Problems with keeping 302Ah in balance

[T3TRIS]

Hello,

I built a 302Ah battery pack using 302Ah CATL grade A cells from EELbattery (if memory serves, the seller was recommended by Will Prowse at the time I purchased). The cells are connected using flexible busbars (same seller), they are dynamically compressed using die springs and 1/8 rubber in between cells.
The cells are monitored by an Overkill Solar 120A 4S BMS.
The cells were top balanced in parallel to 3.65V (multiple times at this point), under compression, using homemade 4AWG cables with hydraulically crimped lugs until less that 0.05A was going into the cells. They have been left connected together for hours before putting them in series.
They stayed close together until I did a capacity test using a 400W electronic load tester. The test was pulling 30A until 11V, 20A until 10.5V and 10A until 9.5V (which it'd never reach of course). The idea was to minimize voltage drop in the wiring (2ft of 6AWG) since this unit doesn't have voltage sense wires.
The BMS is set to stop discharge at 10V total or 2.5V per cell.
Cell # 2 reached 2.5V before the others, though I didn't see the voltage difference.

Then I set my 60A power supply to 14.2V (I was already afraid of cells going out of balance) and let it charge at 50A.
Cell #1 was getting close to 3.65V before the others (cell #3 was about 80mv behind, cells #2 and #4 were about 110mV behind). Cell balancing was happening but the Overkill Solar BMS can't balance at a high amperage.
I let it all sit for a few minutes and turned the power supply to 14.5V (which is what the charger Redarc Manager30 is set to) and turned the power supply back on.
This time, Cells #2 and #4 were ahead and about to hit 3.65V before the others. By this point there were all within 40mV, above 3.61V but still taking 6A.
Cell #2 triggered the BMS 3.65V charge safety parameter and charging stopped.
I let the cells settle, 40min later, cells #1 and #3 were around 3.38V and cells #2 and #4 were around 3.44V.
3hrs later, #1was at 3.348V, #2 was at 3.361V, #3 at 3.341V, #4 was at 3.354V (cell delta was 20mV).
I tried charging again and cell #2 and #4 once again took the lead and triggered the 3.65V limit. Settling looked the same.

This is basically what seems to happen in the rig, with the charger triggering the cell BMS safety parameter before the charger stops charging.

While it might seem like I'm just charging and charging, this stuff happened since the initial top-balance.

I just reconnected them in parallel and got them to 3.65V until they only took 0.05A again and left them connected in parallel. 10hrs later, they are now at 3.592V, still in parallel. I'm hoping this will be a solid top balance but all of this seems over-kill (I didn't have to do that for previous LiFePO4 batteries I built).

I'm going to reconnect the BMS and set the Redarc load disconnect to 11.4V (or theoretically 2.85V on average per cell). I'm hoping this will prevent cells going out of balance, but it also means getting farther away from the advertised capacity.

Additional info:
Cell impedance measurements were (in order): 24mΩ, 23mΩ, 22mΩ, 23mΩ.
I capacity tested each cell and they were all within 1.5% of one another (my 15A tester got close to 299Ah results, my 400W tester got closer to 292Ah results).
When I last top-balanced them to 3.65V, all the cells settled to 3.325V the following day, I thought that was low (charts seem to indicated that 3.325V is around 70% or 80% SOC for LiFePO4). The seller is saying this is normal.

Am I over thinking this? This battery is for a customer and I want to make sure it outlasts their rig!

 

[Alkaline]

CATL 302? Grade B....

Anyway you are probably experiencing some type of self discharge issue, I had this happen all the time with my xuba 280DK cells, the only way I was able to resolve this was to use an active balancer and that kept them in balance. The top balance on these does not stay long...

 

[T3TRIS]

CATL 302? Grade B....

Anyway you are probably experiencing some type of self discharge issue, I had this happen all the time with my xuba 280DK cells, the only way I was able to resolve this was to use an active balancer and that kept them in balance. The top balance on these does not stay long...

These were advertised as grade A from a shop that was recommended by Will… I can’t see the recommendation on his site anymore, so maybe they got worse…

 

[sunshine_eggo]

These were advertised as grade A from a shop that was recommended by Will… I can’t see the recommendation on his site anymore, so maybe they got worse…

"Grade A" is s made up term. There is "EV Grade" and "not EV Grade". "B" grade is something Eve marks on the QR code to indicate "not EV grade"

 

[SupraSPL]

Cells resting at 3.325 are 100% soc for all my batteries I'd consider that normal.

 

[T3TRIS]

This battery cell grading system is so annoying… I’m really tired of this kind of market manipulations (which apparently just happens anywhere).
I know not to trust Grade A from AliExpress, however I was hoping that since it was recommended by Will Prowse, I could trust the darn things (though to be fair, everyone here is kinda speculating but it doesn't look good). This is particularly annoying since it's a customer's project that has been delayed, partly because I can't get these cells to perform right! Under the impression that they were recommended and recognized "grade A," I didn't question the cells too much, instead looking at every other aspect of my build.
Here's a screen shot of Will Prowse's website from 10/28/22 (which is after I bought these cells). These website archive libraries are cool!

I guess I'm going to have to install an active balancer, as recommended by Alkaline (thanks) and be OK with 3.325V 100% SOC (thanks SupraSPL), and reimburse some of the customer's money and do the apology-figure-out-an-outcome dance...

 

[octal_ip]

Cell #1 was getting close to 3.65V before the others (cell #3 was about 80mv behind, cells #2 and #4 were about 110mV behind). Cell balancing was happening but the Overkill Solar BMS can't balance at a high amperage.
I let it all sit for a few minutes and turned the power supply to 14.5V (which is what the charger Redarc Manager30 is set to) and turned the power supply back on.
This time, Cells #2 and #4 were ahead and about to hit 3.65V before the others. By this point there were all within 40mV, above 3.61V but still taking 6A.
Cell #2 triggered the BMS 3.65V charge safety parameter and charging stopped.
I let the cells settle, 40min later, cells #1 and #3 were around 3.38V and cells #2 and #4 were around 3.44V.
3hrs later, #1was at 3.348V, #2 was at 3.361V, #3 at 3.341V, #4 was at 3.354V (cell delta was 20mV).
I tried charging again and cell #2 and #4 once again took the lead and triggered the 3.65V limit. Settling looked the same.

This sounds normal - once your cells are above 3.5v, they're already well above 95% SOC, possibly even higher if only 6A is trickling into them. Regardless of the cell "grade", they'll never be perfectly matched. At some point you'll have one cell reach 3.65v sooner than the others, and as long as the rest are relatively close, your pack can be considered charged.

The fact that you're seeing cell 1 take the lead, but cells 2 and 4 are then ahead on a subsequent charge suggests they're all very close in actual SOC, and they're all very full. I'm guessing during a short rest, cell 1 was drained by the balancer just enough to put it behind 2 and 4, which is a good sign as the Overkill balancer wouldn't have been able to bleed off much energy during this time to make up a difference.

You've tested them to be within 1.5% capacity of each other, so everything you're describing seems correct. They're all just teetering on 100% SOC, and they're very close to each other.

The real test will be how this pack behaves during repeated cycles. If you have cells that are hitting 3.65v while others are below 3.4v, that would be a reason to worry.

When I last top-balanced them to 3.65V, all the cells settled to 3.325V the following day, I thought that was low (charts seem to indicated that 3.325V is around 70% or 80% SOC for LiFePO4). The seller is saying this is normal.

Don't rely on voltage or charts to estimate SOC at rest or during discharge, they'll never be accurate for LiFePO4. The only way to know is to drain the cells and measure the energy they produce. Maybe your next step should be to capacity test then entire pack to see what it can deliver within the high/low voltage tolerances.

 

[T3TRIS]

This sounds normal - once your cells are above 3.5v, they're already well above 95% SOC, possibly even higher if only 6A is trickling into them. Regardless of the cell "grade", they'll never be perfectly matched. At some point you'll have one cell reach 3.65v sooner than the others, and as long as the rest are relatively close, your pack can be considered charged.

The fact that you're seeing cell 1 take the lead, but cells 2 and 4 are then ahead on a subsequent charge suggests they're all very close in actual SOC, and they're all very full. I'm guessing during a short rest, cell 1 was drained by the balancer just enough to put it behind 2 and 4, which is a good sign as the Overkill balancer wouldn't have been able to bleed off much energy during this time to make up a difference.

You've tested them to be within 1.5% capacity of each other, so everything you're describing seems correct. They're all just teetering on 100% SOC, and they're very close to each other.

The real test will be how this pack behaves during repeated cycles. If you have cells that are hitting 3.65v while others are below 3.4v, that would be a reason to worry.



Don't rely on voltage or charts to estimate SOC at rest or during discharge, they'll never be accurate for LiFePO4. The only way to know is to drain the cells and measure the energy they produce. Maybe your next step should be to capacity test then entire pack to see what it can deliver within the high/low voltage tolerances.

Thanks for the detailed response. Yeah, everything seems kind of OK enough, but I'm not happy with the way these cells interact once they are in the system. The Redarc Manager30 charges at 30A and 14.5V and has some charging profiles. The few times I've tried to keep an eye on it, it seems that the BMS triggers the over voltage safety before the charger finishes its thing. I was under the impression that the charger should be the limiting factor for charging, and the BMS is the last resort.
Same with discharge. The charger controls the load disconnect and it's normally set at 10.5V. I'd definitely get some cells hitting 2.5V before the total voltage drops under 10.5V. Fortunately, I can adjust that in the charger settings and I'll set it to 11.4V (with an alarm at 11.8V). I can't adjust the charging profile though, except for just saying it's for lithium.

I did a couple capacity tests on the whole pack. The first one came around 290Ah if memory serves (it was a couple months ago). The last one, a week ago was closer to 280Ah though the Electronic load might have had its cutoff voltage set too high. It was pulling 25A with no voltage sense wires. So when the tester saw 10V, the battery was still at 10.5V.

Anyhow, that's what I'm struggling with. Everything seems relatively OK when diagnosing, but in practice, it goes out of whack...

Regarding EEL battery... I've bought 4 sets of cells from them now.
First set last summer-ish: 304Ah CATL "grade A" cells that capacity tested to 319Ah, great!
Second set shortly after: 304Ah CATL "grade A" cells that capacity tested to 306Ah, pretty good!
Third set in fall: 280Ah EVE "grade A" cells that I haven't tested yet.
Fourth set in late fall: 302Ah CATL "grade A" cells that seem to test under 299Ah, going downhill?

 

[octal_ip]

Thanks for the detailed response. Yeah, everything seems kind of OK enough, but I'm not happy with the way these cells interact once they are in the system. The Redarc Manager30 charges at 30A and 14.5V and has some charging profiles. The few times I've tried to keep an eye on it, it seems that the BMS triggers the over voltage safety before the charger finishes its thing. I was under the impression that the charger should be the limiting factor for charging, and the BMS is the last resort.
Same with discharge. The charger controls the load disconnect and it's normally set at 10.5V. I'd definitely get some cells hitting 2.5V before the total voltage drops under 10.5V. Fortunately, I can adjust that in the charger settings and I'll set it to 11.4V (with an alarm at 11.8V). I can't adjust the charging profile though, except for just saying it's for lithium.

Remembering that your cells aren't ideal and perfectly identical, it's not surprising that charging to 14.5V is triggering the BMS, as that would mean each cell is at an average of 3.625v. You already know that you'll see up to 110mV deviation as your cells finish charging, so you'll need to account for this in your Redarc charging voltage.
I'd suggest lowering the voltage to accommodate your worse case scenario across all cells (about 3.54v per cell, or 14.16v). In practice you'll still get virtually the same usable capacity from the cells, but without the BMS cutoff as you now have a tolerance for the expected cell voltage differences at a high SOC.


I have a 16S pack, and I only bulk charge it to 55.1v or a bit over 3.44v average voltage per cell. It then goes into absorption at a slightly lower voltage to ensure all the cells are topped up and balanced. With this gentle charging profile I get well over 90% capacity from my cells while never getting close to any cutoff. It'll also help extend the cell life as I'm not pushing the voltages to the bleeding edge.

If you're interested in learning more about the SOC and voltage relationship while charging, Andy at the Off-Grid Garage YouTube channel has done some excellent experiments that showed charging to 3.65v really isn't necessary.

 

[T3TRIS]

Thanks! I’ll have to look into that channel.
I agree with your methodology… I just am not sure I can change the charging voltage for this charger. Maybe I’ll look at their other charging profiles to see if one has a lower voltage but same absorption voltage.

 

[time2roll]

I think you are trying too hard and expecting perfection. If the top cell is 3.610 and the others are within 40mV the situation is great. 3.570-3.610 range at the top is perfectly normal and fully charged for all practical purposes. Reduce the in-service charging voltage to 14.2-14.4 range and move forward with the next project.

If the cells don't hit 2.500 volts all together that is perfectly normal also. One will always have slightly less capacity than the others. If the voltage is dropping below 12.000 volts (3.000 Vpc) the battery needs to be charged more frequently or add more capacity in parallel.

 

[Alkaline]

This battery cell grading system is so annoying… I’m really tired of this kind of market manipulations (which apparently just happens anywhere).
I know not to trust Grade A from AliExpress, however I was hoping that since it was recommended by Will Prowse, I could trust the darn things (though to be fair, everyone here is kinda speculating but it doesn't look good). This is particularly annoying since it's a customer's project that has been delayed, partly because I can't get these cells to perform right! Under the impression that they were recommended and recognized "grade A," I didn't question the cells too much, instead looking at every other aspect of my build.
Here's a screen shot of Will Prowse's website from 10/28/22 (which is after I bought these cells). These website archive libraries are cool!

I guess I'm going to have to install an active balancer, as recommended by Alkaline (thanks) and be OK with 3.325V 100% SOC (thanks SupraSPL), and reimburse some of the customer's money and do the apology-figure-out-an-outcome dance...

Yeah put an active balancer and just enjoy life, some people will say ohh it will do this and that... try it for yourself, worked for me and even my grade B were basically always balanced.

 

[The Good Package]

You've fallen into the trap of battery OCD. Active balancers are a must, but I recall that they should only be used periodically and not always. Off Grid Garage just did a video on this with his QSO rig. There's more than one way to automate the periodic balacing. You should only charge to 90-95% SOC to prolong the life of the battery.

This is why I don't build my own batteries anymore - it's a major time sink and you get married to the idea of squeezing out every drop of performance.

 

[T3TRIS]

Thanks everyone, I think this is what I needed to be told a few times before it sinks in... I'll still look at the active balancing. I guess I have to make sure it only happens periodically (thanks for the tip The Good Package).

It's too bad, it's been fun to build these batteries but yeah, it's become less worth it from a making-end-meet perspective, especially when you see what can be bought on Amazon. I just can't get over getting fooled with that whole grade A from what I thought was a reputable shop and paying a premium for this (and asking a premium from my customer). Just doesn't sit right.

Anyhow, my problems kinda still remain: I want the low and high voltage threshold to happen before the BMS gets triggered.
Low voltage: relatively easy, I can just set a higher voltage for the load disconnect.

High voltage: tougher since I can't directly control the Redarc Manager30 charging voltages, however I might look at setting the charger to a different battery chemistry.
Looking at the instructions, it seems like setting the battery type to GEL or AGM will do the trick.
Lithium charges at 14.5V (3.625V per cell) and floats at 13.6V (3.4V per cell)
GEL/AGM charges at 14.4V (3.6V per cell) and floats at 13.5V (3.375V per cell)

I hope the 14.4V would still be under the point at which the cells trigger the 3.65V max and I assume the slightly lower float shouldn't be a problem.
I also assume that the temperatures will be OK. I'm OK with the BMS taking care of temperature limits (which should be pretty rare). The Manager30 has its own temperature sensor that is ON the battery terminal, but it can't override the internal BMS.

As far as charging profiles, the Manager30 offers 2 modes: touring (3-stages) and storage (8-stages).
Touring (per the manual):
Touring mode is designed for use when ‘on the road’. Touring mode offers a 3-stage charging profile consisting of Boost, Absorption and Float stages. In Touring mode, the house battery is monitored to detect only a limited number of faults such as short circuit, over current and over voltage. This allows The Manager30 to operate correctly even when loads are connected to the house battery. This mode will always produce an output (unless a fault condition is detected) and will cycle through the three stages as required to maintain the house battery.
NOTICE: Touring mode will achieve its best charge level if a Storage mode charge has been recently performed.

Storage (per the manual):
Storage mode is designed to charge the house battery to its optimal level and maintain that level while your travel trailer or RV is in storage. This mode requires all loads to be switched off or disconnected from the house battery before charging. It uses a 8-stage* charging profile consisting of Desulphation*, Soft Start, Boost, Absorption, Battery Test, Equalize**, Float and Maintenance stages. Storage mode is designed to detect a wide range of battery fault conditions.
Unlike Touring mode, Storage mode does not continue to cycle through charge profiles. This means that when the charging process is completed, The Manager30 in Storage mode will always remain in either Float or Maintenance stages. Float stage provides the house battery with a ‘trickle’ charge whenever the house battery voltage drops below a predetermined threshold, to ensure that the battery stays charged. Maintenance stage turns The Manager30 output off, but continues to monitor the house battery and will revert to Float stage when necessary.
Valid charging sources for The Manager30 while in Storage mode include AC Shore Power and Solar.
NOTE: If The Manager30 is set to Storage mode and the vehicle is started The Manager30 will automatically switch to Touring mode once it senses an increase in input voltage from the alternator.
NOTICE: When using the charger is in Storage mode, make sure that there is a valid charging source and that all loads are disconnected from the house battery. Failure to do so may cause the house battery to be under charged, give false readings on the State of Charge indicator and possibly cause damage to any loads connected.


If ever the customer puts the Manager30 in storage mode while in GEL/AGM type of battery, it looks like the Manager30 will do Desulphation and Equalize...

Any red flag here?

 

[The Good Package]

Do you have access to a voltage triggered relay? In the past I've used the Victron BMV-712's relay to open/close another 40a relay tied to external charger's output based on voltage. This would allow you to stop charging at a specified voltage and the external charger just thinks you disconnected the battery.

 

[T3TRIS]

Interesting idea... I've wanted to use voltage triggered relays to enable battery heating in the past (when the relay detects that the alternator is going).
I think I'm going to see if the system works on GEL/AGM setting first, along with an active balancer. That seems like the simplest method right now.

 

[time2roll]

The touring vs storage would appear to be meaningless for lithium. Unless the Maintenance at the end is actually a lower float or off. Storage would seem superior for what I would choose. Especially good if there is a trigger to restart at some lower voltage point if the battery actually becomes discharged.

Lithium does not necessarily like to be floated at 100%. Best to let it cycle.

I like the voltages of the AGM-GEL options although float could be slightly lower. Unfortunately the temperature compensation may push things a bit high in cold conditions. A little reduction in the heat will not be an issue.

 

[time2roll]

To avoid all the tedium on the next build.... Assemble the cells direct as delivered with the BMS and put in service. Use a charging system that is more flexible on top voltage to limit the total from the start. If the cells are not well balanced just add an active balancer and let it rip. Just reduce the total voltage to avoid over volt error. As they balance out the voltage can be adjusted.

Skip the top balance and capacity tests. Just assemble and put the battery in service.

I have built three with this method and the active balancer runs full time. No issues the first two years so far.

 

[JaVid]

When you did a capacity test with all four batteries connected, what did you come up with? I didn't see the numbers if you posted them.

Everything looks fine to me. I would lower charging voltage if it was possible to something like 13.7bulk and 13.6float. I also think a balancer would work fine as well. I think you are overthinking it.

 

[timselectric]

Just my 2 cents.
The cells aren't too far off.
Especially that far into the high knee.
Don't charge so high. And if you add an active balancer (a BMS with active balancing would be better). Make sure it's only active during charging. And above 3.4v .

 

[octal_ip]

If you're interested in having the JBD/Overkill BMS activate and disable an external active balancer at certain cell voltages, have a look at the work we've been doing over on this thread: https://diysolarforum.com/threads/active-balancer-make-it-smart-💡.40192/post-731612

 

[T3TRIS]

Alright, new update. I've installed the 5A balancer (so easy!!!) and I think it works, at least the little diode is on.

I put a 25A load on the batter for 20min or so to try to shave off some of that cell saturation, then re-connected the charger. The charger chugged along at 30A, as expected, until cell #3 reached 3.65V. The rest of the cells were at #1: 3.606V, #2: 3.533V (119mV behind) and #4: 3.567V
Total voltage of the pack was 14.35V. The charger is set to AGM, so my assumption is that it will spit out however many amps the battery asks for until they kinda reach some saturation.

If the balancer is working, it would seem that 5A is not enough to even out the cells for the amps coming from the charger to taper down, as it's supposed to do in the charging profile. That's mostly what bothers me... This battery is much more likely to be charged at 100% than to be discharged to 0%. If that 100% happens often, I'd rather have the charger stop the charge than rely on the BMS's mosfet (wouldn't want one of those to fail).
Maybe I'll look into the voltage sense relay that someone else suggested. That'd stop the charge, but again, ideally I'd like the charger to do that as it was designed!
I started a 25A capacity test last night too and got around 281Ah before the BMS triggered (set the value to 2.65V).
Now I'm charging the battery again at 30A. That discharge and the following charge are done with the active 5A balancer. We'll see how that goes.

I did tip the battery briefly and heard quite a bit of gurgling! Got me panicked for a minute until I read a few posts saying that it's mostly normal... Still weird though.

Several questions then:
1. Is Redarc wrong in making their Lithium charge voltage 14.5V?
2. Is there an easy way to check if this works properly?
3 bonus question: Am I still chasing perfection? I'm just not really happy with the end of 100% charge being controlled by the BMS rather than the charger.

 

[time2roll]

5 amp balance would need 500 mV cell differential. At those number closer to 800mA would be the balance current and it will take some time. You are charging 40+ times faster than the balancer can compensate.

Give it a couple weeks at least. Then check in with some new numbers. Or get out that balancing power supply back out and charge the full battery at 1 amp for the last 5% overnight.

 

[houseofancients]

Alright, new update. I've installed the 5A balancer (so easy!!!) and I think it works, at least the little diode is on.
View attachment 141001
I put a 25A load on the batter for 20min or so to try to shave off some of that cell saturation, then re-connected the charger. The charger chugged along at 30A, as expected, until cell #3 reached 3.65V. The rest of the cells were at #1: 3.606V, #2: 3.533V (119mV behind) and #4: 3.567V
Total voltage of the pack was 14.35V. The charger is set to AGM, so my assumption is that it will spit out however many amps the battery asks for until they kinda reach some saturation.
View attachment 141006
If the balancer is working, it would seem that 5A is not enough to even out the cells for the amps coming from the charger to taper down, as it's supposed to do in the charging profile. That's mostly what bothers me... This battery is much more likely to be charged at 100% than to be discharged to 0%. If that 100% happens often, I'd rather have the charger stop the charge than rely on the BMS's mosfet (wouldn't want one of those to fail).
Maybe I'll look into the voltage sense relay that someone else suggested. That'd stop the charge, but again, ideally I'd like the charger to do that as it was designed!
I started a 25A capacity test last night too and got around 281Ah before the BMS triggered (set the value to 2.65V).
Now I'm charging the battery again at 30A. That discharge and the following charge are done with the active 5A balancer. We'll see how that goes.

I did tip the battery briefly and heard quite a bit of gurgling! Got me panicked for a minute until I read a few posts saying that it's mostly normal... Still weird though.

Several questions then:
1. Is Redarc wrong in making their Lithium charge voltage 14.5V?
2. Is there an easy way to check if this works properly?
3 bonus question: Am I still chasing perfection? I'm just not really happy with the end of 100% charge being controlled by the BMS rather than the charger.

lower your charging amps to about 5-10a and give it about 1 hour.
furthermore remember not to leave the balancer connected.
manually start the balancer when cells hit 3.45 high, disconnect it after done.
you could solder an on-off switch to it to make that a bit easier

 

[sunshine_eggo]

3.533-3.650V is a pretty insignificant imbalance.

Lower absorption to 14.2V and increase time to 1 hour.