Almost everyone is charging their LFP wrong!
- Thread starter shvm
- Start date
Active balance at 3.3? I thought balance needs to happen at a higher voltage. Perhaps I’m thinking of only passive balance. I don’t know much about active balance.
sunshine_eggo
Happy Breffast!
LFP cells are primarily balanced in the absorption stage with a little (sometimes a lot) in the float stage.
Are you saying your batteries are never in float?
OK. I get it. Yeah, they get to float. I didn't make the connection to float and balance. Cells are balanced when there's enough PV to power the loads, am I on the right track?
Zwy
Solar Wizard
Watch this video again, I think you missed some details.
Steve_S
Offgrid Cabineer, N.E. Ontario, Canada
Active is moving power not burning it off. I tinkered with this through all of the testing phases & my nasty Thrash tests (pushing to the limit edges" and realized pretty quickly that starting Active balancing at a lower cell voltage actually takes the edge off. With the mix of cells I have, Used EV cells, Bulk, B & A grade, I got to see a bigger picture overall.
If I set to 3.400 to start Active, 2 of my packs will have close to 0.100 deviation BUT when Active starts @ 3.30 by the time we get to float they are 0.020-0.025 so it's a lot less work. I have No Idea WHY people keep thinking that Active wastes the energy, it doesn't, it moves it, with minimal loss.
Lt.Dan
Solar Wizard
A few things I have noticed with cell balancing as well, that many people push hard or talk about;
1) Changing the voltage of when the balancing turns on isn't very important. Think about it, even cells that are severely out of balance, still have relatively the same voltage under 3.4v. So if 1 cell is 90% at 3.38v, and 99% at 3.385v (totally random numbers, not actual values, but just an example), and the BMS is set to not balance under 10mV, then the BMS will not start balancing. So you can set balancing to turn on at 3.2v, or whatever, because the delta wont grow more than 10mV unless the deviation is extreme. So in @Steve_S example, setting to 3.3v I think is no problem. If the cells are under 10mV then there is no balancing, and this will be true even above 3.6v! If the cells are under 10mV, then no balancing will occur.
2) Balancing while charging, In my personal opinion, is not good. Like stated in earlier posts of this thread, the voltage is too volatile while charging, so the balancer might be fighting itself and making the problem worse. If you can program to only balance when current is under .01-.05C, then that would be ideal. This would also rely on the battery not being in use for a period of time, which like mentioned by someone in this thread, is sometimes not feasible.
3) This one is more personal preference. I am not an electrical engineer, nor do I have any proof about this statement, its just how I feel. Charging to a higher voltage (3.5-3.55v) is not as detrimental as some may perceive. I bulk/absorb to 3.5v/cell (56v), and then float at 3.4875v/cell (55.8v). This forces my Seplos BMS to register 100% SOC by hitting the 56v overall pack voltage, and then slightly lowers voltage, allowing for balancing to be easy to identify which cell is high/low. This circles back to what I say in a lot of posts, Calendar aging will kill the pack before cycle life! USE YOUR BATTERIES!
1) Changing the voltage of when the balancing turns on isn't very important. Think about it, even cells that are severely out of balance, still have relatively the same voltage under 3.4v. So if 1 cell is 90% at 3.38v, and 99% at 3.385v (totally random numbers, not actual values, but just an example), and the BMS is set to not balance under 10mV, then the BMS will not start balancing. So you can set balancing to turn on at 3.2v, or whatever, because the delta wont grow more than 10mV unless the deviation is extreme. So in @Steve_S example, setting to 3.3v I think is no problem. If the cells are under 10mV then there is no balancing, and this will be true even above 3.6v! If the cells are under 10mV, then no balancing will occur.
2) Balancing while charging, In my personal opinion, is not good. Like stated in earlier posts of this thread, the voltage is too volatile while charging, so the balancer might be fighting itself and making the problem worse. If you can program to only balance when current is under .01-.05C, then that would be ideal. This would also rely on the battery not being in use for a period of time, which like mentioned by someone in this thread, is sometimes not feasible.
3) This one is more personal preference. I am not an electrical engineer, nor do I have any proof about this statement, its just how I feel. Charging to a higher voltage (3.5-3.55v) is not as detrimental as some may perceive. I bulk/absorb to 3.5v/cell (56v), and then float at 3.4875v/cell (55.8v). This forces my Seplos BMS to register 100% SOC by hitting the 56v overall pack voltage, and then slightly lowers voltage, allowing for balancing to be easy to identify which cell is high/low. This circles back to what I say in a lot of posts, Calendar aging will kill the pack before cycle life! USE YOUR BATTERIES!
Warpspeed
Solar Wizard
From what I have read, there are two potential issues with charging to the higher voltages.
The first is that usually, the battery as a whole is charged to some total final voltage, and we cannot be certain that some cells in the series strings are not reaching much higher voltages than we might be expecting.
A "runner" can reach the point where the voltage begins to rise very quickly into the danger zone.
The lower the overall charging voltage is set, the safer the system will be if one cell decides to suddenly "take off" for whatever reason.
There seems to be the general idea that 3.65v is 100% safe and that 3.66v leads to instantaneous and irreversible damage.
Not absolutely certain its exactly like that.
Published data suggests higher sustained voltages definitely result in a very gradual degradation of capacity over time. You can get away with it, but a few years down the track you might be wondering if those expensive batteries that are supposed to last for xxxxx charge/discharge cycles may be nothing like as good as originally advertised.
Maybe the payoff for mercilessly thrashing batteries is worth it, for the very few percent of extra usable capacity. Or maybe not...
That is a judgement call, and entirely up to you.
The first is that usually, the battery as a whole is charged to some total final voltage, and we cannot be certain that some cells in the series strings are not reaching much higher voltages than we might be expecting.
A "runner" can reach the point where the voltage begins to rise very quickly into the danger zone.
The lower the overall charging voltage is set, the safer the system will be if one cell decides to suddenly "take off" for whatever reason.
There seems to be the general idea that 3.65v is 100% safe and that 3.66v leads to instantaneous and irreversible damage.
Not absolutely certain its exactly like that.
Published data suggests higher sustained voltages definitely result in a very gradual degradation of capacity over time. You can get away with it, but a few years down the track you might be wondering if those expensive batteries that are supposed to last for xxxxx charge/discharge cycles may be nothing like as good as originally advertised.
Maybe the payoff for mercilessly thrashing batteries is worth it, for the very few percent of extra usable capacity. Or maybe not...
That is a judgement call, and entirely up to you.
sunshine_eggo
Happy Breffast!
If you get to float, then you get to a point where your batteries aren't loaded as your PV is essentially direct powering your loads.
Partially. It's voltage-based. When a cell voltage rises above the balance threshold, and the cell deviation is greater than that allowed, balancing begins. It will continue to do so until cells drop below the threshold, or if charging stops (some BMS force this, some BMS allow it as an option).
Good point. It could be counter productive if one set their dV to something silly small.
This I don't get. The balancer can't fight the charge unless the balance current is comparable to the charge. With passive, balance current is probably close to 0.0002C on a 280Ah cell.
IIRC, MANY BMS will not balance unless there is a charge current.
I'm not dismissing it. I'm just not getting it. It doesn't matter whether the charge current is 2.8A-14A or 50A-100A, a balance current isn't going to be a strong influence.
I think there are many who would object to a float of 3.4875V/cell. There are grumblings of low current over-charge @ 3.40V. This guy has a lot to say about cells being fully charged at 3.43V:
How long to leave cells connected after top balancing ?
After the top balance, the voltages slowly settle down to around 3.4V ish. I’m wondering if it’s beneficial or not to leave them connected in parallel during this period? I can’t put them together in series for a week anyway. 16 Lishen 272Ah cells to make 16S battery.
diysolarforum.com
And he's an EE who's always talking 5-10 levels above my head in his detailed battery knowledge.
Search results for query: 3.43
diysolarforum.com
Steve_S
Offgrid Cabineer, N.E. Ontario, Canada
A missed Connected Dots.
LFP Nominal cells voltage = 3.200V which is 50% SOC per manufacturer specs for the chemistry.
The tested & validated working voltage range is from 3.000 to 3.400 , it's a simple fact of the chemistry, regardless.
LFP will always settle post charge regardless, it IS the chemistry and it's behaviour...
NOTE that cells DO start to become runners over 3.400 Volts (outside of the certified test range but within the Allowable Voltage Range).
They will also "run" or fall faster below 2.850 Volts where some will drop quicker than others. Again, outside of the "working range".
I have tested packs using 280AH cells and pushed 140A into them (thrash tests) and that does whacky stuff ! But the tested Genuine A Grade cells actually stayed together really well TILL 3.4v was hit and the divergence began, it was quick to hit HVD. BTW @ 0.5C they do warm up significantly and you can see them fattening too. With the different grades all connected in parallel, I found that 0.2 - 0.3C was the Happiest Place for all of the battery packs in general.
Ahhh well, too many people think they know better than the chemical engineers who make these things... that's okay too as it's all on them and it's their liability if they bugger up... sometimes those expensive hard lessons are the only way a few folks will finally get it. Some people just have to burn their hands on the stovetop, not once or twice but thrice before it clicks.
Good Luck
LFP Nominal cells voltage = 3.200V which is 50% SOC per manufacturer specs for the chemistry.
The tested & validated working voltage range is from 3.000 to 3.400 , it's a simple fact of the chemistry, regardless.
LFP will always settle post charge regardless, it IS the chemistry and it's behaviour...
NOTE that cells DO start to become runners over 3.400 Volts (outside of the certified test range but within the Allowable Voltage Range).
They will also "run" or fall faster below 2.850 Volts where some will drop quicker than others. Again, outside of the "working range".
I have tested packs using 280AH cells and pushed 140A into them (thrash tests) and that does whacky stuff ! But the tested Genuine A Grade cells actually stayed together really well TILL 3.4v was hit and the divergence began, it was quick to hit HVD. BTW @ 0.5C they do warm up significantly and you can see them fattening too. With the different grades all connected in parallel, I found that 0.2 - 0.3C was the Happiest Place for all of the battery packs in general.
Ahhh well, too many people think they know better than the chemical engineers who make these things... that's okay too as it's all on them and it's their liability if they bugger up... sometimes those expensive hard lessons are the only way a few folks will finally get it. Some people just have to burn their hands on the stovetop, not once or twice but thrice before it clicks.
Good Luck
42OhmsPA
What's in a title?
sunshine_eggo
Happy Breffast!
Luscious
In my post, I've provided exactly that.
Sorry for the wall of text.
Except in my case, because I'm charge terminating correctly, my cells are showing below 0.01 V balance at rest even after three charge cycles with the balancer completely turned off.
400bird
Solar Wizard
It might as well be written in hieroglyphics.
No, your cells are still balanced because you balanced them recently and now are only charging them to the very bottom of the voltage knee, where the cell voltage just starts to deviate. There's a wide flat spot on the voltage curve where cell voltage can't be used to infer SOC or capacity.
Yes, everyone charging Lifepo4 batteries should terminate when the current (at the target voltage) drops to spec (0.05c or whatever)
Charging at miniscule currents isn't good for the chemistry in the long term.
But you aren't reinventing the wheel or revealing some long lost secrets. Enter the termination current or just use voltage if your charger doesn't support termination current they're just leaving some capacity on the table to stop based on voltage only.
Even ignoring the termination current isn't likely to cause a drastic hit to battery life, but I've never seen a study on that, so I could be wrong.
So do mine, because my 'float' is off (or at least, at 3.375V or below). This gives your cells the time to settle after a full charge.
The issue with LFP is not so much 100% state of charge. It's keeping them at a high voltage. If you charge to 3.5V or thereabout, and then terminate charging after some absorption time and give the cells time to settle to said 3.375V there are no issues. My active balancer has pretty much nothing to do in summer, only after a long winter (where the battery doesn't see 100% ever but gets close to empty at times) it will have something to do once that first sun charges them to 100% again.
The only reason to even go to 3.5V is to charge faster compared to 3.4V. If you have enough solar and can thus keep your batteries in absorption for longer, you can even just go to 3.4V. The active balancer is only really useful at voltages around or above 3.4V because of the flat charge curve. However, the other reason is that the voltage is higher, faster, if you charge with a higher current (see curve below) - so you need to determine what works best depending on your local conditions.
I haven't reinvented anything, but this is still widely observed mostly because most people don't know this is what their cell's datasheet means.
Even when charge termination is not possible, there are still things to be learnt regarding balancing, absorb times and floating.
Absorb times should be no longer than what it takes for the current to taper down to 0.05 C.
There is no point of Balancing before charge termination. Heck, if charged correctly I don't think balancing is required at all for months even. The process is inherently self-balancing. Even for the recommended float value, it doesn't makes sense to recommend a value above 3.4 V/Cell. Anything above 3.38V is just overcharging.
I still have to charge terminate manually (unless I figure out comms using ESP32 for my JK) but my entire outlook on what to expect has been changed entirely. I no longer have to worry about the ideal bulking or float voltage anymore with this information.
Really? I thought resting voltage at 50% was nearer to 3.3V? All my new grade A EVE cells arrived with 3.29V, which I understood was because they should be shipped / stored at about 50% SOC. Never seemed there was much energy left in them when they dropped to 3.2V either.
Lt.Dan
Solar Wizard
What they are shipped at is never exactly 50%, could be anywhere between 30-70% in my experience.Really? I thought resting voltage at 50% was nearer to 3.3V? All my new grade A EVE cells arrived with 3.29V, which I understood was because they should be shipped / stored at about 50% SOC. Never seemed there was much energy left in them when they dropped to 3.2V either.
Good point, but I'm still not convinced 3.2V is 50%?
Steve_S
Offgrid Cabineer, N.E. Ontario, Canada
Working range is 3.000 (0%) to 3.400 (100%) volts. Mid Point is exactly 3.200 volts. (hence why every manufacturer puts that as Nominal Voltage)
How to reinterpret that is a mystery...
Of course if you run between 2.800 - 3.500 then the numbers are something else.
Funny thing about physics & chemistry, there is not a lot of room for random interpretations... it is not politics or other stoopid crap.
How to reinterpret that is a mystery...
Of course if you run between 2.800 - 3.500 then the numbers are something else.
Funny thing about physics & chemistry, there is not a lot of room for random interpretations... it is not politics or other stoopid crap.
