How to keep Lifepo4 cells balanced without charging to 100%

[87 Dream]

take that to mean you bulk and float while the sun is up, confirm?

Not sure I understand your question but I will try to answer. Most plants I have installed have the Grid available so, depending on the inverter modes, the Bulk and Float will be at those values regardless. Yes by day Solar blending and charging takes place. At night again dependant on inverter mode the battery is at that float. Or alternatively the batteries are being cycled. I work with OEM Voltronics inverters are I'm pretty sure most guys familiar, have heard of the famous premature float bug. Having the bulk and float the same, you eliminate these said issues. Also a float charge is not as important to a LFP pack than to a Lead Acid or AGM.

87

 

[87 Dream]

I fear some people here want to leave their active balancers on all the time and drive their batteries a bit harder than you do.

I have met a few good ppl that do make use of their packs in this way. To each their own. In the method to my madness, I realise that my pack will never reach 100% SoC and will never deplete to zero. I know many are obsessed to get their Shunt or whatever Coulomb measuring device to 100% and then as you say run them hard to maybe 5% or 0%. I only do this at the beginning when I do my first top balance and load test to see capacity. Then another top balance before final 16s assembly into a pack.

87

 

[RCinFLA]

3.45 v for LFP is fully charged if cell charge current drops to near zero level. When you charge to a higher voltage you are doing so to speed up charging time.

When you charge at higher current there will be more kinetic overpotential terminal voltage rise required to support the higher cell charge current.

If you have significant charge current going into cells when it hits 3.45v and remove charger the cell voltage will drop over a few minutes of no cell load or charge current, depending on amount charge current was on cell when you terminated charge and it will end up less than 3.45v, and not be fully charged. So you can fully charge with a maximum charge voltage between 3.45v and 3.65v, it just takes a lot longer time for the charge current to taper off at the lower charge voltage limit.

Balancing below about 3.5v x number of series cells is a problem because most balancers do not balance a cell until it gets above 3.4v. Reason you need to charge above 3.5v is to get some balancing time.

Above 3.4v the cell is at the beginning of reaching full charge and terminal voltage will start to rise quicker so it is easier for BMS to differentiate cells that are at a higher state of charge, so BMS can apply a balance bleed to cell. The trick is when first cell reaches 3.4v it is a race for other cells to catch up on their state of charge before the first highest SOC cell trips BMS cell overvoltage shutdown at >3.65v. Having greater balancing current makes it more likely to win the race. Alternately you could lower the charge rate so ratio of feeble balance dump current to charge current rate is not at such a disadvantage. But this increases charge time.

You should not set continuous float above 3.45v times number of series cells. This pushes cells into overcharge range. This assumes cells are fully balanced which may not always be the case so floating at 3.35v to 3.40v per cell x number of series cells is a safer contiinuous float level. 3.35v no load voltage is still greater than 90% state of charge. This is what most UPS backup units use for LFP batteries.

The other reason to set float lower is many chargers trigger a new bulk/absorb cycle every time the battery falls below float voltage. At 3.4v per cell x number of series cells float setting, just about any moderate inverter load will cause voltage to drop below this level, triggering a new bulk/absorb charge cycle even though battery is near full charge. This increases stress on cells.

Many PV SCC abort absorb cycle based on timer. This is good to give some balancing time but bad if getting hit by repetitive reinitiated bulk/absorb cycles due to inverter sporatic current load on battery. Some chargers use current taper off to abort absorb cycle. This can cause a problem by not allowing much balancing time by BMS.

Best to use a timed absorb cycle to get balancing time and a lower float voltage to reduce repetitive charge bulk/absorb cycles due to random inverter loads dropping battery voltage. It your charger has a third setting that requires battery voltage to be below a certain voltage for a minimum set amount of time before tripping a bulk/absorb cycle all the better.

 

[87 Dream]

3.45 v for LFP is fully charged if cell charge current drops to near zero level. When you charge to a higher voltage you are doing so to speed up charging time.

Exactly what I do in practice. The charger starts it's charge cycle with a Increasing Voltage and CC depends on solar also how much current shines through. When it hits the target 55V it's absorbing the entire time even as the voltage was gradually increasing. The Current also tries to maintain CC and eventually tapers off to zero.

The float is not high at 55V. Because each cell is at 3.43V. remember I don't keep them there for days and months at a time. These batteries are cycled daily. And at best if not cycled when the Inverter goes to Solar/Grid mode blending the charger let's the cells settle.

Great explanation as to good safe practice ????

 

[87 Dream]

My humble apologies to you Will, if in my post #19 is what is described as an affiliate link.

I just attached a link as was requested by another member. Please delete it if it was an infringement.

I am new here and still getting to read through the rules.

My humble apologies ?

 

[Cal]

Balancing at 3.4V makes no sense, so keep the balance voltage at 3.45V, or 3.5V or so. You'll charge to 27.2V or whatever as usual, and any cells that start running will get balanced.

Why doesn't it make sense to balance at 3.4V? 3.4V is well into the knee.

That's what I do and it works great.



As the chart shows, high voltage disconnect is at 3.40V. The DIY BMS determines which cell has lowest voltage (in this case Cell 1) and applies the appropriate balance charge. x-axis is time in minutes.

 

[upnorthandpersonal]

3.4V is well into the knee.

It's well into the knee depending on what your charge current is. With higher charge currents, yes. At lower ones, no (see the graph in my previous post).

 

[Cal]

It's well into the knee depending on what your charge current is. With higher charge currents, yes. At lower ones, no (see the graph in my previous post).

Is that your graph? Is that data consistent with other cell manufacturers? The charging plot I'm showing has 180 AH CALB cells charging at 3A, or 0.017C.

 

[meetyg]

Just a little update:
I ended up setting my charger to 3.5v per cell, JBD BMS balance voltage to 3.45, precision to 10mv, and balance always on mode (won't try to balance if balance voltage not reached).

Seams to work great... I decided on a lower than charge voltage after all, just to give the BMS a little bit of headstart.
Even if the BMS balance can't keep up with the charging (charger is 8A), in due time the cells will be balanced due to the "balance always" mode.

After a charge cycle like this, when testing around 20% discharge, the cells are still well in balance (4mv), but if course that doesn't say much due to the flat voltage curve.

Thanks again for all your help!

 

[upnorthandpersonal]

Is that your graph? Is that data consistent with other cell manufacturers? The charging plot I'm showing has 180 AH CALB cells charging at 3A, or 0.017C.

Not my graph, but the principle remains the same for the chemistry across manufacturers. At high charge currents, you'll hit 3.4V faster than at low currents. Andy from Off Grid Garage noticed this at some point in his videos as well with his EVE cells.

 

[electric]

It's well into the knee depending on what your charge current is. With higher charge currents, yes. At lower ones, no (see the graph in my previous post).

It's the opposite. 3.4V is "in the knee" if your charge current is very very low, making it impractical for most applications, hence the need to go higher in voltage at higher charge rates.

 

[Cal]

Not my graph, but the principle remains the same for the chemistry across manufacturers. At high charge currents, you'll hit 3.4V faster than at low currents. Andy from Off Grid Garage noticed this at some point in his videos as well with his EVE cells.

I'm not doubting the principles of LiFePO4 charging. I was questioning you comment that balancing at 3.40V makes no sense. Regardless of charging current, 3.40V is absolutely in the knee. There's no problem using this voltage to initiate balancing.

 

[upnorthandpersonal]

3.4V can be at 70% state of charge if you charge at a decent rate.

 

[upnorthandpersonal]

It's the opposite. 3.4V is "in the knee" if your charge current is very very low, making it impractical for most applications, hence the need to go higher in voltage at higher charge rates.

Yes, that's what I meant. If you reach 3.4V at a decent charge rate, you might only be at e.g. 70% state of charge.

 

[Smokin]

"How do you keep the cells in balance after a discharge-charge cycle, if you only want to charge up to say 3.4v ?"
You don't. If the cells are new you don't need to worry about balancing. Put them in service. Run um.

 

[John Frum]

If the cells are new you don't need to worry about balancing.

This is far from a universal truth.
AliX mystery meat cells should be top balanced as part of provisioning by default IMO.
Top shelf cells that have gone through and passed the full factory testing procedure are likely good to go.
Also the BMS balance capacity matters.
A Daly has much less balance capacity than a JBD.
Additionally discharging a pack well into the low knee will cause more drift.
As the cycles start adding up, the differential stress on that weak cell will exacerbate its degradation which worsens the situation.

 

[time2roll]

I use an active balancer full time. Mostly charge to 3.45vpc and equalize every three weeks at 3.55vpc.

 

[Ampster]

I use an active balancer full time. Mostly charge to 3.45vpc and equalize every three weeks at 3.55vpc

I also use an active balancer. in my case I have 840 Ahrs of capacity in my pack and my BMS only has a small balancing capacity so I disabled the BMS balancing and use an active balancer with 2 Amps balancing. It is also more efficient because it pulls from the high cell and charges the low cell.
I am familiar with the term "equalize" in the context of Pb batteries. Can you explain the process with LFP? What is the current when you do that?

 

[time2roll]

I also use an active balancer. in my case I have 840 Ahrs of capacity in my pack and my BMS only has a small balancing capacity so I disabled the BMS balancing and use an active balancer with 2 Amps balancing. It is also more efficient because it pulls from the high cell and charges the low cell.
I am familiar with the term "equalize" in the context of Pb batteries. Can you explain the process with LFP? What is the current when you do that?

I set the Morningstar solar controller to equalize at the increased voltage (3.55vpc) for 30 minutes every 3 weeks. Current is not limited. Very low tech, zero science.

I can also push the equalize button at any time to run the voltage up a bit and check the balance.

This is on an RV that is used part time.

 

[Ampster]

I set the Morningstar solar controller to equalize at the increased voltage (3.55vpc) for 30 minutes every 3 weeks. Current is not limited. Very low tech, zero science.

Okay, that sounds like a setting I had on an EV charger. In that case it was a low current setting that gave more time for the balancing function to operate.