I will give this a try and report back to see if it unveils the hidden world of Tesla batteries for me.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
HV Battery Health Test Confusion
- Thread starter Nonlin
- Start date
AlanSubie4Life
Efficiency Obsessed Member
Yep. As expected. 44.6kWh (204mi) was the expectation and that matches exactly the 45.0kWh (206mi).
Most likely your car has 205-206 miles at 100%. (May have changed slightly from the last time when you got 204 miles.)
So SMT matches exactly what is visible from the car UI.
Capacity loss 14% to 18% depending on whether you use 52.5kWh (degradation threshold) or 54.6kWh (EPA result) as the starting point.
Last edited:
Which one does the warranty use? So I guess which one does the battery health test use? Based on the optimistic degradation of 13%. I'm guessing they use the 52.5 KWH?
AlanSubie4Life
Efficiency Obsessed Member
No one knows. I don’t think anyone has hit the warranty threshold so there has been no test of this.
I think 54.6kWh would be reasonable to use for arbitration or court, if necessary.
Since that is what they claimed to the government was the starting capacity.
What the battery health test uses doesn’t matter. It’s all academic. From above it calculates to 51kWh to 52kWh. Just a random number, as one would expect, since it does not matter. But we also don’t know what exact numerator it arrived at for this 100%-15% discharge (it was 38kWh*0.955 but not enough resolution).
Last edited:
AlanSubie4Life
Efficiency Obsessed Member
Actually the 38kWh on the screen has nothing to do with it because it is just the miles added. It’s presumably not used for the battery health check. They must check something on the 100% to 15% discharge. Maybe not - they may be directly doing the same calcs we do but seems unlikely. Anyway no one knows what battery health check means, AFAIK.
Right and that's my issue. They could say whatever they want and we just have to take it at face value? If we want the warranty to mean anything solid we must understand how the health check works, otherwise the fact that no one really has had to use the battery warranty might not even be a good sign.
AlanSubie4Life
Efficiency Obsessed Member
No, none of this matters.
If you think your battery is below the threshold, you have it checked. You don’t need some battery test buried in service mode to check it.
As usual, in order to have warranty made good, you have to demonstrate that your vehicle qualifies. This is on the customer to demonstrate - obviously you cannot rely on Tesla!!! (This is the case for any issue.) It doesn’t matter what they say. All that matters is whether your battery qualifies.
The BMS gives the answer; this is customer facing as you can see. Not sure what the purpose of the health check is (maybe it does other things, but obviously it is not needed for measuring capacity).
Yours does not qualify. It’s not even close and it seems highly unlikely it will get there in the warranty period even if you let it sit at 90% all the time - you’d have to lose 15% of your current capacity! I’m not aware of anyone who has made good on the warranty for capacity loss alone.
Last edited:
I had to spend about 130 dollars to be able to check what Tesla should just be able to tell me in the Battery Health Check in the service menu.
And from what they've told me they use that number to tell if the battery is good or bad.
But now that I've invested in this gear I suppose I can state otherwise if the health check is off from what the OBD data is informing me of.
I am a new owner so don't have much experience but from what I understand, charge it to somewhere around 90-100%, and leave it plugged in. It will take several hours to correct the imbalance. Perhaps good to do night before a long drive.
AlanSubie4Life
Efficiency Obsessed Member
No you did not! I specifically said it was not necessary to purchase that equipment.
All of the battery capacity info is user facing. It is easy to determine what capacity you have (use the energy screen method - it’s a sticky for a reason). Again, SMT tells you nothing about that number that you cannot tell on a 100% charge (it could potentially give one more significant digit but usually not).
There is plenty of other value in SMT, but determining your pack capacity is not one of the selling points.
Sorry if that was not crystal clear - I should have stated more strongly that SMT is only necessary for cross-checks and for other monitoring functions. Not necessary at all to determine pack capacity.
Last edited:
smatthew
Active Member
The car can correct 1mv of imbalance per 24 hours. That's 24 hours where the car is asleep with HV battery disconnected
AlanSubie4Life
Efficiency Obsessed Member
You can have your vehicle sleep at high state of charge and see.
However, I think the impact will likely be at most a couple percent.
It’s normal for batteries to lose capacity. It’s possible yours started a little low, and maybe it has been stored at high SOC, maybe in hot climates.
@AAKEE could help if you knew the entire car history and provided the datapoints. Can be fitted to a model.
But 15% is possible. Anecdotally, these SR+ saw quite a lot of loss, maybe slightly above average. I have no idea why.
Tesla uses LTC6812 BMS chip which has 200mA max balancing current.
SR+ has 96S31P configuration and I think uses 5Ah 2170L cells. If I have to guess, I’d assume Tesla would use 7 LTC6812 chips (which has 15 channels) to reduce cost. In other words, all 31 parallel cells are controlled/measured by single BMS channel. Although 7 BMS chips can measure 105 cells so that doesn’t add up... Extra channels maybe used for redundancy to meet ASIL-D requirements. Though I’d expect the chip is already rated for ASIL-D. Or maybe there are bus-bars and extra channels are used for bus-bar measurements.
31*5Ah=155Ah
100% SOC is 4.2V OCV and 0% is 2.5V. So we have 17mV per 1% of SOC.
Therefore 1mV is ~0.059% (assuming linear relationship between OCV and SOC)
155Ah*0.059% / 200mA =~ 0.46hours
Or more if max 200mA is not used (reduced duty cycle).
Where did you get the 1mV per 24hour figure from?
Last edited:
The last ~ 0.7V (3.2-2.5V) is steep.
So in practical terms we have about 0.1 V / 10% SOC, so 10mV / percent.
From ~95% to ~20% it can be considered linear in this case.
So we end up closer to 0.775h, so ~46 minutes for 1mV.
We can not judge if the imbalance is good or bad from one single point of SOC.
So no need to worry and no need to do anything.
You could charge it to 92-100% and lave the car sleeping (no sentry, no app checking) for ~ 5-6hrs. This would balance the cells at top but depending on the daily charge level you will see another imbalance as the battery is not set in balance for snything else than the 92-100% you used.
(It might even give you a higher imbalance where you usually have the car.
Balancing is normally performed at high SOC, at least not for these numbers. Up to ~20 mV is normal in the end points of SOC, before any balancing has been done.
Imbalance can be because the charge level drifts away between cells, but it is also a normal behaviour because the capacity is different between cells. We can see the different capacity in SMT as CAC, max an min.
If we balance the battery at 90-100% to 0mV imbalance, and then drive it to 0% we will have a higher imbalance at 0% than we would if we did not balance it at 90-100%.
This is because the cells with smaller capacity will discharge faster percent-wise than the larger, so the cells that has 1% less capacity will be at 1% less SOC at low SOC, causing about 10mV imbalance (as 10mV is about 1% of the voltage range).
In the end close to 0% the difference will be higher as the voltage starts to droop faster, so ~20mV is completely normal at 0% even if the imbalance was 0 at top.
If we balance at any SOC in the mid range we will have a imbalance spread both above and below.
Balancing is only adjusting the cell voltage to be about the same, and this will only be at the specific SOC where we perform the balancing.
Top balancing is the normal, as it hides the difference between cells in the buffer and also ensures all cells are at 100% = maximum energy stored.
Balanced somewhere in the middle SOC.
Your CAC number is of, I guess SMT is behind a software change in the BMS data.
The closer the max and min is, the better. But there is nothing we can do as it mainly depends on the small differences in the cell and the coincidence how they get grouped in the manufacturing process.
(My M3P had about 1Ah difference, so 229 Ah min and 230Ah max for example.)
These CAC numbers are all off.
A SR+ pack should be 31 cells in each cell grouping.
4.8Ah new x 31 should reach ~149Ah.
After loosing 10% capacity this should also be about 10% lower.
The expected CAC avg number would be
45.000/ (96x3.7) = ~ 127Ah
But it depends on which nominal voltage the BMS counts with. I used the 3.7V spec, which might be different on a well used battery
We can assume that the difference which is huge, is due to the SMT not being updated to get the correct readings.
(If not, you woukd need a new battery soon. Just kidding, nothing to worry about
) Here is a add-on for the imbalance thing:
The picture below is for about 24hrs from before the charge yesterday and including the charge today.
I use 55% any day I do not need more, and yesterday I did need 80% for a 200 km total drive.
-As we can see during the time with the highest SOC (80% amber/brown line) the imbalance is 4mV, so "good" according to some of the posters.
-We can also see that during the lowest SOC (25% down to the right) the imbalance is 20mV.
- I did add a yellow line that shows that the imbalance is about 12-14mV when the SOC is about 52%
My MSP do not seem to balance the battery when I use 55% daily. The imbalance mostly is about 8-10mV, but as I recently did a few 100% drives and it had the time to balance at high SOC, the imbalance at 55% is slightly higher than it usually is.
Balancing means burning of energy, and there is no need to have the cells at perfect balance at lower SOC's, as it only means the imbalance will be higher at high SOC, wich takes longer time and also ends up with lower than 100% SOC when the balancing is done (as it do this by burning of energy from the high cells).
To sum it off, my MSP 2023, 6 months since pickup shows excellent imbalance at high SOC, but has higher imbalance at lower SOC.
The difference is due to the different cell capacity, and it would be possible to baland it to 4mV at low SOC, but that would only move the high imbalance spread to high SOC.
A single reading of 20mV is not a sign of an issue.
The picture below is for about 24hrs from before the charge yesterday and including the charge today.
I use 55% any day I do not need more, and yesterday I did need 80% for a 200 km total drive.
-As we can see during the time with the highest SOC (80% amber/brown line) the imbalance is 4mV, so "good" according to some of the posters.
-We can also see that during the lowest SOC (25% down to the right) the imbalance is 20mV.
- I did add a yellow line that shows that the imbalance is about 12-14mV when the SOC is about 52%
My MSP do not seem to balance the battery when I use 55% daily. The imbalance mostly is about 8-10mV, but as I recently did a few 100% drives and it had the time to balance at high SOC, the imbalance at 55% is slightly higher than it usually is.
Balancing means burning of energy, and there is no need to have the cells at perfect balance at lower SOC's, as it only means the imbalance will be higher at high SOC, wich takes longer time and also ends up with lower than 100% SOC when the balancing is done (as it do this by burning of energy from the high cells).
To sum it off, my MSP 2023, 6 months since pickup shows excellent imbalance at high SOC, but has higher imbalance at lower SOC.
The difference is due to the different cell capacity, and it would be possible to baland it to 4mV at low SOC, but that would only move the high imbalance spread to high SOC.
A single reading of 20mV is not a sign of an issue.
Steve446
Dusty Crophopper, M3 LR 2021 Midnight Silver
This seems to have happened for my car after it installed 2023.44.1 on Dec 03, 2023. That update seems to have impacted the CAC reporting
. CAC has been off ever since then. I asked Amund Børsand from SMT but at the time he was waiting for that update on his car. Haven't heard anything else since. Has anybody else observed that?Similar threads
