So, I joined the BURN club.

[manson79]

was in my garage, heard a crackle, thought it was something falling, then smelled vapors. walked over there, and saw this.

My system is 3x deye 16kw inverters in parallel, to 5 batteries, connected to a battery breaker box. 5x 125a breakers and 1x 300a main breaker before it connects to the inverters. The batteries are home built, top balanced, and using DALY 250a BMS's with interface boards.

I tried for days, and NEVER got them to talk to the DEYE inverters. but we did leave it in LITHIUM mode, with the BMS"s settings set to protect the batteries.
Also, float setting was not disabled. not sure if that has anything to do with it.

INFO: the 5v on the BMS is because one of the wires melted after the fire.
this is the first time i got these batteries up to 100%
Also notice on Solar man, it was 100% at 12:10, but, the inverters were still sending 13kw for 20 minutes after it was 100%. i'm thinking overcharge? but why would it do that? and why would the BMS's not catch that? why would DEYE overcharge in first place.
Batteries were tightened, didn't detect any obvious loose terminals.

Any thoughts on what could have caused this?

 

[rodrick]

Look like a bad connection on lower terminal the burst disc is still intact less cover or a short or over load on lower cable

 

[crossy]

I seem to recall another fairly recent thread where a Daly BMS failed to cut off on cell over voltage

 

[meetyg]

Is the Deye setting for Lithium correct for Lifepo4? Maybe it's for Ternary Lithium?

 

[manson79]

key info on the data connection. we NEVER got it to talk. maybe 1x or 2x i thought I saw it say 30c instead of the 100c. so maybe there was some communication, not sure if that was random or not though. so now I'm going to try the DEYE on voltage setting. but that's quite annoying with such expesnive and useless BMS's. i know they do balancing as well, but I could have gone cheaper.

 

[manson79]

Either way, I'm surprised that the Deye was still sending current into the battery for 20 minutes when it was already 100%. that feels suspect to me.

 

[manson79]

Look like a bad connection on lower terminal the burst disc is still intact less cover or a short or over load on lower cable

so the silver trianges are a "BURST DISC" is that correct? and where do you see a bad connection? overload seems possible, just don't know how. the deye kept sending 13kw after it was 100%.

 

[fatjay]

Glad nothing else burned.

Tons of people don't have bms communication between inverter and bms, myself included.

Remember, BMS SOC of 100% is an estimate at best. You need to look at battery voltage. This is the key and what people use for running open loop communication.

What are your back to grid voltages on the inverters? Did it hit it? Did the inverters keep sending power back after bms hit top voltage?

Inverters are the primary controllers on when to draw power from battery, solar, and grid. BMS is the backup if the inverters go wrong. I would look at those settings first.

 

[manson79]

i tried to zoom in. does this help?

 

[manson79]

Glad nothing else burned.

Tons of people don't have bms communication between inverter and bms, myself included.

Remember, BMS SOC of 100% is an estimate at best. You need to look at battery voltage. This is the key and what people use for running open loop communication.

What are your back to grid voltages on the inverters? Did it hit it? Did the inverters keep sending power back after bms hit top voltage?

Inverters are the primary controllers on when to draw power from battery, solar, and grid. BMS is the backup if the inverters go wrong. I would look at those settings first.

Well, funny, i did ask that the day before. the DEYE was set to "lithium" mode. which i think then just uses an internal table to send power. I couldn't find anywhere that references those values. When I set the DEYE to VOLTAGE mode, then I can see volt values. I just assumed the DEYE LITHIUM mode had the proper lithium values we all use.

 

[manson79]

A little update. we had an iron rack made. and the welder isn't the biggest attention to detail. So a hypothesis is that when the batteries got fully charged and expanded, it may have expanded into the weld. it's not super sharp, but, maybe it caused the dent?

COUNTER INFO: this system has been up for about 10 days. but never reached 100% until the day of the fire, it got close 3 times, at 90%, but this was the only time it got 100% SOC. not sure if that helps they mystery

 

[rodrick]

so the silver trianges are a "BURST DISC" is that correct? and where do you see a bad connection? overload seems possible, just don't know how. the deye kept sending 13kw after it was 100%.

Yes the little window is a burst disk vent if a cell is over charged typically they will burst i think you had a bad connection or fault internally on that post

 

[manson79]

Yes the little window is a burst disk vent if a cell is over charged typically they will burst i think you had a bad connection or fault internally on that post

ahh.. ok ok.. so you are deducing that it didn't operpressure, since the burst disk didn't pop open. I guess i've never seen what that looks like. did you see the new pictures i posted where it looks like there was damage from the rear? could that hole in the rear have relieved the pruessure? and liquid leaked out there? and then no liquid, would cause an external spark? or spark behind the post?

 

[rodrick]

Yep you had a short and generated a lot of heat you have to use a protective non conducting material between and around the cell where they can contact surfaces

 

[rodrick]

Pressure won’t create a hole like that and not pop the disc the case of the cells are negative just like the terminal

 

[rodrick]

The bottom terminal was positive right?

 

[fatjay]

Well, funny, i did ask that the day before. the DEYE was set to "lithium" mode. which i think then just uses an internal table to send power. I couldn't find anywhere that references those values. When I set the DEYE to VOLTAGE mode, then I can see volt values. I just assumed the DEYE LITHIUM mode had the proper lithium values we all use.

I don't know how DEYE inverters work, but for hte ones i've seen lithium means charging profile. Yours may be different.

For some reason I remember a video talking about lithium meaning bms communications. And if no bms comm's, use voltage. In which case you tell the voltage where you switch back to battery and stop sending the battery power. If that is the case, the answer is a bit obvious, it was waiting for the bms to tell the inverter they're done charging but without comm's, it was impossible, so the inverters just kept sending power.

I would get clarification on what your inverter is doing when in lithium vs voltage mode.

 

[crossy]

I'd second the "short" to the case via your dicey welding. It's imperative that you have insulating material all around and BETWEEN the cells.

 

[manson79]

The bottom terminal was positive right?

this is the left and right side of the 16 cells. I'm deducing that yes, it was positive (black negative being above it)

 

[manson79]

there is 5 batteries in a row, one above each other. these pictures are mainly the middle row, call it #3. But the row below this, had the first cell also have some damage, it would be just below the picture with the black terminal. Not sure if it was caused by blown liquid, or same fire, or total overvoltage for the whole pack. FYI, battery #1, #2, and #5 look fine.

 

[Mrgoodcat]

I'd second the "short" to the case via your dicey welding. It's imperative that you have insulating material all around and BETWEEN the cells.

What material is ideal? Which isn't flammable etc.

 

[manson79]

I don't know how DEYE inverters work, but for hte ones i've seen lithium means charging profile. Yours may be different.

For some reason I remember a video talking about lithium meaning bms communications. And if no bms comm's, use voltage. In which case you tell the voltage where you switch back to battery and stop sending the battery power. If that is the case, the answer is a bit obvious, it was waiting for the bms to tell the inverter they're done charging but without comm's, it was impossible, so the inverters just kept sending power.

I would get clarification on what your inverter is doing when in lithium vs voltage mode.

yeah, i'm suspect of this also. FYI the DEYE is same manufacture as SOL ARK. but for Asia market vs USA. I'm currently in philippines. And yeah, I think you are right, that LITHIUM mode means data-communications, but damn, there is no safeguard on that? it would just blindly send electrons? and then, we do have that DEYE bms. which didn't show any faults either. it technically should have stopped it also.

 

[fatjay]

yeah, i'm suspect of this also. FYI the DEYE is same manufacture as SOL ARK. but for Asia market vs USA. I'm currently in philippines. And yeah, I think you are right, that LITHIUM mode means data-communications, but damn, there is no safeguard on that? it would just blindly send electrons? and then, we do have that DEYE bms. which didn't show any faults either. it technically should have stopped it also.

The BMS is supposed to be the safeguard if the inverter chooses to do what it wants. The bms' i have dealt with are fully programmable.

If it wasn't set up right, ie no cell voltage shutoff, or if it just didn't do it's job, then the battery would cook if the inverter kept sending voltage.

I have JBD BMS, and this is their configuration. Look closely at CellOvp, or cell over protection. It cuts off at 3.75v, and won't release it until 3.55v. Basically it'll shut down the whole pack. Which your Daly BMS did not, either because it wasn't' configured properly, or it just failed to do so.

I believe @crossy said it earlier where a Daly bms failed to do the same thing.

 

[rodrick]

There was no high voltage to stop you had a high discharge that the bms had no way of stopping due to a direct case short

 

[Mrgoodcat]

The settings were set similar to this. Except the cell low voltage was 2.6V (not 2.2V like in this image).

This is from the DALY BMS.