Charges to 84% when charge set to 80%

[GtiMart]

Here's something that could help you understand: it's extremely difficult to estimate a battery's capacity. You can look at the voltage but that's not very precise and it's also not linear. There are parts where the voltage stays almost the same but the energy content changes. It's even worse for LFP. The BMS uses other data to estimate, not just voltage. It helps it when you charge the car to a high percentage for example. In your case you are at low SOC, and maybe you haven't done a full charge in a long while. Total capacity as well as current charge percent are separately difficult things to estimate. The BMS does what it can.
To make things worse, the amount of energy that you can get out of a battery varies depending on the battery's temperature. The BMS doesn't just want to know how much energy you put in, it's trying to figure out how much you can get out of it, and that is a moving target.

 

[AAKEE]

Here's something that could help you understand: it's extremely difficult to estimate a battery's capacity. You can look at the voltage but that's not very precise and it's also not linear.

The definition of SOC is cell voltage.
Regular li ion like Teslas LR / P batteries do not have a very flat voltage curve.
For a car not driving (sleeping or at least Parked) its quite easy to measure the SOC.

Capacity is not that hard to calculate.
I have posted earlier specifically about that, but we can check the SOC before a drive (presumably a bit longer), record the used energy and the end of the drive SOC (preferably about 30 minutes after the drive ends, to allow the car to meeter the true SOC.

Used energy / (SOC before* - End of drive SOC*) = the battery capacity.

The BMS does this as well.
I could actually predict the true capacity at 98.0 kWh despite my new MSP said 95.3 kWh from the BMS.
About three weeks later the BMS had figured this out as well, showing 98.0 kWh.

There are parts where the voltage stays almost the same but the energy content changes.

It do not do this for NCA and NMC/NCMA, the voltage drop at rest is about 0.1V / 10% SOC.

This is not the OCV, but during slow discharge so not very big difference

It's even worse for LFP. The BMS uses other data to estimate, not just voltage.

Any Tesla uses another way than cell voltage during a drive, as the voltage droops during driving it can not be used as a precise measurement.

The car uses the before the drive SOC (voltage measured), and via the estimated capacity this gives a calculated energy.
For example, 80 kWh capacity and 70% SOC = 56 kWh onboard (nomimal remaining).
When driving the car measures the used energy and converts to SOC via before drive SOC - used energy = remaning energy.
Remaining energy / estimated capacity = preliminary SOC, which will be displayed.

After the drive the car can measure the true SOC and also update the displayed SOC.

We can detect a overestimated capacity if the SOC reduces when the car updates the SOC. If the SOC increases after the drive, the BMS is underestimating the capacity.

Same is valid for charging, the BMS can mot measure the SOC during a charge (can not measure the resting voltage) so the number of kWh needed is calculated.
After the charge the SOC is measured and adjusted.
If it drops, there is a underestimate.
If it overshoots, there is a overestimate of the capacity.

It helps it when you charge the car to a high percentage for example. In your case you are at low SOC, and maybe you haven't done a full charge in a long while. Total capacity as well as current charge percent are separately difficult things to estimate. The BMS does what it can.
To make things worse, the amount of energy that you can get out of a battery varies depending on the battery's temperature. The BMS doesn't just want to know how much energy you put in, it's trying to figure out how much you can get out of it, and that is a moving target.

For LFP the real SOC is hard to measure on the middle SOC level, like 20-80% so the car basically use the energy count instead all the time, like all teslas when driving.
Charging full reset the counter if it was faulty/off to 100%, ir the estimated capacity.

 

[GtiMart]

@AAKEE , I can't tell if you're destroying my statement or just helping back it up

Can you give a simple statement for the OP that says that it's normal for the car's displayed SOC to readjust after the charge? Or, if I'm completely wrong and it's not normal, give a short explanation why they should make a service request?

 

[bo3bdar]

Here's something that could help you understand: it's extremely difficult to estimate a battery's capacity. You can look at the voltage but that's not very precise and it's also not linear. There are parts where the voltage stays almost the same but the energy content changes. It's even worse for LFP. The BMS uses other data to estimate, not just voltage. It helps it when you charge the car to a high percentage for example. In your case you are at low SOC, and maybe you haven't done a full charge in a long while. Total capacity as well as current charge percent are separately difficult things to estimate. The BMS does what it can.
To make things worse, the amount of energy that you can get out of a battery varies depending on the battery's temperature. The BMS doesn't just want to know how much energy you put in, it's trying to figure out how much you can get out of it, and that is a moving target.

I'm an EE, so I have better than usual understanding of battery measurements, and how it being a chemical process makes it really hard to get right. However... this is in fact Tesla's job. It should not be my job as a driver to have to second guess the car when it's being flaky and giving me charge measurements that wildly differ. 4% is wildly differing in my opinion and should not be reported. I put in a service request earlier, and their response was that it is 'normal.'

Look at it this way- if I'm at 5% charge, and I think I'm OK because that's 10 miles of range for my car, and I need to drive 5 to get to charge, I'm OK. If I wake up and it's 1% now, with 2 miles of range, now what? If the bottom 5% is total crap and cannot be trusted, then why is this my job to manage that, instead of Tesla keeping a maybe 5% buffer? If the algorithm truly cannot know, then just don't lie to me.

I'm looking for reliability and predictability, when I most need it.

Reframe it as an unknown reserve that can be between 2 miles and 10 miles, but that I can trust it all the way down to 0%. That's de-facto what 5% is right now, but they've made it my job instead of theirs.

 

[GtiMart]

I suggest you also look at your mobile app. There you will possibly see a different SOC than what the car displays. It seems like the car shows the SOC as the amount of energy that it thinks it can pull out whereas the SOC displayed in the mobile app is the "real" SOC. You see a bigger discrepancy when temperatures get colder, and especially when the blue flake appears. It's not a direct answer to your statement but it just shows that a percent value without an explanation of what it is might not be enough as you don't know how to interpret it.

 

[AAKEE]

@AAKEE , I can't tell if you're destroying my statement or just helping back it up

Just trying to add info about how it works

Can you give a simple statement for the OP that says that it's normal for the car's displayed SOC to readjust after the charge?

Yes. In have that both for my old and new Tesla.

Driving from 100-52%. At this time
My BMS grossly underestimated the capacity. ( I know becuase I made a measurement, 100-0% drive at the same time.)
Parked at 52%, when the car started the sleep the SOC was 53%, and when I started next drive it was at 54%, after the sleep:

The next charge after that drive:
Charge level 55%, which was reached ( but estimated) when the charging stopped. Immideately after the level adjusted to 54%, and when the next drive started it was at 53.5%
These 1.5% comes from the BMS being off
So the preliminary/ estimated SOC number is way off and the SOC adjusts after reading the correct SOC.

This is my MSP:
Charged to 59%, adjusted to 58% before the drive.
It normally uses 0.1kWh / a day when parked, so 0.1% each 24 hrs

Or, if I'm completely wrong and it's not normal, give a short explanation why they should make a service request?

Well, in @bo3bdar’s case, it is not consistent. I missed that post earlier.

The 13-17% increase is way to huge to be comming from a overestimate. This as 4% out of the charged 13-17% would be a estimation fault of 50% of the battery capacity.
Also, after next charge the SOC drops 1% instead which means the opposite in therms of under/overestimates.
One time is too little to build a case.

I have two things to say about the increasing SOC:
1) We can not see what happened before the first charge. If the car was driven shortly before it is very probable that the estimated SOC was a bit off, just like I described in my last post. Recently driven and the charger connected directly means that the real SOC could not be measured before the charging was started.
This would mean that the car could have had 4% more SOC than the estimated/displayed SOC and that the OVC/ measured SOC happened after the charging session.
2) I think I remember seeing someting like this once with Teslafi. In my case I think it had to be some kind of data-error, for example the car charging but not getting online properly ( internet connection etc.)

I would not get a service ticket for a on time value like this. I didnt read the post completely if it maybe covered something nfo about always doing this.

 

[AAKEE]

I'm an EE, so I have better than usual understanding of battery measurements, and how it being a chemical process makes it really hard to get right. However... this is in fact Tesla's job. It should not be my job as a driver to have to second guess the car when it's being flaky and giving me charge measurements that wildly differ. 4% is wildly differing in my opinion and should not be reported. I put in a service request earlier, and their response was that it is 'normal.'

Look at it this way- if I'm at 5% charge, and I think I'm OK because that's 10 miles of range for my car, and I need to drive 5 to get to charge, I'm OK. If I wake up and it's 1% now, with 2 miles of range, now what? If the bottom 5% is total crap and cannot be trusted, then why is this my job to manage that, instead of Tesla keeping a maybe 5% buffer? If the algorithm truly cannot know, then just don't lie to me.

I'm looking for reliability and predictability, when I most need it.

Reframe it as an unknown reserve that can be between 2 miles and 10 miles, but that I can trust it all the way down to 0%. That's de-facto what 5% is right now, but they've made it my job instead of theirs.

Do you have that issue when the displayed SOC changes a lot?

Did you do a longer drive before that charging session (just before/to the left of the picture).

Do you use sentry mode?

I take it as you have a 2020 SR, thats with the classic Panasonic NCA cells and not a LFP car…

 

[bo3bdar]

Do you have that issue when the displayed SOC changes a lot?

Did you do a longer drive before that charging session (just before/to the left of the picture).

Do you use sentry mode?

I take it as you have a 2020 SR, thats with the classic Panasonic NCA cells and not a LFP car…

It does move a lot, but generally only 1%. It's relatively rare for me to see a 4% move, but I notice one about every 6 months or so. Mostly I see these at higher states of charge, because it's rare for me to leave the car at low charge when I can charge at home.

I had not driven before that big jump right after a short charge. Battery should have been 'cold' in that it was not in active use, although my actual temps here are 70 degrees. This seems to be different from the expected drop in actual cold weather.

No sentry mode in this case, it's below 20% charge so sentry is disabled. I see these jumps less when sentry is on, presumably because it keeps the car live, but could also be because the consumption is so high that it masks some of these moves.


I've been running it below 10% in the last week because I wanted to see if the BMS might rebalance the cells. It's possible this is related to something like that. Without any info on the BMS results we can't really tell.

The prior month I was running the car at max charge often, which is 90% in my case because this is an SR, and I'm capped at 90%. Always possible that this is such an unusual car that the BMS does not work well if it has never seen a true 100% charge. However, other people see these 4% moves, so that is not very likely.

Just trying to add info about how it works
...
The 13-17% increase is way to huge to be comming from a overestimate. This as 4% out of the charged 13-17% would be a estimation fault of 50% of the battery capacity.
Also, after next charge the SOC drops 1% instead which means the opposite in therms of under/overestimates.
One time is too little to build a case.

I have two things to say about the increasing SOC:
1) We can not see what happened before the first charge. If the car was driven shortly before it is very probable that the estimated SOC was a bit off, just like I described in my last post. Recently driven and the charger connected directly means that the real SOC could not be measured before the charging was started.
This would mean that the car could have had 4% more SOC than the estimated/displayed SOC and that the OVC/ measured SOC happened after the charging session.
2) I think I remember seeing someting like this once with Teslafi. In my case I think it had to be some kind of data-error, for example the car charging but not getting online properly ( internet connection etc.)

Car was not driven before this anomaly, sitting in the garage. I would not expect there to be a huge error in the measurement of SOC, and also I would expect them to do a measurement before starting charging, because otherwise they would not know if it was OK to do so. Measuring SOC should be a fast operation.

Pretty sure my TeslaFi data is accurate because on several occasions, I've gone to check what the car reports on the main screen as a sanity check of SOC, and they matched.


My best guess here, and it's strictly a guess with no backing data, is that the battery SOC can vary this much based on thermal characteristics that are hidden in the battery system. The battery might very well be kept warm until it decides it can no longer afford the expense, and lets it cool to ambient, which could in theory push the SOC off this much.

In this scenario, there is no outside thermal change, my temperature has varied between 68F and 72F in the garage with no change during day. Multiple days of sitting without driving, until I did the small charge shown in the graph. In theory that might have warmed up the battery as part of the process, and thus drift up the 4% as a reaction. All speculation of course.

I just wish it was more predictable, or they'd report on battery temperature if it matters this much. If I was doing the software, I'd either give you the info you need to decide, or more likely hide it behind a 4% range check when I know it might be off from days of sitting. I'd just like to have the certainty you get with an ICE car, where you can drive down to 0% for sure, with no ambiguity. This setup is more like an old VW reserve tank where I can't remember if I flipped the lever back or not.

 

[AAKEE]

It does move a lot, but generally only 1%.

How does it move then in general? The same direction (up / down) or does it vary?

I had not driven before that big jump right after a short charge. Battery should have been 'cold' in that it was not in active use, although my actual temps here are 70 degrees.

As it was not driven the SOC before the charge should be very correct (OCV read).

I've been running it below 10% in the last week because I wanted to see if the BMS might rebalance the cells.

There is no ”rebalancing” of the cells at low SOC.
”Calibration” of the BMS (i.e showing the battery capacity to the BMS) can be done by letting the car getting low SOC, and then sleep (no sentry, no cabin overheat, no summon and no peaking in the tesla app).
It would need to sleep a few hours to get the steady OCV = true SOC number.

Then it need to see the high SOC side fairly close in time, the same is valid for that sleep as the low SOC sleep.

There is no balancing of the cells at low SOC. Balancing is only discharging the high cells to put them at about the same voltage as the lower ones. Tesla balances by burning of energy from the high ones.
There are no reasons to balance at low SOC. It would only create a need for more balancing after recharging, as the imbalance comes from that the cells have slightly different capacity.

A full charge need balancing to actually make all cells reach 100% or close to that.

The prior month I was running the car at max charge often, which is 90% in my case because this is an SR, and I'm capped at 90%.

You should not be capped at 90%.
There is a recommendation for daily chsrge that until recently was 90% (might have been changed to 80% for your car) but when needed you should be able to set 100% and also reach 100%.

If it doesnt, you might wanna contact Tesla for a check.

Measuring SOC should be a fast operation.

Yes, it only takes like 1/100 second.
It will be correct if the car was sleeping.

Pretty sure my TeslaFi data is accurate because on several occasions, I've gone to check what the car reports on the main screen as a sanity check of SOC, and they matched.

I do not mean that the SOC number is recorded wrong but I have seen glitches where the car ”was sleeping” according to teslafi (no data received from the car) but in reality it was not sleeping (charhing or other).
A loss of data due to a connection failure or something like that during a charge would make the SOC jump.

My best guess here, and it's strictly a guess with no backing data, is that the battery SOC can vary this much based on thermal characteristics that are hidden in the battery system.

The true measured SOC doesnt change with temperature. When reading the SOC from the BMS it is not drooping with temperature. It is very constant. This is because of that the lithium battery voltage is more or less constant with varying temperature. The change for a bery big range of temperature might be a mV, which is about 0.1%.
So the SOC number in the BMS do not change due to cold battery.
The car displays a lower SOC though, that is calculated by a formula to reflect the lower output due to higher internal resistance and maybe need for battery heating (active heating during a drive only happens if the battery is sub freezing).

The battery might very well be kept warm until it decides it can no longer afford the expense, and lets it cool to ambient, which could in theory push the SOC off this much

The car never keeps the battery warm when parked.
I did log all BMS data via teslalogger and I also had a tablet in front of the steering wheel on my M3P showing everything.
The screen I setup fot driving included the battery temp.
Data was logged for two years, but when I changed car for the logging I might have caused an erase of the old data unfortunately. Need to check my raspberry pie, I hope it still is there as it shows just how things actually works.

.

In this scenario, there is no outside thermal change, my temperature has varied between 68F and 72F in the garage with no change during day. Multiple days of sitting without driving, until I did the small charge shown in the graph. In theory that might have warmed up the battery as part of the process, and thus drift up the 4% as a reaction. All speculation of course.

At around 70C there are no changes to the displayed SOC due to the temperature.
Teslafi get the real SOC and also the displayed. You can look at the status on your car at teslafi, SOC is shown both as the normal SOC with warm battery (BMS-SOC) and the displayed SOC.

Heres my car at 9C ambient (varied between 4-9C that day). After been parked for 10 hrs, the cell temp was 10C (50F)

Displays 24%

SOC according to the BMS 25.2%

I just wish it was more predictable, or they'd report on battery temperature if it matters this much. If I was doing the software, I'd either give you the info you need to decide, or more likely hide it behind a 4% range check when I know it might be off from days of sitting. I'd just like to have the certainty you get with an ICE car, where you can drive down to 0% for sure, with no ambiguity. This setup is more like an old VW reserve tank where I can't remember if I flipped the lever back or not.

The SOC number will not be of from days of sitting, it will be very accurate becauae of the sleeping gives the real OCV reading.

 

[baloy17]

I have a month-old MYLR and I noticed this past few days that it is charging past the 80% limit that I set. Sometimes it charges up to 85%. Last night it went up to 83%.

 

[AAKEE]

I have a month-old MYLR and I noticed this past few days that it is charging past the 80% limit that I set. Sometimes it charges up to 85%. Last night it went up to 83%.

It should probably be due to overestimating the battery capacity. Not 100% sure but I guess that it is.

If you check the SOC just when the charging is finished (thats, within a minute) it probably show 80%, but after 15-30-45 minutes the SOC is higher if you check.

 

[fholbert]

I have a month-old MYLR and I noticed this past few days that it is charging past the 80% limit that I set. Sometimes it charges up to 85%. Last night it went up to 83%.

My ‘23 MY is spot on 80% every morning using the Tesla wall charger, I use ~42% in my daily commute. Which charger are you guys using?

 

[baloy17]

My ‘23 MY is spot on 80% every morning using the Tesla wall charger, I use ~42% in my daily commute. Which charger are you guys using?

I'm using the Tesla mobile connecting (14-50 Nema).

 

[Bouba]

Can we say....it is good if it charges below what the charge is set to....and it’s bad if it charges to a higher percentage than what the charge is set to ?

 

[AAKEE]

Can we say....it is good if it charges below what the charge is set to....and it’s bad if it charges to a higher percentage than what the charge is set to ?

Ehh, not really.

Over- or undershoot should be a BMS calibration issue. The BMS calib will probably long term be ok, but from time to time it can both over- and underestimate.
It has nothing to do with battery life.

If I have the battery with the lowest degradation the BMS can still overestimate that too, causing the charging to overshoot.

For long trips using low SOC, underestimate is no problem but a large overestimate could get you stranded at low domisplayed SOC. Probably not usual on a newer car though.