New Apexium 280Ah - Eve Cells - Problem?

[Welly]

Hi All,

I would appreciate some experienced opinions as to whether I have some faulty cells in this Apexium 280Ah, LiFePo4 system with JK Inverter BMS.
Having received the self-assembly kit and unpacked the five boxes, I top-balanced the cells to 3.65 VDC and left them parallel connected for 24 hours, all voltages were within a few hundredths volt and internal resistances were within a few thousandths Ohm. Assembled the battery, torqued and checked, powered on, connecting to a Victron Multiplus II 5000/70 through Lynx PowerIn and monitored through Cerbo GX with 5.4kv solar array, Samil & Solax inverters. Everything was fine and I configured the various settings with no issues (having an existing setup with just the Apexium being new).

However, the first cycle has identified some potential cell problems. The attached image (and XL file) shows a screenshot of the JK PC monitoring software with 4 hours captured (I did not include the earlier Discharge/Charge cycles as they show nothing unusual, following the initial top-balance). Cells 2, 10 and 11 are running-ahead of the other cells and causing a premature "Battery Full" message. The Victron ESS is set to 56VDC max and the JK BMS is set to 3.55 (Bat 100%), 3.6 OVP, 3.549 OVPR.

The screenshot shows one or more of those potentially faulty cells causing the charge shut-down, the cell recovers and quickly breaches the OVP again the cycle then repeating over the next 4 hours until the Sun goes down and/or discharge kicks in. The attached XL file shows the full ~4 hours, the screenshot is just an example over the first 8 minutes or so. All the battery cells, including the three identified, balance-up when left inactive or discharging for a relatively short while (minutes, not hours).

I have spoken to the supplier and they have suggested I take the three cells and "externally" charge/discharge them a few times, then match voltages (approximately) and re-install. I don't think that will help at all as they are effectively doing that within the system but will do it anyway.

I would appreciate opinions as to whether this data indicates faulty cells and whether the supplier suggestion is worthwhile.

Ooops, won't allow a spreadsheet... If requested I'll post a link.

Russ

 

[Quattrohead]

Set your charge voltage to 55v instead and see how they do for a few cycles.
You are in the 1% area so they will start to show deviation, but they also do not seem to be matched cells.

 

[Welly]

Set your charge voltage to 55v instead and see how they do for a few cycles.
You are in the 1% area so they will start to show deviation, but they also do not seem to be matched cells.

Thanks for the fast response, I've set to 55v as you suggested, so now will wait and see.

Do you see any point in removing the three cells and charging/discharging them individually, as per the supplier's suggestion?

 

[AshleyL]

Hmmm......can you post a screenshot of JK Inverter BMS wire resistance values?

 

[Welly]

Hmmm......can you post a screenshot of JK Inverter BMS wire resistance values?

Sure, here they are. The connections are all via a central circuit board but I've continuity tested the connections using a DVM and they all return about the same resistance in spite of what is shown on the JK app. Cell 2 was the first to show issues but it also shows a reasonable resistance value...

 

[AshleyL]

Sure, here they are. The connections are all via a central circuit board but I've continuity tested the connections using a DVM and they all return about the same resistance in spite of what is shown on the JK app. Cell 2 was the first to show issues but it also shows a reasonable resistance value...

Strange indeed........cause none of my three JKBMS has more than 0.080 wire resistance values......
Any infrared thermometer? Measure the terminal or busbar temperature of the "offending" cells when the OVP happened.

 

[Welly]

Strange indeed........cause none of my three JKBMS has more than 0.080 wire resistance values......
Any infrared thermometer? Measure the terminal or busbar temperature of the "offending" cells when the OVP happened.

I only have a DVM connected thermal sensor, so not very accurate. All cell values are within a few degrees, very close to the BMS sensor temps, currently between 21 & 24 degrees C with about 40A currently.

Battery voltage has just reached 55v and gone into absorption mode.

 

[Welly]

Strange indeed........cause none of my three JKBMS has more than 0.080 wire resistance values......
Any infrared thermometer? Measure the terminal or busbar temperature of the "offending" cells when the OVP happened.

Yes, my two other batteries use the normal wire connection and their values are all below 0.070.

 

[Steve_S]

Simply put, you are overdriving the cells past Working Voltage Range. Once you cross 3.450 the cells can & will begin to "run".
Factories test & validate cells within the Working Voltage Range which is 3.000-3.400 which most people will use from 2.800 to 3.425 (gross range).
Many get confused with Allowable Voltage Range (causes no harm/damage) which is 2.500-3.650 but that is NOT for daily working voltage. See Below

Recommended CHARGING PROFILE ! (updated to fix typo)

SETTING	V per cell	12V	24V	48V	notes
Bulk / Absorb: (boost)	3.437	13.74	27.5	55.0	Absorb < 60min Transitions to Float
Equalize:	0				NOT for Lithium Based Batteries
Float	3.425	13.70	27.4	54.8	allows cells to balance internally
Min Volts:	2.650	10.6	21.2	42.4	Never allow to drop too low. Safe Buffer
Max Volts:	3.575	14.3	28.6	57.2
Rebulk Voltage:	3.200	12.8	25.6	51.2
End Amps / Tail-Current:					formula: 100AH X 0.05 = 5A

(*1): End Amps is calculated: (100AH X 0.05 = 5A or 280AH X 0.05 = 14A.
EndAmps = TailCurrent
Coulumbic Efficiency / Battery Status Meter Efficiency for LFP = 99%.
This gets the bank charged to full with high amps (Constant Current Constant Voltage) and then float (Constant Voltage, Variable Current) tops off so the cells are at 3.437Vpc. I
! Do Not forget to adjust for Voltage Offsets between Actual Voltage @ Battery Terminal & at Solar Controller.

BMS With Active Balancing should be set to start Active Balancing at 3.420V or possibly lower to 3.410 so that the cells can balance out at the top and not affect the charging process.

Very Special NOTE: Floating & Saturating to 3.437vpc, accounts for the Voltage Settling post Charge of any kind which actually brings the cells to just below 3.400Vpc.
JKBMS - Inverter Edition Settings (Hardware V. 15 / Firmware 15:09) (most current is 15:17)
NB-1 Choose your Cell Type FIRST !
NB-2 When doing the settings, it's better to start from the BOTTOM and work up. This prevents certain "dependent" values from triggering a fault & not saving the data, which will have to re-entered more than once otherwise. Use ONLY the OK next to the value, [enter] on text pad is not reliable.



Enter number of cells in pack

Enter the AH Rating of the cells.



This is populated by default, it will NOT be correct.
See Calibration Notes below.





Start Balancing (anywhere from 3.40 - 3.45)

1.0 or 2.0 depending on BMS model

< Prevent OVP before it overshoots (can be very fast)

Request RCV Voltage (top of working V)







0% SOC should NEVER be too low.



< IMPORTANT, prevents "bricking" BMS by low voltage.

Request FLOAT Voltage to optimal top balancing & low stress.





<- Dependent on BMS model & Cell Types.





< - Dependent on BMS model & Cell Types




!!! DEFAULT TEMPERATURES = WRONG | CORRECT !!!
< Max Temp LFP can take is 55C
<
<
<
<
<
<
<
< Low Temp Cutoff - Cannot charge below 0C - leave margin
<
<
< Charge start @ 5C, below 5C only use Max 0.02C-rate. CAUTION!







< PRE-CHARGE for 5 Seconds when new connect.

< set by DIP switches







User Data - Limited can be extracted by APP
For internal use - tracking etc...


UART & CAN settings to suit your requirements.











DRY Triggers - User Set for particular use cases.














Calibration Notes
This "subject" has been beaten to death. Here is the SIMPLEST & MOST EFFICIENT way to do this accurately.
1) Use a 2 or 3 decimal accurate MultiMeter OR a Battery Cell Tester like a YR1030/YR1035 which an handle up to 100VDC and will provide an accurate 3-decimal accurate voltage.
2) BEFORE connecting cables to Battery Pack !!! Enable Charge/Discharge & turn on breaker (if you have one on the casing). Next take the Voltage Reading at the Battery Pack Terminals and enter that into the Calibration V Field.
3) NO AMPS required because that is used for correction IF the pack is in operation charging/discharging. There is no need IF static.

Once calibrated, the BMS will correct & compensate internally for the cells. (as it should) the display & APPS will now show the correct "Pack Voltage" and report same to connected devices.

EXTRA NOTE: Also remember that from Solar Controller to Battery Pack(s) there is ALWAYS some voltage loss & increased resistance. Every item, fuse, breaker, shunt, lug add to the Loss & Increased resistance. Additionally, the same effect occurs from Battery Pack to Inverter. Using a quality Multimeter check the voltages at the SCC & Inverter Terminals and then at the Battery Packs. Take NOTE of the losses and correct for that in your settings. Lithium Based Batteries ARE MilliVolt & MilliOhm Sensitive. IE: If your SCC think the battery pack is @ 55.0V while charging but there is 0.400V loss the battery will actually be getting 49.600V. IF your inverter is set to cutoff @ 48.0V and if there is a 0.4V offset it will actually cut off the battery @ 48.4 and not 48.0 as desired.

 

[Steve_S]

Sure, here they are. The connections are all via a central circuit board but I've continuity tested the connections using a DVM and they all return about the same resistance in spite of what is shown on the JK app. Cell 2 was the first to show issues but it also shows a reasonable resistance value...

The Cell Wire resistances are too high IMO. They really should be between 0.055-0.075 when properly done.
Double strip the wire, fold it over and then properly crimp (no solder) and cover "tang" with heat-shrink. Those resistances are a flag.

 

[Welly]

The balance leads are all connected through a central circuit board, with only 10cm tails coming off the board. A DVM continuity check showed much lower values. Also, one of the "problem cells" has low resistance (cell 2). However, I take your point and will report to the supplier.

I'll also implement your suggested settings and see what happens, although I would have thought brand new batteries would be better matched and have less deviation through only one cycle.

 

[AshleyL]

The balance leads are all connected through a central circuit board, with only 10cm tails coming off the board. A DVM continuity check showed much lower values. Also, one of the "problem cells" has low resistance (cell 2). However, I take your point and will report to the supplier.

I'll also implement your suggested settings and see what happens, although I would have thought brand new batteries would be better matched and have less deviation through only one cycle.

Fully shutdown the BMS using the power switch.
Then power it on again and JKBMS will re-test the wire resistance.
Check the new resistance value and compare them.

 

[Steve_S]

The balance leads are all connected through a central circuit board, with only 10cm tails coming off the board. A DVM continuity check showed much lower values. Also, one of the "problem cells" has low resistance (cell 2). However, I take your point and will report to the supplier.

I'll also implement your suggested settings and see what happens, although I would have thought brand new batteries would be better matched and have less deviation through only one cycle.

There are many grades of batteries and worse is how vendors represent them as " ??? ".
Properly Matched & Batched cells are tested through the working range with resistance evaluated at test points throughout. These are then batched together so they have consistent IR through the working range. Most Vendors DO NOT DO THIS !

The majority are Voltage/IR Matched using only the static voltage that they were received with & checked to ensure the IR at that voltage is close to others with a bit of variance allowed. These will often diverge through cycles and can generally be corrected with Active Balancers (at te top) but can suffer issues at lower voltages or above Working range.

B-Cells have usually have "defects/blemishes" but passed otherwise. They usually only "just" make the AH rating
Then you have BULK cells which are not fully validated or failed validation but still within base spec. Typically, the may test out at 275-280AH max. These could also be Brokered Cells, New Old Stock or even recycled cells (gotta watch the vendors).

In General, A+, A Cells with Factory Reports that are Matched & Batched will rarely ever cause issues. B-Cells are most of the time fine if from a reputable Known Good Vendor - Bulk - well that's a dice roll.

MOST IMPORTANTLY !
Stay within LFP Working voltage range - even use 2.850 to 3.450 and your in the "happy" place for LFP. Outside that range and things WILL popup and make you pull hair and generate Blue Air your spouse won't like. A moderate & reasonable profile ALWAYS WINS and that is the information I have provided.

LASTLY & IMPORTANTLY.
The JKBMS' can now upgrade the firmware within. The scenes I showed are for V15:09 of the firmware. The most recent is 15:17 and that can be flashed to the BMS with the software & the JKBMS Cable provided (USB to RS485).
Link to the Firmware & Software.

ESS SOLUTIONS

www.jkbms.net

 

[Welly]

Fully shutdown the BMS using the power switch.
Then power it on again and JKBMS will re-test the wire resistance.
Check the new resistance value and compare them.

Very minor differences... And no difference in the suspect cells.

 

[Welly]

There are many grades of batteries and worse is how vendors represent them as " ??? ".
Properly Matched & Batched cells are tested through the working range with resistance evaluated at test points throughout. These are then batched together so they have consistent IR through the working range. Most Vendors DO NOT DO THIS !

The majority are Voltage/IR Matched using only the static voltage that they were received with & checked to ensure the IR at that voltage is close to others with a bit of variance allowed. These will often diverge through cycles and can generally be corrected with Active Balancers (at te top) but can suffer issues at lower voltages or above Working range.

B-Cells have usually have "defects/blemishes" but passed otherwise. They usually only "just" make the AH rating
Then you have BULK cells which are not fully validated or failed validation but still within base spec. Typically, the may test out at 275-280AH max. These could also be Brokered Cells, New Old Stock or even recycled cells (gotta watch the vendors).

In General, A+, A Cells with Factory Reports that are Matched & Batched will rarely ever cause issues. B-Cells are most of the time fine if from a reputable Known Good Vendor - Bulk - well that's a dice roll.

MOST IMPORTANTLY !
Stay within LFP Working voltage range - even use 2.850 to 3.450 and your in the "happy" place for LFP. Outside that range and things WILL popup and make you pull hair and generate Blue Air your spouse won't like. A moderate & reasonable profile ALWAYS WINS and that is the information I have provided.

LASTLY & IMPORTANTLY.
The JKBMS' can now upgrade the firmware within. The scenes I showed are for V15:09 of the firmware. The most recent is 15:17 and that can be flashed to the BMS with the software & the JKBMS Cable provided (USB to RS485).
Link to the Firmware & Software.

ESS SOLUTIONS

www.jkbms.net

Yes, I have the latest firmware and this battery and the cells are from Docan Power, which has a good reputation and spoken of positivity in these forums.

These are the voltages and IR values of the cells, straight out of the box, after overnight acclimation. Do these look like "grade A matched" in your opinion?

 

[AshleyL]

Very minor differences... And no difference in the suspect cells.

Hmm, resistance aside, time to top balance cell individually to 3.65v instead of connecting them in parallel.

Depend on where you connect the power input during your parallel charging......the one nearest to the power source will get more charges/power than the one at the furthest.

Reason : Parallel sucks on certain configuration?

I got a feeling your Cells 2, 10 and 11 are the closest to your charging source......

 

[Welly]

Hmm, resistance aside, time to top balance cell individually to 3.65v instead of connecting them in parallel.

Depend on where you connect the power input during your parallel charging......the one nearest to the power source will get more charges/power than the one at the furthest.

Reason : Parallel sucks on certain configuration?

I got a feeling your Cells 2, 10 and 11 are the closest to your charging source......

I used two power sources connected to cell 1, cell 4, 12 and 16. The source connected to 4 & 12 was just 5A while 1 & 16 was 20A. The connections were pretty much determined by cable length. When completed all cells were very close measured values for both voltage and IR... But I didn't make notes of those values as there was nothing of concern. I'm fairly sure there were no cells over or under charged...

 

[Quattrohead]

I don't like the various resistance values reported by the BMS.
I personally prefer to solder the balance wires to the eyelets.
Don't pull cells out, the BMS/balancer should be able to do the job with the right connections and settings.

 

[Welly]

I don't like the various resistance values reported by the BMS.
I personally prefer to solder the balance wires to the eyelets.
Don't pull cells out, the BMS/balancer should be able to do the job with the right connections and settings.

Right. I've applied the settings suggested by Steve.S (very close to Andy's on OffGridGarage) and will now wait for a few cycles and see what happens.

Although I did the top balance prior to installing, I've never understood why that's any different to letting a BMS with active balance do the job... Assuming of course the cells have at least close voltage and IR. I have a spare 4A NEEE balancer, which I could use as a last resort if current changes make no difference... And if the supplier refuses to do anything.

 

[Steve_S]

Yes, I have the latest firmware and this battery and the cells are from Docan Power, which has a good reputation and spoken of positivity in these forums.

These are the voltages and IR values of the cells, straight out of the box, after overnight acclimation. Do these look like "grade A matched" in your opinion?

These are not Matched & Batched from what you show BUT they are also at different voltages which messes with IR. If you get them all to one voltage IE: 3.333 and then test IR they should be identical +/- "1".

Docan is "okay" but they do have mixed reviews and there have been shenanigans. I've been here since the start of the forum and seen it all - oi too much actually.

Now I am just providing known good reasonable profiles & settings. I am no longer actively supporting JK after having done their last manuals and stuff but still help out the good folks here.

A misunderstanding here to fix:
Top Balance & Top Charge = 2 different things but can be done together.

Top Charge must be done with New Cells or cells that have been in storage for over 6 months. This fully activates the electrolytes and the chemistry within. using 280AH cells as the example, Charge Cells in SERIES to 3.650V and allow amps taken to drop to 5A or lower (won't go farther than 2.5A on average before "stalling". The moment charge stops, the cells begin to settle towards Max Working Voltage around 3.45# ish and will usually get there within 24 hrs +/-. I tend to do that in batches of 4 cells. Once the "SET" is ready be it 8 or 16 cells tp charged, then Charge the "lot" again to 3.650 and allow the entire lot to settle while still connected. Each cell will have same voltage right across (if not something is wrong). Now instal in proper order and they should be good to go. Top Balancing can be done at anytime AFTER the cells are fully activated and can be done at any voltage provided that it runs till Amps Taken drops low. This is NOT done with cells assembled in a pack.

Active Balancers MOVE energy from High Volt cells to lower volt cells. Passive just burns it off very slowly. With fully matched cells, Active Balancing should happen during at Absorb & Float and not take very long at all (less than an hour). With lesser cells Active Balancing may take long (on average) but should be ok. If Active Balancing can't sort out the top, the odds are a weak cell or two and nothing will fix that.

An FYI: I use a Bench Power Supply that can provide up to 45A @ 3.650V.
Fresh Cells in a set of 4x280AH can take up to 10 hours to reach 3.650V and take <5A.
Cells that have been in use, set identically can take about 2.5 hours to reach 3.650 <5A taken.
It just happens that I finished 3 packs worth of cells doing just that. I've been busy building new packs and upgrading in-use packs and I am matching everything fully to ensure a healthy large bank (7 packs in parallel for my own house).

Hope it helps, Good Luck.

 

[Welly]

Your posts have been a great help, so thanks for that.
Scanning the QR codes shows that the cells were all produced on the same day, example attached. So they should all be "activated", when I top balanced them, left them connected in parallel for 24 hours, then checked the voltage and IR values, they were pretty much as you described, 3.45v within 0.01v and very similar IR (can't remember the actual value but certainly very close).
That all changed after the first cycle.

 

[Welly]

Now the battery is discharging, all cells now balanced within 0.003v.
I guess I'll see what happens over the next few days.

 

[AshleyL]

Now the battery is discharging, all cells now balanced within 0.003v.
I guess I'll see what happens over the next few days.

Well, hate to break it to you. But Lifepo4 maintains a very consistent flat voltage when the real SOC is between 5% to 95%......If you discharge it all the way without recharging again, you will find some cells will drop immediately below 3.0v first

 

[kommando]

The only test is at 3.45V, what is the difference at that point. For my 2 two 16 cell batteries the diff at 3.45V is 0.010V. Currently at 3.365 and the delta is 0.004V but that does not tell me much, when I had a runner and a lagger it showed the same until 3.40V+ when the delta expanded, then had to leave the balancer to sort it out.

 

[Welly]

Well, hate to break it to you. But Lifepo4 maintains a very consistent flat voltage when the real SOC is between 5% to 95%......If you discharge it all the way without recharging again, you will find some cells will drop immediately below 3.0v first

Sorry, I don't understand the relevance of your comment to mine, I didn't mention SoC, just that the cells were now balanced when discharging as opposed to some running away when charging, prior to implementing SteveS settings.