How does BMS cell balancing work?

[fafrd]

I’ve built an 8S 90Ah LiFePO4 battery using a bare-bones China-sourced BMS probably similar to the Daly:



I’ve started running some charge/discharge cycles on my new battery and realize I really don’t understand how balancing works.

During the charge cycle, does ‘Balancing Turn-On 3.55V’ mean that when individual cells reach 3.55V, 76mA are pulled out of them (and dumped to ground) in an effort to ‘hold them back’ / ‘slow them down’ until the other cells catch up?

Is balancing typically active whenever cells are above this ‘Balancing Turn-On’ threshold or only when the battery is actively charging?

I’ve charged my battery to 28.6V (average of 3.575V/cell) and left the charger connected overnight (with no apparent charge flowing based on the color of the indicator light), but the next morning, the ‘high’ cell was just as high as before.

Would like to understand how the balance function on these bare-bones BMSes as well as higher-end BMSes works so I can decide what BMS capability I want for the new 280Ah 8S battery I am planning...

From a quick check, the Chargery BMS has a much more significant balance current of > 1A. Is that balancing active throughout the discharge cycle or only during charging and only at the high-end like my bare-bones BMS?

 

[snoobler]

Passive balancing uses resistors to bleed off current from the high cells.
Active balancing actually transfers charge from high cells to low cells.

Balancing BMS generally have programmable threshold below which balancing does not occur to ensure the BMS doesn't continuously balance the cells when they are inactive at lower states of charge.

76mA is a TINY amount of current compared to even a small battery. That's 0.076A - would take 13 hours to bleed off 1 Ah of capacity.

The pictured BMS will only balance during charge when cell(s) are above 3.55V. You won't get resting voltages at or above 3.55V for very long.

 

[fafrd]

Passive balancing uses resistors to bleed off current from the high cells.
Active balancing actually transfers charge from high cells to low cells.

Does a basic BMS like the Daly use active balancing or passive balancing? I guess for the price I paid (<$40) I just assumed this bare-bones BMS used passive balancing (cheaper).

[quote{
Balancing BMS generally have programmable threshold below which balancing does not occur to ensure the BMS doesn't continuously balance the cells when they are inactive at lower states of charge.
[/quote]

Yes, I can see why a threshold makes sense to conserve charge/energy when it won’t make much difference. Balancing has the biggest impact near the top of the charge curve (and also near the bottom, at least in the case of active balancing that can boost/extend the empty cells).

The bare-bones BMS I have offers no programmability, so a threshold of 3.55V is what I’ve got to live with (though I can see the benefit of a programmable threshold for my next BMS).

76mA is a TINY amount of current compared to even a small battery. That's 0.076A - would take 13 hours to bleed off 1 Ah of capacity.

Yes, a higher balance current like the >1A that the Chargery supports would be nice. But being active even when charging has ceased seems like it would be even more important.

If balancing remained active when my charger had charged the battery to 28.6V and was maintaining that voltage, balancibg even 1% / 1Ah of current overnight would be useful to me, but thus bare-bones BMS appears not to be doing that. Once charge current stops flowing because battery has reached 28.6V, no evidence of any balance charge even after 12+ hours...

The pictured BMS will only balance during charge when cell(s) are above 3.55V. You won't get resting voltages at or above 3.55V for very long.

Yes, understand that the few cells I have bumping up above 3.55V as the battery reaches 28.6V win’t stay there very long.

What I’m confused about is why the cells continue to drift down even though the charger remains connected (supposedly in voltage/float mode since the ‘charging’ light has switched from Red / current flow to Green / no current flow).

I suppose a very low current charger like 1A may be more effective to allow this bare-bones BMS to have any impact on cell balance but the MPPT charge controller I plan to hook this battery up to will charge at even higher current (30A).

For perfectly matched and top-balanced cells, I can see how these bare-bones BMSes suffice, but if there is any drift to deal with at all, a more capable BMS seems like a necessity...

 

[Craig]

Does a basic BMS like the Daly use active balancing or passive balancing? I guess for the price I paid (<$40) I just assumed this bare-bones BMS used passive balancing (cheaper).

[quote{
Balancing BMS generally have programmable threshold below which balancing does not occur to ensure the BMS doesn't continuously balance the cells when they are inactive at lower states of charge.

Yes, I can see why a threshold makes sense to conserve charge/energy when it won’t make much difference. Balancing has the biggest impact near the top of the charge curve (and also near the bottom, at least in the case of active balancing that can boost/extend the empty cells).

The bare-bones BMS I have offers no programmability, so a threshold of 3.55V is what I’ve got to live with (though I can see the benefit of a programmable threshold for my next BMS).


Yes, a higher balance current like the >1A that the Chargery supports would be nice. But being active even when charging has ceased seems like it would be even more important.

If balancing remained active when my charger had charged the battery to 28.6V and was maintaining that voltage, balancibg even 1% / 1Ah of current overnight would be useful to me, but thus bare-bones BMS appears not to be doing that. Once charge current stops flowing because battery has reached 28.6V, no evidence of any balance charge even after 12+ hours...



Yes, understand that the few cells I have bumping up above 3.55V as the battery reaches 28.6V win’t stay there very long.

What I’m confused about is why the cells continue to drift down even though the charger remains connected (supposedly in voltage/float mode since the ‘charging’ light has switched from Red / current flow to Green / no current flow).

I suppose a very low current charger like 1A may be more effective to allow this bare-bones BMS to have any impact on cell balance but the MPPT charge controller I plan to hook this battery up to will charge at even higher current (30A).

For perfectly matched and top-balanced cells, I can see how these bare-bones BMSes suffice, but if there is any drift to deal with at all, a more capable BMS seems like a necessity...
[/QUOTE]
The charger has built in features that stops it from charging for longer than probably necessary. SO what happens is even if it drifts down which all cells will when they are above the knee, it may not start charging immediately. Even the Chargery's 1+ amp balancing wont deal with a really bad cell that is where active balancers work better. But i=once things are in sync it is plenty sufficient.

 

[snoobler]

What I’m confused about is why the cells continue to drift down even though the charger remains connected (supposedly in voltage/float mode since the ‘charging’ light has switched from Red / current flow to Green / no current flow).

LFP cell voltage are "spongy" like a 12V FLA/AGM battery. You charge a lead-acid up to 14.4V-ish. Remove current, and voltage will drop to 12.6-12.9 ish in 24 hours. In float mode, it settles to 13.2V with a small current maintaining it.

LFP is similar. Reduce or remove current, and the cell voltage will decrease. Let LFP cells sit, and they'll drift down to somewhere around 3.3-3.4V, which is where they should be "floated".

 

[fafrd]

LFP cell voltage are "spongy" like a 12V FLA/AGM battery. You charge a lead-acid up to 14.4V-ish. Remove current, and voltage will drop to 12.6-12.9 ish in 24 hours. In float mode, it settles to 13.2V with a small current maintaining it.

LFP is similar. Reduce or remove current, and the cell voltage will decrease. Let LFP cells sit, and they'll drift down to somewhere around 3.3-3.4V, which is where they should be "floated".

With my 28.6V charger, battery charged to 28.73V before the light switched from red to green, then drifted down to 28.14V over the next hour with the charger still connected but green. I’ll check the voltage from the charger next time it goes green - I’d just assumed it was 28.6V but perhaps it ‘floats’ at a lower voltage, as you suggest...

28.14V is an average of over 3.52V/cell, so I’ll let it float overnight next time to see if the cells drift down closer to 3.4V...

 

[snoobler]

Float should be set somewhere around 26.4V

 

[fafrd]

The charger has built in features that stops it from charging for longer than probably necessary. SO what happens is even if it drifts down which all cells will when they are above the knee, it may not start charging immediately. Even the Chargery's 1+ amp balancing wont deal with a really bad cell that is where active balancers work better. But once things are in sync it is plenty sufficient.

That’s kind of what I figured (but appreciate the confirmation).

For my next (280Ah) battery, I may characterize how well the cells are matched before selecting a BMS.

For this first 90Ah 8S battery, with another cycle or two I believe I can get the one low cell balanced up close to the others (by draining using my 50W 2.5ohm resistor), so it should this simple BMS will hopefully be good enough for jazz.

These basic BMSes really just protect your cells from getting damaged but do very little to compensate for unbalanced cells.

This being my first build, I assembled the battery before balancing and I’ve had to screw around with top-balancing after charging using a dump resistor.

This first battery was built before I discovered this board.

For the new 280Ah, I now understand the benefit of top balancing using a 3.65V charger and will probably invest more effort up-front.

 

[fafrd]

Float should be set somewhere around 26.4V

Just checked my charger when disconnected and it reads 28.8V, but that is with 0mA of current (just the multimeter). The BMS has some consumption so I probably need to read at the end of a charge cycle to really measure the float voltage...