
I take your point, but arguably the optics of burning up every last drop of fossil fuel are worse.
Already, 96% of mammal biomass is humans and livestock. No one has the power to stop this process, to do what it would take. Enforcing a strict max-2 child policy on the whole world and/or telling people to live in the pods and eat the bugs. Hopefully, once our civilization goes interplanetary, we can try to restore a bit of what was lost.
Fertility rate is already below replacement for all of the world except sub-Saharan Africa, and it's dropping there too. We're headed for population decline within a century unless something changes dramatically.
But still better than growing corn or soybeans on the ground.
agrivoltaics requires more space between racks to provide sufficient light for the plants to grow. that's neither a good or a bad thing - it's just a statement of fact about the paired compromises that arise from combining PV generation with agriculture. It is still likely a good idea in many places.
Interesting idea. I've been toying with the idea of doing a ground mount install at home, but maybe I should try this instead to get more density. I love toys...
I do not believe their claimed savings. Steel is a very cheap building material and at the scale of solar farms, the cost should be very low. (As an alternative to stainless steel on the other hand, it may make more sense)
> From 2010 to 2021, the levelized cost of installing utility-scale solar fell 88% .... But in the last couple of years, supply-chain issues have halted these price declines globally

How many "couple of years" have there been between 2021 and now?

I suppose if we're charitable, maybe "From 2010 to 2021" is programmer style [2010, 2021), i.e., ending at 1 Jan 2021. Then all of 2021 & 2022 form "a couple of years". (The book is more or less closed on 2022. Given that my "2 day shipping" from cyber Monday took 8 days, I'm not holding my breath on "the supply chain" fixing itself prior to people giving it up for the holiday retreat.)
The two time periods you have quoted are not necessarily non-overlapping time periods. Nothing from your quote implies that the couple of years begins at 2021.

Falling between two dates does not imply that it fell uniformly or that it fell for every year between those two dates.

At scale there is probably a fair bit of dead time if you want to get 100+MW of solar panels. So, it might be well over 2 years at this point.

Aka companies had agreed to deliver in 2020 for X$, but took longer to actually deliver.

I think "couple" is pretty typically accepted as being "two", at least in most US settings; I can't speak for international variations. In my home area of central Pennsylvania "couple" is a more general, meaning anywhere between 2 and 4, but I think that usage is rather the exception than the rule.

Nonetheless, we can probably give them a pass for saying "couple" instead of "nearly three" since January 2020.

A rare instance of xkcd being just plain wrong. ;)

"a couple" = 2

"a few" = 3 (maaaybe 4)

"several" = 4 to 7

"a handful" = context dependent, usually ambiguous

A few: count of 2 to 10% of a sample in my book.

There are only a few people who understand quantum mechanics, for example

Yeah I guess that works for me.
Also from central PA: my dad and I had this discussion years ago and he argued the syllables made the count: (1)A (2)Few and (1)A (2)Coup (3)le. I was in the other camp that a couple is two (husband and wife) and a few was much more loose on definition (2-5).
There has been one couple of years since the beginning of 2021 and now. Quite exactly like the text you quoted says.
Will Prowse did something similar (much smaller scale) on his YouTube channel last year. He set up a 6000 watt ground mount array in about 40 minutes and discusses the trade offs here:

https://youtu.be/71vME5k-oiw

His is more or less temporary.

What's great about him though is he already has a link up to the power company, so as he sees fit he can just reduce his bill by flaying them out in his courtyard. That or just mine btc/etc...

He's still an awesome dude and has helped me set up my own solar in the number of videos. Even better is his forum:

https://diysolarforum.com

Temporary solar panels are interesting to me - where I live there's a large chunk of the year where weekly solar output plummets. Stacking them in the garage overwinter is appealing, I might try this soon.
I would assume the risk of moving them each year is much bigger than any reduction in loss of function?
Meh. Throw a moving blanket over it and bungee it on. Could weld up a cart to roll them all at once sideways. Could use spring casters like for gates as a suspension. Or hold them in place with low durometer rubber. Or both. Probably want them sideways for footprint in my garage anyhow. It would be fun to optimize the process more each year.
Why bother stacking them in the garage during winter. Are you afraid of winter panel damage?
My property is very small and I currently use that back space to park my nice car during the winter so the daily is in the garage and the front driveway is easier to plow. Sounds weird to me typing it out, but I promise this arrangement makes more sense when living here.
(comment deleted)
Generally if you're grid-tied your setup needs to be approved by the power company and they have a surprising amount of say in what specific equipment by what vendors you can use. If you watch his videos on his specific home setup, he has a separate grid-tied system and his testing system and they don't touch. He even remarks that some of them wouldn't have been his choice.
I watched a video from the 8-bit guy about tying solar panels to the grid. He said that if you do that, your power fails at the same time as the grid's power, so you don't get any backup electrical power. For that reason, he recommended against tying your solar to the grid.
It’s completely possible to have backup power and be grid-tied. The main issue is your panels feeding the grid during a blackout, which can be dangerous for people trying to fix stuff. All that’s required (requirements will vary widely per region) is an automatic disconnect that disconnects your setup from the grid when the grid goes down. For example, Tthe new Enphase EQ8 micro inverters have an optional controller to do this. AFAIK the Enphase systems are the best selling micro inverter systems out there so new installations should have no issues being self sufficient during outages.
Anything that can be legally connected to grid has to be certified to stop supply back to grid in case of blackout. Is there any region where this isn't the case? I doubt it.
(comment deleted)
Definitely love me some Will Prowse. I bought one of the LiPo batteries he recommends for my campervan and have been super happy with it.
Interspersing standing panels with crops looks to be an effective strategy, some crops suffer from too much sunlight and require a little shading for optimal production (many vegetable-type crops fall into this category). Some strategies employ vertical bifacial panels. See:

https://youtu.be/lgZBlD-TCFE

Solar Panels Plus Farming? Agrivoltaics Explained

Vertical panels strike me as needing quite robust supports for windy days.
Don't do singular poles, mesh them together?
Hail enters the chat

Maybe they should prop them up on cinder blocks at a minimum?

What about flooding?

If they lie on a flat, impact absorbing material (i.e. sand), then hail would do less damage than for standing panels. One could walk on them. If, as you propose, put flat rigid cement below, I'd thing you could drive on them.
(comment deleted)
My panels were cheeper than the mounting equipment... If I had tons of space I would scale up panels and put them on the ground instead of mounting/tilting gear.
The price of panels has fallen so much that installation is now a significant proportion of the cost. So this is a good idea.

But without any airflow behind the panels they will heat up, which will reduce efficacy. This is the main reason BIPV (solar roof tiles, in this case) has failed for decades. So this is a bad idea.

Which is it? I suspect, based on the BIPV example, that this will probably not work. It would be cool if this suspicion turned out to be wrong!

In this case, I think the earth acts as a heat sink. Or so they claim on their website (https://www.erthos.com)

Neither their press release nor the article says where exactly in Texas it is, but I bet it would make sense to put it in the desert where it gets cold at night.

Earth is a much worse heat sink than moving air. You can check it while stepping on the earth on a sunny day.
What has driven the cost of panels down? Do we have more mines now, or cheaper labor, or something else?
Economies of scale, better energy efficiency in smelting polysilicon and making ingots of purified silicon, better ingot quality, more wafers per ingot, more usable cells per wafer, new cell manufacturing that reduces the number of steps and produces better cells, bigger panels so there is less dead area around the edges, better glass that is stronger so can be thinner and less reflective so the panels perform better,optimized wiring inside the panels, better cheaper wiring connectors...

A zillion tiny improvements, at every stage in the manufacturing process.

The next quantum leap is coming soon: two-layer cells with an efficiency jump from the current 22% to over 30% sunlight-electricity.

multijunction pv cells with better than 30% efficiency have been around since the 01970s; the best research cells are over 40%

those are used almost entirely in space applications because they cost so much more per watt

is there some reason to expect that to change

Yes, high-efficiency cells have been around for a long time for cost-no-object applications, mainly space.

What's changing is commercialization. Two or three of the big Chinese manufacturers have pilot projects going for two-layer cells. Of the order of 10 MWe, that sort of size. (I can't remember which companies, sorry; it was a few weeks ago I read about this. Probably at least one of Jinko, JA Solar, or LONGi is in there, as well as one or two of the second tier.) Also in the West there are a few startups working on two-layer cells, either perovskite on silicon or perovskite on perovskite.

(Perovskites are more easily "tunable" in terms of which frequencies of light they absorb, apparently--that's one of their attractions.)

The big change in recent years is the explosion of perovskite-family cell materials. There's a huge variety of perovskite materials offering tunable band gaps and they can be processed at low temperature, so they can be used as top cells over silicon. That promises a big cost reduction from conventional multijunction cells built with III-V compound semiconductors on germanium [1].

There are a few problems so far:

- The perovskite materials containing organic moieties tend to be sensitive to degradation by moisture and/or oxygen. They need to demonstrate 20+ years of service life to match silicon.

- The purely inorganic materials like cesium lead iodide are more stable but have yet to attain high cell efficiency.

- There is no proven high-volume way to deposit the thin films of perovskite materials, which have different handling characteristics than anything previously used in solar manufacturing.

I would say there's a good chance of silicon/perovskite tandem cells taking off this decade but it's not yet a sure thing.

[1] https://eom.umicore.com/en/germanium-solutions/markets/multi...

thanks for the explanation!

i'd say they need to demonstrate 50+ years of service life to match silicon, but at 10+ years they'd be marketable at conventional utility discount rates, and beyond 20 years the npv doesn't change noticeably

it'll sure be interesting to see what happens here, but it's going to be really tough to match silicon's cost per watt, much less beat it, unless you can dispense with the glass or something

Because of performance degredation panels are rated over 20 or 25 years and generally need to be replaced after that.
while they are indeed rated for 20–25 years, as you know, in silicon panels most of the degradation happens in the first couple of years; additional degradation in the following 30–25 years is measurable but fairly minimal

as i understand it, panels are rated for 20 years not because they need to be replaced then but because 30 years ago nobody knew what would happen over that time, and also manufacturers didn't want to set themselves up for unlimited liability

usually it's more advantageous to add more panels than to replace the existing ones at that point, though rooftop installations are often an exception due to the extreme space limitations

i think if you sold a perovskite hybrid panel that cost half as much per watt as existing silicon panels, but degraded down to 70% of its rated capacity at 10 years and rapidly down to 50% after that, i think it would still sell in a lot of markets

They could do water cooling underneath. Probably not worth it though.
With the ground covered, evaporation is greatly reduced, so the soil stays moist. Water is a good heat conductor.

Having the underlayer for your roof solar tiles be full of water is considered undesirable by most permit-issuing authorities, so roof tiles must use air cooling.

From the land topography fitting perspective, it's too bad panels are triangular. Then one could essentially create a triangular mesh overlay of panels over all kinds of uneven terrain.
Really I don't care where or how the panels are placed.

All movement towards renewables is good, no matter how "inefficient" it is vs fossils

Yeah I don't understand people getting hung up on efficiency with solar panels. The sun is always up there producing the exact same amount of energy with zero input from us. We're not doing any work to get the input energy for the panels.

Efficiency would only matter if we'd already covered all the available area with panels and needed to start replacing existing ones, otherwise $/watt hour is the only metric that's important.

until only a few years ago solar panels were enormously more expensive per square meter and per peak watt, so doing things to increase their efficiency was a great way to reduce $/watt hour
Not sure where they are building these but I’ll tell you hwhat, fire ants love electronics and they are found in the majority of this state. Putting these right on the ground is just inviting destruction by critters.
thats really interesting, thanks for sharing! a google search for fire ants electrical equipment brings up a website on ants conducting electricity and shorting circuits. I had no idea that was an issue
So we are fighting fire with electricity?
Pikachu > Charmander
Bizarrely, ants were attracted to our Tesla power connector on our previous car. We googled it and it was a common problem, ants seem to love electricity.
Solar inverters are always water proof, sealed, and potted. The ants will have their work cut out for them.
an install is expected to last 10-25+ years... that's a long time....
Can't tell from their website but I really hope they're not using per-panel "solar inverters" at all. The whole ground-touching array should be DC with no electronics whatsoever (except maybe for diodes, and with no significant shadows they might not even need those).

Any inverters present should be large, few, and in their own weatherproof housings above the ground.

I think per-panel inverters are a stupid idea, especially for large utility-scale installations. Sorry if I misinterpreted your comment and you meant something else.

even worse: Rasberry crazy ants
i swear those buggers sense where the volts are
"Erthos claims it can build a solar power plant in half the time on one-third of the land" sounds like red flag to me. They have not tripled the solar flux or the efficiency of their solar panels.
On most solar farms, the panels only cover about a sixth of the land area, or less. The rest is taken up with accessways between the rows of panels, and roads around the outside. Check out some photos, for example https://energy-cc.co.uk/wp-content/uploads/2020/09/Benbole-S...

The reason the rows are widely spaced is to minimize shading by a row of the one behind, when the sun is low in the sky.

With the panels being flat on the ground there is no shading problem. Laying them side by side eliminates the gaps and they plug together so there is no wiring to do. Omitting the racking eliminates its cost, the time to install it, and the wiring between rows.

Etc.

A SIXTH? I don't see how that can be necessary. They might use that much land because it's worthless desert or something like that.

Where I am there is not exactly a solar farm, but there is a maybe half megawatt array on a flat canopy over a parking lot. This is in an area of expensive real estate, and there is almost no wasted space.

As for spacing to avoid shading when the sun is low, obviously that is an optimization problem and in all likelihood they have worked it out. Don't forget that when the sun is low, the flat array produces almost no power.

Sounds great, but the land underneath is absolutely killed as if it were paved over, no grass, ground cover, crops or anything. That said, it seems to be less than half the amount of land.

Nevertheless, it seems that the innovative German installation method of using vertical panels in between agricultural rows is better [0].

[0] https://cleantechnica.com/2022/07/25/new-research-says-verti...

No, not as if it were paved.

No aggregate is laid down, nor is asphalt put on top of the aggregate. No toxins are put in, nor does the topsoil need to be scraped off.

End-of-life reinstatement of the land shouldn't take more than ten years of seeding a sequence of plants that specialize in colonizing and re-aerating bare compacted soil, and revitalizing the soil ecology. (Weeds.)

Yes, EOL recovery could be quicker than if it were paved (assuming that paving strips the topsoil completely and not just paves over it) but much slower than if it remained growing all the time.

The problem is that soil is not just inert, and completely covering it like that kills the entire microbiome — the fungi, bacteria, and myriad of multicellular micro-critters that make the soil good for plant growth will be long gone when they take up the panels.

Who cares? Honestly, a hundred acres of scrubland turned into green energy is a steal compared to everything else we do on this planet for food, energy, and material
Sure, on a scale of comparative bad things, e.g., vs strip mining...

But if you don't have to kill the entire soil microbiome, why do it? Just questioning whether this is as good as vertical panels + farming.

I agree, if it really is nearly devoid of vegetation, it seems better to use the least possible amount land and also not use the metals for the racking.

Birds and little critters to be relocated to affordable housing in San Marcos.

But seriously why can't we just start building more net-zero natural gas plants that take up 1/100 of the land area? This is insane.

How can natural gas be net zero?