The uniformity in your graph (reproduced below) is remarkable. After a year and a half, the cars with the worst degradation have a displayed range that's just a few miles less than the best.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Keeping an Infrequently Driven LFP Happy
- Thread starter PianoAl
- Start date
Why is this surprising? Storing the car at a constant 50% vs 100% results in a difference of 2% of capacity after 9 months (according to research charts above). We don't know the storage SOCs of the cars in this hand-picked chart, but they probably aren't at the extremes of 50% permanently vs 100% permanently. A 1% measurable difference seems reasonable based on real-world usage.
Just want to reiterate that cycle degradation and supercharging degradation are basically negligible for LFP and can basically be ignored.
Just want to reiterate that cycle degradation and supercharging degradation are basically negligible for LFP and can basically be ignored.
Another thing that you could do is to drive a 100-0% single drive and see how much caqpacity you get from the battery.
Driving 100-0%, at normal speeds (like 55mph), check the indicated rest range in km and then let it sleep for one hour or two should indicate how much energy that is left. Then we should have an indication about the capacity in your battery. It most probably would tell us if your battery has more capacity left Than the ones always parked at 100%.
ucmndd
Well-Known Member
This and other reasons lead me to believe that Tesla is much more assertive about managing displayed range and capacity on these cars than they have been in the past. I’ve had a 3RWD for about 2 months and 5,000 miles and the range is perfectly pegged at the rated 272 every full charge. This was absolutely not the experience I had with my 2016 Model S.
I strongly suspect Tesla is steadily decreasing displayed range based on estimated calendar aging and placing much more weight on that metric in the algorithm than observed capacity/cycles.
Smooths things out, gives a more consistent experience to owners, dramatically reduces the number of service complaints and other freak-outs by new owners.
My (admittedly completely uninformed) guess is that over time the weighting will shift and the BMS will place more importance on actual pack observations vs. calendar aging estimates.
This would make sense to me. Interestingly, I came across a random comment on YouTube claiming very matter of factly that displayed range in LFP Teslas drops .01% per day, and that they could calculate manufacture date based on that, and when I generated a curve based on that math and compared it to the values I've logged for my own car, sure enough, my own car matches that curve exactly. But this couldn't possibly be accurate long term, because it doesn't allow for degradation to slow much, as we know it should, and would lead to a 25% loss of range after 8 years, which sounds too high for an LFP battery IMO. So maybe, like you said, after the first 2-3 years, the cars will start shifting weight more to actual pack observations.
Last edited:
Prediction:
2 years in (do we have any 2 year + LFP cars?) the displayed degradation rate slows. One day after warranty expires the algorithm changes.
Nevertheless, the battery chemistry is as the science says and keeping state of charge low will lower degradation no matter what tricks are done to hide that fact.
I think other auto makers bypass this problem mostly by including only % SOC and no absolute predictions of range.
Interesting.
Once I visited a friend who had a Health-o-meter electronic scale. I weighed myself three times and got the exact same reading, to a tenth of a pound.
I was impressed and bought one for myself. I soon realized that if two successive weighings are within about two pounds of one another, it displays the first reading for each new weighing. That is, it lies to make it seem more accurate.
Are you thinking that Tesla could be doing something analogous to this?
ucmndd
Well-Known Member
I don't know that I'd call it "lying" so much as "smoothing the data for a more consistent user experience" or something like that.
But yes, that's what I'm thinking.
I think with LFP the state of charge calibration and the estimation of the capacity is more erroneous because of the long flat mesa in the charge vs voltage curve. So the conventional algorithm was giving more noisy results in the field and causing customer complaints & spurious service requests. You can say this is a variance vs bias tradeoff, the usual issue in any statistical modeling.I don't know that I'd call it "lying" so much as "smoothing the data for a more consistent user experience" or something like that.
But yes, that's what I'm thinking.
So to some degree the "smoothed" model based prediction may be more accurate (on average). Also consider that merely estimating state of charge with LFP in that flat area requires two things (1) accurate coulomb counting, which they're good at already (2) knowledge of true battery capacity. They will observe the 100% high voltage part frequently but very rarely the bottom when SOC v voltage changes.
Their estimate for battery capacity with the old algorithm was probably off too often and that would cause errors in state of charge estimates and that resulted in people getting pissed off when they get stranded, or the system erroneously told them they ran out much sooner than expected.
So probably for normies, this approach is the right one even if it doesn't satisfy the super nerds like us.
What happens in 5 year old LFP, the Tesla team will figure out the model to use when it comes.
Very encouraging. When I look at the storage degradation charts for LFP I'd have expected more degradation especially given the high ambient storage temperatures in AZ (unless your garage is air conditioned)
The degradation charts for calendar aging is all made with battery cells from a few years back.
-We can not know the exact rate for calendar aging of the cells of today. But we know that the basic behaviour has changed very very slow and that the basics probably still rules (low SOC is better thab high, abd low temp is better than high.).
-As it looks NCA cells for the LR/P still degrade by the same rate.
-For LFP the calendar aging might be lower than before but using for example teslalogger and looking at the range loss after 50K km the difference is not very big.
LFP did loose 20km (440-420km in 50K km) = 4.5%
M3 LR
467/499km = 6.4% loss, this includes the cycles that will show noticable degradation on a NCA cell, but more or less neglible on a LFP cell.
If we think that the cyclic degradation on a LFP is negligble, the cyclic degradation on the NCA is about 0.5% / 10K km if the calendar aging is about the same.
One possible thing is that Teslas with LFP shows the degradation on a time schedule as it might be hard to measure the true capacity.
RichAZ/CapeCod
Member
Garage is not air conditioned. Did check on the M3 from time to time over the summer but never found a garage temp (at least as shown by the M3) over 95 degrees, even during a 105 degree day in Tucson. For the record the M3 was/is plugged into a 120v outlet.
Rich
OP's issue seems to be the inconvenience of sharing one charging socket with two vehicles.
With only little use, there is no need for a 240V charging socket. The lesser used vehicle could simply be plugged into any convenient 120V household socket in the garage. This could easily maintain a 100% charge.
Nice first World problem.
With only little use, there is no need for a 240V charging socket. The lesser used vehicle could simply be plugged into any convenient 120V household socket in the garage. This could easily maintain a 100% charge.
Nice first World problem.
Here is an example LFP voltage / SoC graph for a 12V battery:
Note the flatness in the middle part of the curve. If the BMS does not see 100% (or very low) SoC for a while, it would lose track of the actual SoC. If it overestimates, it may result in the driver seeing 20% and then a sudden drop to 10% or so when the SoC finally enters a range with a significant voltage drop.
It is probably better to park the car mostly at 70% or lower, although LFP batteries tend to degrade less and slower than NCA batteries, so charging habits may make much less difference with LFP batteries than NCA batteries.
Looking at this chart, it seems to me that we should advise people to recharge at 10% SOC to maybe 30 or 50% if your regular/infrequent use are short distances. Charge to 100% when planning a trip that would use at least 30% SOC.
Actually, sharing one cable between the two cars has never been an issue. Tessie just doesn't get plugged in much.
We actually keep Tessie at about 70% most of the time, but if someone's going to drive her, we'll bring her up to 100%.
New 2023 Model 3 RWD owner here in Northern Virginia. Picked it up on 12/2. I've read far too much on LFP battery charging.....
Like the original poster, I am an infrequent driver. I work from home 4 days per week and the day I go into the office, it is only a 10 mile round trip commute.
I have an attached, unheated garage. For the time being, I am planning to charge using a 15A 120v plug that is very conveniently located. I am seeing a steady 12A while charging and see 2% charged added per hour. With my driving, I am planning to charge daily from home and only super charge when needed.
One question that I have not seen answered is whether the advice on not keeping the battery at a higher SOC applies the same during the winter months. I have seen the charts that show the degradation but those tests were at 77 degrees F and 122 degrees F. Does anyone have any data at 20 - 50 degrees F? It seems that the winter months will show some range degradation and the extra charge capacity would be useful.
Ideally, I'd like to set the charge limit to 100% and plug it in every day. Any advice for this practice in colder weather?
Like the original poster, I am an infrequent driver. I work from home 4 days per week and the day I go into the office, it is only a 10 mile round trip commute.
I have an attached, unheated garage. For the time being, I am planning to charge using a 15A 120v plug that is very conveniently located. I am seeing a steady 12A while charging and see 2% charged added per hour. With my driving, I am planning to charge daily from home and only super charge when needed.
One question that I have not seen answered is whether the advice on not keeping the battery at a higher SOC applies the same during the winter months. I have seen the charts that show the degradation but those tests were at 77 degrees F and 122 degrees F. Does anyone have any data at 20 - 50 degrees F? It seems that the winter months will show some range degradation and the extra charge capacity would be useful.
Ideally, I'd like to set the charge limit to 100% and plug it in every day. Any advice for this practice in colder weather?
Not LFP specific but Table 3 has data on NMC vs stored temp.
Calendar aging = Time x state of charge x Temperature.
Low temperature is good, and reduces calendar aging.
I guess you can convert Celcius to F by your self.
You can see that they all behave about the same including temp.
Here’s a 10C graph included. Its from NCA chemistry but you can use the difference for LFP. Not a very big difference from 25C, but at least better.
Another 10C graph, for reference:
Not a problem, unless you strive to keep the degradation at an absolute minimum!
Thank you for this reply. I think I will go with the plan of charging to 100% daily during the winter. Now I'll have to see if my garage gets too cold to charge during the winter. Luckily really cold (< 15 degrees F) days are pretty rare and the garage should be a bit warmer and shielded from the wind.
I wouldn’t charge to 100% daily if you don’t need it. Doing so will increase your degradation compared to maintaining at a lower SOC.
In your situation, I would charge it to 70% daily and bump it up to 100% once per week, the night prior to your planned usage.
Depending on how little you actually drive, you could probably get by with charging it to 100% once per week, and leaving it unplugged the rest of the time.
Similar threads
