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

[tanglewood]

Originally Posted by OceanSeaSpray

Yes, that is also what I would expect. However, the longer you leave it, the tougher it gets to actually reach that full charge. So the risk is that you might think that you are reaching a full charge without quite succeeding and if this happens, then the issue starts compounding itself and capacity shrinks. Technically, I should have been "fully recharging" frequently each summer, but the verdict today is that it hasn't worked.

In my case, winter essentially prevents full recharging. This is now the 5th winter for this bank and it is still operating fine because there is still more than enough capacity for routine daily cycling, but reserve capacity is down. If I hadn't taken the trouble of looking into it and measuring it, the situation wouldn't look much different than when the cells were new.

To get a truly full charge when there have been a lot of partial cycles before, you need really good termination control and fairly solid charging parameters, not the kind of stuff some people talk about with low voltages and timed absorption.
Charging to 3.65V/cell and C/30 residual current as per the cells specs would no doubt get me a little further and it is likely to be the next step while data logging.

This memory phenomenon that you describe is interesting if for no other reason than you seem to be the first to identify it. To that end, can you provide some details on how you are assessing all this? All you've said is that you can only recharge to 70%, but how are you determining that? How are you determining SOC both at the bottom and the top your operating range? How are you determining that the missing capacity is at the top of charge vs the bottom of charge? You might be on to something, but you might also just be misunderstanding or misinterpreting what's going on. Right now, I can't tell and all we have is an assertion, so it's hard to take very seriously.

[tanglewood]

I'd like to suggest that this thread be closed.


It's not because the information is bad or has gone sour. It's really just the opposite. It has become so overloaded that really valuable discussions are getting lost, or at risk of getting lost. It's just too daunting to even thing about reading the whole thread. It would take weeks.


At this point, I think we would all be much better served if new topics were started in new threads, titled to reflect the issue or subject. That's why we have threads, after all. This recent assertion of memory effect in LFP is a great example of a topic deserving it's own thread rather than getting buried in this galactic thread.


What say ye?

[senormechanico]

tanglewood,



You have a point.

The thread is so large, the same questions keep getting asked over and over because newbs don't want to read all of it.
It's a problem with several threads on CF as well, some with as few as 30 posts !

[newhaul]

Originally Posted by senormechanico

tanglewood,



You have a point.

The thread is so large, the same questions keep getting asked over and over because newbs don't want to read all of it.
It's a problem with several threads on CF as well, some with as few as 30 posts !

it is starting to look like a climate thread in that respect .
Perhaps it is time to divest .

[Checkswrecks]

At 440 pages on this thread, I'd agree with tanglewood.

As for what OceanSeaSpray wrote, partial cycles really relate to depth of discharge and it appears there is use of the terms memory and capacity interchangeably. It's easy to think that reduced capacity is the cell or battery developing a charge memory and nobody here would be the first or last to do this. But as tanglewood noted, there are just so many unrecorded or unmentioned variables, it's hard to respond. Recharging rate, temps, DOD, max charges imparted, and all over a number of cycles. I'm not criticizing, because I personally know that it is all too easy to think you have one thing till digging through a lot of data. Here's a limited but more complete discussion which could shed some light on the subject:
https://batteryuniversity.com/learn/...ased_batteries

[john61ct]

The measure of SoC via specifying a charge profile - certain voltage CV setpoint at a certain current amps C-rate in Bulk, finally a stop-charge endAmps at a certain C-rate taper

is actually defining a 100% SoC for purpose of daily-use charge cycles, maybe not actually hitting it all that often but that defines your goal

and should be much lower / gentler than the

mfg specs which are absolute maximums, stress ratings, the limits to which the devices can be stressed without causing permanent and irreparable damage.

These limits are not recommended for normal functional operating.


Operating at the absolute maximum ratings for extended periods will produce reduced reliability / lifetime longevity.

_______
Then we have the only way to actually measure **capacity** which %SoC is only relative within.

That is using controlled constant current load testing. To get super accuracy requires precise control and measuring instruments, but a reasonable measure is neither very expensive nor rocket science.

Some may even rely on an Ah counter, but timing a constant current load is better.

If that capacity is permanently reduced by wear and tear, cycled usage / abuse and calendar aging, IOW cannot be recovered, then the relative % capacity compared to that benchmarked at commissioning time is called SoH%, State of Health.

Balmar SG-200 claims to automate that calculation, but likely will be affected by this memory effect, if any.

I first came across this "memory issue" from "the man from Mannum" in SA, whose postings were largely deleted when he (and any mention of him and his business) were banned from the relevant Aussie forums, several years ago.

Now, to the extent any capacity losses are temporary, and easily remedied by access to power on demand and a powerful charging setup, where voltage and amps are easily adjustable as needed,

which is also just as much required for **any** routine testing and maintenance protocols

They are NBD, but certainly easier to conduct off shore power, and in some cases on each lowest-level component separately, ideally at the individual cell level. For most that means while the bank is not in use as it's been broken down, disassembled from its usual xPyS layout.

The **key** takeaway here is, just as with lead chemistries, keep on top of what's going on with your expensive large bank, know what you're doing, do not just rely on automated set-it and forget-it equipment.

If longevity, getting maximum value for money is important to you.

For those happy to just get a decade of use then replace,

and/or willing to spend incremental thousands on proprietary packaged systems + installation / system design help & training from a Victron Mastervolt or Lithionics dealer,

most of these finer-point issues can just be ignored.

____
For the record, closing this thread and starting a new catch-all is I think a great idea.

But for at least a year or so, noobs should be referenced to it and encouraged to parse through it all to jump-start their learning process.

[tanglewood]

Originally Posted by john61ct

For the record, closing this thread and starting a new catch-all is I think a great idea.

But for at least a year or so, noobs should be referenced to it and encouraged to parse through it all to jump-start their learning process.

I'm not suggesting a new catch-all. That's what we already have with this thread. I'm suggesting that new threads be started with their particular topic. More narrowly defined topics are easier to search and provide more lasting useful information than a dumping ground of every thought on LFP. It had it's place when LFP was new enough, but we are past that. This thread is now the equivalent of a thread dedicated to something "diesel engines" and just becomes rubble on the floor.

[weavis]

Hold hard me hearties.

Before this thread is closed, The moderators wish to point something out to the participants.

Our job is to ensure the smooth running of the CF under all circumstances and trying conditions that are thrust upon us.

We have noticed that there has been and is continuing, a mob action with regard to one member participant. Without taking sides or allowing our personal opinions to intervene, we are under obligation to apply the rules in a fair and even manner.

To this end, we suggest the following to be cleared up immediately.
A member has requested an apology for being called a fanboy. Would the participant who made the statement please rectify the matter.

To the other members who continually make personal and snide remarks regarding one CF contributor, please realise that CF is the host venue and that the rules of conduct need to be adhered to. Such behaviour will not be tolerated.

The entire admin is in unity over good manners being displayed.

The administrators are the arbiters of who is permitted to post or not post, and as such work closely with the site guidelines to ensure that no rules are broken. You don’t have to like people, you may feel someone is talking out the wrong orifice or is completely wrong, yet, like us, each member needs to work within the framework. We suggest putting an individual on ‘ignore’ if you are over irritated by his or her comments. A demeaning or rude remark will end up in tears for the respondents.

Should the above not be adhered to, an offending poster will be placed in a ‘review before publication’ situation. Every post from that member will require admin approval. This may cause a delay in continuity to a thread due to when the post is reviewed, but we don’t see any other way to deal with the situation other than removing a member.

You made indeed have a good reason and be justified in your opinions of situations. Write the mod team, NOT respond on the thread concerned. You will risk censure and infractions if it becomes a personal attack.

It’s a chat board folks, not life or death. No one died. Put it behind you and move on.

[Don C L]

...and with nearly 1,100,000 views we must have set some kind of record... we can't stop now!

[Cpt Pat]

Originally Posted by newhaul

the 2 biggest killers of Lfp are being under charged ( discharge below a specific set point significantly reduced service life.
The other is over charge . Hence the bms has both HVD and LVD
It has been estimated using real world conditions that when kept between 20% and85%
The extrapolation shows a likelihood of well over 5k to 7k ( or more) cycles in service .

Agree. But... HVD is a last resort form of protection. LFPs can be over charged with a terminal voltage as low as 13.6 volts (4S pack). Voltage CAN NOT be used to determine a fully charged SOC. You must do coulomb counting instead (amp/hours in and out). This factor, depending on HVD, is the major cause of overcharging.

[nebster]

Originally Posted by weavis

Hold hard me hearties.
Without taking sides or allowing our personal opinions to intervene, we are under obligation to apply the rules in a fair and even manner.

I think you're also under an obligation to preserve the intellectual integrity of the forum. In extreme situations, that should actually trump individual sensitivities or fairness.

What should a majority of the knowledgeable posters on a forum do when they have concern about a single poster filling the halls with noise? Do we have no recourse, weavis?

[john61ct]

I had no intention to direct the fanboy modifier at anyone in particular.

Just meant that I was doing my best to take owner preferences, space available, overall power needs, battery chemistry etc

into account, tempering my default tendency to blindly advocate for solar input.

[newhaul]

Originally Posted by Cpt Pat

Agree. But... HVD is a last resort form of protection. LFPs can be over charged with a terminal voltage as low as 13.6 volts (4S pack). Voltage CAN NOT be used to determine a fully charged SOC. You must do coulomb counting instead (amp/hours in and out). This factor, depending on HVD, is the major cause of overcharging.

pat I'm sure you understood that I only stated about the HVD as being in the bms control to protect from over charging as a safety feature not as a normal stopping point .

[Don C L]

Originally Posted by nebster

I think you're also under an obligation to preserve the intellectual integrity of the forum. In extreme situations, that should actually trump individual sensitivities or fairness.

What should a majority of the knowledgeable posters on a forum do when they have concern about a single poster filling the halls with noise? Do we have no recourse, weavis?

The best recourse is to state verifiable facts, evidence and qualifications and let the readers decide whom they care to listen to. Everyone is free to stop following a thread if they find it less than useful or informative.

[john61ct]

Allowing discharge to a certain low voltage but no further has nothing to do with charging.

Would be nice if automation allowed that V setpoint to be varied by load's current level, but we're not there yet.

Same with the other way, as when charging, the bank's netAmps acceptance rate (left over after feeding concurrent loads) being low enough, as is usually the case with solar,

changes the voltage setpoint needed to get to (whatever working definition of) 100% SoC without needing to go into CV.

Yes when netAmps are higher, going into CV allows stop-charging based on endAmps, regardless of the specific amps rate available during Bulk,

but very few source regulators allow for that.

Originally Posted by Cpt Pat

Voltage CAN NOT be used to determine a fully charged SOC.

Well, then even less so between 10-90% SoC!

I do of course agree, within the caveats I pointed out in the above discussion.

However with regard to the top say 5-10% range (compared to vendor maximums, gladly sacrificed by me to extend longevity),

just to play Devil's Advocate - if net amps into the battery are relatively stable,

fortunately in normal daily use, extreme levels of stop-charge precision is simply not required, getting to "just enough" SoC to handle our daily budget is fine, with defaulting to a safer (conducive to longevity) lower SoC point,

for when charge currents are known to be low, say under 0.15C.

Memory effect (if any) and balancing issues can be treated as periodic maintenance issues for when sufficient power and precise controls are available.

> You must do coulomb counting instead (amp/hours in and out).

I think that's overstating things a bit personally. Ah counting meters have proven to be less accurate than I would prefer when it comes to something as critical as stop-charging. Maybe if a known-reliable HVC is also there as a working backstop (not counting the usually crazy high last resort HVC included with most BMS)

By doing load testing to precisely determine the actual objective SoC levels resulting from given profile choices

net amps + V setpoint (endAmps / CV optional at low currents)

as long as the netAmps into the battery is pretty consistent in a given use case,

to me that is preferable to basing stop-charge on an Ah counter.

I agree when currents in/out are highly variable, your approach may be necessary, but too many sources I respect say coulomb-counting is just not accurate enough yet for relying on for stop-charge control, at least in most cases.

Of course SG-200 may prove itself to be a real game-changer in that regard.