‘Normal’ LiFePO4 post-charge settling?

[fafrd]

I’ve built a first 90Ah 8S battery and am planning on building a 280Ah 8S battery soon.

The 90Ah is my first battery build and I’m trying to characterize the battery to figure out if it is behaving normally or not.

I’m charging with a 10A 28.6V LiFePO4 Charger that gets the battery up to 28.75V before shutting down (3.6V/cell avg).

After charging, I’m not applying any load and am disconnecting the BMS balance wires (to eliminate the possibility of any balance currents in case that is impacting the results).

I’m characterizing the post-charge settling I’m seeing and in the 3 hours immediately after charging, the battery is dropping to 28.15V (avg of 3.52V; avg loss of 75mV).

In the following 12 hours, the battery has dropped to 27.82V (avg of 3.48V; avg loss of 80mV).

The cells are still higher than the 3.35 - 3.40V others have reported as the typical ‘settled’ voltage post-charge, so perhaps my cells haven’t reached their final ‘settled’ voltage yet, but my current thoughts are to let the battery settle for another ~48 hours to see if the cells stabilize once they reach 3.35V or keep settling beyond that.

If anyone else has characterized what happens in the hours and days after their cells are charged to 100%, I’s appreciate any insight.

Once I’ve confirmed that my cells are behaving normally as far as retention, I’ll reconnect the BMS sense wires to see if that dramatically changes the results or not.

 

[snoobler]

Normal. Rate doesn't matter as much as CONSISTENCY does.

It's analogous to 12V lead-acid charging. Charge at 14.4V disconnect, and they'll settle to 12.6-12.8V in 24 hours.

 

[fafrd]

Normal. Rate doesn't matter as much as CONSISTENCY does.

It's analogous to 12V lead-acid charging. Charge at 14.4V disconnect, and they'll settle to 12.6-12.8V in 24 hours.

Yeah, I think you’re probably right.

So charge, let settle, and measure.

Repeat 4-5 times and see how similar the results are.

If you were doing a consistency check like that, what period would you give for ‘settling’ - 24h? 48h? More?

 

[snoobler]

Rather than multiple discharges/charges, I would use one charge and a single long waiting period of 7 days.

 

[fafrd]

Rather than multiple discharges/charges, I would use one charge and a single long waiting period of 7 days.

Thanks.

 

[Ampster]

I have noticed that these larger cells take longer to settle. I have also observed that they settle around the same voltage whether I charge them to 3.65 or 3.40 volts. Therefore I am not sure there is any scientific value it watching them settle. The important issue to me is if they are close in voltage when they are in the knee of the charge curve and still close together when they are close to the knee of the discharge curve. What they do in between is not as important.

 

[fafrd]

I have noticed that these larger cells take longer to settle. I have also observed that they settle around the same voltage whether I charge them to 3.65 or 3.40 volts. Therefore I am not sure there is any scientific value it watching them settle. The important issue to me is if they are close in voltage when they are in the knee of the charge curve and still close together when they are close to the knee of the discharge curve. What they do in between is not as important.

The ‘Dishcharge curves’ in the sticky thread indicated a SOC of ~93% at 3.4V, so I agree that if you are only aiming for 80% capacity to extend battery life,, there is no much point to charging above 3.4V.

Seems as though that last 7% is relatively short-lived in any case.

The only real benefit I’ve been able to see from charging to higher voltages is that it’s easier to top-balance above the hockey stick.

 

[Ampster]

Seems as though that last 7% is relatively short-lived in any case.

No it is not short lived. The cell does not self discharge 7% as the surface charge dissipates. That is why is is often called a surface charge. To understand just look at a discharge curve of a freshly charged cell and you will see voltage drops very fast but Ahrs drop slowly. That only reinforces the knowledge that voltage of Lithium cells is a poor indicator of SOC or remaining capacity.
That was my whole point is saying that if you charge to 3.65 or 3.4 eventually the cells will settle to very close to the same voltage. Whether it is 3.350 or 3.346 is below the accuracy of most multimeters.
I am not sure which "Discharge curve you are referring to in the stickie thread. If I were trying to measure capacities of various charging strategies I would want to look at Charge curves.

 

[fafrd]

No it is not short lived. The cell does not self discharge 7% as the surface charge dissipates. That is why is is often called a surface charge. To understand just look at a discharge curve of a freshly charged cell and you will see voltage drops very fast but Ahrs drop slowly. That only reinforces the knowledge that voltage of Lithium cells is a poor indicator of SOC or remaining capacity.
That was my whole point is saying that if you charge to 3.65 or 3.4 eventually the cells will settle to very close to the same voltage. Whether it is 3.350 or 3.346 is below the accuracy of most multimeters.

I certainly see that but if the ‘surface charge’ is not being lost but merely absorbed/settled into the cell, I’d expect the cell to immediately jump back to 3.65V if the charger is connected again.

Instead, my charger riuns at 10A for a good 4-5 minutes before the cell voltages recover to 3.65V.

That’s only about 0.75% and I suppose I should repeat that process over a few days to see if the ‘resettled charging’ starts to decline with repeated cycles...

 

[Ampster]

Instead, my charger riuns at 10A for a good 4-5 minutes before the cell voltages recover to 3.65V.

How many Ahrs is that? Do you notice if it tapers? Maybe it is one or two Ahrs? That is less than 1% as you mentioned, not 7%.

 

[John Frum]

The ‘Dishcharge curves’ in the sticky thread indicated a SOC of ~93% at 3.4V, so I agree that if you are only aiming for 80% capacity to extend battery life,, there is no much point to charging above 3.4V.

Seems as though that last 7% is relatively short-lived in any case.

The only real benefit I’ve been able to see from charging to higher voltages is that it’s easier to top-balance above the hockey stick.

The reason to charge to 3.65 volts per cell to get a full charge quicker.

 

[John Frum]

Anyone think of surface charge like the foam on a pint?

 

[snoobler]

Anyone think of surface charge like the foam on a pint?

Not until now, but that's awesome.

 

[fafrd]

How many Ahrs is that? Do you notice if it tapers? Maybe it is one or two Ahrs? That is less than 1% as you mentioned.

5 min = 0.083h x 10A = 0.83Ah or 0.925% of my 90Ah battery.

The charger runs at full current until it hits 28.7 Volts and then has a low-current voltage mode that continues until it hits 28.75V (usually 1-2 minutes).

I’ll start getting more rigorous about actually timing the 10A charge current phase to see if it decreases with repeated cycles...

 

[Ampster]

Anyone think of surface charge like the foam on a pint?

Great analogy. Makes me thirsty thinking about it.

 

[fafrd]

The reason to charge to 3.65 volts per cell to get a full charge quicker.

Chargers are typically current-limited until they hit target voltage.

My 28.6V LiFePO4 Charger is charging at 10A until the battery reaches ~28.7V (when it then switches to a lower-current voltage mode until voltage reaches 28.75V (or an average of 3.6V per cell).

I think what you guys are saying is that if I turn off the charger after battery voltage passes 27.2V (avg of 3.4V per cell), difference in settled SOC 24 or 48 hours later is unlikely to be noticable...

 

[fafrd]

Great analogy. Makes me thirsty thinking about it.

Well, if is is like foam on a pint, that foam settles into some minuscule volume of liquid beer eventually, meaning next time you decide to top-up the glass, there will be less foam...

Methinks an experiment is called for .

I’d volunteer but I’m still not certain my cells and BMS are behaving as they should.

Someone with a proven battery should try topping up once a day for 10 days or however long it takes to get to below 1/10th on the initial top-up charge.

If that never happens, we know the top-up charge is actually lost.

If it does actually taper close to 0, we know the battery has been charged above 100% (perhaps to as much as 107%) and a capacity test will prove the battery holds more charger after 5-10 charge cycles (w/o any discharge in between) than it does after being charged to 100% just once...

 

[John Frum]

Chargers are typically current-limited until they hit target voltage.

My 28.6V LiFePO4 Charger is charging at 10A until the battery reaches ~28.7V (when it then switches to a lower-current voltage mode until voltage reaches 28.75V (or an average of 3.6V per cell).

I think what you guys are saying is that if I turn off the charger after battery voltage passes 27.2V (avg of 3.4V per cell), difference in settled SOC 24 or 48 hours later is unlikely to be noticable...

I don't think that is exactly what I'm saying.
A lifepo4 batteries' coluombic(sp) efficiency is close enough to 100% as make no odds.
Its the couloumbs(s) that fill up a battery the voltage is just a means to an end.
I'm saying that you will spend more time is cc mode and less time in cv if your charge voltage is a bit higher.

 

[fafrd]

I don't think that is exactly what I'm saying.
A lifepo4 batteries' coluombic(sp) efficiency is close enough to 100% as make no odds.
Its the amps in that fill up a battery the voltage is just a means to an end.
I'm saying that you will spend more time is cc mode and less time in cv if your charge voltage is a bit higher.

Yeah, I agree - the higher the cc-to-cv transition threshold is set, the more (faster) charge you’ll get into the cell before that transition (to lower-current cv mode).

If my charger went to 3.65V before switching to CV mode instead of 3.59V, it would be a bit faster (but not much - from 3.59 to 3.6 in cv mode takes about 2 minutes if I used a similar mode to get all the way to 3.65V, it would take less than 12 minutes versus a 100% charge cycle of over 9 hours or 540 minutes in cc mode).

 

[Ampster]

Methinks an experiment is called for .

A simpler experiment would be charge to 3.4 volts, note the cumulative Ahrs or reset Ahrs and change charger to 3.65 volts and note the Ahrs it took to get there.
The unknown variable in your repetitive top up experiment is the amount of self discharge over time. Your experiment would take considerable time. And each iteration would be incrementally smaller. Kind of like going half the distance to a wall with each step. I am happy with my conclusion and am not interested in subjecting my cells to being maintained at 3.65 volts for 10 days.
I am not sure how you get to 107% full.

 

[snoobler]

At 10A, you're barely over 0.1C, so there's almost no CV charge phase relative to CC.

 

[Ampster]

I’ll start getting more rigorous about actually timing the 10A charge current phase to see if it decreases with repeated cycles...

My suggestion would be to use a meter that records Ahrs or even better Watthours. If you have one that can turn those numbers into a graph that would be even easier to visualize.

 

[fafrd]

A simpler experiment would be charge to 3.4 volts, note the cumulative Ahrs or reset Ahrs and change charger to 3.65 volts and note the Ahrs it took to get there.
The unknown variable in your repetitive top up experiment is the amount of self discharge over time. Your experiment would take considerable time. And each iteration would be incrementally smaller. Kind of like going half the distance to a wall with each step. I am happy with my conclusion and am not interested in subjecting my cells to being maintained at 3.65 volts for 10 days.
[b{I am not sure how you get to 107% full.[/b]

Spec for 100% charge is a single cycle to 3.65V.

According to the sticky thread data, 3.4V corresponds to about 93% SOC.

So if repeated topping-off cycles gets voltage to settle closer to 3.65V rather than 3.4V, that ‘100%/93%’ difference translates to 107.5% of the reference single-cycle full charge...

But I agree with you, this subject is not important enough to risk overcharging / damaging cells...

 

[fafrd]

My suggestion would be to use a meter that records Ahrs or even better Watthours. If you have one that can turn those numbers into a graph that would be even easier to visualize.

I’ve got a discharge meter on the way but have no means to measure Amphours other than an ohmeter and a stopwatch.

 

[Ampster]

So if repeated topping-off cycles gets voltage to settle closer to 3.65V rather than 3.4V, that ‘100%/93%’ difference translates to 107.5% of the reference single-cycle full charge...

I think it would be easier for you to come to terms with the fact that all batteries "settle" including Pb and if there was actually something there to be discovered it would have been figured out over 40 years of Pb usage. I am distinguishing between settling and long term self discharge over time. We do know that Pb self discharges significantly over time and that is why they are often Float charged to maintain them. Lithium batteries have significantly less self discharge.