• Have you tried out dark mode?! Scroll to the bottom of any page to find a sun or moon icon to turn dark mode on or off!

diy solar

diy solar

Local Pittsburgh Solar Battery Fire

I'm sure that plenty of inspectors wouldn't catch it, if the plywood has a fire rating stamp on it.
That's what they are used to looking for.
 
The Sunnyboy 4.5kw each can charge at 110 amp continuous / 140 peak, yet the PHI batteries are 28 ah maximum each so in 5p2s specified would be 140 amps continuous.
Looking at the different pics, it looks like 2 separate 5P battery banks going into the "combiner box" above the batteries. Additionally, it looks like the center battery of each 5P bank is what is connected to the combiner box.
So these "center" batteries would be getting LOTS of current.
 
Ten 75AH, 48v, 16 pouch, batteries in parallel. I wonder if anyone checked current distribution balance between the ten batteries? If one BMS overheats and shuts down it pushes more current to remaining batteries. Every BMS should have shut down in a domino cascade.

This assumes the original ignition was batteries. It could have just as well been junction box wiring, or breakers. Overheat the batteries and they will pop.

LFP electrolyte is no different than most Li-Ion batteries. They will have thermal runaway and burst. Only difference is LFP cathode does not produce the large amount of oxygen the other Li-Ion types cathodes produces. That does not stop ignition of electrolyte if there is enough heat and oxygen in the environment. With the lack of oxygen vented, LPF will not 'blow torch' out the vent port like other Li-Ion chemistries.
 
Last edited:
If this was not LiFePO4, and instead a cobalt based lithium ion battery bank, the temperatures created would destroy everything you see in the photo. It's LiFePO4 for sure. Looks like the electrolyte did combust though, but when cobalt based lithium ion does a thermal run away, the temperatures created would melt everything off the wall. Especially this many batteries. Look up videos of what a single cell can do. This many batteries would make this scene look much different.
 
Last edited:
Here's a great video if what I assume is a cobalt based battery going full rocket ship. Right before the fire you can see the metal box bulge up.

 
I would think the BMS of the remaining batteries would have gone into shutdow due to high load demand when other batteries had gone into LVD mode.
 
I see the Sunny Boy is still truckin though !!!

Lights on! (but nobody home)

Reviewing the 4.5kw Sunny Island and the 1.4 PHI batteries there appears to be a design flaw in the installation

The Sunnyboy 4.5kw each can charge at 110 amp continuous / 140 peak, yet the PHI batteries are 28 ah maximum each so in 5p2s specified would be 140 amps continuous.

So the grid charges the batteries at 220 amp but the batteries can't handle more than 140 amp in that configuration. Add in the issue with wire length differences ( the manufacture warns against this ) would allow more current to one battery and it's a time bomb.

Just because they can, doesn't mean they have to.
The Sunny Island should be programmed to regulate battery current to desired amount.
That's the beauty of this sort of system. Mine has AGM, and charge rate is about 1/3 of available PV power. So it sustains that over a wide range of load.
 
Just because they can, doesn't mean they have to.
The Sunny Island should be programmed to regulate battery current to desired amount.

Yes, I agree and the permit package would have those details as stamped by the PE.

With the 5p configuration, just unequal length wires would be the cause and as I previously posted showing the closet battery was the one that flamed out first.

Update to post:

I clipped a frame from the video of the fire, and added lines to highlight the approximate positions of the batteries. It's clear in the video the top, second from the left battery was the source of the fire, which is also the shortest to the combiner in the box above. You can also clearly see the electrolyte on fire as it is running down and on to the floor.

Sealed metal boxes prevents this type of spread and why UL9450A batteries are all in liquid tight metal containers. Also note the plywood backing / support which is NOT to code is spreading horizontally the flames.


ignition-source-png.132623
 

Attachments

  • ignition source.png
    ignition source.png
    609 KB · Views: 222
Last edited:
My guess is the litigation for who was at fault will go on for awhile :(
The Food Rescue will blame the people they bought the system from, That company will blame the manufacturer of the batteries, The manufacturer of the batteries will blame the installer. The installer will blame the Food Rescue for not operating according to instructions. The various insurance companies and Bond issuers will blame everybody and deny payment. The lawyers will get together and have long martini lunches that they bill their clients for.

Me, I blame the gods. Probably Thor's handiwork since he is the God of Thunder.
 
I'm trying to look closer at their design in the hopes that I can continue to reassure the folks I'm working with right now that our new system won't have a similar flaw.

I've found a web page that purports to tell all about the system but the bottom of the two links goes to a file that's been made private. Did anyone happen to cache it while it was still up? I'd really like to see the system design. I've circled the page I can not get to.


1675260019403.png
 
I'm trying to look closer at their design in the hopes that I can continue to reassure the folks I'm working with right now that our new system won't have a similar flaw.

I've found a web page that purports to tell all about the system but the bottom of the two links goes to a file that's been made private. Did anyone happen to cache it while it was still up? I'd really like to see the system design. I've circled the page I can not get to.


The other linked pdf, looking at the properties is a the author which is on LinkedIn : He maybe able to share the supporting work for this project which may have been from this organization
 
I grabbed the video showing the system when it was originally built but was not able to copy down that PDF as it was already private.
To be fair to the system it had been running for 3 years before it all went wrong. My guess is that whole website and all technical information will disappear if it hasn't already.
Ultimately I would be asking big questions of the city and their inspectors as well as the installers. However it looks like there's a lot of DIY going on there.
 
I grabbed the video showing the system when it was originally built but was not able to copy down that PDF as it was already private.
To be fair to the system it had been running for 3 years before it all went wrong. My guess is that whole website and all technical information will disappear if it hasn't already.
Ultimately I would be asking big questions of the city and their inspectors as well as the installers. However it looks like there's a lot of DIY going on there.

Permits require wet stamp from a PE, showing it is engineered to code and the one that if shown missed things is all on them.

Looking at the names involved from web-site and LinkedIn, none I've seen are ESS designers, looks like allot of well meaning people trying to help in this community project. Fortunately, it is only property damage and likely mostly donated time and materials.
 
From what I could gather from the video, the whole system seemed odd.

They keep the massive array on the roof DC all the way to the basement where they convert it to AC to feed a rectifier 10 feet away that converts the AC back to DC to tend to the batteries.

Then those batteries seemed to power DC lighting circuits throughout the building made by a company called SigmaLuminis who's website says they can talk the BACnet protocol but they are unknown to the BACnet community (and haven't registered for a unique BACnet ID).

Odd lighting vendor aside, I can't wrap my head around why they decided to set the system up this way. All of the long runs are DC. Both the huge cables coming from the roof arrays and the presumably smaller ones feeding DC lighting throughout the building.

But the short connection, the 8 feet from the power-in on the right side to the battery tending rectifiers on the left side they chose to convert it to AC and back.

Can anyone chime in on what the thinking must have been? I have no direct experience in grid-tied systems. Is this type of design common? It almost appears as though they went out of their way to make it inefficient and complicated.
 
The solar should be high voltage DC, so run length is less important. I don't see why that's an issue.
 
Can anyone chime in on what the thinking must have been? I have no direct experience in grid-tied systems. Is this type of design common? It almost appears as though they went out of their way to make it inefficient and complicated.

The attached document that was linked in a post above, explains the system. The Sunny Islands are what charge the battery, AC coupled to the Sunnyboy's which converters to Solar DC to AC. Just a run of the mill, been out for 20 years SMA storage solution, except for the batteries.
 

Attachments

Sorry for a thread necro, I thought SimpliPhi chemistry and build quality was high? I was looking at purchasing half a dozen of their 3.8 kWh batteries for my off grid. Could anyone point me to a few top 24v 3-4 kWh batteries for off-grid use?

EDIT: just checked and the SimpliPhi 3.8's are • UL, CE, UN/DOT and RoHS compliant components - UL 1973, UL 1642

 
Last edited:
9540A is the fire rating to look for. we have pushed our EG4 packs through it and exceeded the standards with the fire arrestors just to ensure that the batteries are hardened against this.

the main thing to watch for is improper torque on busbars paralleling batteries. a perfectly safe battery will go through a lot of thermal strain with bad torque
 
9540A is the fire rating to look for. we have pushed our EG4 packs through it and exceeded the standards with the fire arrestors just to ensure that the batteries are hardened against this.
Do you have 9540A listing now, or is it still being formalized?
 

diy solar

diy solar
Back
Top