LFP memory effects thread

[tanglewood]

The irony in all this is inescapable. We need to do periodic "equalization" charges of our LFP cells. I didn't see that one coming.

[nebster]

Originally Posted by tanglewood

The irony in all this is inescapable. We need to do periodic "equalization" charges of our LFP cells. I didn't see that one coming.

I'm not sure that follows, yet.

The two packs being discussed in this thread appear to have large numbers of charge cycles landing them consistently at pretty low SOCs. (70%?)

I'm not sure why someone would set up a system that returns such an expensive battery to such a low SOC so often... although I definitely understand why lack of access to energy could lead to that result in practice, at least some of the time.

It seems to me that most people would do one or more of (a) reduce their pack size, (b) increase their charging capability, or (c) tune their charging parameters. Any of those would fill up a battery much further and be more efficient in up-front costs, even before a memory effect comes into play and makes things worse.

Or, am I missing something obvious?

[OceanSeaSpray]

Originally Posted by Cpt Pat

My conclusion: Terminating charging using only a charge termination cut-off voltage will result in a progressive development of under charging.

Using coulomb counting on the other hand, if properly calibrated...

Everybody knows that lithium cells require absorption. MaineSail has done 10 years of cycling with termination based on voltage and residual current with no memory effects.
"If properly calibrated", or the elegant way of relying on the impractical to offer a solution. CC drifts out after just a few cycles and becomes a random number. Please get over it instead of filling the thread with it and - incidentally - you have no experience with memory effects.

Originally Posted by Cpt Pat

I can't say if charging to 100% SOC every 25 cycles, the coulomb counting method, or the two combined prevented memory effects.

Well, fortunately we can. MaineSail has returned his pack to 100% SOC every 50 cycles or less, without relying on any CC, and it has had no capacity issues. In other words, standard charge termination is perfectly adequate if it can be achieved often enough.

[OceanSeaSpray]

Originally Posted by evm1024

...The next series of tests will likely charge (starting at) a 200 amps to 300 amps rate (my 14.4 volt charging capacity is limited to about 300 amps) and discharge at 100 amps.

That series would be designed to move all active particles into a single phase regime.

I'm looking for a log jam break type of change - say a change of capacity from 500 AH to 600 AH - that would be nice. But I will take a number of 20 AH changes. Just gotta find a way to make the change.

The problem is not with the lithium you can move, that one is easy to shift and will give high currents, but with the one that won't shift.
Below this point, it is almost like operating a normal, but smaller, battery.

I have discharged to empty several times now and recharged starting at about 0.7C and hit the wall each time. Try however, by any means.

Now I changed my charging parameters to terminate absorption at 0A, which is only going to mean "absorb for as long as possible", because I doubt I will manage to reach no current at the moment. The objective is seeing whether this regime begins reclaiming capacity or not and I should be able to observe something before too long.

[evm1024]

I see what you mean by hit the wall now. Yes, that is true. I "hit the wall" at about 70% of the sticker capacity. I would love to break the wall down.

As that there are 2 cells that appear to be the cause. It would be nice to get my hands on 2 "normal" cells to see what the capacity was.

And you are right - I have a perfectly normal 500 AH pack. Still not a bad thing.

[CatNewBee]

Don't mix up the memory effects based on PSOC all the time with unbalanced cell packs based on ignorance of a properly configured BMS with balancing capabilities.

It is true, cheap top balancer BMS are useless if you never charge the battery to a voltage, where they start balancing (often 3.55...3.6V, what translates to 14.2...14.4V).

Just operate the battery to the specs.

[Cpt Pat]

Originally Posted by OceanSeaSpray

Please get over it instead of filling the thread with it and - incidentally - you have no experience with memory effects.

True. No experience. Three years in service, over 500 cycles, and I see no evidence of detrimental memory effects in my LFPs. My LFPs still deliver 100% capacity. Everyone should be so "lucky."

"Experience keeps a dear school, but fools will learn in no other."
-Benjamin Franklin

[john61ct]

“Only a fool learns from his own mistakes. The wise man learns from the mistakes of others.”
― Otto von Bismarck

[senormechanico]

So THAT'S why you don't own any, but just give advice on what/how to... !

[Cpt Pat]

Originally Posted by senormechanico

So THAT'S why you don't own any, but just give advice on what/how to... !

Not sure who that quip was addressed to. I have a 4 series/two parallel bank (8 total) of GBS 100 AH cells that entered service in late 2015.

CatNewBee above is right, but I'll go further to assert that cheap cell balancers are worse than useless: one malfunction, and that bad balancer board can discharge the cell to destruction. In my "fractional C" installation, my cell balances slip out of balance by about 50 millivolts per year. I just use a 2.7 ohm cell shunt resistor to manually rebalance. Even then, it just makes me feel good to rebalance the cells - it's never been really necessary. Since, with low current charging, my cells never exceed 3.55 volts at 100% SOC, the cell balancers would do nothing, as CatNewBee points out.

But then, I don't have the luxury/liability of an engine driven alternator shoving high currents into my pack. My boat is a sailing vessel - not a wind-assisted motor vessel.

All the balancing complexity comes into play when the user impatiently shoves large charging currents into the pack. I don't mind spending 10 hours on shore power charging my pack at 0.1C before a sail, and keeping it topped off with solar underway. If I were in such a hurry, I'd buy a motor boat.

Cell balancers are necessary for electric vehicle applications where charge and discharge currents are at or above 1.0C, and the energy is being recovered during braking by shoving large charging currents back into the cells. Hopefully, house banks on boats are treated much more gently.

[senormechanico]

It was addressed to john61ct, and I thought it would be obvious.

[Delfin]

Originally Posted by tanglewood

The irony in all this is inescapable. We need to do periodic "equalization" charges of our LFP cells. I didn't see that one coming.

That is my take away as well. And with further irony, that is essentially what the manufacturer of my bank recommends. Who would have thought someone who builds these packs and supplies then to governments know what they are talking about?

[Cpt Pat]

Originally Posted by OceanSeaSpray

Well, fortunately we can. MaineSail has returned his pack to 100% SOC every 50 cycles or less, without relying on any CC, and it has had no capacity issues. In other words, standard charge termination is perfectly adequate if it can be achieved often enough.

I believe we may be arguing the same position here: that regular routine recharges to 100% SOC will perform a "memory release." My schedule was every 25 cycles, I interpret (from the above) that MainSail's schedule was every 50 cycles. I performed my 100% SOC charges by manually monitoring the current taper ("tail current") to 0.02C. That low current value is "on the transfer knee" so further charging adds no more than and amp/hour or two.

There are many people (from my readings) who avoid ever charging above 80 or 90% SOC in the belief that practice will extend battery cycle and calendar life. Never charging to 100% SOC seems to have the opposite effect on battery life in view of memory effect.

I ran a load test over the weekend: 224 amp/hours at a discharge rate of 10 amps (0.05C) from my 200 amp/hour batteries to a pack voltage of 11.0 volts (there's a little amp/hour gain due to peukert function at that low discharge rate).

[Cpt Pat]

I'm going to make all of you who have tried to shout me out of this thread happy and close out my posts with this definitive statement: If your BMS does not factor in charge current - if it doesn't even have a current measurement shunt to measure charge current to your LFP battery - then it is incapable of being adequate and will eventually contribute to degrading your LFP battery. Those who are terminating charge by detecting only terminal voltage will see a progressive accumulation of lost capacity due to memory effect from undercharging or destruction of your cells from overcharging.

Quoting from Marine lithium batteries in operation | Nordkyn Design

LFP cells simply don’t really charge at voltages up to 3.3V [per cell] and then fully charge already at 3.4V and upwards. The transition is so abrupt that claiming to control the charging process by adjusting the voltage is purely and simply bound to fail.

Charging at reduced voltages, down to 3.4V/cell, only increases the absorption time and therefore the overall charging time, but achieves strictly nothing in terms of preventing the battery from getting fully charged and then overcharged. It only takes longer for this to happen. Furthermore, low-voltage charging opens the door to severe longer term performance issues which arise from memory effects in the cells.

(...)

In all instances where significant charging currents are present or where the battery has seen a large number of interrupted and partial charge cycles, correct termination can only be obtained by monitoring both current and voltage to make an informed decision.

I personally am using coulomb counting for charging while underway from low current variable charging sources (PV), along with periodic full charges to 100% SOC by monitoring for charge current taper using a constant charging source (current-limited constant voltage battery charger on shore power). It has worked for me for over three years without any evidence of memory effect degradation.

Most of the off-the-shelf BMS systems are junk, and only accomplish fulfilling market demand for conformance to common LFP mythologies. I had to build my own BMS. If you are actually using sun and wind - and not dead dinosaurs as underway energy sources, I recommend coulomb counting for controlling charging with those variable-current sources.

Sorry if you took the lead-acid battery approach to charging and bought into the "terminate charging with terminal voltage detection" snake oil. Use something better after you replace your ruined batteries.


Maybe I'm a heretic. Just get over it.

[tanglewood]

The memory effect described in the research papers is a bump in voltage during the charge cycle when SOC reaches a previous charge termination point. It's not a big bump - 20mv sticks in my mind - and it's hardly a "wall" that you can't get past. If you continue to apply current, the battery continues to accept it all the way through and past the "bump".


So I can see how if a charge protocol terminates solely on voltage, and does so immediately after the target voltage it attained, you will have charged your battery to a slightly lower SOC than you thought.


But if you continue to apply current (traditional absorb mode) long enough to get past the voltage bump, you will be back on track for normal charging. The absorb time to achieve this could be strictly time based, or it could be return current based. Either way as long as it gets you past the hump, you are back on the normal charge curve.


And I haven't seen anything in the research that suggests any sort of wall that you can't charge past.


To me the interesting question is how long do you have to continue charging to get past any memory bumps. It would seem that is the min absorb time required.


What am I missing?