Dead cell balancing on JK BMS

[jontm]

Hi!

Currently doing testing and trying to get my new 16S EVE LF280K battery working.
In the start everything was working as intended. Got balance current up to 2 A and system was charging with a low 30 A.
Turned off the balancing for much of the charging as it was not really needed. As I was approcing 3.35 V per cell i turned it on again, only this time there was only A very tiny balancing current. 20 mA or so. As I got to top of charge, a couple of the cells started to shoot up in voltage and tripped the over voltage protection at 3.6 V.
I figured i might as well restart the BMS and see if I could get the BMS to do the top balancing.
However, when I disconnected the BMS and connected it again, it no longer recognice the balancing leads.
All leads show 0 ohm and I get an alarm for high balancing lead resistance. I have checked the connections and they look ok.

Anyone have had similar issues with the JK BMS or JK balancer?
I have a B2A25S60p. This uses external relays.

 

[upnorthandpersonal]

However, when I disconnected the BMS and connected it again, it no longer recognice the balancing leads.

How do you turn on the BMS? I would turn it all off, double check the connections, reconnect without the charge controller or loads, and then power on the BMS.

 

[jontm]

That is one other thing that is a bit strange. Before the BMS turned on once i connected it up. Now it requires me to connect the charge detection lead to battery pluss for the BMS to turn on.

I have tried to disconnect the BMS fully and leaving it overnight.

See the attached picture showing how things are connected up. Previously i was not using the K+ lead. It powered up from V+ and V-. Now it requires me to connect it for it to power on.

 

[upnorthandpersonal]

I'm not familiar with that particular BMS - I tend to use their MOSFET based ones. Assuming other operations are the same, the charge voltage (or a voltage over battery voltage in case of the MOSFET BMS, or I'm guessing the charge detection lead in your case) is needed to turn it on.

 

[jontm]

Yes I have read about that, for some reason connecting the battery is enough to power this on and previously that was not even needed.
However powering it up is not the issue. The problem is that it does not do any balancing and thinks balance wires have too high resistance. It does however apparently measure the cell voltages correctly.
I guess this is not something others have experienced before.
Guess I will need to measure out all the leads and check the fuses on the board. It looks like all of the balance leads are internally fused.

 

[jontm]

Have been trying to figure out what might be wrong and I think it must be related to the circuit for measuring the balance current. I have measured all the leads and everything measure correct. I assume that there should be a shunt somewhere, but have not found any yet. Have measured the body diodes on all of the mosfets and have not found anything wrong yet. I can however see that the mosfets that switch in the super capacitors are operating. There are however not any balance current on the balance leads. The balance current is however controlled, so it might be that it is controlled to be zero if the circuit for measuring balance lead resistance has a active alert.

Only thing i have found so far that might indicate something that is wrong is a 5 V rail that measure 4.5 V. Will use a thermal camera on it too see if I can see anything pulling the 5 V down.

 

[MisterSandals]

From your pic, the cycle count is 0 and the charging amps are 0. I think the internal resistance values will (can?) only be calculated by the BMS after a certain amount of charging (possibly discharge, i dunno).

It sounds like you are trying to use your BMS to perform a top balance on your battery. I am a firm believer in performing a top balance with all cells in parallel (or as some others recommend, one at a time).

Many recommend having balance turned on only during charging and starting at 3.4V (and higher).

Sounds like you are highly capable of probing around and analyzing things but i would try some of the above and get a couple cycle counts in to see if that helps.

 

[jontm]

Since the cells are brand new and matched I was planning to skip the top balance. At least til 3.4 V, all the cells where at the same voltage even without balancing.
The balancing was working as intended the first couple of hours.
I am just testing the system now before I pack it in a case and transport it up into the mountains. Don't really have the option to run several charge discharge cycles on it here at home, but I don't think it should matter.
The balancing circuit worked before, now it unfortunately don't.
It does not seems like there are any dead mosfets or diodes, not really surprised about that since all the components seems really overbuilt. The mosfets are massive for just having to conduct 2 A.
In my opinion it has to be something with the balancing current measurement circuit. Since it is locked at zero. The micro probably also uses the balance current to calculate the resistance in the balance leads.

 

[jontm]

I have now measured pretty much everything i can think of. I simply cannot find any components that measure bad. I have also recrimped and measured all the balance leads.
I am starting to think it also could be a software bug or something. Looking at the log, it seems like the clock/timer is very wrong as well. It seems like it has suddenly just jumped something like two days ahead. The logs also have not updated at all since the balancing stopped working.
It also gives me an error message if I try to set the resistance of balance lead 4 manually. All the other ones it let's me adjust.
Is there any way to factory restore the unit ? Or reset the software in any way ?

 

[TorC]

With LFP, voltage is not a reliable measure of SoC. From ~3.1-3.4V is almost all the capacity, and from 3.4-3.65V is tiny. Now that you have all cells near charged, it's time to properly top off the charge. Parallel charge at 3.65V is probably the easiest, not necessarily least work. With the battery disconnected you can sequentially charge each cell to 3.65V without disconnecting the main connections you have.

Once done the current runners should settle into place. Or turn on balancing on charge at more like 3.3V to start with and it will probably reach balance in a few weeks to couple months. Set your SCC to cut off just as the high cells approach 3.65V if you want to speed this up a bit, and monitor it.

 

[jontm]

Yep. I am aware that I am currently at close to 100 % SOC. Not really worried about the top balance, my main issue is that the active balancing circuitry of the BMS seem to be dead and gives me errors for too high balance lead resistance. Unless ofcourse this is a bug with the BMS, that it fails to read balance lead resistance at 100 % SOC.
Anyone with the JK BMS experienced similar issues when near 100 % SOC ?

 

[upnorthandpersonal]

Anyone with the JK BMS experienced similar issues when near 100 % SOC ?

No, never seen it. What version of the app are you running?

 

[jontm]

4.3.2

I did see that the balancing current was steadily decreasing as it was approcing 100 % state of charge, even if the difference between the cells was starting to increase. Disconnected the BMS and balance leads in a hope that it would restart and increase the balance current again, but as mentioned the BMS then gave me this error when l reconnected everything.
Had a thought this morning that maybe the algorithm for measuring the balance lead resistance measured voltage of a cell with and without a short pulse of balance current. At high state of charge the internal resistance of the cell would be higher and the voltage would increase faster when a pulse of current is supplied so maybe that throws the algorithm off and detect a high balancing lead resistance based on this. It is a bit impractical to discharge the battery currently as it is just in a test setup in my living room, but I guess I will need to try to discharge down to 80 % or so to test this theory.
Does it seem like the JK BMS decrease the balance current when the battery is fully charged or does the balance current stay high to the end ? In theory it should increase as it is approching 100 % state of charge to compensate for increasing differences in cell voltage, but mine did the opposite.

 

[jontm]

I think i was able to fix it. At least get rid of the error :D

If anyone have the same issue, this is how i think i was able to fix it.
After slightly dischariging the battery to about 95 % SOC i noticed that i got a short blipp of balance current if I changed the number of cells from 17 and back to 16. I did however get the same error and no balance current after 1 secound. I figured that it might be that it is throwing this error due to not being able to reach the balance current within a short test interval during initialization. I therefore tried to adjust the balance current down to the minimum allowed number, 0.3 A. If I then reinitialised the balancing by changing to 17 cells and back to 16. It was then able to reach the 0.3 A goal within the initialization period and calculated correct balance lead resistances.
I think it might be as I suspected that the BMS is not able to initialize and calculate balance wire resistance at full SOC. The voltage required to push 2 A into the cells during full SOC tricks the BMS into thinking there is a high resistance in the balance leads. My theory is that it calculates the balance lead resistance by comparing the voltage when supplying the set balance current to when not supplying current. If you know the current, you can calculate the balance lead resistance based on this. To do this you however have to assume that the balance current does not increase the measured cell voltage. This assumption is probably correct for low balance currents and/or most SOCs. For 99+% SOC the assumption is probably not correct. Especially at 2 A balance current, due to the near exponential increase in internal resistance of the battery cell when approaching 100 % SOC.

 

[jontm]

Tried charging to 100 % SOC again and it seems like there is a limitation with the design of the active balancing circuitry of the BMS. As you approach 100 % state of charge, it seems like the active balancing circuitry is not able to output the voltage required to push decent balance current. The active balancing circuitry result in very low cell voltage differences when at 80-90 % SOC, but when charging fully, it looks to decrease the balance current with increasing SOC. I have stopped charging now as one cell has reached 3.52 V while most other is at 3.39 V. The maximum cell delta is 130 mV, but still the balance current is not higher than 4 - 9 mA.
Is this normal behavior on all JK active balancing BMS devices ?
Maybe i will do a top balance after all Right now i cannot charge with more than about 54.8 V, any higher will cause a couple of the cells to drift significantly away from the others.

 

[upnorthandpersonal]

Is this normal behavior on all JK active balancing BMS devices ?

I've never seen that behavior.

 

[jontm]

I've never seen that behavior.

Hmm ok. You have full balance current all the way up to full SOC ? My BMS seems only to balance full 2 A when there is a significant difference in cell voltage and only when SOC is lower than 95 % or so.
Maybe there is something wrong with the BMS after all. Considering to purchase one of the JK mosfet BMS units to have in spare, but I probably should really have something closer to 300-400 A for my 10 000 VA Victron Quattro inverter/charger. Having external relays makes it much easier to implement a automated precharge. A bit strange that it seems like none of the mosfet BMS systems come with automated precharge, would be so simple to implement.
Pain in the ass with seperate charge and discharge relays when I have a inverter/charger combo. Probably have to add large power diodes in parallel with the relays and a relay controller that makes sure to turn on discharge relay when charging and opposite.

 

[upnorthandpersonal]

You have full balance current all the way up to full SOC ?

Full SOC at around 3.5V per cell (I never charge to 3.65V), but to be honest I haven't had a big cell balance deviation in a long time...

I have stopped charging now as one cell has reached 3.52 V while most other is at 3.39 V. The maximum cell delta is 130 mV, but still the balance current is not higher than 4 - 9 mA.

Yeah, at those deltas I balance at full balance current.

 

[jontm]

Yeah, then I think it might be something wrong with my BMS. Just tried to reboot my BMS by disconnecting and connecting it again. Got the same issue as before error for high balance lead resistance. Did the trick that worked for me last time and then suddenly the BMS balanced with good balance current for a little bit before falling to almost zero again. Just 4-14 mA spikes. Tried to disconnect it and connect it again, but this time I got the same issue as before and was no longer able to fix the issue with the trick. Probably at too high state of charge again, at least that is what I hope.
Will probably be fine to use until I get something to replace it, but seems like there is something wrong with the balancing circuit on mine.
Fear it probably will be impossible to get them to replace it since the issue is not that easy to explain. 200-300 $ down the drain, but I guess that is what you get with trying to get a good deal by importing it yourself instead of buying from a western seller with warranty and such

 

[upnorthandpersonal]

Where are you located? If you don't need it/can't get warranty or replacement/... I would like to have a look at it. I can pay shipping. If I can fix it, you get it back. If I blow it up, well, so be it.

but I guess that is what you get with trying to get a good deal by importing it yourself instead of buying from a western seller with warranty and such

Well...

This could be interesting

So, instead of just doing the usual and buying some Victron and Batrium equipment and calling it a day, I went shopping on Alibaba and Aliexpress to see if I could build my entire 32x280Ah and 10kW solar panels system for under 10K Euro, keeping in mind the power and charge requirements - so not...

diysolarforum.com