LFP memory effects thread

[newhaul]

biggest issue on the thread is all of the round and round technical bs . I have only been using my Lfp for a year and charge daily with solar . I have not noticed any reduction in capacity but that could be due to my settings on my solar controller . I run at a stop voltage of 13.8. (3.45 at cell level) as of last I still had 112 ah of capacity in my 100 ah bank. When installed it showed 113ah but that was with a stop charge voltage at 14v (3.5 at cell level. )
just wish someone that is experiencing capacity issues would write a simple post
Like this

My bank was x ah when installed charged to x vpc. After x usage it now tests at x ah so trying charge at x vpc to try to recover lost capacity.

Results of higher vpc charge gained x ah back or did nothing to ah capacity.



Does that make sense?

[OceanSeaSpray]

Originally Posted by rgleason

What would happen if you did a controlled charge routine on each cell without the BMS I wonder? I suppose this is a lot of work.

That would all depend on what the controlled charge routine would look like. At the moment, I see no benefit in breaking up the pack or taking the BMS out.
My bank appears quite happy to keep charging beyond 76% SOC, it is just that the process is so slow due to the low acceptance rate that so far I haven't managed to actually do it, but I will.

Charging the cells individually wouldn't change that because they are so resistive above 76% SOC and I would lose the data logging capability of the BMS. I frequently see the state of cell balance near the bottom and it looks reasonably even, within what one would expect for a top-balanced pack. Balance is not the issue, not after charging only 76% of nominal capacity beginning with cells near 3.0V.

I am also interested in the problem from a BMS point of view and this prompts dealing with it without reconfiguring anything, in an algorithmic way.

[AGM]

Originally Posted by newhaul

biggest issue on the thread is all of the round and round technical bs . I have only been using my Lfp for a year and charge daily with solar . I have not noticed any reduction in capacity but that could be due to my settings on my solar controller . I run at a stop voltage of 13.8. (3.45 at cell level) as of last I still had 112 ah of capacity in my 100 ah bank. When installed it showed 113ah but that was with a stop charge voltage at 14v (3.5 at cell level. )
just wish someone that is experiencing capacity issues would write a simple post
Like this

My bank was x ah when installed charged to x vpc. After x usage it now tests at x ah so trying charge at x vpc to try to recover lost capacity.

Results of higher vpc charge gained x ah back or did nothing to ah capacity.

Does that make sense?

Yes, very good idea.
I've added a few extra clues to mine.

Age of bank: 6 years
Advertised capacity: 3.6Ah
Capacity test done when new: 3.4Ah
Most recent capacity test result:
6 months ago, 2Ah
Approximate number of mixed cycles: 30
Kept at 40% SOC for most of the time.

Charging termination:
mostly around 3.6V at various termination currents

Float charging: none

Trying to regain lost capacity:
Charging to 3.7Vpc, termination 0.02C (70mA), disconnect charger.

Voc after resting for 10 hours: 3.33Vpc
Capacity recovered: none
Irreversible capacity fade: 42%

Age of bank: 5 years
Advertised capacity: 40Ah
Capacity test done when new: 47Ah
Most recent capacity test result:
6 months ago, 38Ah
Approximate number of mixed cycles: 50
Kept at 40% SOC for most of the time.

Charging termination:
mostly around 3.6V at various termination currents

Float charging: none

Trying to regain lost capacity:
Charging to 3.7Vpc, termination 0.02C (0.8A), disconnect charger.

Voc after resting for 10 hours: 3.52Vpc
Capacity recovered: none
Irreversible capacity fade: 20%

[newhaul]

Originally Posted by AGM

Yes, very good idea.
I've added a few extra clues to mine.

Age of bank: 6 years
Advertised capacity: 3.6Ah
Capacity test done when new: 3.4Ah
Most recent capacity test result:
6 months ago, 2Ah
Approximate number of mixed cycles: 30
Kept at 40% SOC for most of the time.

Charging termination:
mostly around 3.6V at various termination currents

Float charging: none

Trying to regain lost capacity:
Charging to 3.7Vpc, termination 0.02C (70mA), disconnect charger.

Voc after resting for 10 hours: 3.33Vpc
Capacity recovered: none
Irreversible capacity fade: 42%

Age of bank: 5 years
Advertised capacity: 40Ah
Capacity test done when new: 47Ah
Most recent capacity test result:
6 months ago, 38Ah
Approximate number of mixed cycles: 50
Kept at 40% SOC for most of the time.

Charging termination:
mostly around 3.6V at various termination currents

Float charging: none

Trying to regain lost capacity:
Charging to 3.7Vpc, termination 0.02C (0.8A), disconnect charger.

Voc after resting for 10 hours: 3.52Vpc
Capacity recovered: none
Irreversible capacity fade: 20%

that is exactly the type of data that I find most helpful . And it seems that we all will at some point feel the effects of diminished capacity now it seems we should progress to the discussion of how to mitigate the effects as long as possible .

Or is it an inevitable fact that needs to be factored in upon the initial sizing of the bank.

Say if I need 100ah bank should I start with a 125ah bank so in 10 years I have that 100 ah capacity ?

[AGM]

Originally Posted by newhaul

Or is it an inevitable fact that needs to be factored in upon the initial sizing of the bank.
Say if I need 100ah bank should I start with a 125ah bank so in 10 years I have that 100 ah capacity ?

Very good question.
It's probably a tossup between not having to modify a somewhat over-sized bank for a long time vs buying smaller and being prepared to add to it at some stage (with all the implications).
Also, keep in mind you can't lose what you don't have, so maybe buying smaller again and replace the lot when you feel the available reserve capacity has become unacceptable?

[newhaul]

My bms disconnects at 3.65v here is an idea would this seem like a good idea to run up to bms disconnect 14.6v at the bank. Actually charge to 14.5 or 3.625v at cell which is at .05v below bms disconnect say on a monthly or bimonthly basis .

Opinions?

[CatNewBee]

My settings are:

BMS

Balance voltage start 3.375V
Balance Voltage bulk/end 3.382V

End of charging signal for Quattro and BMS sync 3.65V
Hysteresis 0.250V
End of discharging signal Quattro 3.050V
Hysteresis 0.100V

Over voltage protection cut off 3.750V
Hysteresis 0.250V

Under Voltage cut off 2.900V
Hysteresis 0.100V

Solar charger settings to LFP profile, no temp compensation., no equalizazion or smart management.

Absorption 14.3V, 5min (at the cell level it translates to 3.52..3.55V)
Float 13.5V (3.37V at the cells)

Stop charge and over voltage protection are different beasts, you dont want your BMS to brute force disconnect the busses, it is a protection measure like a fuse, yu must allow some play for the balancers to equalise the cells.

[AGM]

Originally Posted by newhaul

My bms disconnects at 3.65v here is an idea would this seem like a good idea to run up to bms disconnect 14.6v at the bank. Actually charge to 14.5 or 3.625v at cell which is at .05v below bms disconnect say on a monthly or bimonthly basis .

Opinions?

According to the literature, memory may accumulate every time charging is terminated below 100% SOC.
100% SOC is guaranteed by charging to manufacturer's specs for instance 3.8Vpc @ 0.02C termination (Sinopoly).
Terminating there not only achieves 100%, it also erases any memory which may have formed during a previous incomplete charging session.

If you decide to charge to a lower voltage it may take longer for the current to taper off sufficiently so there's potential for premature termination and gradual build up of memory.

In each incomplete charging session, a thin voltage layer is added which masks an increasingly lower SOC causing more memory.

So yes, charging to a higher level at least occasionally should help.
But you need to ensure every single cell gets to the same higher voltage which requires adequate balancing/shunting circuitry to prevent over-shoot in single cells.

[newhaul]

Originally Posted by AGM

According to the literature, memory may accumulate every time charging is terminated below 100% SOC.
100% SOC is guaranteed by charging to manufacturer's specs for instance 3.8Vpc @ 0.02C termination (Sinopoly).
Terminating there not only achieves 100%, it also erases any memory which may have formed during a previous incomplete charging session.

If you decide to charge to a lower voltage it may take longer for the current to taper off sufficiently so there's potential for premature termination and gradual build up of memory.

In each incomplete charging session, a thin voltage layer is added which masks an increasingly lower SOC causing more memory.

So yes, charging to a higher level at least occasionally should help.
But you need to ensure every single cell gets to the same higher voltage which requires adequate balancing/shunting circuitry to prevent over-shoot in single cells.

that makes sense and I'm good then because my camel brand cells are spec at 3.65 top balance and a .5c charge rate. 100ah 4p4s

[Delfin]

Originally Posted by AGM

According to the literature, memory may accumulate every time charging is terminated below 100% SOC.
100% SOC is guaranteed by charging to manufacturer's specs for instance 3.8Vpc @ 0.02C termination (Sinopoly).
Terminating there not only achieves 100%, it also erases any memory which may have formed during a previous incomplete charging session.

If you decide to charge to a lower voltage it may take longer for the current to taper off sufficiently so there's potential for premature termination and gradual build up of memory.

In each incomplete charging session, a thin voltage layer is added which masks an increasingly lower SOC causing more memory.

So yes, charging to a higher level at least occasionally should help.
But you need to ensure every single cell gets to the same higher voltage which requires adequate balancing/shunting circuitry to prevent over-shoot in single cells.

That's a great summary, and I think correct based on my experience. I would only add that I think deep discharges may also be desirable at least occasionally for optimum maintenance of capacity.

[Dev0n60]

To answer a question previously posted in this thread re Chargers: The Victron Phoenix Charger with the Mark 2 interface and the VE connect app will allow one to control the charger output as one reaches full SOC. I use the app to turn the power (I) down and maintain CV this way when near 100% SOC. I am now shutting charge off at .02C. So far this protocol seems to be working for the memory effect I have become of aware of only recently. Per Nordkyn Design's comments on this topic, I will be repeating this.
This all good on shore power but how do I control my regulator's field current output to the alternator in order to control output at CV in a similar way? Mostly we are at anchor or underway.

[john61ct]

Wakespeed, and American Power Systems APS-500.

Both outgrowth's of Al Thomason's SAR open source Very Smart Regulator / "Smart Alternator Regulator" project.

Sorry this KB dump isn't clickable

[newhaul]

It seems that a monthly top balance will forestall any memory loss of ah capacity issues from cropping up . For my bank that means a couple days with my solar set to 14.45.
And let my bms do its job.

[Delfin]

Originally Posted by newhaul

It seems that a monthly top balance will forestall any memory loss of ah capacity issues from cropping up . For my bank that means a couple days with my solar set to 14.45.
And let my bms do its job.

That's pretty much what I've concluded from this excellent thread. If you continuously undercharge your LFP bank, and if memory effects are real, that is when you'll see them. So don't continuously undercharge your LFP bank. My BMS starts shunting at 3.65 vpc, a little higher than yours. Lithionics would say that 3.61 vpc will get you to 98% charge, but what I have also concluded is that some period of absorption during these full charge events is also likely to be beneficial. For me, that's 30 minutes at the shunting voltage. Hope that's cool....

[ghrt]

I wonder whether periodic deep discharges are also advised

I have an unused 400ah Sinopoly bank, sit for 2 years.
In the first cycle, discharging @10amps, I saw VPCs dropping 0.2mV and then getting back 0.1-0.15mV. Measured by CellLogs, which are not very precise but it's a delta, not absolute value.
So it looked like memory effect on discharge and that's why I'm here; LFP life is suddenly more complicated, as I planned of bottom balancing and PSOC.

I wonder if OPs had time and patience to follow methods suggested, ie charging with low currents for long time, and what the outcome is.