Model 3 SR+ LFP Battery Range, Degradation, etc Discussion

[FARKLE!]

I daily charge to 100%, and I'm now only getting 249 miles at 100%. It started declining from 253-254 around 2 weeks ago. 2400 miles on the odo. I'm going to stop daily charging to 100% and instead charge to 70% - 80% and see if there is any improvement. But yeah, I agree...most people probably don't need to charge to 100% every night.

Also, I have a mobile service appt tomorrow, will ask them if they've seen this decline on other LFPs around me.

EDIT: I just got a phone call from Tesla Service. They ran diagnostics on my battery, and didn't find anything wrong or any actual degradation. He blamed the lower displayed range on algorithms, third party apps, and driving habits. He said the battery was 'just a hair' below 100%. He recommended that I uninstall any third party apps, as well as change my Tesla account password because they are preventing the car from going into sleep mode (I think he said about 60% of the time? ). Basically it sounds like the apps could be keeping some systems on, sapping a little bit of energy? Half of what he said was basically Greek to me, but that was my overall take.

Also, he said while you can charge to 100% daily, he doesn't recommend it unless you really need the range. And that even the LFP pack can 'learn' a lower range. Also recommended me to drive to below 30%, then charge to 100%. I don't think I've ever gotten below 30% except for when I took delivery (the car only had 70 miles of range)

 

[Dave EV]

I am attaching two captures. 99.4% SOC with 31kW regen and another one with 85% SOC with 12kW regen.

The difference was the battery pack temperatures, 62F vs 47F average temps. Since I timed the full charge before I leave home, I had 62F and after 8.5 hours parked, temp dropped from 66F (@work) to 47F (leaving work).

When I got home that day, pack temp was 55F with 22kW regen.

Some interesting tidbits from your screenshots:
1. Max pack voltage is ~355V.
2. Max cell voltage is ~3.35V
3. Should be 106 cells in series.

The max cell voltage here is a bit lower than what most references to max LFP voltage are - 3.5-3.6V. So quite possibly 3.35 is OK for extended periods of time with these cells, but it's hard to say for sure unless you actually test the cells. Regardless of the lithium chemistry, lower states of charge will result in slower rates of capacity loss - it's just a question of how much.

 

[pipp]

From what I understand from @AAKEE, it is probably (he is not 100% sure on this latest version of LFP) detrimental to leave LFP batteries at 100%, for calendar aging. I understand the need to do it periodically for LFP, and I know Tesla does not limit or recommend against it. And I understand limited regen is nearly a non-issue on these packs (are you still finding that to be true in cooler weather near 100%?).

But it sounds like people are routinely doing it (rather than once a month or whatever), even if they don’t need the energy daily? Why?

I guess it will probably matter less than for NCA cells, but still…if I had one of these cars, and I was doing short daily trips, I’d daily charge (though deep cycles are apparently not an issue for LFP) it to 60-70% most of the time, and only occasionally bump it to 100%, to get the BMS estimate correct.

I say this as someone who frequently used 90% (there were reasons for it - free charging at work being one) as a charge level for my NCA, but after 10% capacity loss, I’ve changed to 60-70% daily level, with daily charging to ensure shallow cycles, because for NCA it seems like it can only help.

My guess is people who leave their LFP batteries at lower SOC most of the time will see generally better results long term (but there is going to be a distribution, and some people who do this might still have problems). Obviously if you need the full range, you just don’t worry about it and probably LFP will still do much better than NCA in that daily 100% to 0% use case.

Sorry AlanSubie4Life, I did go rogue about charging to 100% every night, and I do sweat a little bit over possible long term degradation that might cause to my battery. But I’m not committed to this approach I have taken, and may swerve away from the dark side in the near future. I think I would feel better about charging to less than 100% if I had a Tesla wall connector installed and was able to charge the car up from 0 to full in any given single night. And especially now that you have pointed out the charging inefficiency of level 1, I am further motivated. But I still might dawdle a bit more out of inertia than any other reason…

 

[Streetfish]

Sorry AlanSubie4Life, I did go rogue about charging to 100% every night, and I do sweat a little bit over possible long term degradation that might cause to my battery. But I’m not committed to this approach I have taken, and may swerve away from the dark side in the near future. I think I would feel better about charging to less than 100% if I had a Tesla wall connector installed and was able to charge the car up from 0 to full in any given single night. And especially now that you have pointed out the charging inefficiency of level 1, I am further motivated. But I still might dawdle a bit more out of inertia than any other reason…

Directly from the Tesla website:

Charge the battery to the appropriate charge limit for your vehicle based on the installed battery. To adjust the charge limit for your vehicle, open the Charging screen on your touchscreen and then touch ‘Set Limit’ or open the Charging screen in your mobile app and drag the slider.

• For Rear-Wheel Drive vehicles,
  If the image of the battery displays ‘50%’ and ‘100%’: Tesla recommends that you keep your charge limit to 100%, even for daily use, and that you also regularly charge your vehicle to 100% . If your vehicle has been parked for longer than a week, Tesla recommends driving your vehicle as you normally would and charge to 100% at your earliest convenience

 

[Dave EV]

Directly from the Tesla website:

Charge the battery to the appropriate charge limit for your vehicle based on the installed battery. To adjust the charge limit for your vehicle, open the Charging screen on your touchscreen and then touch ‘Set Limit’ or open the Charging screen in your mobile app and drag the slider.

• For Rear-Wheel Drive vehicles,
  If the image of the battery displays ‘50%’ and ‘100%’: Tesla recommends that you keep your charge limit to 100%, even for daily use, and that you also regularly charge your vehicle to 100% . If your vehicle has been parked for longer than a week, Tesla recommends driving your vehicle as you normally would and charge to 100% at your earliest convenience

You have to read that very carefully. It sounds like some RWD cars (LFP?) don't come with a slider and come with either a 50% or 100% toggle.

Since you can't choose something in between 50 and 100%, they suggest regularly charging to 100%.

Honestly, I still wouldn't charge to 100% daily on a LFP car with only 50/100% SOC charge limits unless I found the pack to get out of balance significantly.

 

[AlanSubie4Life]

Directly from the Tesla website:

Charge the battery to the appropriate charge limit for your vehicle based on the installed battery. To adjust the charge limit for your vehicle, open the Charging screen on your touchscreen and then touch ‘Set Limit’ or open the Charging screen in your mobile app and drag the slider.

• For Rear-Wheel Drive vehicles,
  If the image of the battery displays ‘50%’ and ‘100%’: Tesla recommends that you keep your charge limit to 100%, even for daily use, and that you also regularly charge your vehicle to 100% . If your vehicle has been parked for longer than a week, Tesla recommends driving your vehicle as you normally would and charge to 100% at your earliest convenience

Yes, it's interesting. Tesla recommends that, possibly due to better BMS estimates.

It'll be interesting long term to see what is actually best for longevity. Tesla doesn't necessarily care too much about that - they want a car that is usable for owners, with the least complexity, and they want to ensure that the battery life is not compromised so that capacity loss doesn't go below 70%.

I think this recommendation is likely pretty optimal for that. But to be clear this recommendation does not imply in any way that this is best for battery capacity loss. It's very very good for many other reasons though! LFP is a great, simple chemistry for drivers.

Remember Elon Musk implied that SOC % didn't matter much (honestly in this tweet he kind of didn't address the topic), which hasn't proved to be true (though there's debate about whether it is pack-to-pack variation or other usage factors). And this type of use for good battery life (meaning capacity loss, not time to failure) is not supported by the data.

https://twitter.com/x/status/1118225459416682496

To be clear, I'm not privy to Tesla's chemistry details and I'm guessing. It's just that the data on LFP and NCA don't support these use scenarios as being optimal for capacity loss.

In any case mostly you should just charge to what you need and enjoy your car, and be aware of the underlying details if you care to explore them.

It sounds like some RWD cars (LFP?) don't come with a slider and come with either a 50% or 100% toggle

Yes it's an odd phrasing.

 

[Baluchi]

From what I understand from @AAKEE, it is probably (he is not 100% sure on this latest version of LFP) detrimental to leave LFP batteries at 100%, for calendar aging. I understand the need to do it periodically for LFP, and I know Tesla does not limit or recommend against it. And I understand limited regen is nearly a non-issue on these packs (are you still finding that to be true in cooler weather near 100%?).

But it sounds like people are routinely doing it (rather than once a month or whatever), even if they don’t need the energy daily? Why?

I guess it will probably matter less than for NCA cells, but still…if I had one of these cars, and I was doing short daily trips, I’d daily charge (though deep cycles are apparently not an issue for LFP) it to 60-70% most of the time, and only occasionally bump it to 100%, to get the BMS estimate correct.

I say this as someone who frequently used 90% (there were reasons for it - free charging at work being one) as a charge level for my NCA, but after 10% capacity loss, I’ve changed to 60-70% daily level, with daily charging to ensure shallow cycles, because for NCA it seems like it can only help.

My guess is people who leave their LFP batteries at lower SOC most of the time will see generally better results long term (but there is going to be a distribution, and some people who do this might still have problems). Obviously if you need the full range, you just don’t worry about it and probably LFP will still do much better than NCA in that daily 100% to 0% use case.

I'm trying to rotate mine each day between 60, 70, 80, 90, and 100% charges. That way the car gets opportunities to sleep at various SoCs. It's not a hard-and-fast rule I'm following, and I'll always adjust upward if I think there's a remote chance I'll need the extra range. But going to 100% every day for me isn't necessary, and I'd rather not do it all the time. This is especially true for the weekends when I might not drive the car for a day or two--don't want to let it sit with a full battery for that long.

 

[USBSeawolf2000]

Regardless of the lithium chemistry, lower states of charge will result in slower rates of capacity loss - it's just a question of how much.

LFP is not sensitive to depth of charge/discharge. You'll get 3,000 cycles doing 0-100% and retain 90% of the original capacity. For me, l do 20-100% cycle ~per week so 52 cycles per year so it is not worth worrying about. NCA and NMC are very sensitive to DOD.

LFP is more sensitive to the temp but even 35 deg C will get over 3,000 cycles and retain 80% original capacity. Would be a good idea to park in the shade or garage.

The rate of charge/discharge (C) is interesting as LFP likes 1C and 2C more than 0.5C. This would mean LFP will last longer with Supercharging at 55-110 kW rather than 0.5C at 27 kW. Supercharging at 3C 165kW (I think max seen is 157kW) is bad.

Tesla WC is 7.8 kW so I charge at 0.14C.

 

[AlanSubie4Life]

LFP is not sensitive to depth of charge/discharge.

I don’t think he was implying that. The way I read that, he was not talking about the use of the battery. He seemed to be addressing the effective storage SOC (which is presumably a major contributor to any capacity loss, since that’s what is happening most of the time, though clearly depends on magnitudes of the effects).

 

[USBSeawolf2000]

I don’t think he was implying that. The way I read that, he was not talking about the use of the battery. He seemed to be addressing the effective storage SOC (which is presumably a major contributor to any capacity loss, since that’s what is happening most of the time, though clearly depends on magnitudes of the effects).

I get that charging to 100% would have the battery sitting at higher SOC through out it's life. I don't think it is something to worry about as there are other factors that could affect more.

I have 13 years old Black & Decker VPX tools with A123 LFP batteries and they still work! My usage is to fully charge it and recharge when needed, pretty much 0-100%.

 

[AlanSubie4Life]

there are other factors that could affect more.

It’s user dependent. For a high mileage user, calendar aging is less important.

Separately LFP may have different calendar aging than NCA. (There is data that suggests it is the same order of magnitude though.).

If you look at the datapoints in the various range loss stickies, I think the balance of the evidence suggests that calendar aging is actually the dominant factor for most users. It is hard to say though, and tough to separate from temperature factors, pack-to-pack variation, and other factors that are not easy to determine for each report.

Unfortunately TeslaFi does not allow plotting in three dimensions (would be nice to see a surface plot of capacity loss vs. age and mileage).

Whether this calendar aging will actually be dominant for LFP, and how strongly dependent it will be on temperature, is obviously TBD. What we do know is that to the extent it exists, lower storage SOC slows it down.

I don't think it is something to worry about

I’m not advocating worry. I think users should not worry about it, and simply be aware of what is likely the best for a given goal. There are lots of different goals and they are not the same for everyone.

 

[Dave EV]

I have 13 years old Black & Decker VPX tools with A123 LFP batteries and they still work! My usage is to fully charge it and recharge when needed, pretty much 0-100%.

A123 batteries are awesome from the data that I can recall. High power, long cycle life. But they were expensive. Not sure what happened to them after they went bankrupt. How much capacity is left in them after all this time?

On the topic of LFP capacity loss over time, found a nice study from NREL using Sony US26650FTC1 Lithium Iron Phosphate cells:

https://www.nrel.gov/docs/fy18osti/70616.pdf

Temperature and SOC have major influences on rate of capacity loss, as expected. The lower the temperature and the lower the SOC, the slower the rate of capacity loss.

 

[Dave EV]

Another interesting chart from the above study - highlights how stresses increase on the battery during storage significantly at SOCs above 75% for these LFP cells. I would opt to keep the storage SOC at 70% or lower. Interestingly the difference between 90-100% is not significant. Storing this LFP at 25C and 100% SOC is the same as storing it at 75% and 35C.

Also interesting - cycling the battery when cold (below 25C) also increases the rate of capacity loss of the battery and gets very bad at 0C. Unsure if that's a result of charging or discharging, though - I'd have to guess it's a result of lithium-plating that can occur when charging at very low temperatures.

 

[AlanSubie4Life]

Unsure if that's a result of charging or discharging, though - I'd have to guess it's a result of lithium-plating that can occur when charging at very low temperatures.

If they aren't controlling for that in this test and aren't warming the batteries for charging during their cycles, it also would invalidate the tests at 10C/15C. As I understand it, you can't charge batteries at arbitrarily high rates when they are cool, even if they are above freezing. Tesla's very careful about this. It's not like something magical happens at freezing (obviously at some temp the electrolyte will freeze which would be "magical", but that is very unlikely to be at 0C). There's just kind of a curve as far as I can tell, where the rate-limiting step of intercalating gets slower and slower (and exponentially slower as temperature decreases with a "knee" in the vicinity of 0C). I could be wrong but that's how I understand it.

In any case Tesla has looked at this carefully, so I definitely think the cycle tests in the second picture are probably not really too applicable (I'd look at just the best case curves, and maybe the extremely high temp results are also somewhat applicable - though Tesla will help a bit there as well if they attempt to cool the battery).

 

[AAKEE]

From what I understand from @AAKEE, it is probably (he is not 100% sure on this latest version of LFP) detrimental to leave LFP batteries at 100%, for calendar aging. I understand the need to do it periodically for LFP, and I know Tesla does not limit or recommend against it. And I understand limited regen is nearly a non-issue on these packs (are you still finding that to be true in cooler weather near 100%?).

I havent read very much about the latest LFP batteries. I did read ONE reserch reports that had som different conclusions compared to all other/earlier I have seen. That reports shows results in the line of what Tesla says.
On the other hand, it is only one report. And they did some things in the test setup that makes one think that they do not have datapoints enough to make the conclusions certain. For example, the test setup had three datapoints in the temperature range, and very high up(40/47.5/55 degrees C), and they only hade three datapoints in the SOC range, and these all was with 55C temperature.
All conclusions is made up by constructing a graph where these three datapoints match. This isnt such a dumb idea, but three datapoints is not good enough to be sure that the graph is valid. At least one more point would be needed to validate the assumed graph, and they didnt do that. I would like to have seen two points as checks that veryfi the assumed graph.
Also, datapoints very high in the temperature range(40-55C) might not be valid for true research results at 10-20C etc.

The research sure looks like the tested LFP like to stay at high SOC and that medium SOC causes more calendar aging than 100%.
If this is true, or correct….I do not know. As for the things I state in my posts, nothing comes from a single research report. I draw my conclusions from several research reports where it is possible to see that the same behavior is valid in these.
I have seen some wrong conclusions in some reports, most often because of ”faulty” test setups like to few datapoints that ”covers” behaviour. One example is the clear step in calendar aging around 57-60% SOC for NCA cells that isnt visible if the test setup has datapoints at 40 and 60% SOC without any points between.

A side note: The report I mentioned did use three cells at every point, and I think all degradation etc. was within 2% or less between the cells. This is about the same as other reports, and it show us that batteries in general is very predictable and that the ”battery lottery” shouldnt really exist. Unless Tesla/Panasonic is a worse battery manufacturer than most other.
I personally guess that the differences mainly comes from the BMS that is affected from a lot of things which set us up with small differences in NFP.

The LFP test

 

[Lephturn]

LFP is not sensitive to depth of charge/discharge. You'll get 3,000 cycles doing 0-100% and retain 90% of the original capacity. For me, l do 20-100% cycle ~per week so 52 cycles per year so it is not worth worrying about. NCA and NMC are very sensitive to DOD.

LFP is more sensitive to the temp but even 35 deg C will get over 3,000 cycles and retain 80% original capacity. Would be a good idea to park in the shade or garage.

The rate of charge/discharge (C) is interesting as LFP likes 1C and 2C more than 0.5C. This would mean LFP will last longer with Supercharging at 55-110 kW rather than 0.5C at 27 kW. Supercharging at 3C 165kW (I think max seen is 157kW) is bad.

Tesla WC is 7.8 kW so I charge at 0.14C.

I know I'm off-target here - but for NCA packs (I presume that is what I have in my Sept. 21 M3SR+ (I know it is not LFP) this indicates that keeping the pack between 40-60% SOC leads to vastly improved number of cycles, while LFP packs go the other way and last much longer with 20-80 and 0-100% SOC charging. Am I reading that right? (Middle set of graphs)

It also shows that for LFP charing at 2C results in a longer life, but frankly the massive charge cycles LFP is capable of makes this less important. For those of us with NCA packs there is a significant improvement in capacity by charging at 2C. I was under the impression that supercharging would reduce life, but this implies that charging at 2C is significantly better, and doing so between 40-60% charge would compound those benefits.

Does this mean that for LFP packs you should Supercharge at 2C to 90-95% regularly and can run it down to a much lower state of charge? For NCA packs I hope you live close to a Supercharger because it's best to charge from 40-60% at 2C as much as possible, if I understand these graphs.

 

[AAKEE]

No, lower charging speed is better for NCA.( for most batteries also).
Slow charging is often defined by 0.25C or less, this keeps the cells fine. At least 1C or less for longelivety.

NCA will live longer the further down in SOC the cycle is.
A cycle between 25-35% causes roughly 1/4 of the wear of a cycle at 65-75%, same currents and temps etc.

 

[USBSeawolf2000]

Also interesting - cycling the battery when cold (below 25C) also increases the rate of capacity loss of the battery and gets very bad at 0C.

Yup, that's why regen is limited. Per my screenshot, 18.4 kW regen at battery temp of 50 deg F (10 deg C). That's 0.33C.

The test in the paper was done at 1C and 10 deg C got 2,700 cycles with 10% capacity loss. Tesla is being very cautious by allowing only 0.33C.

I know I'm off-target here - but for NCA packs (I presume that is what I have in my Sept. 21 M3SR+ (I know it is not LFP) this indicates that keeping the pack between 40-60% SOC leads to vastly improved number of cycles, while LFP packs go the other way and last much longer with 20-80 and 0-100% SOC charging. Am I reading that right? (Middle set of graphs)

Yup, the middle row and right most is the NCA. It is very sensitive to SOC as it ranges from 250 to ~1,500 cycles.

1,500 cycles sounds good but you are only getting 20% depth (40-60%) so you are only getting 300 equivalent full cycles.

700 cycles at 20-80% yields 420 equivalent full cycles.

 

[AAKEE]

Yup, that's why regen is limited. Per my screenshot, 18.4 kW regen at battery temp of 50 deg F (10 deg C). That's 0.33C.

Low charging rates do not really hurt the battery at low temps( but lithium ions should not be charged at freezing levels!)

Cycling tests of NCA with 1C/ charge and 1C discharge show mire or less no change in cyclic degradation for NCA in at least some tests. And 1C is quite high.
Max charge power with the UMC for a M3 is 11kWh, thats about 0.15C.
Tesla heat the battery before it starts charging the battery.

I left home today with 15.7C cell temp and got 73KW regen. I had the car preheated at work before the drive home with the message ”The car has the wished temperature”. Cell temp 6C, and max regen 52.9kW from the CANBUS. The regen test gave me 56kW, but the cell temp had of course climbed slightly, allowing a higher regen. Cell temp 13.5C = 72.4kW.

When I leave home and the charging was ready about 1-1.5h earlier, the cell temp often is about 17C. The heat pump use the battery heat and it comes down to about 12-13C before I am at work.
When I lesve work, cell temps often 5-6 degrees preheated snd reaches about 12C whennat home. It seems the battery is heated to that level during a drive in winter. ( battery inlet has higher temp then outlet and batt cell temp). As it cools towsrds 12C and is heated towards 12c, 12C seems like the target with SOC about 35-45%.
Most certainly not dangerous for the battery if Tesla set 12C as the target.
Oke year ago, when I got the car, the same drive stabilized the battery at about 6-7C. Im sure they increased the temp target as the car got really slow with too cool battery. I do not think Tesla thought 6-7C was bad for the battery.

Attached pics from todays drive.

 

[Lephturn]

Yup, that's why regen is limited. Per my screenshot, 18.4 kW regen at battery temp of 50 deg F (10 deg C). That's 0.33C.

The test in the paper was done at 1C and 10 deg C got 2,700 cycles with 10% capacity loss. Tesla is being very cautious by allowing only 0.33C.

Yup, the middle row and right most is the NCA. It is very sensitive to SOC as it ranges from 250 to ~1,500 cycles.

1,500 cycles sounds good but you are only getting 20% depth (40-60%) so you are only getting 300 equivalent full cycles.

700 cycles at 20-80% yields 420 equivalent full cycles.

It seems to me that in the middle row, right most, it shows much faster degradation at .5c while 2C seems like the optimal charge rate. Or am I reading that wrong? Probably doesn't really matter much since I don't have a DC fast charger capable of 2C in my home or anywhere easily accessible.

So the overall lesson is to keep SoC as low as possible given my use case and driving needs. The big change is that previously I had thought that the best way to go was to keep the car at 60-80% SoC as much as possible. I work from home and don't drive more than 20k most days, so I can easily let the car sit at 20-55% and be able to safely get to a Supercharger if I suddenly need to drive a longer distance.