LFP batteries: How balanced is balanced?

[tanglewood]

I've got my battery bank wired up and instrumented with per-cell voltage metering. The bank is 2P16S for 320Ah @ 48V nominal. The batteries are as-delivered from CALB with open circuit voltages of 3.260 volts, give or take 7mV. I ran a brief charge @6A and voltages came up a little bit, and all tracked within 7ms. So far so good.


All this is in prep for top balancing the bank prior to putting it in service. Although the current voltages are good and closely matched, I don't think it's really very indicative of balance at their current moderate state of charge. I think I need to charge until the cells start to climb up the voltage knee and balance them there. I'm thinking somewhere in the 3.5 to 3.6vpc range. That's not what I would normally charge them to, but I think for balancing it makes sense. Any required balancing will be done by hand - my BMS does monitoring and control, but not balancing.



Does this all sound right? Any suggestions or tips?

[john61ct]

It's fine to occasionally go higher than you normally would during normal cycling (for me 3.45Vpc is "my in-usage 100%"),

for maintenance purposes like manual top balancing.

3.60-3.65V will be high enough to get up into the steeper shoulder area of the curve.

Maine Sail actually tapered trailing amps (Absorption CV stage) to 0A in a recent thread, as a benchmark for (what I call) vendor / theoretical 100% SoC and for load testing.

That obviously needs a decent user-adjustable power supply and if all done together in parallel, should be at a pretty high C rate.

Then let them rest as charged for say overnight, then separate them out and verify they're within .01V of each other, or if needed, make it so, can be a PITA with a substandard or worn second-hand set.

Then put into your parallel / serial configuration for production use.

As long as you stay well under that top shoulder area in normal use, and C rates aren't too high, most people can just check cell-level balance once in a while, and find (with a good set of cells bought new) that re-balancing is never or at least very rarely needed.

[tanglewood]

I understand many people balance cells by charging them in parallel, but in my case it's not practical to do so. My bench power supply can only put out 6A. But it can do it at up to 60V. So in series I can charge the bank in about 2 days. Then I can charge individual cells that need to be caught up. If I charge all in parallel it will take around 30 days. Not going to happen. So I figure I'll charge in series until I start to see cells climbing the curve, then bring them into balance.


BTW, what is considered an acceptable voltage difference? 5mv? 10mv? 50mv?

[coopec43]

Is this relevant?

CB-LFP - Cell Balancer Only for 3.2V LiFePO4 cells - EV Power BMS ...
https://evparts.com.au/ev-power-bms/...nce-board.html
This small device allows LFP battery packs to self balance. Small differences in internal resistance between cells causes LFP batteries to become unbalanced over time. ... - Allows LiFePO4 cells to be safely configured in 12V applications.

Clive

[CatNewBee]

Originally Posted by tanglewood

I understand many people balance cells by charging them in parallel, but in my case it's not practical to do so. My bench power supply can only put out 6A. But it can do it at up to 60V. So in series I can charge the bank in about 2 days. Then I can charge individual cells that need to be caught up. If I charge all in parallel it will take around 30 days. Not going to happen. So I figure I'll charge in series until I start to see cells climbing the curve, then bring them into balance.


BTW, what is considered an acceptable voltage difference? 5mv? 10mv? 50mv?

That is OK for the initial charge. Bring them up in a serial configuration / manually to around 3.65V and then parallel them for the end balancing / initialization to 3.8V, not necessary to go all the way to 4V. It will be done quite quickly, because of the steep curve in that area. Then let them rest connected in parallel with no charging for a day to balance out and than create your battery serial configuration. (The voltages above are for the Yellow Winston LiFeYPO4 cells, check datasheet for the CALB cells for your required settings)

[john61ct]

Originally Posted by tanglewood

I understand many people balance cells by charging them in parallel, but in my case it's not practical to do so. My bench power supply can only put out 6A. But it can do it at up to 60V. So in series I can charge the bank in about 2 days.

No problem, get them all to 2.5V or so by hook or crook, then do the higher parts one or a few at a time with what you have.

Will be a **lot** more tedious, depending how you value your time, a decent used power supply on eBay can be $100-200.

.01V or 10mV is what I use.

[john61ct]

Originally Posted by coopec43

https://evparts.com.au/ev-power-bms/...nce-board.html

Link doesn't work.

If this is for "active balancing" while the bank is in use, totally different topic.

> Allows LiFePO4 cells to be safely configured in 12V applications.

Marketing BS

> Small differences in internal resistance between cells causes LFP batteries to become unbalanced over time.

I have found not actually the case, as long as gentle House usage rather than EV, and your 100% SoC in daily use is defined under the top shoulder of the voltage curve. I use 3.45Vpc.

If I need to manually balance once or twice a year, NBD but so far no need.

[rgleason]

Originally Posted by john61ct

Link doesn't work.

If this is for "active balancing" while the bank is in use, totally different topic.

> Allows LiFePO4 cells to be safely configured in 12V applications.

Marketing BS

> Small differences in internal resistance between cells causes LFP batteries to become unbalanced over time.

I have found not actually the case, as long as gentle House usage rather than EV, and your 100% SoC in daily use is defined under the top shoulder of the voltage curve. I use 3.45Vpc.

If I need to manually balance once or twice a year, NBD but so far no need.

https://evparts.com.au/ev-power-bms.html


https://evparts.com.au/ev-power-bms/...ter-units.html


Each to his own John. The verdict is still out, in my opinion.

[coopec43]

Originally Posted by john61ct

Link doesn't work.

.

Google the title then to get the link

CB-LFP - Cell Balancer Only for 3.2V LiFePO4 cells - EV Power BMS ...

Clive

[tanglewood]

Just to confirm/reiterate, I'm looking to do a single, initial balance of a new bank. So not looking for on-going balancing devices. I've decided not to do that, for better or worse.