LiFePO4 battery capacity decreasing

[CatNewBee]

You are welcome.

The cells indeed reach full capacity at 3.65V (or 14.6 for the battery) and the BMS balancing usually starts working at 14.2 ... 14.4V. If you stay way off all the time, the cells not get full, and may get out of balance, what leads to even a lower usable capacity.

So charging to full and giving the balancers the chance to do their work from time to time is a good idea if you really insist to use the battery only in the PSOC area.

[john61ct]

Still don't like the "drop-in" concept, but did some research, look like a solid company and product line.

[Epicurean]

CatNewBee, I got it to the point where the Soc reads 100% by setting the absorption cutoff voltage at 14 v. do you think I should periodically go higher, e.g. 14.2 or even 14.6? My fear is permanently damaging the battery by charging at too high a voltage. Under what circumstances c

[john61ct]

The question as to how many lifetime cycles you are sacrificing by going above 3.5Vpc is not a proven settled question.

The key is holding above the point where the BMS is configured to start balancing.

The goal there is to get the cells balanced.

No other useful function is served by going higher, the actual additional Ah pushed into the bank **actually stored as useful energy** is insignificant.

Personally I would prefer a balancing method that did not require going that high in normal daily cycling usage.

Ideally I would want a battery set and usage pattern that did not show any significant imbalances at all, for many months at a time, ideally years.

[mbartosch]

I recently performed another capacity test on our 400 Ah LiFePO4 battery bank (all the details are described here).



The bank was installed in 2015 and used extensively during a one year cruise, but less so in the last 1.5 years.



Charged to 13.8, stopped charging. Discharged to 11.9 V, measured 396 Ah. Not great but very close to the last capacity test which measured something around 400 Ah IIRC.



The bank was balanced initially in 2015. The BMS performs passive balancing (starting at 3.55 V cell voltage) but it never gets the chance as my charging set points are 13.8 for the alternator regulator and inverter/charger and 13.9 V for the solar chargers (no voltage sensing).

As my bank reaches 13.8 V without tripping a HVC (3.6 V cell voltage) I am confident that the cells are within their comfy zone.



My observation and opinion is that balancing is usually not required unless one wants to actually squeeze out the last 10 % of the bank. Using conservative set points for charging and discharging may yield spot-on or slightly less capacity (my measurement tells me that I might get a few more Ah from my bank if I took the time to rebalance), but I don't see the point.



Other users on this forum have different philosophies on charging voltage and balancing but this works well for me.

[CatNewBee]

Originally Posted by mbartosch

I recently performed another capacity test on our 400 Ah LiFePO4 battery bank (all the details are described here).

The bank was installed in 2015 and used extensively during a one year cruise, but less so in the last 1.5 years.

Charged to 13.8, stopped charging. Discharged to 11.9 V, measured 396 Ah. Not great but very close to the last capacity test which measured something around 400 Ah IIRC.

The bank was balanced initially in 2015. The BMS performs passive balancing (starting at 3.55 V cell voltage) but it never gets the chance as my charging set points are 13.8 for the alternator regulator and inverter/charger and 13.9 V for the solar chargers (no voltage sensing).
As my bank reaches 13.8 V without tripping a HVC (3.6 V cell voltage) I am confident that the cells are within their comfy zone.

My observation and opinion is that balancing is usually not required unless one wants to actually squeeze out the last 10 % of the bank. Using conservative set points for charging and discharging may yield spot-on or slightly less capacity (my measurement tells me that I might get a few more Ah from my bank if I took the time to rebalance), but I don't see the point.

Other users on this forum oppose this, pushing charging voltage higher and advocating balancing.

13.8V is about 15% below 100%SOC, so you can perfectly charge the battery and discharge the whole 400Ah afterwards. 13.8V translates to 3.45V cell voltage, but it depends on the voltage sag on the wiring if the cells really are at 3.45V when your controller measures 13.8V.

Your battery is simply not being full before discharging.

[Maine Sail]

Originally Posted by CatNewBee

13.8V is about 15% below 100%SOC, so you can perfectly charge the battery and discharge the whole 400Ah afterwards. 13.8V translates to 3.45V cell voltage, but it depends on the voltage sag on the wiring if the cells really are at 3.45V when your controller measures 13.8V.

Your battery is simply not being full before discharging.

If you stop at 3.45VPC, once you hit 3.45 VPC, the cells will be below 100%. If you holder 3.45VPC, and allow current to taper, the cells will reach 100% SOC even at just 3.45VPC or 13.8V for a 12V pack...

[CatNewBee]

Originally Posted by Maine Sail

If you stop at 3.45VPC, once you hit 3.45 VPC, the cells will be below 100%. If you holder 3.45VPC, and allow current to taper, the cells will reach 100% SOC even at just 3.45VPC or 13.8V for a 12V pack...

You can try it...

keep the battery at 3.45V infinite until no current goes in and than increase to 3.65V, measure the current and watch how it decreases over time, showing, that the cells absorb more energy.

[john61ct]

Originally Posted by CatNewBee

13.8V is about 15% below 100%SOC
…
Your battery is simply not being full before discharging.

If a low enough current rate, say .1C or less, even with zero holding Absorb (charge **to** termination voltage), the additional Ah actually stored above 3.45Vpc can be a lot less than that, even under 5%.

With a high C rate of charge, say over .4C, yes just stopping at 3.45Vpc may be below 90%.

Which of course causes no harm either.

Holding Absorb for a few minutes, endAmps at say .05C will determine a very precise & predictable 100% point for say setting an SoC meter if needed.

[john61ct]

Originally Posted by CatNewBee

You can try it...

keep the battery at 3.45V infinite until no current goes in and than increase to 3.65V, measure the current and watch how it decreases over time, showing, that the cells absorb more energy.

Just because you see current flowing from the charger, does not mean you're actually going to be getting it all out.

And I am very sure going that high is losing some cycle lifetime, and perhaps causing imbalances that then needs to be burned off.

Capping at 3.45Vpc is IMO much better

[Maine Sail]

Originally Posted by CatNewBee

You can try it...

keep the battery at 3.45V infinite until no current goes in and than increase to 3.65V, measure the current and watch how it decreases over time, showing, that the cells absorb more energy.

Not about trying it, I have done it, and measured it, under controlled conditions many times.

[CatNewBee]

Originally Posted by john61ct

Just because you see current flowing from the charger, does not mean you're actually going to be getting it all out.

And I am very sure going that high is losing some cycle lifetime, and perhaps causing imbalances that then needs to be burned off.

Capping at 3.45Vpc is IMO much better

HI,John,

it is not about losing cycles but about capacity and SOC. The cell is full when it cannot absorb and store more charge, and this is at 3.65V. Of course you can stop earlier, BUT if you want to measure the capacity of a cell, you HAVE to charge it to full and discharge it to empty to be able to tell, what capacity she has stored, it is as simple as that.

It is like yout fuel tank, if it has a capacity of 20gal and you refil it for longevity only to 15gal and than claim it has detorated in capacity, while you only can drain 15gal, it is simply not true. fill it up to the top and you can take 20gal off again.

[nebster]

Originally Posted by CatNewBee

You can try it...

keep the battery at 3.45V infinite until no current goes in and than increase to 3.65V, measure the current and watch how it decreases over time, showing, that the cells absorb more energy.

Please show us your data.

Mine show that the cell will charge completely at 3.45V.

You may have an active BMS in circuit, effectively altering the true voltage that the cells see. Or, you are not measuring what you think you are.

[john61ct]

Originally Posted by CatNewBee

it is not about losing cycles but about capacity and SOC.

What is your referent for "it"? Maybe you don't care about that factor, I personally consider it important.

> The cell is full when it cannot absorb and store more charge, and this is at 3.65V.

Flat out wrong, in many ways, energy storage does not act like a tank holding a fixed volume of liquid.

You can get to exactly the same SoC point from a myriad of termination profiles.

The only way to tell how much of your continued attempted input the bank is actually **storing** as energy you can get back **out** is through precise load testing.

Continued input past that point is mostly dissipated as heat, and harms the battery (as in, lost lifetime cycles). As you try to get that last few Ah in, only a small fraction of the flowing current is actually getting stored.

There is **no** reason in daily cycling to try to get all the way to **any** definition of Full, only downside risk.

Yes for SoH capacity / load testing, you want to define a precisely standardized "Full" stop-charge point, and

that can safely be defined a bit higher than the one you'd want to use for daily use since such tests are only occasionally performed.

But something like "3.5Vpc held until current tapers to .03C (3A per 100AH)" is certainly a precise enough definition to use for that purpose.

Of course if you want to use "3.65Vpc and hold that Absorb until .005C" as **your** definition of "SoH testing full" that's fine, but

I'd bet the **actual usable** Ah difference between the two is well under 1% SoC, and

if you're using something like that as your daily cycling stop-charge point, I'd bet you're sacrificing hundreds if not thousands of lifetime cycles, maybe only getting the vendor rated lifetime, which too me would be a shame, given your investment of many thousands of dollars.

[john61ct]

Originally Posted by nebster

You may have an active BMS in circuit, effectively altering the true voltage that the cells see.

Yes there's a BMS there, CNB documented their install quite meticulously in a dedicated thread.

They are taking the vendor charge specs as if optimum, plus the BMS requires a high (too high for me) Absorb voltage before it can even **start** its "live" cell balancing routines.

Which are at a very miniscule balancing current level, so takes a very long time, holding at that high V all the while, and

which would likely not even be necessary if the usage cycle were kept away from the voltage shoulders in the first place.

I consider this a faulty design myself, but it is pretty standard for most BMS, they usually do not allow for any adjustment of setpoints at all.