Amazon Took All US Solar Rooftops Offline Last Year After Flurry of Fires, Electrical Explosions

Become a fan of Slashdot on Facebook

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way. Snubber For Boiler

Amazon Took All US Solar Rooftops Offline Last Year After Flurry of Fires, Electrical Explosions - Slashdot

Industrial size solar is rather poorly understood, especially by the so-called professionals that think theyâ(TM)re dealing with 12/24VDC not understanding the currents involved with the current going through even just a handful of panels. Ideally youâ(TM)d convert every panel immediately to 240VAC. There are some companies starting the micro inverter systems but at $250/inverter (which handles 2 panels) theyâ(TM)re not very widespread and make the system about 50% more expensive.

The problem with solar generation is that the power input to the system is difficult, if not impossible, to remove. Other generating systems will scram the reactor, dump the hydroelectric turbine water, feather the wind turbine. But short of going up on the roof and throwing a giant tarp over the arrays, the power keeps flowing. And feeding into a fault.

If the -voltage- is low enough, all you have to do is trip a few relays between panels and the flow stops. Not terribly hard... if you installed the relays in the first place.

The the voltage is too high, well, that's where things get interesting. With high voltages, starting around 480 volts, the flow doesn't want to stop and can break the air between separating contacts down into plasma over which the current will arc like lightning.

But that's not an arc flash explosion. An arc flash explosion is a differ

The video you linked just shows a slow motion view of an Arc Flash. I think the video from Fluke tools is a little more informative and corroborates most of your information. https://www.youtube.com/watch?... [youtube.com] I linked to the video at about 35 seconds to bypass most of the OSHA disclaimers.

Badk in my telephone exchange days, I saw a shifting spanner vapourise when dropped onto 50V busbars, fed by of 5000A batteries. Rather spectacular in fact.

I'll bet anything you want you didn't see an arc flash. 48 to 56 volts in telco systems isn't enough to create or sustain a plasma arc after the metal conductive path breaks. I'm sure it melted the spanner and to some extent the busbars, splattering them everywhere but that's not an arc flash.

all you have to do is trip a few relays between panels and the flow stops

From the linked article in TFS, it sounds like the fires were in the panels on the roof. That might be ahead of any place one could put protective relays or circuit breakers.

A solar panel has to be designed to be inherently safe. That is, it won't cause damage to other equipment or structures even if the input power remains connected. The sun, in this case. For solar panels, that would probably involve incorporating large enough heat sinking surfaces in them so they could continuously dissipate short circu

Individual panels don't generate that much power or that much heat. To get useful amounts of power you have to connect them together.

If you connect them in parallel then they generate lots of amps at low voltages. Lots of amps need very large and very solidly connected wires or else they heat up and burst into flame. But that's not what TFS described, TFS described an arc flash.

If you connect them in series then you don't have to use expensive, thick wire. The amperage stays low but the voltage becomes high

At any rate, at low voltages all you need is something that breaks the circuit if a short or over-current is detected so that it never gets hot enough to burst into flame. Not heat sinks on a solar panel.

High voltages are more difficult because, as noted, the circuit doesn't want to break and may resist breaking with an extreme increase in temperature.

Ideally you'd convert every panel immediately to 240VAC. There are some companies starting the micro inverter systems but at $250/inverter (which handles 2 panels) they')re not very widespread and make the system about 50% more expensive.

Today micro inverters are everywhere thanks to rapid shutdown requirements in NEC 690 and solar installers looking to make more money. Even if you want a simple string DC system you are forced to spend money on MLPE doubling the number of interconnections anyway. Few are opting for strings anymore especially not in the US "in or no buildings".

Personally not a fan of the idea of large numbers of high current electronics for every (1-4x installed panels) sitting outside in the elements unattended for decade

Some companies starting? Ive had micro inverters for 5 years on my panels.

Lol nice trolling, but sorry the article doesn't actually say that. Unless "third party vendor" means "union".

Are they not able to hook into city power when the solar goes down?

Surely they're not 100% dependent upon solar...?

The places were only shut down*during* the fires, and as a precaution they shut down all of the solar systems until they were inspected. They continued to operate on mains power. From TFA, “ Out of an abundance of caution, following a small number of isolated incidents with onsite solar systems owned and operated by third parties, Amazon proactively powered off our onsite solar installations in North America, and took immediate steps to re-inspect each installation by a leading solar technical expert fi

Why its right there inTFA. Pity you didnt bother to read it. Late last year, while CEA was still conducting its inspections, it informed Amazon of one critical and 259 major findings across Amazon’s rooftop solar portfolio. Problems included mismatched module-to-module connectors, improper installation of connectors, poor wire management and evidence of water intrusion in the inverters, internal documents said. Username checks our.

Late last year, while CEA was still conducting its inspections, it informed Amazon of one critical and 259 major findings across Amazon’s rooftop solar portfolio. Problems included mismatched module-to-module connectors, improper installation of connectors, poor wire management and evidence of water intrusion in the inverters, internal documents said.

And how does that remotely address my point that Amazon's problems might have been caused by Amazon themselves? All your statements says is that Amazon's third party auditor found problems with their solar panel systems. Duh. Of course Amazon wants to blame it on their vendors.

Does the article address if Amazon hired unqualified or inexperienced contractors to install their systems because they were the cheapest bid? No. Does the article address whether or not Amazon selected substandard parts like inverter

Cheap labour or not, the workers clearly did something wrong to cause the issues. If they used unqualified people to do the install, thats a failure of your legal system safety protections for electrical work. In Australia, we don't allow unqualified people to install potentially dangerous things. Any breaches of the standards are taken very seriously.

Cheap labour or not, the workers clearly did something wrong to cause the issues.

Did you even remotely address my point that Amazon KNEW exactly what kind of labor they were getting? No. Did you address the point if the contractor installed the system the exact way Amazon instructed? No.

If they used unqualified people to do the install, thats a failure of your legal system safety protections for electrical work.

Please cite which electrical safety protection was violated? I did not say that any of that. I said "unqualified" meaning the person doing the work had not done the work previously or not certified to do the work.

In Australia, we don't allow unqualified people to install potentially dangerous things. Any breaches of the standards are taken very seriously.

Can you cite in the article where Amazon found these problems right after install and the co

Cheap labour or not, the workers clearly did something wrong to cause the issues.

No, some of the problem could be due to cheap foreign imports that are of low quality. Bad cable management, water intrusion, etc.

Bad cable management is something that could be on the contractors

The design of a device can help or hinder various installation issues. I'm just saying its premature to blame the installers.

NOT correct: "Amazon Took All US Solar Rooftops Offline Last Year..." Should be: "Amazon Took All of the Company's US Solar Rooftops Offline Last Year..."

Maybe the editor didn't select an Amazon Day for word delivery so they weren't all shipped together and some have been delayed ...

NOT correct: "Amazon Took All US Solar Rooftops Offline Last Year..."

You underestimate the power of Alexa

Well, there was "the incident". https://www.facebook.com/CaptM... [facebook.com]

Krikey! I had not seen that one before, exactly on target!

To be fair though I would actually like the mini Bellagio hot tub.

The installations are experiencing Arc Flash due to wiring faults. Could be the crimps, could be damaged wires, could be corrosion. Most likely the panels are fine, the wiring was probably done by the lowest bidder contractor with little to no experience in DC solar panels.

https://www.youtube.com/watch?... [youtube.com]

Hmm, those sort of events sound like back-fed mains from non-isolated DC-AC inverters. If they failed to use transformer isolated inverters then they surely got the expected outcome.

I'm not sure I understand, but I want to. How would a transformer help?

And, I'm pretty sure all DC-AC inverters have a transformer in them, pretty much by code.

Wish I could edit / remove my other reply. I did some research and I'm unhappy to find that some inverters are NOT isolated. I'm still pretty sure there's a transformer in there, but maybe some cheap out and just use a simple inductor in a "flyback" boost circuit? No way would I allow that if I was in charge of anything (maybe I should be...)

But anyway, if you see my other posts you'll see that isolation isn't the problem- the cheapo inverters are flaming up.

I'm hoping for more detailed information and exact cause. Pure speculation but my (strong) hunch is on burrs / sharp edges- in conduit, conduit connections, conduit-box connectors, lack of threaded bushings, etc. I've done some of this work, and other electrical in conduit (up to 480 3-phase) and I remove all burrs (deburr) and countersink all conduit. But I don't know anyone else who bothers to.

TFA mentions water intrusion into inverters, some connector mismatches, and I think that was all they gave for

The installations are experiencing Arc Flash due to wiring faults

I remember wiring a shed long time ago learning just this lesson. You get to a particular level of voltage/amperage and even just stapling the wires down wrong can be bad news for you. I'm no expert, but I've done enough to run a new line from breaker box to outlet boxes or fixtures and "learned a thing or two". I'm guessing that all the various ways wires can arc on you is the reason NEC 2020 mandated all branch circuits to be on combination GFCI/AFCI now?

Okay, I didn't want to write anything until I got the inside scoop. I happen to know a guy... He said they (Amazon compelled the electrical contractor) used the cheapest (Chinese) inverters and they're the problem. They're overheating and don't seem to have thermal shutdown built-in.

And they're probably mounted under the PV arrays, and cause a catastrophic domino effect when they flame up.

Amazon and engineering company now require inverters to be installed on concrete pads, with some separation, so if they do flame up, they won't cause a catastrophic "domino" effect.

Walmart is having the same problems, but I didn't see it mentioned in TFA.

Are you suggesting Amazon's contractors bought their inverters on Alibaba?

This makes more sense than the idea that some warehouse worker was sneaking off to take a pee.

For several (obvious?) reasons I'm hesitant to divulge too much, but I've known and sometimes work for a guy who is an engineering project manager (and very hands-on when needed) for PV systems (and many other things). He started his own PV install company maybe 13 years ago, and we did a couple dozen residential and a couple of commercial systems.

With Amazon he started out maybe 7 years ago, through a contract employment agency working for the main electrical contractor. He's quite dynamic and motivated

Re. inspections: I'm an EE but (and) do much hands-on work, including wiring, conduit, etc. (also IT, sw, analog electronics, etc.) My first PV system around 2010 boss hired an inspector (required of course). Being an EE I know how to wire things. Inspector failed me on many totally stupid nit-pick things, because PV was new and people don't understand, just parrot. I had used some industrial terminals in breaker panel for grounding, and he didn't recognize them and wanted a more typical residential looking ground terminal block.

Fast forward- boss being probably the smartest, certainly shrewdest efficient manager I've ever met, hires a different inspector for the rest of the jobs we did (in that geographic area anyway). This guy was HUGE, so big he couldn't get in and out of his car easily. He would drive up, write up a bunch of stuff, and hand us the official inspection sticker to apply to the panels.

Point is, and sorry for the extreme cynicism, but many people are lazy, some extremely lazy. Even the best of them are not going to run a borescope through a conduit and see if every joint was deburred. If you've ever pulled any wire in conduit, you'll find out quickly how the burrs and edges will gouge the insulation. I can't tell you how many times I've pulled wire out of a conduit- usually removing something that's no longer needed- and find sharp edges everywhere and wire all gouged up. PVC is less of a problem that way, and even if copper is exposed, obviously you won't have a short-circuit like you will with metal conduit.

In fact, because of this they've (NEC) gotten more particular about grounding metal conduit. Box connectors must have a little grounding paint-piercing screw, and/or extra terminal for actual grounding wire. IE, just the steel metal contact isn't good enough- paint may be enough to keep conduit non-grounded (bonded is the word they like), and sometimes rust forms such that conduit is now insulated. But PVC isn't always allowed, there are many factors.

They were pretty much all being built in a short time, and kind of followed a standard design / template.

The problem with Amazon is everywhere: a bit too much capitalism. To be fair, and I'm just playing devil's advocate because I don't agree with this line of thinking, but Amazon just said: "hey, we can get the same functionality for less $, so we'll buy the cheaper ones." If they were the same specs and obviously failed, either they're not up to spec, or the design requirement was too lax. My bet is on

He said they (Amazon compelled the electrical contractor) used the cheapest (Chinese) inverters and they're the problem.

So Amazon finally realised that people shouldn't buy things from Amazon?

BTW, thank you for that video link.

A little embarrassing, but a huge learning experience: on my first PV install, we had 3 "panels" of 8 PV "modules" each, so about 300 VDC coming into the inverter. Open-circuit voltage test is done, no problem. Short-circuit current measurement is also a legitimate test.

Well, I've worked with many things, including lots of repairing CRT TVs and monitors, high voltage, etc. So I know how to be safe. So I do the short-circuit current test, and it's good.

But, as I pull the meter probe away, an ever-increasing arc happens- just like the video, but even bigger because 300 volts. I pulled probe away fast, but it sure scared the you-know-what out of me.

Bottom line- you'd need some pretty special switches to safely break that flow and arc.

"Bottom line- you'd need some pretty special switches to safely break that flow and arc."

Your typical $0.50 industrial toggle switch handles that kind of voltage all day. Nothing special about it.

"Bottom line- you'd need some pretty special switches to safely break that flow and arc."

Your typical $0.50 industrial toggle switch handles that kind of voltage all day. Nothing special about it.

It might be rated for 400VAC - the key word being "AC".

Load breaks for AC and DC are wildly different - because AC you can rely on the fact that voltage will drop to zero at some point in time and if you design your arc break correctly, it just has to quench in that time to stop the arc. It's what allows your circuit breaker at home to handle fault currents of 10,000-20,000A without completely blowing up. The breaker might die, but it will handle and break that load current.

DC load breaks can't count on this behaviour - the current will keep flowing as long as the arc path is maintained and it can be maintained for a while because it maintains an ionized air conducting path.

The DC breaking capacity is often much lower and you have to rely on a lot more suppression circuits to starve the arc of current so it dies out quickly. That's why on DC switching circuits you often have snubber networks - when a switch contact opens, the snubber networks serves to rob the contacts of current to maintain the arc, so when the contacts are fully open thats when the snubber network stops conducting which lets full current flow to the contacts again but by now the gap is too big to conduct.

You use a switch and you'll end up with welded contacts.

DC loadbreaks don't have that luxury

Exactly right. Typically, a 10A AC rated toggle switch can only handle 2A DC. Thats 5X worse! Geez, I expect better basic technical knowledge from Slashdot posters.

You might be surprised by how few people understand this stuff, even some engineers. It's somewhat specialized knowledge, even in EE.

But again, as "tlhIngan" wrote, it's the voltage, being DC, that will "draw" a huge (long) arc. Your switch needs to physically open maybe 3 inches (8 cm?) to break the arc.

What they do with high-voltage switching like this would need is a snap-action mechanism so you're not drawing a growing arc, and then an insulator that gets slammed into the gap to be sure to break the a

"Typically, a 10A AC rated toggle switch can only handle 2A DC."

Funny, we use a 10A AC switch with 10A DC when we test the LASER modules we make at work. It works just fine.

120V is the input voltage, the LASERs themselves run at 16.8VDC.

You use a switch and you'll end up with welded contacts.

Or a non-stop arc that will cause some kind of fire, things melting, possible catastrophe.

As I mentioned below, some of the switches and relays for higher voltage DC insert an insulator in the arc path to be sure to break the arc.

"Most likely the panels are fine, the wiring was probably done by the lowest bidder contractor with little to no experience in DC solar panels."

The wiring is a simple barrel-plug MC4 assembly, so doing the wiring wrong is literally fucking impossible.

I can tell you don't build or manufacture solar panels. I did for Sunspark Technologies for two years.

MC4 connectors are anything but simple. Compatibility is not guaranteed between manufacturers, and it is easy for them to not be properly seated. They are also ineffective at load-break. Field crimps make things worse, and improper wire insulation makes the overall system a hazard.

Bottom line is that these jobs were done by low skilled workers with low quality equipment... and poorly commissioned and maintained.

"MC4 connectors are anything but simple."

They literally operate like plastic buckle clasps you find on backpacks and tote bags. It's dead fucking simple. If you're fucking that up, you don't need to be a solar installer. They're also standardized, so they'd better be fucking compatible between manufacturers.

"They are also ineffective at load-break."

That's literally what a piece of plywood covering one panel out of a string is for. You kill one panel, you effectively kill the string right there, making it sa

MC4 connectors are more complicated than they appear at first glance. That is why UL and the NEC are requiring additional labeling/certifications... specifically because they lend themselves to poor connections often enough to pose a hazard. Your parallel to snap clasps on backpacks is appropriate-- just because they are the same size and function doesn't mean they fit together properly. Accidentally trying to disconnect them with current flow makes problems worse.

Yes, they should be dead simple, but in

"There were a lack of tragic scenes as the event unfolded today, the surrounding region was unaffected and everyone went on with their day, mostly unaware. The community remains intact and no large scale evacuations were needed, people are also able to remain in their homes or go outside if they so choose for the forseeable future. The building remains intact and people are able to go about their business as usual."

"Apart from some minor financial inconvenience the accident was benign and no long term h

Can't blame it all on solar flares though. Even if the flares induced a current in the transformers in the inverters, they are supposed to have protection there that will prevent this sort of thing.

Nope. Most solar inverters are actually transformerless. The IC 4047 is one of the most popular solar charging ICs available (but it has shit thermal tolerance... guess what's in most inverters?)

They should have chosen Amazon as the seller. It is the smart thing to do.

Presumably they didn't want to buy cheap Chinese crap and blacklisted Amazon as a potential vendor.

There is no mention of micro inverters on each panel they are string setups, and they are far more vulnerable to fire due to high voltages and currents. Its most likely a poor quality installation that caused these issues. Who modded this dribble up?

Outsourcing significant plant design/installation/maintenance really doesn't work This stuff ain't tinker toys kids

Rooftop solar already costs a lot of money. If it is going to start fires then that is only going to make it look worse. https://en.wikipedia.org/wiki/... [wikipedia.org]

For some reason mention of the Lazard study triggers people, it angers the pro-nuclear people for making nuclear power look bad and it angers the pro-solar people for making rooftop solar look bad. Ignore it on the chart then. Look at the other numbers. Rooftop solar still looks bad in the other studies.

Put the solar collectors on the ground so it is ch

Anybody with the slightest bit of actual knowledge knows that plants happily co-exist with solar panels. See this link https://www.wired.com/story/gr... [wired.com]. Facts, so inconvenient. In your desperation to push the nuke barrow, you dont bother to do the slightest bit of research on other options do you? When you do you, its only on nukes and you quote someones blog with little credibility.

As for the danger of rooftop solar, we have here in South Australia one of the highest adoption rates for home and carpark ro

As for the danger of rooftop solar, we have here in South Australia one of the highest adoption rates for home and carpark rooftop solar in the world, and there has not been ONE reported example of flash over or fire.

https://theconversation.com/so... [theconversation.com]

I still havent seen a report of such incidents in South Australia, there is no mention in the article of SA.

Anybody with the slightest bit of actual knowledge knows that plants happily co-exist with solar panels.

How much does that cost? With more land, material, and labor needed for the same electrical output I can only imagine the costs are higher than if the solar panels were packed in tight.

We are living proof you dont need nukes.

There's plenty of dispute on that point among those in the industry and among policy makers.

Why not read the link and find out, panel spacing doesnt change, you need to be able to access them for maintenance anyway, nice try, but fail. Theres no debate here, either in industry or policy. We dont want or need nukes.

Why not read the link and find out,

I did read the linked article, it didn't answer my questions on costs. That is why I asked.

panel spacing doesnt change, you need to be able to access them for maintenance anyway, nice try, but fail.

Bullshit. It's in the article that the spacing of the panels needs to change to accommodate the specific crops. If it has to change to match the crops then it's not going to be the same as that used if there are no crops.

Theres no debate here, either in industry or policy. We dont want or need nukes.

If there's no debate then you have nothing to worry about. If there's no debate then why even bother to reply to my comments?

If this is all so obvious then how is it that planting crops among sola

...by small scale solar everywhere, that is in the constant leakage of sp6 into the environment.

A catastrophically bad chemical for warming, 23000x as potent as co2, ubiquitous, and per the bbc just the solar installations in EU in2017 caused as much warming as A MILLION cars on the road.

And solar has vastly expanded since then.

https://www.bbc.com/news/scien... [bbc.com]

As even your crappy citation says, sodium hexafluoride is used for power connections for ALL TYPES of power generation. Here you are pretending it's leaking out of solar panels, when more grid connections are LEAST likely to be associated with solar power, because it is MOST likely to be installed at the point of production, where it DECREASES the amount of interconnections necessary.

If we add ANY kind of power which is NOT residential, we need more grid connections. And the switchgear increases in size and

The small-scale solar project near me off the Santa Ana River bike trail has none of what's mentioned in the article. I'd know, I helped make the solar panels for the thing and was present when it was opened and shown to the public.

That sort of thing is in larger-scale farms where arcing is quite a concern.

I have a better idea. Let's simply not mod up an ignorant dolt who thinks SP6 has a direct link to small scale solar and just move on with our lives.

I died, but I got better!

Last time I checked, solar sytem is owned and operated by gravitation law.

There may be more comments in this discussion. Without JavaScript enabled, you might want to turn on Classic Discussion System in your preferences instead.

Cruise Recalls Robotaxies After Passenger Injured In Crash

Major VPN Services Shut Down In India Over Anti-Privacy Law

Variable Spring Supports Anyone can do any amount of work provided it isn't the work he is supposed to be doing at the moment. -- Robert Benchley

Amazon Took All US Solar Rooftops Offline Last Year After Flurry of Fires, Electrical Explosions - Slashdot