Most useful article I've ever found on LFP battery charging

[Maast]

LFP has no memory it is NOT NimHi, NiCad or those other horrible batteries of the past which did set a charge memory.

Yeah, it does, not nearly as bad as the others but it exists. It was a shock to me too. Don't take my word for it - read the first link, and then the PDF study I posted, then google the subject and read more papers. I just posted the two best documents I found on the subject but I found many many many more.
The good news is it's reversible if you carefully do a perodic full charge without overcharging it - Dont watch the voltage watch the current go to 0 or near-0 and then stop charging. Don't float.

 

[grizzzman]

Yeah, it does, not nearly as bad as the others but it exists. It was a shock to me too. Don't take my word for it - read the first link, and then the PDF study I posted, then google the subject and read more papers. I just posted the two best documents I found on the subject but I found many many many more.
The good news is it's reversible if you carefully do a perodic full charge without overcharging it - Dont watch the voltage watch the current go to 0 or near-0 and then stop charging. Don't float.

Not always. If allowed to under charge long enough, it may become a permanent loss.

 

[electric]

Oh my God, red line has slightly deviated from the black line, call the Nobel Prize committee, we discovered that sky is falling on all LFP batteries.
This will be chewed, regurgitated, chewed again and again until no longer recognizable in all social media for the next 10 years.
You guys are a funny bunch...

 

[upnorthandpersonal]

From the text:

The memory effect was found to strengthen with the number of incomplete charge cycles performed before the erase cycle. It was also strengthened when a partial charge was followed by a shallow discharge, rather than a deep discharge.

and

These latter aspects have proved to be of key significance when considering the longer term performance of LiFePO4 batteries in house bank applications, because incomplete charge cycles are common when relying on renewable energy sources and shallow discharge cycles are also frequently experienced. These have the potential to render battery banks near unusable after as little as 2-3 years in regular service in the absence of memory-releasing cycles. Ineffective memory-releasing cycles are very common in DIY installations where the charging process is not properly controlled and/or configured incorrectly by fear of overcharging or due to widespread mythologies.

Are they minimal? Probably. The sky isn't falling. Are they good to know about so systems can be designed better and last longer? I would say, definitely.

 

[Maast]

I asked the original author and got an answer. It appears a completely full charge using current taper to find the termination point once or twice a year is sufficient. My current equipment can't do it so it'll have to be a manual "memory wipe" but at only a couple times a year this is acceptable.

Am I making much ado about not much? I dont think so and the science behind it is more than solid even though it's not common knowledge, maybe because people dont WANT to believe it.

I've got more money invested in my battery bank than my first three cars I owned - combined! I saved for years to build my system. You damn well better believe I'm going to do everything I can to get full use out of it for as long as is achievable.

I’m a little late to the party here, but I have to ask: how often should a full charge be accomplished to wipe memory effects before they become permanent? Once a month? 6 months? a week? I’ve got 3 55A Midnite Classic 250s charging my 50kwh LFP bank w/ a Radian GS8048 inverter. I can adjust the bulk and adsorption voltages & time, and turn off the float.
Nothing I have allows me monitor current and get that information into the Midnites or Radian in a useful way.
The only way I can see accomplish a periodic full charge is to set the equalize function to 3.6v/cell and the time to a value I establish w/ a current monitor to when the charge current gets to 0 or near 0 and then have it repeat once a week/month/etc.
If you have any ideas I’d love to hear them.
Thanks

Eric Bretscher says:
30 December 2019 at 7:46 am

Memory effects don’t become permanent, they just become more difficult to overcome. Your question is difficult to answer because there is not much data available and it depends too much on the operating regime of the bank. Performing a full recharge once or twice a year would appear to be quite enough in most cases.
If you can’t monitor current and use this information, you can’t achieve a correct charge termination. The absorption time varies a lot with the condition (and age) of the cells, so trying to deterine a fixed absorption time is a flawed strategy. You must use voltage and current.
Trying to charge LFP banks with lead-acid chargers doesn’t work properly. The BMS must measure the current and be able to control the charge termination.
Best regards,
Eric

 

[gnubie]

So best thinking is to pull the battery up to 3.55v / cell, and unless you have the ability to automatically monitor tail current and terminate charging that way, stop charging once that voltage is reached and twice a year manually monitor tail current to really bring the cell to do a near full charge.

Once charging is complete and the battery is not being left idle, ie it's an active system with the battery being drawn down daily, hold the cell ('float') at under 3.4v.

?

 

[Dzl]

I asked the original author and got an answer. It appears a completely full charge using current taper to find the termination point once or twice a year is sufficient. My current equipment can't do it so it'll have to be a manual "memory wipe" but at only a couple times a year this is acceptable.

Can anyone explain to a newbie what it means to fully charge based on current as opposed to voltage? Does this mean setting a dc charger or power supply to the desired voltage but instead of turning it off when it hits the desired voltage you hold that voltage until current drops to zero or near zero? Something similar to these steps (taken from this article):

1- Wire the cells in parallel
2- Set the power supply to 3.400V and 80% or less of the rated amperage (80% to not burn it out)
3- Turn on power supply and charge cells to 3.400V
4- When current has dropped to 0.0A at 3.400V turn off the power supply & set it to 3.500V
5- Turn on power supply and charge cells to 3.500V
6- When current has dropped to 0.0A at 3.500V turn off the power supply & set to 3.600V
7- Allow current to drop to 0.0A (or very close) at 3.60V
8- Done, pack is balanced.

 

[Maast]

Thats pretty much it according the above sources:

- Set up your ammeter
- Set your charging source to 3.6v/cell using a good quality digital multimeter before you connect it to your pack. Keep the current under .25C (25% of the packs capacity in amps)
- Connect your pack and charge it while watching the ammeter, when the ammeter drops to zero or within a amp or so disconnect the charging source. Don't let it 'float' at max capacity as that degrades the battery.
- Watch the voltage while charging, if it goes over 3.6v/cell reduce it.

I can tell you from testing I've done on large individual cells your charging current will sharply drop near the end of the charge cycle once the pack starts hitting 3.4v/cell or so, after the current drops you're within a half hour of being done.

 

[Maast]

So best thinking is to pull the battery up to 3.55v / cell, and unless you have the ability to automatically monitor tail current and terminate charging that way, stop charging once that voltage is reached and twice a year manually monitor tail current to really bring the cell to do a near full charge.

Once charging is complete and the battery is not being left idle, ie it's an active system with the battery being drawn down daily, hold the cell ('float') at under 3.4v.

?

From what I'm reading dont let it float at all! Bring it up to charge then stop charging altogether. Personally if I can't figure out a way to completely turn off float I'll set it to float at 3.2v/cell.
Since I'll be grid tied I'm going to set the inverter to 'sell' at 3.3v/cell so it'll hover around 70% SOC daily. It's got a manual equalize so I figure I can use that while keeping an eye on the current going in. Don't know if I can limit its current through settings though.

 

[gnubie]

In practical terms you either have some floor voltage or you are constantly cycling the battery at the top 5%. Remember this is not grid tied, its a system where the battery is being actively used daily.

Float charging is a hold over term that many people understand even though we are not actually float charging. In this application the purpose is not to be constantly pushing current at the cell and the voltage indicates that.

 

[Dzl]

Since I'll be grid tied I'm going to set the inverter to 'sell' at 3.3v/cell so it'll hover around 70% SOC daily.

I have a very basic and mostly unrelated question. What does State of Charge refer to and how is it measured. My understanding is that it is a % value derived from a voltage relative to the minimum (2.5) and maximum (3.65) voltage, but my understanding of the term is very fuzzy.

 

[Maast]

I have a very basic and mostly unrelated question. What does State of Charge refer to and how is it measured. My understanding is that it is a % value derived from a voltage relative to the minimum (2.5) and maximum (3.65) voltage, but my understanding of the term is very fuzzy.

State of charge refers to just how full your battery is, you should go to the original post and read the linked article. It'll answer this and a whole lot more.

Keep in mind that LFE cells charge/discharge curve is fairly flat so using voltage isnt a very good way to tell SOC except at the top and bottom end.
From that article take a look at this:

If you're charging 70% SOC is about at 3.32v/cell.

 

[Maast]

In practical terms you either have some floor voltage or you are constantly cycling the battery at the top 5%. Remember this is not grid tied, its a system where the battery is being actively used daily.

Float charging is a hold over term that many people understand even though we are not actually float charging. In this application the purpose is not to be constantly pushing current at the cell and the voltage indicates that.

You and I understand that, but most new people who are looking at this from a lead-acid understanding don't. I wouldnt use 3.4v/cell as your float, it's too high and at the fully charged level and you'll still be pushing electrons at it when it's fully charged and degrading the battery.
In the article I linked and the other article that Dzl linked stated several times to charge it up and STOP charging completely and there is a absorbtion time needed but its only 10-30 mins. There is no appreciable self-discharge to deal with so why would you want to anyway?

 

[gnubie]

Sounds nice to say stop charging but in a system where you have a load on the battery all the time and the SCC is doing what it should be and is supplying current to that load it is a practical impossibility to so finely balance the current that the battery will not be in a position where it has current available from the SCC on its terminals. Even if you run your SCC at a lower voltage the battery will draw down to that point and once again be in the same position.

So what I said, sans specific voltages, is what is going to happen in practice. Charge then hold some lower voltage, that or you put your battery in the cupboard once charged.

 

[SolarRat]

We had the float argument a while ago: https://diysolarforum.com/threads/float-charging-lifepo4.1719/

The fact is, this is new technology and anyone who claims to know it all is a fool. I don't think anyone really does...not enough years have past and not enough documented testing has been done yet.

 

[Max Bender]

Came across this as a reference in a thread, it's possibly the most useful article I've ever came across about LFP charging regimes. Evidently there actually is a well documented memory effect on LFP cells and that always holding charge voltages low to avoid overcharging the cells is counterproductive. The author recommends a periodic full charge and discharge to 'recondition' the cells.

Marine lithium batteries in operation | Nordkyn Design

This article details the relations between voltage, state of charge, current and temperature for lithium iron phosphate (LFP or LiFePO4) battery banks.

nordkyndesign.com

Very useful, and thank you - it turns out that LFP cells are even better than I had originally thought.

 

[Steve_S]

A Word of Caution here folks....

We are talking about LFP batteries here, which are generally very similar but not all the same.. These are made by many different companies with variations in the chemistries, anodes and cases right. Also one must keep in mind that some are made to handle High Discharge while others are not, (for some reason this is neglected in common discussion). I must point out that as a result there are variations on what the voltages are for the different brands, IE SinoPoly, vs Calb, vs Winston ,AYouPower and so on and even within the various "Brands" there may be differences. Always try to follow the spec sheets for the particular cells you are using. Unfortunately the "Spec Sheets" can be quite elusive, therefore I would recommend being on the Conservative side with charging profiles.

 

[krby]

Re Memory effect. If I plan to store a LiFePO4 bank for weeks or months how should I treat it? Full-charge (using the paper's definition of full-charge, a CC phase followed by a CV phase that ends at C/20 or lower current)? I thought leaving LFEs at full charge shortened their life? Full charge followed by a shallow discharge (maybe down to 90%?)

 

[SolarRat]

Re Memory effect. If I plan to store a LiFePO4 bank for weeks or months how should I treat it? Full-charge (using the paper's definition of full-charge, a CC phase followed by a CV phase that ends at C/20 or lower current)? I thought leaving LFEs at full charge shortened their life? Full charge followed by a shallow discharge (maybe down to 90%?)

About 40-60% SOC. That hasn't changed.

 

[krby]

About 40-60% SOC. That hasn't changed.

So...from a discharge, charge to 50% (let's say), then I leave it for weeks. Isn't that exactly the condition described in paper as creating the memory effect? That's the source of my confusion.