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

[john61ct]

Just talking about a hour, and infrequently. Less damage than fast charging in cool temps.

[newhaul]

Originally Posted by john61ct

Just talking about a hour, and infrequently. Less damage than fast charging in cool temps.

Please post the technical details that support your assertion.

[evm1024]

Funny how we stopped talking about the Battleborn video and the points that were made.

I guess that I will have to be the one to watch it again and itemize the questions and answers. Then dig out the pop corn

[Checkswrecks]

Originally Posted by john61ct

As long as not **too** far below, but me, that low I'd stay under 0.1C, or pre-warm the bank.

Like so many issues, hard to nail down hard numbers when the damage done is lost cycles off the back end.

Actual now-noticeable damage, and certainly rendering the bank instant scrap, only happens long past the point where longevity was being reduced.

Most of the recent research studies show the effect very clearly without destroying the cells, but

each chemistry, each model battery will have different numbers.

Only the minimum "no charging anywhere near this temp or instant scrap" spec is on the data sheet.

In truth you usually can charge at a low C-rate around 0°C

But in general, being conservative I'd start pre-warming below 10° if going over 0.1C

25° should be fine for 0.2-3C

Faster charging than this, anybody's guess, impossible to detect the factors leading to lost lifecycles without forensic autopsy, studies show different numbers for even slightly different chemistries.

So for 0.5C and above, I'd go to 30° first to maximize longevity.

Of course in many contexts people just take the hit, if they're even aware of the issue.

John's making some very good points that I'd like to back up with some data. These test results were from our testing the Boeing 787 cells which are cobalt and not iron based. The reason I'm showing this data rather than the LiFePo4 data I've seen is that the LiFePo4 data were proprietary with similar form to the curves. The plots shown below are from here:
https://www.ntsb.gov/investigations/...sic_Report.pdf
fwiw - When at the NTSB we reverse-engineered the 787 battery, worked out with Yuasa and Boeing what needed to be redacted (blacked out text in 1st chart), and everything shown is now in the public domain.

First off, discharge is generally not the problem because the ions are being released from the negative electrode surface layer (anodic SEI), rather than being driven into it. The decrease in efficiency is largely from the characteristics of the electrolyte shown in the next chart.

Reason I'm showing this - If using a high C discharge to warm the cells, do it in short cycles and pause at least a few minutes before hitting the starter or microwave to let the heat distribute through the cell or you are concentrating the discharge in a smaller area close to the current collectors. (Report Figures 95-103) In the 787 cells it melted portions of the separator film between anode and cathode which degraded the cell real fast.

Here's what is going on with the electrolyte as temperature changes:


Most of your commercial LiFePo4 cells are running thinner electrolytes and have additives which will move these curves slightly to the left. But here is where I am using this chart to emphasize something John wrote. Nobody but the engineers at the manufacturer (who are under proprietary restrictions) will know what these curves are or how they relate to charge rates.

Charge rates -
The third chart is called an EIS curve (electrochemical impedance spectroscopy) and understanding it goes into heavy-duty nerd-dom really fast. Basically EIS testing passes high frequencies between the cathode & anode to characterize how well the ions intercalate into the SEI. When the charge rate at any temp is too high, the ions are not able to work their way into the SEI layer and collect on the surface of the anode as dendrites. (Report Fig. 1) These ions deposit their lithium as metal which can not go back into solution on discharge, so once that tripping point is reached the cell rapidly begins to degrade both in capability and life. In other words the degradation is the point at which the charge rate exceeds temperature-dependent rate of intercalation and a high enough charge rate can cause this at any temperature, including well above freezing.

Enough of these lithium dendrite deposits in a cobalt cell can lead to thermal runaway with sufficient charge rate. The iron-based LiFePo4 doesn't have the reactance of the cobalt so while the fire risk is greatly reduced, the deposits will still very rapidly degrade/kill your expensive cells. (High enough rate or voltage can still lead to runaway in LiFePo4.) As with how fast the curve changes below freezing in the above electrolyte curve, you can have the onset of dendrite formation and that is what the EIS chart relates:


The bottom line in reading this chart is that you do not want to see the charge rate create the C-shaped curve exhibited by the right two plots. In our testing we were using a fixed charge rate and if the charge rate is reduced then the ions are able to work their way into the SEI layer and you'd not have the C-shape develop till colder. We saw the degradation begin with only a few charge cycles in cold tests. (Fig 93) The reason I'm showing the nerdy EIS plot is that it points out how close the bottoms of the left three curves are, how fast the curve changes near freezing (2nd & 3rd curves), then there is a big step over to the -20C (-4F) curve. The data is similar for LiFePo4 cells which basically means that somewhere in that tight spot between the 2nd and 3rd curves is a tripping point for the charge rate. This chart nicely shows the point at which the cell is damaged is a cliff and not a straight line relationship.

If you started with minimal charge you could add a little bit more and a little more and then suddenly the rate of lithium deposition would rapidly increase. The point is that the loss of usable life of the cell will match this. This balance between charge rate and temperature is going to be very specific and unique to each manufacturer's chemistry but the manufacturers can't go too far when below freezing because there's only so much they can do with the additives.

I personally don't see it as a matter of "I got away with charging at X temperature with a rate of Y." Only lab testing like this or a knowing very specific usage variables over long period of time will tell how much the life of the cells were degraded. Without those, the manufacturer recommendations are all we have to go by for safe and longest cell life.

[newhaul]

Originally Posted by Checkswrecks

John's making some very good points that I'd like to back up with some data. These test results were from our testing the Boeing 787 cells which are cobalt and not iron based. The reason I'm showing this data rather than the LiFePo4 data I've seen is that the LiFePo4 data were proprietary with similar form to the curves. The plots shown below are from here:
https://www.ntsb.gov/investigations/...sic_Report.pdf
fwiw - When at the NTSB we reverse-engineered the 787 battery, worked out with Yuasa and Boeing what needed to be redacted (blacked out text in 1st chart), and these are in the public domain.

First off, discharge is generally not the problem because the ions are being released from the negative electrode surface layer (anodic SEI), rather than being driven into it. The decrease in efficiency is largely from the characteristics of the electrolyte shown in the next chart.

Reason I'm showing this - If using a high C discharge to warm the cells, pause at least a few minutes before hitting the starter or microwave to let the heat distribute through the cell or you are concentrating the discharge in a smaller area close to the current collectors. (Report Figures 95-103) In the 787 cells it melted portions of the separator film between anode and cathode which degraded the cell real fast.

Here's what is going on with the electrolyte as temperature changes:


Most of your commercial LiFePo4 cells are running thinner electrolytes and have additives which will move these curves slightly to the left. But here is where I am using this chart to emphasize something John wrote. Nobody but the engineers at the manufacturer (who are under proprietary restrictions) will know what these curves are or how they relate to charge rates.

Charge rates -
The third chart is called an EIS curve (electrochemical impedance spectroscopy) and understanding it goes into heavy-duty nerd-dom really fast. Basically EIS testing passes high frequencies between the cathode & anode to characterize how well the ions intercalate into the SEI. When the charge rate at any temp is too high, the ions are not able to work their way into the SEI layer and collect on the surface of the anode as dendrites. (Report Fig. 1) These ions deposit their lithium as metal which can not go back into solution on discharge, so once that tripping point is reached the cell rapidly begins to degrade both in capability and life. In other words the degradation is the point at which the charge rate exceeds temperature-dependent rate of intercalation and a high enough charge rate can cause this at any temperature, including well above freezing.

Enough of these lithium dendrite deposits in a cobalt cell can lead to thermal runaway with sufficient charge rate. The iron-based LiFePo4 doesn't have the reactance of the cobalt so while the fire risk is greatly reduced, the deposits will still very rapidly degrade/kill your expensive cells. (High enough rate or voltage can still lead to runaway in LiFePo4.) As with how fast the curve changes below freezing in the above electrolyte curve, you can have the onset of dendrite formation and that is what the EIS chart relates:


The bottom line in reading this chart is that you do not want to see the charge rate create the C-shaped curve exhibited by the right two plots. In our testing we were using a fixed charge rate and if the charge rate is reduced then the ions are able to work their way into the SEI layer and you'd not have the C-shape develop till colder. We saw the degradation begin with only a few charge cycles in cold tests. (Fig 93) The reason I'm showing the nerdy EIS plot is that it points out how close the bottoms of the left three curves are, how fast the curve changes near freezing (2nd & 3rd curves), then there is a big step over to the -20C (-4F) curve. The data is similar for LiFePo4 cells which basically means that somewhere in that tight spot between the 2nd and 3rd curves is a tripping point for the charge rate.

If you started with minimal charge you could add a little bit more and a little more and then suddenly the rate of lithium deposition would rapidly increase. The point is that the loss of usable life of the cell will match this. This balance between charge rate and temperature is going to be very specific and unique to each manufacturer's chemistry but the manufacturers can't go too far when below freezing because there's only so much they can do with the additives.

I personally don't see it as a matter of "I got away with charging at X temperature with a rate of Y." Only lab testing like this or a knowing very specific usage variables over long period of time will tell how much the life of the cells were degraded. Without those, the manufacturer recommendations are all we have to go by for safe and longest cell life.

while all great information what if anything does it have to do with charging lifepo4 banks at a low ( below .2C ) rate when the bank is sub freezing ? Which was the subject of the I interview with the CEO of Battleborn that I posted.

[newhaul]

Originally Posted by evm1024

Funny how we stopped talking about the Battleborn video and the points that were made.

I guess that I will have to be the one to watch it again and itemize the questions and answers. Then dig out the pop corn

derailed again???

[Checkswrecks]

Originally Posted by newhaul

while all great information what if anything does it have to do with charging lifepo4 banks at a low ( below .2C ) rate when the bank is sub freezing ? Which was the subject of the I interview with the CEO of Battleborn that I posted.

Newhaul -

Sorry but I wasn't responding to your question, as your post was already two pages back in this 484 page thread. I generally go back a page or so but the thread moves too fast to try to read every post between visits. As stated in my opening, I was responding to John61ct's post. However, a couple of thoughts now that I went back to find yours and saw the video that Will Prowse posted earlier today that you referenced. . .

My post was not about Battleborns' product but everything I wrote does apply.

You can see from the electrolyte curve I posted that there's little change from freezing to 25F (-3.88C), which is where the Battleborn shutoff occurs.

Mr. Phares cited the .2C rate only in passing as an example, not as a specific target, and then says that it "might not be great" to charge at 1C and that "you might start to do some damage." (video time 2:20-2:52) As in what I wrote and he said, there's a point between those where the damage begins. How tightly do you control your charge rate over temp? Most people have no clue.

So now that you asked and he did not relate more specifics about charge rate vs temp, what I wrote does go back to the video because is that rate .21C or is it closer to .99C at 25F? Again, only the manufacturer knows.

btw - Since this also keeps coming up on a motorcycle site and does every year I cleaned up what I wrote and put it as a download on my website at www.HowItBroke.com on the Information page.

Regards,
Bob

[newhaul]

Originally Posted by Checkswrecks

Newhaul -

Sorry but I wasn't responding to your question, as your post was already two pages back in this 484 page thread. I generally go back a page or so but the thread moves too fast to try to read every post between visits. As stated in my opening, I was responding to John61ct's post. However, a couple of thoughts now that I went back to find yours and saw the video that Will Prowse posted earlier today that you referenced. . .

My post was not about Battleborns' product but everything I wrote does apply.

You can see from the electrolyte curve I posted that there's little change from freezing to 25F (-3.88C), which is where the Battleborn shutoff occurs.

Mr. Phares cited the .2C rate only in passing as an example, not as a specific target, and then says that it "might not be great" to charge at 1C and that "you might start to do some damage." (video time 2:20-2:52) As in what I wrote and he said, there's a point between those where the damage begins. How tightly do you control your charge rate over temp? Most people have no clue.

So now that you asked and he did not relate more specifics about charge rate vs temp, what I wrote does go back to the video because is that rate .21C or is it closer to .99C at 25F? Again, only the manufacturer knows.

btw - Since this also keeps coming up on a motorcycle site and does every year I cleaned up what I wrote and put it as a download on my website at www.HowItBroke.com on the Information page.

Regards,
Bob

agree with your assessment of the cobalt banks in our 787's and as I stated its all great info.

Now I have noticed on here those of us using Lfp as house banks are usually charging at well below .5C most are actually at in the vicinity of .25C due to the size of their banks any more charge infrastructure makes little sense. I'm the one that can get close to .5C
( 240ah bank ) most of the other guys are running somewhere between 700 and 1000ah that's what I have noticed anyway.

[newhaul]

Latest video from will about Battleborn
https://youtu.be/hOvcgxiSs_Q

[evm1024]

Originally Posted by newhaul

derailed again???

Here is a quick write-up of the main points. I take credit for all omissions, mis-spellings, mis-representations and other errors.

This is worthy of discussion.
-------------------------------


Denis Phares interview – Low Temperature charging of LiFePO4

As the temperature lowers Li ions do not move very fast while electrons still move very fast.
You end up with more electrons than Li ions at the anode. This imblanace causes the Lithium ions to be reduced to lithium metal. “Lithium metal plating”
Plating is bad for a number of reasons. The worst case is Li dendrite formation that penetrates the separator and cause an internal short. Also, plating causes a loss of capacity by making areas of the anode inaccessible.
What is a safe temperature to charge LiFePO4 batteries? It depends – on the charging rate. You can charge a very cold LiFePO4 battery, just you need to charge it very slowly.
Battle Born batteries have a low temp cutoff of 25 F (-3.89 C). Typical Battleborn customers charge at around C/5 (20 amps for their 100 AH battery) and this is a good low temp cutoff with a margin of error built in. Charging at 1C and 25F may be pushing the batteries low temp charging limits.
Low temp discharge is not a problem. That is to say no Li plating happens on discharge at low temperatures. The lithium ions are still slow (and electrons fast) but that leads only to reduced performance at low temps on discharge.
Battleborn derived the 25 degree F cutoff from internal testing and expected use cases.

In regards to high temps – if the battery is being charged and discharged consistently over 100 F then it is suggested that you could see a 30% reduction in the battery life. Over 100 F now and then is OK. The elevated temp causes the natural aging of the battery to happen at a faster rate. “Sped-up calendar aging”
Battleborn does not do any high temp cutout until the internal temp of the battery reaches 140F. (Is is not sure if the 140F cutout is from the internal BMS or otherwise) Battleborn wants to prevent excessive aging of the battery. At internal temps of around 200 F there is a chance to melt the internal separator which will toast your battery. Small cells are safer in that failure mode. (They have placed their batteries in ovens)
Your batteries like to be at the same temps that you like. They accept a wide range of temps – no issues between 40 F and 80 F.
Battleborn wants its batteries to be fully charged. This is due to the passive balancing that they use. At 100% SOC individual cell voltages diverge enough to cause one cell to balance to the others. Also, LiFePO4 is more resilient at “high” voltages than other chemistries.
Battleborn has no problem with leaving their battery at 14.4 overnight. Problems happen when at this voltage for weeks to months.
LiFePO4 does not need to be floated. Float is so that the “charger” can supply house loads while charging is available. Battleborn batteries can float at 100% SOC at 13.4 to 13.8 without undue degrading. “13.8 for months at a time, you are OK”
There is no problem with charging to 100% SOC and then leaving the battery unconnected for months. The discharge rate is 2 or 3% per month (due to internal BMS?). Battleborn says do not “store” at low SOC.
Cold Cranking Amps and Marine Cranking Amps have no meaning with LiFePO4.

[Q Xopa]

Originally Posted by newhaul

Sorry but i would have to take Denis Phares word over your word he has skin in the game and has to honor his company warranty.

Yes Denis has lots of experience and in house testing. Another great information video from Will Prowse.

Unlike others who disagree with him, I haven't seen him say too many unfounded or outrageous things.

I'm not sure where others get their 'truths' from?
I agree before making such statements I want to know what 'qualifications' are behind them.

[Q Xopa]

Originally Posted by john61ct

As long as not **too** far below, but me, that low I'd stay under 0.1C, or pre-warm the bank.

Like so many issues, hard to nail down hard numbers when the damage done is lost cycles off the back end.

Actual now-noticeable damage, and certainly rendering the bank instant scrap, only happens long past the point where longevity was being reduced.

Most of the recent research studies show the effect very clearly without destroying the cells, but

each chemistry, each model battery will have different numbers.

Only the minimum "no charging anywhere near this temp or instant scrap" spec is on the data sheet.

In truth you usually can charge at a low C-rate around 0°C

But in general, being conservative I'd start pre-warming below 10° if going over 0.1C

25° should be fine for 0.2-3C

Faster charging than this, anybody's guess, impossible to detect the factors leading to lost lifecycles without forensic autopsy, studies show different numbers for even slightly different chemistries.

So for 0.5C and above, I'd go to 30° first to maximize longevity.

Of course in many contexts people just take the hit, if they're even aware of the issue.

Do you have any references to support this opinion?

[Dockhead]

I use my boat in cold places, yet I can't imagine this ever being a problem. Charging at -5C? This temperature is not anywhere close to being inside the operating range of the boat itself. As long as the boat is in water (as opposed to being frozen in sea ice or on the hard in very cold weather), it will be at least plus something inside, even if no supplementary heat is being provided. And in normal operating conditions, we will be heating the boat in cold weather and it will be at least +10C.


So for actual practice, I can't imagine anyone would need anything other than a temperature sensor in the battery box and mental note not to charge at high rates on the odd occasion that the batts might be at less than +5C.


Since leaving the batts in low SOC is not a problem I can't imagine you'd ever be in such a big hurry to charge if you are coming onto the boat after leaving it without heating and it gets unusually cold inside. You'd heat it up first, then think about charging. In an emergency, you could do a short heavy discharge first to warm the batteries up, but can't imagine why you would ever need that.


Or am I missing something?

[newhaul]

Originally Posted by Dockhead

I use my boat in cold places, yet I can't imagine this ever being a problem. Charging at -5C? This temperature is not anywhere close to being inside the operating range of the boat itself. As long as the boat is in water (as opposed to being frozen in sea ice or on the hard in very cold weather), it will be at least plus something inside, even if no supplementary heat is being provided. And in normal operating conditions, we will be heating the boat in cold weather and it will be at least +10C.


So for actual practice, I can't imagine anyone would need anything other than a temperature sensor in the battery box and mental note not to charge at high rates on the odd occasion that the batts might be at less than +5C.


Since leaving the batts in low SOC is not a problem I can't imagine you'd ever be in such a big hurry to charge if you are coming onto the boat after leaving it without heating and it gets unusually cold inside. You'd heat it up first, then think about charging. In an emergency, you could do a short heavy discharge first to warm the batteries up, but can't imagine why you would ever need that.


Or am I missing something?

Nope not missing anything and thats my take on us . Cold ( sub freezing ) is not an issue we need ro really pay much attention to .

[Checkswrecks]

Originally Posted by newhaul

agree with your assessment of the cobalt banks in our 787's and as I stated its all great info.

Now I have noticed on here those of us using Lfp as house banks are usually charging at well below .5C most are actually at in the vicinity of .25C due to the size of their banks any more charge infrastructure makes little sense. I'm the one that can get close to .5C
( 240ah bank ) most of the other guys are running somewhere between 700 and 1000ah that's what I have noticed anyway.

The battery design and manufacturing world is pretty small so I've talked to Denis at a conference and a couple of trade shows. Nice guy, definitely an engineer (I R 1 2 ), knows his product, and from what I've seen as a non-customer their product seems to have all the right stuff in a conservative design. A client is probably going with a number of them in a design we are working on.

Denis said in the video the same that you noticed and I'm just going to repeat, in that most people are not charging at high rates and that's a good thing. It makes sense when one realizes that even with a 30A shore charger on a single 100 Ah battery, the rate is only .3C. Absolutely no problem when above freezing for pretty much any battery.

If charging off of solar, I've noticed a circular relation between batteries and panels which keeps rates down in the same range. People using solar tend to have 200+ Ah of storage so getting more than .2C would need 40+ amps off the panels. Most monohulls just don't have the real estate and cats with big roofs tend to have all the electric appliances so already have bigger battery banks.