One of my Winston cells is lame. Will a high voltage charge bring back the capacity?

[Unlikely]

I made a 12V battery from 4x Winston 700Ah cells. Been managed by REC-BMS with Active Balancer for 4 years. Raspberry Pi + Victron.
The cells were new direct from China factory with consecutive serial numbers.
You might call them A Grade these days, this was the way they did cell matching back then.

As new they measured 770Ah capacity each, on a rating of 700Ah, not bad!
With my EBC-A40L on a constant current 40A discharge test, I get 655Ah from 3 cells, but one is 615Ah
I have operated a 3.65V max voltage, 2.80V min voltage, balance from 3.45V upwards on the REC BMS.
I can see that some of the cells have bulged, as I could never manage to fit the supplied aluminium compression plates, being too small.

I am looking at the Tech Spec's for Winstons, and they have charge to 4.0V and low voltage to 2.8V
So that is quite different from my other lithium cells, which are EVE LF280K cells.

Before I try it.....

I'm planning on putting the battery through a full test cycle with the EBC-A40L Battery tester, going to 4.0V and down to 2.8V
Perhaps it will restore some capacity ??
And perhaps parallel up all 4 batteries for a day and night with Constant Current to 4.0V to see if I can get some of those lithium ions to go home.

What do you reckon? ;-0

Or should I just call it a 610Ah battery from now on, and cut my losses??

 

[sunshine_eggo]

HELL NO!!!

There is no mechanism for capacity restoration. It only goes down. More extreme ranges increases wear. Charging is where most damage occurs.

That manual is from 2007.

What has happened consistently as time has progressed with LFP in general is lower charge voltage for increased longevity. The current "best practice" is to charge to 3.45V to capture 98%+ rated capacity with a longer absorption period - to avoid the level of deterioration you're seeing.

As you can see from the charts you supplied, you only get about 3% at a 0.5C charge rate between 3.65V and 4.00V:



Are you charging at 350A? If not, then the benefit will be even less. 40A is WAY less than 0.5C

You're seeing a 15% deterioration on 3 cell and 20% on #4. That's an end-of-life number.

I recommend you stop charging to 3.65V and charge to 3.45V at 0.2C max with a two hour absorption. Float at 3.375V.

 

[sunsurfer]

I have Victron SCCs. They default charge to 3.55 and hold till current drops to near zero(?), not sure if its a timed thing or not. Float is 3.375. I thought that was in the safe zone but maybe not.

 

[sunshine_eggo]

@toms likely has a very meaningful position on this topic. Hopefully, it's mostly in line with my post.

 

[Nobodybusiness]

I made a 12V battery from 4x Winston 700Ah cells. Been managed by REC-BMS with Active Balancer for 4 years. Raspberry Pi + Victron.
The cells were new direct from China factory with consecutive serial numbers.
You might call them A Grade these days, this was the way they did cell matching back then.

As new they measured 770Ah capacity each, on a rating of 700Ah, not bad!
With my EBC-A40L on a constant current 40A discharge test, I get 655Ah from 3 cells, but one is 615Ah
I have operated a 3.65V max voltage, 2.80V min voltage, balance from 3.45V upwards on the REC BMS.
I can see that some of the cells have bulged, as I could never manage to fit the supplied aluminium compression plates, being too small.

I am looking at the Tech Spec's for Winstons, and they have charge to 4.0V and low voltage to 2.8V
So that is quite different from my other lithium cells, which are EVE LF280K cells.

Before I try it.....

I'm planning on putting the battery through a full test cycle with the EBC-A40L Battery tester, going to 4.0V and down to 2.8V
Perhaps it will restore some capacity ??
And perhaps parallel up all 4 batteries for a day and night with Constant Current to 4.0V to see if I can get some of those lithium ions to go home.

What do you reckon? ;-0

Or should I just call it a 610Ah battery from now on, and cut my losses??

Wow. 15-20% degradation. That is a lot for 4 years use.

 

[Unlikely]

HELL NO!!!

There is no mechanism for capacity restoration. It only goes down. More extreme ranges increases wear. Charging is where most damage occurs.

That manual is from 2007.

What has happened consistently as time has progressed with LFP in general is lower charge voltage for increased longevity. The current "best practice" is to charge to 3.45V to capture 98%+ rated capacity with a longer absorption period - to avoid the level of deterioration you're seeing.

As you can see from the charts you supplied, you only get about 3% at a 0.5C charge rate between 3.65V and 4.00V:

...

Are you charging at 350A? If not, then the benefit will be even less. 40A is WAY less than 0.5C

You're seeing a 15% deterioration on 3 cell and 20% on #4. That's an end-of-life number.

I recommend you stop charging to 3.65V and charge to 3.45V at 0.2C max with a two hour absorption. Float at 3.375V.

These batteries have rarely seen voltage above 3.55V;
I set REC-BMS to balance above 3.45V and charge completion at 3.50-3.55V.
I occasionally push up the Max Charge Voltage to 3.65V for a single balance cycles (using solar) and then restore it afterwards.
I keep screen grabs in a Telegram chat, so I can see the situation over the whole period, because I know I am prone to forget.
For a long time, I set the Max Charge Voltage to 3.45V effectively stopping balancing, as I noticed the cells evened out themselves.

So, I reckon that means overcharging is not a big thing here.
And since the Victron Multiplus inverter is turned off my Raspberry Pi also, over VE.Bus,
I have always noticed that the inverter turns off just before the REC would do it's beep-beep for battery low.

Living off-grid, you kind of cuddle your power system.
I know there were 2 occasions when the voltage on Cell 2 went to circa 2.77 V which is when it alarms.
So I reckon that's not it either.
I'm trying to be honest here to diagnose my system, so happy to confess my occasional cockup (!)

So, that brings me back to compression, and the reference in the manual to using a compression jig to squash the batteries back to shape.
But this sounds like some black science to me

???? Did I just open a can of worms, hope not !!!

Some Details
The solar is controlled by DVCC and CANbus, so there is no potential for operator error (thankfully!)
I just use the Bluetooth app to alter the max voltage for the balancing moments.
(For those that don't know, DVCC means my Victron Solar Controller has no target top voltage, it just receives a message on the CANbus to tell it what current to output (through the CCL message.) The BMS reduces charge current as it notices the lowest cell reaching it's individual limit.)
That also means there is no notion of float voltage, as the system manages the battery current to zero and continues to supply power to the house from the solar system. If you turn on the kettle, the battery gets a quick discharge, and then when it turns off the system will work it out.
So, essentially the BMS is completely running the show, based on the lowest / highest cell voltage.

 

[sunshine_eggo]

Slightly familiar with the REC BMS, but i have a Batrium on a Victron system, so I'm familiar with the inner-workings of BMS/DVCC interaction, and I've worked with several systems utilizing Pylontech-type BMS.

I think what you're seeing is the consequences of low current over-charge, i.e., you're never letting the cells fall back after hitting full charge until charge power is removed completely. By charging AND holding 3.50-3.55V/cell, you're stressing them. IMHO, this is why REC and even Batrium are bad for LFP but great for 3.7V chemistry. The Batrium has some features where you can simulate a float voltage, but it's inexact and requires some experimentation. Basically, you can define a global SoC vs. charge current relationship, e.g., you tell it that at and above 98% SoC, you limit charge current to 0.1A thus avoiding holding at 100% and likely never actually hitting 100%.

If you can do something similar with the REC, that's what I would do.

Alternatively, set to balance above 3.40V and charge completion to 3.45V.

 

[Unlikely]

Wow. 15-20% degradation. That is a lot for 4 years use.

Only 500 cycles according to REC BMS. They were on a boat for first few years, before we moved them to the land where they would have regular daily cycles.

 

[sunshine_eggo]

Only 500 cycles according to REC BMS. They were on a boat for first few years, before we moved them to the land where they would have regular daily cycles.

If they were held at peak SoC on shore power, that's likely another contributor to degradation.

 

[Nobodybusiness]

I think what you're seeing is the consequences of low current over-charge, i.e., you're never letting the cells fall back after hitting full charge until charge power is removed completely.

Can you explain why this is bad for the cells?

The cells never really get a chance to fall back to anything.
As soon as the charge is stopped the inverter starts drawing again unless you are using PV to cover the loads I guess.

Also are you saying that if you aren’t charging at .5c you are damaging the cells?

 

[sunshine_eggo]

Can you explain why this is bad for the cells?

The cells never really get a chance to fall back to anything.
As soon as the charge is stopped the inverter starts drawing again unless you are using PV to cover the loads I guess.

But the charge never actually stops.

Also are you saying that if you aren’t charging at .5c you are damaging the cells?

No.

The concept is similar to lead acid. You have to charge to an elevated voltage, but only as long as is needed to fully charge.

If you hold a lead acid battery at 14.4V all day, you will degrade it. Once full, you drop to a float voltage where the cells maintain 100% or very nearly 100% SoC while there is NO current going in them.

If you hold a LFP cell at elevated voltage, it causes analogous degradation. LFP cell voltages "settle" to a lower voltage after charge to a peak voltage. Not allowing this "settling" means you are still sending very small amounts of current and causing damage.

@RCinFLA has probably posted far more technical commentary on why holding LFP cells at elevated voltage is not good for them.

 

[Unlikely]

Slightly familiar with the REC BMS, but i have a Batrium on a Victron system, so I'm familiar with the inner-workings of BMS/DVCC interaction, and I've worked with several systems utilizing Pylontech-type BMS.

I think what you're seeing is the consequences of low current over-charge, i.e., you're never letting the cells fall back after hitting full charge until charge power is removed completely. By charging AND holding 3.50-3.55V/cell, you're stressing them. IMHO, this is why REC and even Batrium are bad for LFP but great for 3.7V chemistry. The Batrium has some features where you can simulate a float voltage, but it's inexact and requires some experimentation. Basically, you can define a global SoC vs. charge current relationship, e.g., you tell it that at and above 98% SoC, you limit charge current to 0.1A thus avoiding holding at 100% and likely never actually hitting 100%.

If you can do something similar with the REC, that's what I would do.

Alternatively, set to balance above 3.40V and charge completion to 3.45V.

Edit 1:

Yes, I can see from this graph that the highest cell was held at 3.50V for the whole afternoon, and yet the lowest cell never balanced up.

The system is totally solar, and in the tropics, so the solar turns off around 5.30pm

Edit 2:

I would have to say that this is a very unusual day for our system as we are living off the Inverter, so if the battery goes to 100% then the breadmaker goes on, or something else is consuming the power. I think I mentioned we "cuddle" the system!!

I'll get a graph for VRM of average usage.

 

[sunshine_eggo]

Looks like the REC is smarter than the Batrium then. Is there a way you can get it to drop to 13.50V? That's 3.375V/cell.

What I find very concerning is that your low voltage cell is happily settling to a very favorable voltage (less than 3.375V), but your high voltage cell is pegged at 3.50V, and your charge current, while near zero is still slightly positively. It should be zero.

 

[Nobodybusiness]

But the charge never actually stops.



No.

The concept is similar to lead acid. You have to charge to an elevated voltage, but only as long as is needed to fully charge.

If you hold a lead acid battery at 14.4V all day, you will degrade it. Once full, you drop to a float voltage where the cells maintain 100% or very nearly 100% SoC while there is NO current going in them.

If you hold a LFP cell at elevated voltage, it causes analogous degradation. LFP cell voltages "settle" to a lower voltage after charge to a peak voltage. Not allowing this "settling" means you are still sending very small amounts of current and causing damage.

@RCinFLA has probably posted far more technical commentary on why holding LFP cells at elevated voltage is not good for them.

Gotcha. Thanks!!

 

[Unlikely]

But the charge never actually stops.



No.

The concept is similar to lead acid. You have to charge to an elevated voltage, but only as long as is needed to fully charge.

If you hold a lead acid battery at 14.4V all day, you will degrade it. Once full, you drop to a float voltage where the cells maintain 100% or very nearly 100% SoC while there is NO current going in them.

If you hold a LFP cell at elevated voltage, it causes analogous degradation. LFP cell voltages "settle" to a lower voltage after charge to a peak voltage. Not allowing this "settling" means you are still sending very small amounts of current and causing damage.

@RCinFLA has probably posted far more technical commentary on why holding LFP cells at elevated voltage is not good for the

View attachment 180763

Looks like the REC is smarter than the Batrium then. Is there a way you can get it to drop to 13.50V? That's 3.375V/cell.

What I find very concerning is that your low voltage cell is happily settling to a very favorable voltage (less than 3.375V), but your high voltage cell is pegged at 3.50V, and your charge current, while near zero is still slightly positively. It should be zero.

These batteries have had their 2.5 years on land, driving a MultiPlus 3k inverter. I have taken delivery of a shiny new GobelPower GP-SR1-PC200 14.3kWh 48V battery and an 8k Quattro so we can make bread any time the day or night. I will run that on a VERY modest setting.

The Winston's will go back to the boat and have a leisurely life at sea, with a gentle DC fridge / freezer system that will take 2-3 days to discharge them. So life is good for the yellow bricks. But kinda annoying for the guy that paid for 5000+ cycles, and already can see 12% drop from nominal on the weakest cell.

But my geeky nature makes me ask...

1. Is there anything I can do to give them the best shot before they go back to sea?
2. Is there anything we can learn from this to save some other offgrid liver?

 

[Unlikely]

..

I think what you're seeing is the consequences of low current over-charge, i.e., you're never letting the cells fall back after hitting full charge until charge power is removed completely. By charging AND holding 3.50-3.55V/cell, you're stressing them.

Whilst on the land, the cells have enjoyed 1900W of Solar, so they are certainly getting a daily workout. 140A from 2x Victron SmartSolar, less house loads. And we recently started using electric hob, etc which we use more often when solar is available. The inverter was a Victron MultiPlus 12/3000, so limited to 2-2.5k on the peaks.

So voltage may have been held too high might occasionally possible, but sometimes the currents are reasonably high in and out (although never near 0.5C)

 

[Unlikely]

Here are some graphs showing longer term usage profile - the peaks and troughs are a bit flattened here, more detailed below.



And the rainy month of September, where we have patches of sunny days and patches of cloudy days ....



This shows better than I explained how overcharge is unlikely to be the issue!!

You'll notice that we turn off the inverter at night when if it reaches 12% SOC so that we can use the coffee machine first thing in the morning. There are certain standards that must be maintained!!

 

[timselectric]

I don't see anything wrong with your settings.
What temperatures have these cells been subjected to?

 

[toms]

Have a look at what the REC is saying the IR of the cells are, this will be a good indicator of the extent of damage.

The Winston cells do not like high temperature (above 35°C) when at high SOC. The cell bulging you see is consistent with electrolyte decomposition which is characteristic of high temperature charging.

Cycling from 4.0V to 2.7V won’t get you any cell capacity back. That cell voltage was a theoretical voltage, in practice the cells should never be held above 3.4V.

There are youtube demonstrations of removing the vent caps and releasing the cell pressure, this won’t reverse damage but will square up the cells and help with installation fixture moving forward.

If the cell internal resistances are significantly different you can use an active balancer to maximise your useable capacity.

 

[Unlikely]

I don't see anything wrong with your settings.
What temperatures have these cells been subjected to?

Steady 25-35 C in the tropics, nothing more than that.