Range Loss Over Time, What Can Be Expected, Efficiency, How to Maintain Battery Health

[AAKEE]

Meaning when my car is sleeping in the garage am I only getting calendar aging?

Yes.

And as we can not stop time, calendar aging happens all the time, as we can not stop the time.

Or another case would be if I am supercharging and flowing energy in to a hot battery and subsequent drive after, am I only getting cycling aging at the elevated battery temperature? Or does time at temp still apply here as a double whammy.

The increased calendar aging is there for a preconditioned battery during supercharging.

But as preheat + charging and after heat might be only for three hours, you get 3/8760* = 0.00034 parts of one year of valendar aging at that temp.
If the sverage SOC is 80% durijg this three hours and the avg cell temp is 40C, one year would cause about 9-10 % degradation or so, which give us 0.003% of calendar aging induced degradation for tha three hours.
Not heating the battery might cause lithium plating which is not goig at all. Of course Tesla counters this by reducing the charging speed but it sure looks like non heated batteries suffer from that.
I always precondition.

A preconditionen 48-50C hot battery will suffer the least from supercharging.
The other option is to not supercharge.

*One year is 8760hrs

Another case would be if I have my car in garage and sentry mode on, car is pulling some load, am I only getting cyclic aging?

No, calendar aging is still present and you add a little part of a cycle each day with sentry mode on.
Obe year of constant sentry would perhaps add 20 Full Cycle Equivalents each year but these are very nive to the battery as it is s very low current ( C-load)

Seems its a blended aging effect but wonder if there is even further best practices I can do to reduce any aging further even beyond lowered SoC, especially where I live in the hot summers. The data research always run tests on calendar aging or cyclic aging but not together and its impact, if any.

There is calendar sging during cyclic tests
Some tests specify the difference.

As batteries are quite good tou vcan not perform cycle and calendar aging tests in the same rate as the cars use the battery. It would take as long as it takes for a EV to break the battery.

Calendar aging is tested for 1-2 years and then the commonly accepted formula for calendar aging is used to project the calendar aging further on.

Cyclic agijg is performed in a much faster way than real driving (many cycles each day), minimizing the impact from calendar aging.

Some researchers combine this by deducting calendar aging from the cyclic test and then combine the net from cyclic with the calendar aging.

Then we get graphs for how a battery will hold up long term.

Each of this, of course show us that low SOC and small cycles wins tge long life contest.

 

[arghx7]

I posted this in another thread, might as well add it here:

Here's my battery degradation report from TeslaFi for my 2019 long range AWD Model 3, starting maybe 4 months after I got it through today. You can see the logarithmic scale on the X axis. Basically no degradation (and a BMS recalibration I think) for the past year and a half or so. Also note that I changed my daily charging to max 70% over a year ago, as I typically use no more than 30-40% in a day and I want the battery to last as long as possible. At this point I have it charge from the grid in the morning, with a scheduled departure of 7 AM.

Factory rated EPA miles was 310, and I'm at 285, so about 8% degradation.

 

[AlanSubie4Life]

Factory rated EPA miles was 310, and I'm at 285, so about 8% degradation.

Technically 10%, of course.

 

[AAKEE]

It would be interesting to see what you find

Here’s a recent research report that carries Jeff Dahns name.

ShieldSquare Captcha

They cycled a bunch of NMC batteries, both 0-25, 0-50, 0-75 and 0-100 (“LC, Low cutoff” that 0% SOC at end of discharge. (They also cycled with UC Upper cutoff, 100-0, 100-75, 100-50 and 100-0%.)
Not really anyting strange in this report, it follows just about the same path of findings as the other good ones.

The picture shows LC tests:
Green is 100-0%
Red is 75-0%
Blue is 50-0%
Black is 25-0%

The black, blue, and red data points near the top of the panel are the data for the 100% DOD C/10 checkup cycles.

A lot of cycling down to 0% and the 25-0% did best as we can see below.

 

[johnny_cakes]

Here’s a recent research report that carries Jeff Dahns name.

ShieldSquare Captcha

They cycled a bunch of NMC batteries, both 0-25, 0-50, 0-75 and 0-100 (“LC, Low cutoff” that 0% SOC at end of discharge. (They also cycled with UC Upper cutoff, 100-0, 100-75, 100-50 and 100-0%.)
Not really anyting strange in this report, it follows just about the same path of findings as the other good ones.

The picture shows LC tests:
Green is 100-0%
Red is 75-0%
Blue is 50-0%
Black is 25-0%



A lot of cycling down to 0% and the 25-0% did best as we can see below.
View attachment 932491

Great work, thank you!

 

[AAKEE]

Great work, thank you!

I just read a couple of other reports from research groups that included Jeff Dahn.

They cycled cells in lower Voltage regions, down to 3.0V in extreme temperatures and the results where ”excellent” according to that research group. (I think so also, good work and good findings).

Going below 30% didnt kill these cells either.

 

[Candleflame]

I just read a couple of other reports from research groups that included Jeff Dahn.

They cycled cells in lower Voltage regions, down to 3.0V in extreme temperatures and the results where ”excellent” according to that research group. (I think so also, good work and good findings).

Going below 30% didnt kill these cells either.

There was some other trial where they discharged the cells at 2C (Model 3 max is like what... 1.5C driving 250km/h?) and even discharging at 2C from 20 to 10% was better for longevity than discharging 50 to 40% at 2C.

 

[DownshiftDre]

I just read a couple of other reports from research groups that included Jeff Dahn.

They cycled cells in lower Voltage regions, down to 3.0V in extreme temperatures and the results where ”excellent” according to that research group. (I think so also, good work and good findings).

Going below 30% didnt kill these cells either.

Please post the reports if you can! Have a flight tomorrow and will make for good read.

 

[DrChaos]

FYI the link to the paper (open access) is : ShieldSquare Captcha

The paper is remarkably clear for a professional scientific article about what is known well and what is not. The explicit discussion of hypotheses and internal thinking is not usually written in so much detail in a written article (which often just reports results) as compared to personal discussions or conference talks.

An interesting fact is a simple model proposed by someone they cite (Deshpande and Bernardi)

is capacity
Q(t) = Q(0) * (1 - n*B*DOD^2 - K*sqrt(t))

n is cycle number and DOD is depth of discharge as a fraction (i.e. <= 1). So low depth of discharge really really helps on cyclic aging. The calendar aging model is with square root of time.

Any simple model is a very oversimplified approximation to the complexity of the chemistry actually happening in a real cell. It's remarkable how much is not fully understood despite decades of work and lots of effort. By contrast, nuclear physics was generally solved (for good enough approximations in common uses) in 25 years.

Another model they propose is

Q(t) = Q(0) * (1 - A0 * DOD + B) * sqrt(t) as fitting the data better to the previous model. Note that A0 and B are refitted---for what we care about the change in A0 and B depending on other parameters also matters but these aren't discussed as much. There's of course a time averaged state of charge which should be in there somewhere, but that is indirectly related to DOD.

Furthermore they go on to say that there's another additional 'impedance' term necessary for capacity loss due to impedance growth (different from the SEI film growth that's the usual loss of capacity mechanism on the graphite), which becomes more important at higher charging rates, and that model appears to be the primary result of this paper. That model depends on coefficients which are not observable outside a lab, the general trend though is for higher charging speeds to make more of a contribution.

Figures 8, 14, 15a and 15b showed that a simple empirical model based on the square root of time model, Eqs. 4 and 5, was more accurate at describing the data at C/10. This model was able to describe the data by assuming that the parameter A was linear vs DOD. At C/5 and C/3, it was shown that an impedance term needed to be included to account for the capacity loss that resulted from impedance growth. While the impedance term used in this work doesn't account for non-ohmic contributions, it does give a useful first-order approximation of capacity loss due to impedance growth.

Note that C/10 means recharging the whole battery pack over 10 hours. That's a typical L2 charging rate. It is compatible with the notion that too much supercharging will lower capacity faster.

 

[yesimon]

Lithium batteries used in EVs seem to have less degradation than batteries tested in the lab with continuous charge-discharge cycling, even temperature-controlled and comparing the same number of equivalent full cycles. I think this is likely due to regen braking breaking up the discharge reaction and adding a bit of charge, thus mimicking some of the benefits of shallow discharge cycles, which might be inherently protective of the battery.

If you actually think about it, driving an EV from 100%-0% is actually charging/discharging greater than 1 full charge cycle due to regenerative braking. You would expect degradation to be even worse, not better, than lab-tested batteries.

 

[johnny_cakes]

I just watched a presentation from Jeff Dahn (

)

He did not say anything strange.
He showed the work on lithium batteries that could do very many cycles.
All was tested down to 3.0V (which was considered 0% on those cells). He is impressed himeself.

View attachment 932309

I watched this as well. Interesting how Dahn got NMC to degrade much less than LFP with the right mix of chemistry, and when treated right, his battery cells lasted longer than everything around it, including the charger.

 

[AAKEE]

Lithium batteries used in EVs seem to have less degradation than batteries tested in the lab with continuous charge-discharge cycling, even temperature-controlled and comparing the same number of equivalent full cycles.

Nope, I would say they do not. I see good correlation between the research and the real world.

If looking at cycling only, for example this test of panasonic NCA: Look at these graphs:

3.7V = 50%
3.9V = 70%
4.1V = 90%

70-90% charging target and 20-40% depth of discharge degrades about 10-12.5% after 1000 FCE cycles.
1000 FCE cycles on a LR would be about 400.000 km.
This is about 0.25% per 10K km or about 0.5% per year for an average car.

If we add calendar aging to this, for example two years

Two years with mostly 90-60% SOC at 25C would be about 7.8% ( 10 months recalculated to 24).

If we combine 7.8% with 60K km cyclic aging (0.25%/ 10K km) we get 9.27% degradation.
M3P 2021 had a 82 kWh battery but the EPA test (and most cars initial value) was about 81 kWh.
81kWh -9.27% is 73500 Wh/km which correlate to 463km range.
(73500/ the charge constant, 158.7 Wh/km)
This is our calculation for a average M3P at two years age and 60K km.

A look at Teslafi graph, at 60K km and about two years age, the average range in Teslafi is very close to this.

If you actually think about it, driving an EV from 100%-0% is actually charging/discharging greater than 1 full charge cycle due to regenerative braking. You would expect degradation to be even worse, not better, than lab-tested batteries.

You should see it the other way, regen makes the daily cycles smaller.
I have about 10% regen so the cycles had been 10% larger if no regen.

The cycles we drive get from charging get very smal sub-cycles, and very small cycles cost very low degradation.

 

[AAKEE]

I watched this as well. Interesting how Dahn got NMC to degrade much less than LFP with the right mix of chemistry, and when treated right, his battery cells lasted longer than everything around it, including the charger.

There is more than one research finding with very good results.
It always takes time to enter the market and sometimes the findings is not usable in real life.
So, while it looks good, we do not need to jump too high until its mounted below the floor in our cars.

 

[MP3Mike]

A look at Teslafi graph, at 60K km and about two years age, the average range in Teslafi is very close to this.

You do realize that that "fleet" average is based on only 8 vehicles right? That seems like a fairly small fleet/sample size.

 

[dansev]

This thread is pretty amazing. I wish I could spend the time to actually dig in to what you engineer-types are saying. But I would like to summarize my take-aways in lay terms. Can somebody confirm that I have it right?

Best practice is ultimately to charge your car to the lowest percent necessary to allow you to drive where you need to go. Range anxiety shouldn’t be dismissed (it’s can’t be so long as humans are driving.) Charging should take place as close to departure as possible. Better to have a very low battery SoC vs. high. Tesla’s 80-30% advice is based more on marketing and keeping people from worrying, rather than actual battery best practices.

I used to charge to 80% immediately upon getting home. That was what Tesla told me was best practice. Obviously, they say that so that the average owner always has plenty of range in their car for unexpected trips.

I now charge to between 60 and 70% starting at 4am on typical days, and only 80 or 90% when I absolutely know that I will need it for longer commutes. When I went on vacation recently, I left it at 50% for 10 days.

I charge to a minimum of 60%, basically because, with a family of 5, with three teens, you never know when you’re going to need to take a drive. 60% feels like I can go most places without needing to charge, under most circumstances.

I start my charging at 4 AM. Unfortunately, my workdays are too unpredictable to set one departure time. 4 AM feels like a good compromise. A lot better than starting to charge at 6 PM like I used to do. Right?

My 2019 M3 dual motor has 79,012 miles. I do not have a garage, and live in the northeast, where the temperatures can range from -3 to 101 F. I rarely supercharge. The Stats app tells me my rated range is between 279 and 286. When my car was new, I think the max at 100% was around 305. This seems like decent/acceptable/typical range loss. All things considered, I’ll take that!

I plan to keep my car for quite some time. At this point, I will only get a new car when Tesla comes out with something truly significantly better. Right now we’re just seeing iterations of the same thing. I’m hoping my range remains somewhat close to where it is now, but even another 5% or 10% or so wouldn’t be the end of the world. Thanks for all the info and data in this thread!

 

[Bouba]

This thread is pretty amazing. I wish I could spend the time to actually dig in to what you engineer-types are saying. But I would like to summarize my take-aways in lay terms. Can somebody confirm that I have it right?

Best practice is ultimately to charge your car to the lowest percent necessary to allow you to drive where you need to go. Range anxiety shouldn’t be dismissed (it’s can’t be so long as humans are driving.) Charging should take place as close to departure as possible. Better to have a very low battery SoC vs. high. Tesla’s 80-30% advice is based more on marketing and keeping people from worrying, rather than actual battery best practices.

I used to charge to 80% immediately upon getting home. That was what Tesla told me was best practice. Obviously, they say that so that the average owner always has plenty of range in their car for unexpected trips.

I now charge to between 60 and 70% starting at 4am on typical days, and only 80 or 90% when I absolutely know that I will need it for longer commutes. When I went on vacation recently, I left it at 50% for 10 days.

I charge to a minimum of 60%, basically because, with a family of 5, with three teens, you never know when you’re going to need to take a drive. 60% feels like I can go most places without needing to charge, under most circumstances.

I start my charging at 4 AM. Unfortunately, my workdays are too unpredictable to set one departure time. 4 AM feels like a good compromise. A lot better than starting to charge at 6 PM like I used to do. Right?

My 2019 M3 dual motor has 79,012 miles. I do not have a garage, and live in the northeast, where the temperatures can range from -3 to 101 F. I rarely supercharge. The Stats app tells me my rated range is between 279 and 286. When my car was new, I think the max at 100% was around 305. This seems like decent/acceptable/typical range loss. All things considered, I’ll take that!

I plan to keep my car for quite some time. At this point, I will only get a new car when Tesla comes out with something truly significantly better. Right now we’re just seeing iterations of the same thing. I’m hoping my range remains somewhat close to where it is now, but even another 5% or 10% or so wouldn’t be the end of the world. Thanks for all the info and data in this thread!

I think that is a fair summary…one other factor is the Tesla advice to be always plugged in…which in effect puts your minimum long term home storage at 50%

 

[dansev]

I think that is a fair summary…one other factor is the Tesla advice to be always plugged in…which in effect puts your minimum long term home storage at 50%

And, your point is, that is not ideal? The problem with not having it plugged in is, if I forget, and the car doesn’t charge overnight, I’m in trouble the next day. I guess it’s all about trade-offs.

It doesn’t happen very often, but when I go on vacation, should I leave it at something like 20 or 30%?

 

[derotam]

This thread is pretty amazing. I wish I could spend the time to actually dig in to what you engineer-types are saying. But I would like to summarize my take-aways in lay terms. Can somebody confirm that I have it right?

Best practice is ultimately to charge your car to the lowest percent necessary to allow you to drive where you need to go. Range anxiety shouldn’t be dismissed (it’s can’t be so long as humans are driving.) Charging should take place as close to departure as possible. Better to have a very low battery SoC vs. high. Tesla’s 80-30% advice is based more on marketing and keeping people from worrying, rather than actual battery best practices.

I used to charge to 80% immediately upon getting home. That was what Tesla told me was best practice. Obviously, they say that so that the average owner always has plenty of range in their car for unexpected trips.

I now charge to between 60 and 70% starting at 4am on typical days, and only 80 or 90% when I absolutely know that I will need it for longer commutes. When I went on vacation recently, I left it at 50% for 10 days.

I charge to a minimum of 60%, basically because, with a family of 5, with three teens, you never know when you’re going to need to take a drive. 60% feels like I can go most places without needing to charge, under most circumstances.

I start my charging at 4 AM. Unfortunately, my workdays are too unpredictable to set one departure time. 4 AM feels like a good compromise. A lot better than starting to charge at 6 PM like I used to do. Right?

My 2019 M3 dual motor has 79,012 miles. I do not have a garage, and live in the northeast, where the temperatures can range from -3 to 101 F. I rarely supercharge. The Stats app tells me my rated range is between 279 and 286. When my car was new, I think the max at 100% was around 305. This seems like decent/acceptable/typical range loss. All things considered, I’ll take that!

I plan to keep my car for quite some time. At this point, I will only get a new car when Tesla comes out with something truly significantly better. Right now we’re just seeing iterations of the same thing. I’m hoping my range remains somewhat close to where it is now, but even another 5% or 10% or so wouldn’t be the end of the world. Thanks for all the info and data in this thread!

I will only make one comment here concerning the WHEN of charging. This is a whole big YMMV based on all kinds of things, but...at least in the winter time charging directly after driving can be a more efficient(power and time wise) option. Charging a cold soaked battery in colder climates can increase charge time and waste lots of energy in the process.

 

[DrGriz]

I will only make one comment here concerning the WHEN of charging. This is a whole big YMMV based on all kinds of things, but...at least in the winter time charging directly after driving can be a more efficient(power and time wise) option. Charging a cold soaked battery in colder climates can increase charge time and waste lots of energy in the process.

I figured that out as well. Always charge a warm battery if you can. Especially for relatively small charges. Efficiency can be 50% or 100%. It's a lot of kWh to throw away if you charge a cold battery when you don't need to every day.

 

[AAKEE]

You do realize that that "fleet" average is based on only 8 vehicles right? That seems like a fairly small fleet/sample size.

No, it’s not. It’s more than 30 cars in total.

Each range point has a number of ”hits”
It is more then 30 cars, but a progressive drop of cars. iIn the far end to the right its fewer, as not every car did reach 61K km.

Doing the same but at a lower range in the graph hit quite well also, but there need to be accounded for cars with low range as time degrades more than miles.