LiFePO4 Batteries: Discussion Thread for Those Using Them as House Banks

[john61ct]

Just knowing the cells used are from a known-good maker like GBS is IMO worth **a lot**, likely a huge longevity advantage.

Not so much in paying much higher prices, but simply refraining from buying from "assemblers" whose sources are unknown and can vary in quality without customers knowing.

Ability to access individual cell voltages, to bypass / replace / disconnect the BMS are also important.

[JmanC]

Originally Posted by mikedefieslife

Surely since it's a drop-in you don't have access to the cells. You shouldn't need to balance them on first use as they should be well matched and balanced from the factory. All you can do is rely on the internal BMS

The BMS phone app shows cell voltages & a few more cheap brands are having this feature. The very first few full charges I did on this battery and the 4th cell was out of balance so I was wondering if maybe a sign of mismatched or average cells?

[john61ct]

Please post a link to the BMS and especially its documentation if you have that. If not, google for "xiaomi" "jbd" or maybe "ant" together with

4S BMS blutooth

see if you can determine what you have. Dozens of quite differently configured boards all use the same app, and same/similar firmware.

Also please link to the specific model of drop-in and the seller.

Does the app show you explicitly when balancing has started? Is the start-balance voltage documented?

If you have an adjustable PSU or charger, and assuming it is LFP rather than some other LI chemistry,

try holding a low current, say 5A, at a voltage just above the start-balance setpoint.

If that is adjustable, set it to 3.4Vpc, and hold your charge at 3.45V

If you don't know the start-balance setpoint, you may have to go another volt higher, or even 3.6Vpc. Try to figure it out by observing the per-cell voltage behaviour, and dropping V in 0.02 or 0.01 increments, until you see the balancing activity stop, then going just above that threshold.

If the cells are far apart, it may take **days** held at that charge voltage to bring them together, since the balancing (resistance bleeding) current rate is likely well under 1A. If that parameter is adjustable, set it to as high as it goes.

Are there "balance leads" available to give you access to the individual cells internally?

[JmanC]

Originally Posted by john61ct

That is the fatal flaw with drop ins, you must rely on their internal BMS, and lose the ability to balance manually.

Fantastic you can at least see the individual cell voltages to become aware of imbalance problems.

You should find out the "start balance" voltage, and if possible the "bleeding "balancing current rate mA.

I think it's 14.2-14.3V, but manufacturer wasn't clear.

Are these published? If not call and ask.

Since you have a user-adjustable charger, set it to 0.1V above that start balance voltage point, but nowhere near that 14.4/3.6Vpc, going that high is likely **causing** the cells to go out of balance.

I don't think this is the cause, most of the time I only charge to 14V. Only twice gone above that. Not exactly sure how going high 14.4 can cause imbalance issues unless held until taper current is right near 0.

Drop battery's SoC first by say 10A to reduce the risk of overcharging.

I would have thought the opposite. At full charge the amount of current flowing will be small like a few amps, unless it's different when imbalanced.

If you have control over the current rate, drop that down to 500mA (half an amp), or just at the same rate as the balancing current.

My sterling only goes down to 12A. But I thought about running a 11A load to hog all the chargers capacity. But it looks like doing so drops the cell voltages & may be undoing the balancing.

That is now the closest you can get to a "balancing protocol", let it sit there a long time, you should see the cell voltages (very slowly) converge, as the weaker / faster / higher cells trigger a stop-charge, are then bled by resistors at the (very slow) balance rate, then charge resumes, etc.

Inherent in that (common but IMO very stupid) balancing method, is that once cells are greatly out of balance, getting back to proper balance can literally take days.

So with an unbalanced dropin holding 14.4V float for a few days should not be a problem? Will all cell voltages stay at 3.6V & no higher?

Hopefully once that has been done, it will happen during the last few minutes of normal cycle usage charging as it is supposed to in theory.

You can help **prevent** getting too far out of balance by, while normal cycle usage charging,

stay at a healthier lower voltage setpoint, as low as possible just over the BMS' start-balance voltage.

Anything over 13.8V will get you to 100% Full, but likely that is too low for the BMS.

Basically, the profile "top goal" should be to stop as soon as the cells are balanced, keeping on pushing to any higher SoC serves no constructive purpose, only reduces lifespan and causes balance problems.

If you are concerned about a so-called "memory effect" then do go up to the mfg-spec stop-charge spec (14.4V) say once or twice a month, but don't worry about precise balancing at that higher point, just watch amps taper to 0.01C then stop.

Also, be aware that the BMS' "stop-bleed-balance" method may be causing you to think the tapering amps has dropped low enough, when in fact it will jump back up to 0.05C or above, after the bleeding cycle has finished.

Thanks
My response is in blue.

[JmanC]

Originally Posted by CatNewBee

The long float phase usually gives enough time at low current to keep the cells balanced using the internal BMS.

Are you saying when charging to the balancer's activation voltage (14.4) the balancers will still be operating even when the charger steps down into float 13.6V?

[john61ct]

Originally Posted by JmanC

Thanks

My response is in blue.

That quoting style makes it hard for me to address the details.

Also use per-cell voltages rather than pack-level for clarity.

The great imbalance in that case is due to not going high enough - over the start-balance threshold - for long enough.

That is more critical than going too high, but **just above** is better than sitting at 3.6V too long.

The much lower current rate is also optional, voltage is the key, getting familiar with how that specific BMS works.

The sellers often are not familiar with these levels of detail, no English technical staff available.

[john61ct]

Originally Posted by JmanC

Are you saying when charging to the balancer's activation voltage (14.4) the balancers will still be operating even when the charger steps down into float 13.6V?

No, unless you can adjust the setpoint down that low (under 3.4Vpc).

There may be a separate "stop balance" setpoint, but I've never seen that

[CatNewBee]

Originally Posted by JmanC

Are you saying when charging to the balancer's activation voltage (14.4) the balancers will still be operating even when the charger steps down into float 13.6V?

What I am saying is, the balancer kicks in at a specified cell voltage. To make it clear, i will now give you an extreme example.

Assume there is a charge current of 5A, balancer burn 1A max at 3.65V, while balancing current at 3.55V will be 0.
Assume cell voltages 3.0V, 3.0V, 3.0V and 3.55V, with start balance at 3.55V

Balancing would start for the cell at 12.55V already,
Cell 1,2,3 and 4 will be charged with 5A,

Cell 4 will reach soon 3.6V resulting the balancer to burn 0.5A, so cell 1,2,3 will be charged with 5A, cell 4 will be charged with 4.5A and 0.5A will be burned in heat. Pack Voltage may be then 3.05 x 3 + 3.6V = 12.75V, balancing running only on cell 4 at 50%.

We continue charging and voltage rises by 0.05V on all cells (just as example), cell 1,2,3 have all 3.1V, cell 4 has 3.65V.
Balancer is running at 100%, pack voltage is then 12.95V, cell 1,2,3 accepting 5A, cell 4 accepting 4A, 1A burned by balancer. Voltage increases further about another 0.2V per cell, cell 1,2,3 have now 3.3V, cell 4 has 3.85V, pack is 13.75V, balancer still burning 1A, cell 1,2,3 accepting 5A, cell 4 charging with about 3.5..4A, balancer burning 1.5A...1A, overheating, over voltage protection disconnecting the pack.

What I am trying to show is, the pack voltage is irrelevant for the balancer starting to do something. It is the individual cell voltage that triggers the balancing.

The bigger the imbalance is, the sooner a balancer will kick in.

[john61ct]

> So with an unbalanced dropin holding 14.4V float for a few days should not be a problem?

Well I would never balance using a BMS that worked that way myself, but unless it's user replaceable you may not have a choice.

That's why I'm saying find the lowest possible voltage high enough to keep the balancing going.

Once you know what you're doing, don't let them go out of balance so much, and under 15min each time may be enough to maintain.

> Will all cell voltages stay at 3.6V & no higher?

Whatever the top V setpoint that BMS is set to, that's when it halts bulk charging and starts resistance-bleeding the high cells.

Then restarts charging, loop repeat.

But there are dozens of balancing algorithms, really just guessing here based on the fact that's how 99% of the cheap units work.

[john61ct]

> I don't think this is the cause, most of the time I only charge to 14V.

Excellent, if only your BMS could be set to start balancing at 3.45Vpc that would be ideal.

But if it starts at 3.55V, you must go over that.


> Not exactly sure how going high 14.4 can cause imbalance issues unless held until taper current is right near 0.

No, pushing up the V curve past the shoulder greatly exacerbates imbalance problems.

If you have other ways of getting cells balanced IMO 3.45Vpc is ideal for normal usage cycling.

Owners of high-quality banks can do that for years, balancing never required at all, just monitor and if needed balance manually.

That standard "stop charge / resistance bleed" balance method is IMO completely moronic, contributes to the problem that the device is supposed to be solving.

But cheap to implement, and creates a lot of higher turnover revenue for the industry.

[Delfin]

Could you kindly provide your data source for your declarative statement that pushing voltage past the "shoulders" exacerbates imbalance, since no LFP bank supplied by any manufacturer, or assembled by any person with half a brain lacks circuitry to balance at voltages at, or above said shoulders.

[john61ct]

Originally Posted by Delfin

pushing voltage past the "shoulders" exacerbates imbalance

This is obvious, from easy observation, just go back and read this thread, read Maine Sail's article and hundreds of posts from people with no axe to grind

> since no LFP bank supplied by any manufacturer, or assembled by any person with half a brain lacks circuitry to balance at voltages at, or above said shoulders

Yes top balancing needs to be done at high voltages.

If you are forced to use a stupid BMS that only balances at damagingly high voltages, you should not let it do your balancing. Just don't allow voltage to climb above the start-balance setpoint and you'll be OK.

Balancing can be done in many alternative ways, **if** it ever needs to be done at all.

But with quality well-matched cells in good condition, it does not need to be done frequently, or at all, for many years at a time

**if** you avoid going into the shoulders in your normal cycling.

Unfortunately, some people get hoodwinked into buying drop-ins that do not allow for proper care, and, along with the fact that the quality of the cells themselves is questionable, that will inevitably result in relatively poor longevity.

Yes, then people jump in, say who cares, don't worry about these persnickety details, long as we get X years we'll just replace them.

Which is fine I guess, but not what I'm about.

How about this:

Obviously it is unmatched capacity, differing ESR levels that are the primary cause of cell imbalance, especially as they wear from cycling and age.

Even if you do not believe that pushing end-charge up to high voltage/SoC exacerbates cell imbalances,

surely you can see that they do not **manifest** as a practical problem if you usually stay away from the shoulders?

And do you really believe that pushing past the shoulders reduces lifetime cycles?

That in itself means worse imbalance problems arriving sooner. . .

[donradcliffe]

Having spent some time recently with a drop-in 200 amp-hr Life, I would agree with John. My experience was with a depleted 12v 4S 1P drop in, which came to me well onto the bottom shoulder with cell voltages ranging from 2.92 to 3.05. I started by bottom balancing the lower voltage cells, then switched to charging the whole pack.

At this point the light came on. On the shoulders, the voltage vs SOC is much steeper. so imbalanced cells are more apparent. Once you get into the normal operating range, the voltage vs SOC is very flat. As long as you stay inside that range, you should see very little difference between cell voltages unless you have a really weak cell.

[evm1024]

Imbalances can only occur when the intercalation/de-intercalation efficiency is different between cells in a battery pack. The (de-)intercalation efficiency rate is very difficult to derive but can be assumed to be near the same for cell of "like" parameters (age , condition etc).

Moving a string of cells up into the "knees" does not by its'self cause the cells to have differing efficiencies in "moderation". There are other factors such as the actual current which brings into play "log-jam" effects.

The quickest way to change the charge efficiency is to have a BMS that starts shunting....

The point I am making is that arm-chair experts often underestimate the complexity of the systems that they are talking about. And, that their lack of experience has prevented them from finding the limits of their understanding. They are right often enough to think that they are right all the time. Thus, they Know that they know.

[CatNewBee]

It is pretty simple.
Cells drift and get out of balance with each cycle.

John's idea, derived from readind Mainesail's excerpts is, stay at low soc, dont go to the shouldes, better even don't use the battery and cells will stay balanced.


It is true that you better see the imbalances at the shoulders, and also efficiently can fight them there what you cannot in the medium SOC, because the difference of lets say 5..10% soc between 2 cells result at 50% SOC in few millivolt, while at the shoulder you see a 100mV and higher difference.

If you want to get them equal, you must go that high or that low. If you do balancing at low SOC chances are you make the situation worse, because that small voltage differences may result from contact resistance, cell temperature or simply measurement errors.

Also the whole concept of BMS balancer with low currents is based on the idea, that you charge your battery to full, and this most of the cycles and not cycle at low soc. This allows small adjustments to one cell and approximation to the others over a few cycles, also keeps the cell in balance regardless of small drifts on each cycle. Avoiding the shoulders - effectively meaning not charging the battery to full - only causes larger drifts, that cannot be handled by the BMS in one or two cycles, but need either manual correction or many cycles with low currents, making the battery unusuable during this time.

There is a reason why the technical data points out, what the proper set points should be and how to charge the battery.