6000xp gone crazy, and why a BMS is mandatory for LiFePo4

[sollap]

I was doing some balancing today, and had my 6000xp set for 56.2v charge and 56v float.
Charging from PV. AC charging disabled.



For some reason, the XP let it run all the way up to 59.8v!

I'm sure that's open circuit voltage because my BMS's would have disconnected before then.
Still, when I shut off the PV and checked the batteries (still connected to the XP) they were showing 57v on my meter. Much higher than I set the charge for.

Everyone here already knows it's a terrible idea to run LiFePo4 without a BMS, but this is a great example for the new folks. I can only imagine how bad this could have been if the BMS's hadn't stepped in to protect the cells.

I'm on ccaa-190D0E. I think the latest firmware includes a battery voltage fix. Hopefully that will address this in the future.

 

[MisterSandals]

If you have a lithium iron polonium battery, why does it look like you are expecting lead acid settings to be the culprit?

The 59.8V could be a momentary spike when/if charging cut off unexpectedly. 57V battery could very likely have a cell over volt disconnect.

 

[sollap]

The 59.8V was more than a momentary spike, the battery circuit was sitting at that voltage according to SA. I didn't check it with a meter.

I see your point though, that with charging disabled (PV off and AC charging not enabled), the 57v, and 59.8V, could be no-current readings from the charging circuit if the batteries were out of circuit due to BMS disconnect.

If that were the case, I would expect the voltage to drop suddenly when one of the BMS's kicked back in. I didn't see that happen. But, I also wasn't watching it like a hawk. Now I wish I had been.

 

[MisterSandals]

But, I also wasn't watching it like a hawk. Now I wish I had been.

Watching like a hawk for the first few cycles is standard operating procedure if everything goes perfectly. It sometimes takes scaling back the voltages drastically and slowly inching upwards to where its good enough and still plenty reliable/predictable.

I used to start out with high voltages but its far easier to start low and gradually increase, mostly based on the need for more storage.

 

[sollap]

I forgot about the charts in SA. They show what I didn't see.


I incremented voltage from 55.6V to 56V over 20 minutes. (12:32- 12:52)

It then floated at 56V for 26 minutes. (12:52-13:18)


After that it began a slow climb on it's own. Over 20 minutes, it incremented from 56V to 58V. It never should have done this.


That's when my BMS's started going off.
It went from 58 to 59.2 in 2 minutes.
2 more minutes and it was at 59.8


These chart times are not exact, and are affected by SA's 1 minute refresh times.
But, it shows the significant difference between the slow unauthorized voltage climb, and the sharp jumps caused by the BMS's dropping out.

It looks to me like the 58V (at least) was very real charging voltage being sent to the batteries, even though the XP was set to stop at 56.2

 

[SunDave]

On the computer web browser version of solar assistant you may be able to zoom in. Mine goes to 10 second increments. Just in case you were interested in more detail. Might be an option for you. Might not, depending on your settings.

 

[sollap]

Awesome, that does help. Now I can more clearly see that slow ramp up in voltage, but I can also see some crazy spikes after the "close encounters" plateau. Based on the time it happened, that's when I switched off the PV.