The fact that there exist in our world flat rocks that make lightning when you point them at the sun is one of the most unappreciated bits of wizardry in this modern age. As hackers, we love all this of techno-wizardry–but some of us abhor paying full price for it. Like cars, one way to get a great discount is to buy used. [Backyard Solar Project] helped a friend analyze some 14-year-old panels to see just how they’d held up over the years, and it was actually better than we might have expected.
The big polycrystalline panels were rated at 235 W when new, and they got 6 of them for the low, low price of “get this junk off my property”. Big panels are a bit of a pain to move, but that’s still a great deal. Especially considering that after cleaning they averaged 180 W, a capacity factor of 77%. Before cleaning 14 years worth of accumulated grime cost about eight watts, on average, an argument for cleaning your panels. Under the same lighting conditions, the modern panel (rated to 200 W) was giving 82% of rated output.
That implies that after 14 years, the panels are still at about 94% of their original factory output, assuming the factory wasn’t being overoptimistic about the numbers to begin with. Still, assuming you can trust the marketing, a half a percent power drop per year isn’t too bad. It’s also believable, since the US National Renewably Energy Laboratory (yes, they have one) has done tests that put that better than the average of 0.75 %/yr. Of course the average American solar panel lives in a hotter climate than [Backyard Solar Project], which helps explain the slower degradation.
Now, we’re not your Dad or your accountant, so we’re not going to tell you if used solar panels are worth the effort. On the one hand, they still work, but on the other hand, the density is quite a bit lower. Just look at that sleek, modern 200 W panel next to the old 235 W unit. If you’re area-limited, you might want to spring for new, or at least the more energy-dense monocrystalline panels that have become standard the last 5 years or so, which aren’t likely to be given away just yet. On the gripping hand, free is free, and most of us are much more constrained by budget than by area. If nothing else, you might have a fence to stick old panels against; the vertical orientation is surprisingly effective at higher latitudes.
Reminds of an article showing the popularity of used panels in other countries. Kind of like what happened with used clothes.
And used tires.
and second-hand tobacco (https://en.wikipedia.org/wiki/Here_Come_the_Warm_Jets side b track 2)
The best use of free used panels is to run last-year’s crypto miners. But I don’t think you can base an industry on free stuff unless you are selling bottled water.
Remember, the lifetime given by manufacturers and installers is not a drop-dead date. It is the age at which x% of panels will lose >y% of their capacity.
I think the take-away is that you should expect to keep using your solar panels well beyond their “lifetime”. The marginal improvement from replacing an old panel is probably never worth the expense.
If building an array from older used panels, I might lean towards microinverters instead of string inverters, so that the total output of the array is not limited by the most-degraded panel.
The flip side is, your roof will have a finite lifetime so if they’re installed on a building, that’s probably when they’ll be replaced (and i’m guessing is why there’s already a good supply of used panels being thrown away).
There are teams trained to pull off all the panels from the old roof, then reinstall on the new roof. (Source: I did that)
Like hard drive’s MTBF. I have a number of hard drives that are well past 100,000 hours in my ancient Macintosh that are still working (some of them were originally estimated for 25k or so). And I’ve had new hard drives that failed before reaching just 100 hours (less than a week continuous) An average is just that, not a hard life span.
Except that hard drive MTBF numbers are worthless. Most of them are rated for 1,000,000+ hours, which is over 100 years. Completely made up bullcrap. I don’t think the manufacturers even use the same methodologies to come up with their advertised ratings.
BTW, I don’t recall ever seeing hard drives rated for 25k hours, so I think you might have overlooked a few zeros. It also looks like Seagate and WD are only posting MTBFs for their enterprise grade drives now.
I’ve often wondered about that – does MTBF assume constantly accessing the drive, as in an online database situation, or does it factor in idle time and even spindown. I’ve never had a spinning disk failure since the IBM Deathstars, but I attribute that to the fact that 90% of the time they are not even spinning.
Spindown or shutting the drive off doesn’t typically count towards the MTBF. If the power is off, or mostly off except for some basic IO, that counts towards shelf-life.
MTBF is not shelf-life.
When I worked on hard disk reliability years ago, we found different failure modes from different sources. We developed start stop tests and continuous run tests. Read/write was what had to happen to say it was still working. The specific head design, material, air bearing and disk design , materials, hardnesses and thicknesses and more were all key variables to be controlled. Had to pass all tests to be a good system.
Really old sub-40MB hard drives were rated about that low.
Because the steppers aged and the track alignment drifted.
A low level format fixed it.
It’s literally in the name: mean time between failure. If you take a bunch of hard drives and run them to failure, the MTBF is the average of that. If other units in the test still survive, the figure becomes the average minimum expected lifetime.
Another misunderstanding is that the MTBF always represents the total lifetime of a device. Calculating it actually assumes we’re operating in the middle of the “bathtub” curve, i.e. it accounts for random failures like a weak cell in a solar panel developing into an open/short circuit – not early failures or the inevitable end-of-life failures like a solar panel fading below X percentage points. If someone is using MTBF to count total operating life, they’re using it wrong.
What it means in practice, if you have a bunch of solar panels, you may expect to see some of them fail at N hours. Some will fail early, all will fail eventually, and the MTBF is used to estimate the need for maintenance/replacement in between.
Mean and average are two different things, even if they may happen to be the same number.
And knowing this is fundamental to understanding – and then being able to draw the correct conclusions – in everyday life, from science to politics!
“Mean” and “average” are indeed the same thing. Add up all the values of all the samples, divide by number of samples.
You’re thinking “median”, which is the central value, the 50% percentile. Line them all up shortest to tallest, pick the one in the middle.
If you have a skewed distribution, the mean/average can be quite different from the middle/median. The middle person doesn’t earn the average wage, for instance, because a few with very high earnings push the average higher.
Please get your facts right Mr facts matter.
Mean and average are the same. Median is different.
Out of the set 1,5,10 the mean/average is 5.333 the median is 5.
Median and s the middle number in the set, or in an even numbered set the average of the two middle numbers.
Again, don’t use that name and be so confidently wrong. Especially on a hacker forum.
I’d have to push back against those calling Mean and Average synonyms, as they are not. The Mean is the most commonly used form of averaging, and often treated like a synonym but there are many ways to average your samples – Mode, Median, Mean (possible others I’m not thinking of too) and they are all an average you can justifiably call in natural language simply the average of the data.
So while I agree FactsMatter isn’t really correct in this context either, they are correct Mean and Average are not the same – as Mean is a more specific definition for a type of Average – “Mean” is a member of the group “Average”, but it isn’t the only member, so you will at times describe something as Average when it is actually the Modal Average entirely correctly.
You can push back all you want. Youre still incorrect,
The mean is the average (sum of all numbers divided by the count), It is NOT a member of the group “Average”.
the median is the middle number in a sorted dataset,
and Modal average,” more commonly called the mode, is the value that appears most frequently in a data set.
So in the dataset 1,1,2,1,5,7,8
The mean is 3.571428571428571
The median is 2
and the modal average, or mode is 1.
The only one of these that is an AVERAGE of the dataset is the MEAN. There is NO other way to compute the average of the dataset. There are other things you can do with the dataset, but those operations are not averaging.
Number theory wise here at least we describe all ways of averaging a dataset as TYPE AVERAGE or just AVERAGE if the method is obvious or already defined earlier. The job of condensing the dataset into a simple point that represents it for comparison to other datasets is done by all the methods, they by definition are averages.
So here at least strictly speaking a Mean isn’t exactly the same as Average at all. Mean is by far the most common and usually most useful, so it is the average you expect when it hasn’t been typed clearly so folks do often call it average, especially at the youngest age ranges when the concepts of other forms of average are not being introduced yet – but strictly it isn’t.
By far the most common language definition for “average” is the arithmetic mean. The rest is like saying “I’m gay” means that I’m elated and generally happy with my life.
MTBF is a totally reasonable metric for a hard drive, and a super-weird one for things like solar panels or car tires, which wear out at a roughly constant rate.
In an ideal world, you’d get both MTBF — how long it will last before catastrophic failure — and some kind of time-until-80%-output (or whatever%). The former is probably related to build quality, and the latter is material science.
Of course, we do this with all kinds of stats, so I’m old-statistician-yelling-at-cloud here. But MTBF was originally conceived for things that fail, not things that wear out. Mixing the two concepts together gives you a stat that doesn’t accurately describe either.
A failure in terms of MTBF on a solar panel would be like a corroded bond wire in one of the photocells, left weak due to random manufacturing errors. It would work for some time, then fail open and the solar panel or some areas of the panel would drop dead.
The MTBF should count the number of hours until these kind of defects typically develop into failures in the affected units – after you remove the early failures that were clearly defective.
Bypass diodes are cheaper, simpler, and equally or more reliable. And they avoid the conversion losses from DC to AC, then back to DC again for your batteries
I’ve been getting free 2nd-hand panels for a while. My installer does insurance jobs where damage to one or two panels means the insurance company says “replace the lot”, so my roof has three strings of panels of varying ages, none of the free panels were more than 3 years old when I got them. I just pay for labour.
The newer panels feed one charge controller and the older panels feed another controller.
My sister’s ex has been in solar for years and sees this happening more and more.
My friend uses a good trick to get use out of a single damaged panel he got for free.. Instead of using it to cut his electric bill, he uses it just to charge a 12V lead acid car battery, which he uses just to run his modem, router, wifi, and maybe a little NAS box. Where he lives, power outages are somewhat more common than telco outages so it gives him more reliable internet. And one 200W-ish panel is plenty for keeping a battery topped off even with a number of small computers on it.
[Kris De Decker] runs the webserver for LowTechMagazine.com that way.
When you stop and think about it, you can probably come up with quite a few DC-only loads you could run on old solar to avoid inverter losses and grid-tie paperwork.
Meanwhile nuclear output level is the same for decades.
Kudos for inventing an industry which degrades over time as energy demand increases thus you need to keep making more panels to keep up.
It’s like you created supply and demand for shareholders.
Good thing the world is being forced onto your business model, by….. shareholders.
Nuclear power plants maintain output for a few decades, then it drops to zero when the plant is retired. Building a new one is very expensive.
For solar, replacing the panels without rebuilding the rest of the infrastructure would cost a lot less than a whole new array.
first you spent a decade or two building one, while borrowing a few billions a year. The you sell electricity for few decades, the you spent and decade and some more billions decommissioning it
The first five years are spent in jumping red tape and lobbying, acquiring permits, doing environmental impacts studies, convincing authorities and investors to let you build. Then five years into the building half of your industrial partners bail out and delay the project for another five years.
Then as the construction nears completion, the next five years are spent getting permits to operate the plant, waiting for the officials to perform certifications and inspections with work on hold until each part gets tested with demands for changes and modifications. Then you wait for the permission to start loading fuel, then a couple more years to run tests before getting the permits to turn on full power. Then it will probably break down somewhere, gets fixed, and then you can start running for real.
For example, the EPR in Olkiluoto waited from 2018 – 2021 for the permission to fuel it up for testing at 5% power. The fuel was there on the site and the reactor was already “hot tested” (i.e. run through with heated water) and just waiting for the radiation safety officials to decide it’s safe.
It’s all incredibly bureaucratic because the officials are doing exactly one thing at a time and you can’t proceed until you get the go-ahead to start doing the next thing. If you pour a concrete slab, you wait six months for the officials to react, inspect, write their report that says it’s not good enough and needs to be fixed, then you tear it up, re-do, and wait another six months for the officials to react.
I’m all for nuclear power but, in fairness, nuclear power plants have a lifetime, as well. Many of the plants in France are reaching end-of-life and apparently they don’t have plans to bring up enough replacement power.
A nuclear power plant is not unlike a car: as long as you keep fixing it, it can have an infinite lifespan. It just depends on whether it was maintained and modernized along the way – if it was neglected for 50 years then the problem becomes finding anyone who knows how it even works. At that point the cheaper option is to tear it down and replace it.
What has happened is, the entire nuclear industry was left to languish and money allocated elsewhere, so now there’s not enough expertise or industrial capacity to keep them going and the plants themselves have been neglected for too long for it to make sense.
Except that you can’t fix everything. When the concrete cracks, you loose. And there’s only 2 types of concrete, the cracked concrete and the concrete that’ll crack.
Exactly like rust with cars – you weld in a new panel and keep going.
The slab under the reactor is a bit more difficult to fix, but if you’re replacing the reactor vessel you might as well pour a new one.
Also, there’s no such thing as un-cracked concrete. Every concrete pour has cracks in it once it cures and settles for a couple years. It’s just a matter of how many, how deep, and do they even matter?
Point being, concrete does not work under tension, so it is always designed to operate under compression, which closes the gaps, which means cracks do not actually signify a failure of the structure as such. It’s the same as laying stone upon stone – the fact that they aren’t one contiguous material doesn’t mean a stone wall can’t stand up.
Solar has it’s place. While I am a huge advocate for nuclear, a couple hundred watts of solar for my off grid sites is awesome.
Has good environmental effects too.
https://sustainablepowernews.com/desert-solar-panels-thriving-ecosystems/
Still have to replace spent uranium rods. If too many spent rods aren’t replaced, the output can eventually drop due to decreased heat output.
Less than 5% of the energy is used before uranium is replaced. The fission products include some “neutron poisons”, like Xenon, which block too many neutrons before a small fraction of the fuel is used. The fuel can be reprocessed but at this time it is cheaper to store the used rods and get new ones.
France has a big storage and reprocessing facility where spent fuel can be stored until the economics favor processing. They handle fuel from the large number of power plants in France and for several other countries.
How much does nuclear fuel cost these days? Oh yeah, has the US figured out where to store the nuclear waste that was generated half a century ago?
Nuclear waste is a solved problem. It’s a myth that it’s a difficult thing to deal with. Deep geological storage is the answer and is perfectly safe.
LMAO
“Deep geological storage is the answer and is perfectly safe.”
Yeah just bury it and forget about it for a few thousand years. We havent actually been storing it for more than a few decades but SURE perfectly safe. No problems will ever arise. Totally proven okey dokey!
There are Gen4 reactor that’ll happily runs with your “waste”. A nuclear fuel for negative price (since other countries pays you for disposing their waste) is a good motivation to me for storing them. But sure, it’ll likely contaminate the storage site in 2000 years. Hopefully capitalist will have eaten them by the time.
Demanding total proof or total safety is a nirvana fallacy that can be used to shut down everything.
The question is simply, when it fails, what are the worst outcomes? For deep geological storage, aside for a literal volcano erupting on the spot, not a whole lot can happen. If trace amounts of radionuclides end up on the ground level a few thousand years from now, big whopee, we have them already all around us.
Is this some new thing where “Big Solar” is now replacing “Big Oil” in bloody stupid internet comments?
No.
Besides heat (and nuclear materials) in our planet’s crust, all energy is coming from the sun, converted by / extracted from plants, solar panels, wind or water circulation.
And time, a lot of time, in the case of fossil fuels. Time we might not always have.
Why you gotta be like that? Trolling make you tough guy? Nukes are for suckers. Crazy expensive to build, years to build, thousand of miles of wire to get it to your house, then you pay whatever the utility company feels like charging you. My solar paid for itself in four years, and I’ve enjoyed free power for six years now, with a 25 year warranty on German panels. Nukes make billionaires richer. Solar roof makes me richer.
So you go right ahead and build a nuke on your roof. My panels are mine to catch the money that falls on my roof every day. You do you
You could take advantage of these bigger panels and create a roof for a parking spot.
Nuclear power is good if available, but nothing can beat the solar power generation for a off-grid ranch. if you could make the solar power work. I would suggest to keep using the old ones as long as it would generate
Assuming you have enough wood and fuel oil, or propane, for keeping the place warm and cooking and making hot water all year round.
You can get used panels off Ebay for cheap. Somewhere around $20-$30/250w panel including freight shipping.
+1 million for the “on the gripping hand” reference
Too many nerds read this for it to be an “obscure” reference. Some of the best science fiction of my lifetime!
Glad somebody caught it. “The Mote in God’s Eye” is probably my all-time favourite sci-fi novel.
Well, here in Italy the bulk acquisition of used panels appears at first glance a great deal but after calculating the watts per euro there was no advantage. So I bought new more powerful panels that occupy less space and are sure to give a good output for many years. Unless the cost of pallets of used panels comes down significantly I don’t see the justification of investing time and installation cost and eventual disposal costs.
Well, I am in Italy and this week I am putting up my “new” old panel string. At 15€ x 300W they are way below the best marked price I found for new panels (61€ x 400W). I also managed to get “brand name” ones, albeit used, rather than a random brand name new panel. Over provisioning of the string takes care of possible breakages for the future, and honesty, being my 4th setup of used panels, I couldn’t be happier.
Do you have a link for these used and new. I’m in Italy too
So, for new panels, I generally rely on Leroy, where as of today, you have cheap options like 61€ x 400w, or branded ones for 100€ x 600w. If not available near you, remember that you can order from the website with delivery to you or to a close shop. For used ones, Facebook Marketplace and eBay are your friends, but sites like galvah.com are great for finding deals on used panels every now and then.
When you can now buy PV panels from Amazon that can output 100 w when there’s 1.2 KW a square metre of solar irradiance for 33 quid, I fail to see the point of bothering with second hand ones.
Free is 33 quid cheaper though – if someone has a large plot or a large roof and they can get a load of panels for free they could be making a decent amount of electricity for nearly free, saving them 33 quid x N panels which could add up to quite a lot.
On no the new one is slightly better than the old one that was available for free…
On one side you have a few grand of extra cost as you have to buy all the panels not just the mounting hardware and infrastructure on the other it cost you that extra100 quid to buy a few more mounting rails than you’d have needed so your system can have the same ballpark of performance… Or cost you nothing extra as you just filled your space using the same mounting hardware either way, so simply are getting slightly less energy out of it – which with the usual rates domestic producers tend to get paid for exporting energy probably doesn’t matter fiscally (though your location may vary)
The panels on my roof come with a 25 year guarantee, and seems to be standard practice. Did you know most of the wind turbines that are decommissioned are refurbished and sold on the second hand market?
Does anyone know some good cheap ways to ground mount panels or make shade structures like gazebos or even greenhouses?
If I could get some old panels I’d love to give my grass some help with the sun. Or make some nice outdoor area to enjoy or grow plants.
On Amazon, you can find some stands for ground mount relatively cheap, I usually add a small plastic bag to the leg, fill it up with cheap cement, and get a very stable setup. The structure usually bolt together and can be extended making install super easy. For solar pergolas, I had success buying a used flat gazebo with a damaged cover, since people don’t bother to replace it for a new one, because it could be expensive or strange measurements, and bolting panels directly to it. Facebook Marketplace is a godsend sometimes…
My panels were built in the early 1990s and are still making 83% of their rated output. I have 12 100w Siemens panels.