Off grid cabin— frozen 48v battery in low voltage and BMS won’t start up

[Hardergamer1]

I'm wondering if the reason for the total battery drain was something to do with the RPI, or the RPI wiring etc, as it sounds like it may have had an internal short, or wired wrong? And when you brought the batteries back up, something had to give...

 

[upnorthandpersonal]

What type of current is needed to get that affect?

Jävligt mycket/en jävla massa I think is the right term

As @Hardergamer1 just mentioned, I think it's time to do a close up on the wiring diagram. This together with you noticing it stopped charging for some reason, and the battery completely draining,...

 

[timselectric]

Ditto
Something is definitely not right.

 

[HRTKD]

It's pretty bad when we have to put fuses on communication wires.

 

[timselectric]

It sounds like current leaking between the two systems. Which means that there is another connection between them to complete the circuit.

 

[Hardergamer1]

Can you test the output signal voltage from the smart shunt? As I'm wondering if it is giving out full battery voltage which is feeding into the RPI.
And you did say the battery/shunt is not connected to the RPI power supply?

 

[Messier11]

Can you test the output signal voltage from the smart shunt? As I'm wondering if it is giving out full battery voltage which is feeding into the RPI.
And you did say the battery/shunt is not connected to the RPI power supply?

Yes, now, when the BMS is “open”, it’s just app 2.5-3v between the communication port of the smart shunt and (1) the minus pool of the battery, (2) the USB port of the RPI and (3) a minus pole of the 12v system.

It sounds like current leaking between the two systems. Which means that there is another connection between them to complete the circuit.

Yeah, how, when and where I don’t know.

The 12v system is one option, it’s connected through (a) one of the RPIs getting it’s power from it, and then there is a COM/MBUS cable from the EPEver charger to the other RPI that runs Home Assistant. This RPI is powered by the 48V system (230 cable). In addition, the RPIs are connected with com cables to the MPPT, the BMS, the MP2, the SmartShunt (the cable that burned up), etc.

But with both RPI surviving without a scratch, I wonder if the power must not have gone from one USB cable to another?

This is how it looks now, but it’s of course impossible to get a detailed overview from a picture.

Jävligt mycket/en jävla massa I think is the right term

As @Hardergamer1 just mentioned, I think it's time to do a close up on the wiring diagram. This together with you noticing it stopped charging for some reason, and the battery completely draining,...

Haha, exactly my thought, would not have expected that if I shorted a car battery…

Yeah, I need to check it up.

Where should I start? Any measurements I can make? Should the BMS discharge be off?

 

[timselectric]

I would start looking at every known connection between the two systems. And work my way out from there.
Although, it could also be a rare occurrence. That may only happen when battery voltage is extremely low.

 

[Hardergamer1]

You could start with checking to see if there is any resistance/continuity between the shunt/battery bank and the RPI power supply?

 

[Messier11]

Will do! I am charging now, but if I turn off everything in the BMS later, I should be able to check if there is any phantom currents later, right?

 

[upnorthandpersonal]

I would start by disconnecting anything remotely non-essential, such as the Pi, anything to do with the 12V system, etc. Make sure the battery works fine, that cells hold their voltage, etc. Then before doing anything or plugging in anything, draw out what you're going to reconnect and how, and analyse if there are potential issues (creating any loops?, potential reverse polarity?, etc.) and go from there, one step at a time. Whatever you do, don't rush!

 

[HRTKD]

Is there a fuse on the power to the rpi? I can't tell from the most recent picture. My Cerbo GX has an inline fuse. I vaguely recall that is was maybe a 2 amp fuse.

 

[Messier11]

I would start by disconnecting anything remotely non-essential, such as the Pi, anything to do with the 12V system, etc. Make sure the battery works fine, that cells hold their voltage, etc. Then before doing anything or plugging in anything, draw out what you're going to reconnect and how, and analyse if there are potential issues (creating any loops?, potential reverse polarity?, etc.) and go from there, one step at a time. Whatever you do, don't rush!

Yeah, thinking about it, the last thing I got was a com cable for the EPEver (12v), that I connected to a home assistant RPI that also is connected to the communication port of the BMS, to be able to remotely monitor both the 12v system and the BMS. It’s not many days after I got that working and the MPPT stopped charging. That could be it.

Knock on wood, I don’t think anything is mistakenly connected, but if power for example can ‘leak’ (if that makes any sense) from for example a communication port — then the number of erroneous circuits that can be created is almost endless…

 

[sunshine_eggo]

What does the JK report now?

My biggest concern is the inability to report IR on prior screenshots. Deep discharges of LFP, if they can be recovered at all, typically exhibit notable increases in internal resistance which notably worsens voltage sag when discharging. They may have been rendered unusable due to increases in internal resistance. They will have reduced cycle life, and the likelihood of spontaneous cell failure is dramatically increased.

 

[Messier11]

What does the JK report now?

My biggest concern is the inability to report IR on prior screenshots. Deep discharges of LFP, if they can be recovered at all, typically exhibit notable increases in internal resistance which notably worsens voltage sag when discharging. They may have been rendered unusable due to increases in internal resistance. They will have reduced cycle life, and the likelihood of spontaneous cell failure is dramatically increased.

It is looking good in that there is a fairly consistent progress, have been charging for what 8 hours now:


BTW, should I try to off-set the 13th balancing cable to get the balancing going? If so, should it be minus 0.048? It’s perhaps not that big of a deal.

 

[sunshine_eggo]

It is looking good in that there is a fairly consistent progress, have been charging for what 8 hours now:

BTW, should I try to off-set the 13th balancing cable to get the balancing going? If so, should it be minus 0.048? It’s perhaps not that big of a deal.

There is zero point balancing below 3.40V. It may actually do more harm than good.

Hopefully, when you are able to discharge, the IR values may populate.

Do you happen to have any screenshots pre-event?

 

[Messier11]

Is there a fuse on the power to the rpi? I can't tell from the most recent picture. My Cerbo GX has an inline fuse. I vaguely recall that is was maybe a 2 amp fuse.

No, it’s just an ordinary USB to USB C cable.

You could start with checking to see if there is any resistance/continuity between the shunt/battery bank and the RPI power supply?

When the power cable is disconnected, I don’t get any reading between the power cable and the smart shunt/or the 48v system, it just calibrates all the time. It’s the same when I measure towards an iron radiator in the room. I do get connectivity between the RPI and the SmartShunt (even when the RPI doesn’t have any power). So electricity can obviously loop over the RPI and the USB cables connected to it.

 

[Messier11]

There is zero point balancing below 3.40V. It may actually do more harm than good.

Hopefully, when you are able to discharge, the IR values may populate.

Do you happen to have any screenshots pre-event?

Aha, sorry, brain fart by me, didn’t figure out that the only one with a correct reading is 13 and the other are off when they are at zero… Thought 13 had issues.

 

[Messier11]

Done charging for today, family is fed up with the sound of the generator, ran it for pretty exactly 10 hours at 28A. This is how it looks now:


Do you guys think the 89%/remaining capacity of 249a figures could be accurate?

There is zero point balancing below 3.40V. It may actually do more harm than good.

Hopefully, when you are able to discharge, the IR values may populate.

Do you happen to have any screenshots pre-event?

This is how it looked during the summer:

Considering this is shown as 0, a power reset of the BMS might fix this. But this might also be related to the low cell voltages. See if they get back up and stay up first...

How do you restart the BMS? Should I just pull out and replug the balancing cables?

 

[upnorthandpersonal]

How do you restart the BMS? Should I just pull out and replug the balancing cables?

Yes, just disconnect from the battery and then start it up again with the 9V battery method.