JK BMS cell UVP ‐ Prevents charging via AIO inverter

[meetyg]

A short intro:
I have a 24v DIY Lifepo4 pack, 8x105Ah EVE cells.
While I got these cells from a reputable supplier (grade A), it seems one cell is a bit faulty. If I recall correctly, there was one cell in this order that made a slushing sound when I shake it, it might be this one. There was no visible leakage, so I just built my pack up.
Anyways, I top balanced all of the cells and the pack has been running fine for a few months now, with a JK_B2A8S20P BMS.

But recently when draining down the pack to low SOC, I noticed that this cell may have a bit less capacity then the others.
At high SOC there are only a few mV differences, but as the SOC goes down, it's about 20 mV. Then at even lower SOC, it produces a large delta of around 200mV.

Anyways, I have this pack connected to a 24v AIO inverter, without BMS comms to the inverter (only to SolarAssistant).

I set the inverter to stop discharging battery at 25v, which is more than set at the BMS (3v per cell).

But what happens is that the JK goes into cell UVP (2.9v of that problematic cell), while overall voltage is above 25v (3.2v per other cells) and cuts off the discharging MOSFET.

So I manually forced my inverter to charge (changing settings), but the BMS doesn't show current flow.
If I enter "emergency mode" on the JK, which enables the discharge MOSFET, charging occurs properly.

I have the inverter set to charge @ 27v float, 28.9v absorption.

I think this issue is related to what Andy shows on this video:

His solution was to up the power supply voltage by 0.6v.
But what I don't understand is why his solution doesn't work when the inverter tries to charge?
As I have the inverter charge voltages set higher than the current voltage of the battery by more than 0.6v.

Any Ideas to a solution ?
This behavior certainly doesn't happen with JBD BMSs, which automatically enable the discharge MOSFET when charge current is present.

Of course I plan on replacing the faulty cell, but this issue worries me and I would like to find a solution for now.

Of course I could increase the "stop discharging" voltage on the inverter even more, but then I loose even more usable capacity.

Thanks in advance.

 

[Skypower]

That’s pretty early for that amount of delta at 3.2 volts average. My usual call is to harp about a loose buss terminal connection. Since you have had a good top balance and you noticed an oddity, I don’t believe in coincidence. This may be the rare bad cell that bad luck dropped in your lap.

 

[Skypower]

You do have the low voltage protection a bit higher than some, not that it’ll make much difference in capacity. That and a bad cell, perhaps the BMS doesn’t know how to deal with it. From a pamphlet I found in a CATL shipment. The absolutes are a bit extreme though

 

[Steve_S]

ALL LFP cells make a slushing sound when shaken, that s perfectly normal, that is the electrolyte.

You cutting off the BMS too early.
Check out the settings I use & Recommend for JK @ 24V

Allows for Voltage Sag when low to prevent cutoff.

Never let the BMS hit 2.500. Allows for charging right away.

<--- START @ 3.42




















<<<---
< Correct TEMPS; Factory defaults are wrong for LFP
<
<
<
<
<
<
<
<<<---

 

[meetyg]

Thank you all for your inputs.
I know that my UVP settings are somewhat conservative, but that's on purpose for cell lifespan.

Still I think that the JK BMS has a bug (as Andy shows), that could prevent a system from running if UVP is reached. It's kinda like a deadlock: Battery has UVP, and therefore discharge is disabled, but then charge can't commence and UVP never goes away.

I find it odd that not many have mentioned this bug in the JK BMS.
Again, JBDs don't have this bug, AFAIK.

 

[Skypower]

We, (you) are the beta testers
So far the bugs have been remote fixable. May want to first check/adjust your settings a tad and see if it makes any difference. If not notify JK.