How to get the most accurate reading on LiFePo4s?

[Rohan]

Last month, I installed a bank of 4x LiFePo4 batteries in parallel. Basically got the kinks worked out now except for one. How to accurately know their charge?

I have 3 ways to ordinarily check, without having to use a multimeter. A built in analog voltmeter, a digital one that I plug into a 12v outlet, and another voltmeter on my solar charge controller, which should be the most accurate one since it's wired directly to the batteries. That one tends to show slightly lower than the other two, but they are always pretty close to each other.

According to the instruction sheet with the batteries, they are 100% charged at 13.46v, 50% at 13.23v, 30% at 13.21v, 20% at 13.06v, and 10% charged at 12.85v.

But sometimes their charge seems almost random. I can charge it for hours off shore power and it still only reads 13.2v. Then I flip the battery charger off, then on again, and it temporarily jumps to 13.4v before dropping to 13.2v again. A 60% drop! I temporarily turn off everything using the batteries to give me an accurate reading, but usually jumps up only .1v. Sometimes it even shows slightly LOWER which doesn't make any sense.

I realize that LiFePo4 have a very flat discharge curve and so a standard voltmeter might not be up to the task. Is there another consistent way to read them?

[markcouz]

I suggest that you install a Victron BMV712, this will give you several benefits;

1. The BMV will measure the voltage at the battery post, not some other voltage subject to error due to voltage drop.

2. Once you have correctly synchronized the BMV at 100% state if charge you will get a continuous accurate readout of the state of your batteries.

3. If you link the BMV to a Victron Smart MPPT (optional) then the MPPT will use BMV's accurate voltage measurement for charging info, rather than its own incorrect measurement which is subject to error due to volt drop between charger and battery.

This last item is particularly beneficial when using LFP because of their sensitivity to high voltage.

[markcouz]

If you charge your batteries to 13.8v (3.45v per cell) they will be fully charged, if you then stop charging and start discharging the voltage will quickly drop to around 13.4/13,2v

Most of the discharge energy provided will be between 13.2v and 12.4v (3.1v to 3.3v per cell) - measured at the battery post and nowhere else.

[SV Confianza]

Voltage isn't a good measurement for LFP because its voltage curve is so flat. You'll need a battery monitor that measures amps in and amps out.

Until I get to the project of getting a cerbo gx set up to monitor everything, we're just using a victron smart shunt to keep track of our state of charge.

Here's a whole writeup of our lithium setup if interested.

[SOLAR SUPPORT]

While the average Soc of a Lifepo4 battery is between 10-90%, although the voltages of parallel connected batteries are the same, their charge levels can be very different. There are many reasons for this. Although the energy entering and exiting batteries connected in parallel varies little, the differences in the energy level of each battery may increase as it accumulates over months of use.

The ideal situation is when you simultaneously increase the parallel connected batteries to 100% Soc level. However, to achieve this, you must charge each battery until the series connected cells within it reach an equal voltage value of 3.65vpc. When you measure the 4s battery voltage as 3.65x4=14.6V, this soc is 100% true. If the BMS stops charging the battery before it reaches this level, it means that the balance between the cells is disrupted.

My own batteries are 2px4s sealed type. It was not possible to balance the cells within them with standart BMS. I managed to solve the voltage balance problem, which accumulated 0.2vpc in a year, by adding a smart active balancer to each battery.

The economical solution to the problem is possible by choosing the battery with the largest cell. Choosing batteries that contain BMS with smart active balancing feature is very meaningful, especially when you create parallel or serial arrays.

[markcouz]

Originally Posted by SOLAR SUPPORT

While the average Soc of a Lifepo4 battery is between 10-90%, although the voltages of parallel connected batteries are the same, their charge levels can be very different. There are many reasons for this. Although the energy entering and exiting batteries connected in parallel varies little, the differences in the energy level of each battery may increase as it accumulates over months of use.

The ideal situation is when you simultaneously increase the parallel connected batteries to 100% Soc level. However, to achieve this, you must charge each battery until the series connected cells within it reach an equal voltage value of 3.65vpc. When you measure the 4s battery voltage as 3.65x4=14.6V, this soc is 100% true. If the BMS stops charging the battery before it reaches this level, it means that the balance between the cells is disrupted.

My own batteries are 2px4s sealed type. It was not possible to balance the cells within them with standart BMS. I managed to solve the voltage balance problem, which accumulated 0.2vpc in a year, by adding a smart active balancer to each battery.

The economical solution to the problem is possible by choosing the battery with the largest cell. Choosing batteries that contain BMS with smart active balancing feature is very meaningful, especially when you create parallel or serial arrays.

I reach full balance at 13.8v (3.45v per cell) with a similar setup to you but at lower charge voltage, here is how...

I set bulk and float to 13.8v, with the result that my batteries float at that point for several hours each day in the Caribbean allowing plenty time for the active balancer to do its job.

3.45v is just into the knee point where the cell voltage increases rapidly. If the battery approaches this level at an out of balance state then the high cell will quickly begin to rise resulting in the 13.8v being reached and the mppt ending charge.

For example 3.6, 3.45, 3.45, 3.3 = 13.8v

In the above example I have a 0.3v imbalance (pretty radical) with all cells in a safe state and the balancer able to do its job over a few hours. Do this for several days and the balance improves each day to the point where the imbalance is gone.

There is negligible additional energy between 3.45v and 3.65v

[CaptainRivet]

Originally Posted by markcouz

I suggest that you install a Victron BMV712, this will give you several benefits;

1. The BMV will measure the voltage at the battery post, not some other voltage subject to error due to voltage drop.

2. Once you have correctly synchronized the BMV at 100% state if charge you will get a continuous accurate readout of the state of your batteries.

3. If you link the BMV to a Victron Smart MPPT (optional) then the MPPT will use BMV's accurate voltage measurement for charging info, rather than its own incorrect measurement which is subject to error due to volt drop between charger and battery.

This last item is particularly beneficial when using LFP because of their sensitivity to high voltage.

good write up. Plus the BMV712 relay output can be used additionally as last line of defense for your LFP. if you connect its 2nd measuring input and connect to the midbank point (when you have acess to the cells) you can measure the delta of pack 1+2 to 3+4 incase of imbalnce.


the only way to get a precise SOC is having a battery monitor that combines several methods of measuring and coloumb counting must be one of them.
The Victron BMV712 is known for being as good as it gets if you calibrate it right, much better then a lot BMS out there.
forget the voltage readings or alike, the victron MV710 or 712 is the only senseful and economical way to get precise(!) SOC if your BMS is not delivering that.


regarding active balancer is only needed if you charge only from time to time to put your bank back to balance BUT you can only add it if you have access to cells.
The calendar aging of cells has a much bigger effect then what you save when you charge your cells only to 3,45V instead 3,55V which is the commonly accepted best compromise for end of charge parameter between bank balancing, avaliable capacity and life span. but yes only going to 3,45V needs an active balancer as bank has no time to balance as below 3,4V balancing makes no sense and is contraproductive.
Charging to 3,65V depends on the cell manufacturer, with some you should avoid it completly to get max lifespan, other need it all 4 weeks minimum as reference and to not get out of balance as they only have an passive balalncer.

[SV Confianza]

Markcouz alluded to the charging voltage, but I think it's worth asking what the charge profile is for that charger. Some lithium charge profiles set float all the way down to 13.2, so that could be the strange jump when it's flicked on and off. It ramps up, increases the voltage on the bus, then realizes it should be in float, and drops down to 13.2. Just a guess.

Also @Markcouz, I think your float voltage might be a bit high for most people's battery setups. I've read multiple drop-in type LFP manufacturers that don't recommend a float voltage above 13.6.

[UFO]

I have a BMS that has multiple small motherboards that sit on the terminals for each Parallel set (I have 4P4S) which directly monitors each set of cells voltages and also does passive balancing - It is also connected to a highly accurate AMP in/out reader so you know the SOC.


All the info/settings are available Via a Bluetooth App.


I would recommend getting something similar for your system.


https://123electric.eu/products/123smartbms-gen3/

[CaptainRivet]

Originally Posted by UFO

I have a BMS that has multiple small motherboards that sit on the terminals for each Parallel set (I have 4P4S) which directly monitors each set of cells voltages and also does passive balancing - It is also connected to a highly accurate AMP in/out reader so you know the SOC.


All the info/settings are available Via a Bluetooth App.


I would recommend getting something similar for your system.


https://123electric.eu/products/123smartbms-gen3/

For a 4p setup the passive balancer is mostly underdimenioned as 4p requires 4x the balancing current if they go out of balance. Most passive balancers have 100-200mA max where they bleed of the highest cell which is limit if you used 4x100AH.
bigger capacity cells in 4p i highly recommend to additionally get an active balancer.

[CaptainRivet]

Originally Posted by markcouz

I reach full balance at 13.8v (3.45v per cell) with a similar setup to you but at lower charge voltage, here is how...

I set bulk and float to 13.8v, with the result that my batteries float at that point for several hours each day in the Caribbean allowing plenty time for the active balancer to do its job.

3.45v is just into the knee point where the cell voltage increases rapidly. If the battery approaches this level at an out of balance state then the high cell will quickly begin to rise resulting in the 13.8v being reached and the mppt ending charge.

For example 3.6, 3.45, 3.45, 3.3 = 13.8v

In the above example I have a 0.3v imbalance (pretty radical) with all cells in a safe state and the balancer able to do its job over a few hours. Do this for several days and the balance improves each day to the point where the imbalance is gone.

There is negligible additional energy between 3.45v and 3.65v

In your setup i would also set it to 13,8V as your charge sources just see total voltage and 3.65V can be easily reached and even float overcharging a cell. You perfectly explained what happens in case of an imbalance.

If you have the BMS control and manage all your charge sources which knows each pack voltages you can easily go to 3.55V as end of charge as no cell voltage will ever be above 3,6V. So charging is optimized also in the upper knee.

[sailingharry]

Originally Posted by CaptainRivet

For a 4p setup the passive balancer is mostly underdimenioned as 4p requires 4x the balancing current if they go out of balance. Most passive balancers have 100-200mA max where they bleed of the highest cell which is limit if you used 4x100AH.
bigger capacity cells in 4p i highly recommend to additionally get an active balancer.

Do you have real life experience to back that up (in other words, have you used passive balancing and found that the cells get too far out of balance)? Some have indicated that if you routinely hit full charge and get perhaps an hour of day of balancing, passive balancing is sufficient. Then I hear others, such as yourself, indicating that much more balancing than that is needed even with good cells, and recommend active balancing.
It is sometimes very difficult to pick out the difference between theoretical benefits and real life benefits. And as an engineer, I am sometimes too eager to fall into the trap of theoretical perfection.

[CaptainRivet]

Originally Posted by sailingharry

Do you have real life experience to back that up (in other words, have you used passive balancing and found that the cells get too far out of balance)? Some have indicated that if you routinely hit full charge and get perhaps an hour of day of balancing, passive balancing is sufficient. Then I hear others, such as yourself, indicating that much more balancing than that is needed even with good cells, and recommend active balancing.
It is sometimes very difficult to pick out the difference between theoretical benefits and real life benefits. And as an engineer, I am sometimes too eager to fall into the trap of theoretical perfection.

yes i have. Its a simple math game. The upper knee of a 250-300AH cell above 3,4Vis about 6-8AH. 80mV deviation that can easily happen if your bank spends a lot time without being in the balancing range or eg big loads.
80mV is about 1-2AH deviation per cell, if you have 4 its 4x so 4-8AH that 200mA passive balancer should bleed off.


I am running 4p4S Lishen 272AH grade A+ with 1088AH and Electrodacus BMS with a passive 300mA Balancer that also balances before 3,4mV with a special algorythm if deviation is high. Due to the creator of Electrodacus BMS 100mV deviation on 4p4S EVE 280AH takes 5 weeks in normal operation to get back to balance.
Budget 250-300AH cells Lishen, Basen,EVE, Hitech are basically all the same from hardware side, very small differences in Lishen being the once that can handle more C then the other without lifespan reduction because of lowest internal resistance of them but a bit less capacity.
so i added an active 5A Heltec balancer as my BMS can steer it. If not I recommend the newest NEEY active balancer.
my system gets thrown off sometimes by 100-130mV due to i have an oversized bank for my needs (around 250AH average usage a day, so 4 days without 1 amp charge possible) that runs normally between 50 and 80% SOC and not often gets into balance area and 2nd i run 6kw inverter @12 V so 500A is regularly pulled that puts a bit stress on cells.



the standard cell size in drop ins are 100AH and there a passive balancer works well if bank was top balanced.

[UFO]

Originally Posted by CaptainRivet

For a 4p setup the passive balancer is mostly underdimenioned as 4p requires 4x the balancing current if they go out of balance. Most passive balancers have 100-200mA max where they bleed of the highest cell which is limit if you used 4x100AH.
bigger capacity cells in 4p i highly recommend to additionally get an active balancer.

I have not looked at active balancers... Do you attach to the Parallel sets or to all the individual cells - So for me that would either be 4P Sets or 16 Individual cells

[CaptainRivet]

Originally Posted by UFO

I have not looked at active balancers... Do you attach to the Parallel sets or to all the individual cells - So for me that would either be 4P Sets or 16 Individual cells

An active balancer is connected paralell to the voltage sense cables, sonif you have a 4S BMS you also get a 4S balancer for Lifepo4, if 8S an 8S balancer. Number of cells parallel don't matter but the higher they are the higher the blancing current of the balancer should be.
But balancing should only be happening when above 3.4V and not.permanent. so if your BMS can steer an active balancer you can take the cheap heltec capacitor based ones. If your BMS can't you need the neey active balancer, its the only that properly switches on above 3.4V or an adjustable threshold you set and off if all is balanced.