Clarity around CAC imbalances, voltage imbalances, bad bricks, and battery warranty - 2015 90D

[aerodyne]

I actually have a dumb question. I have SMT and I'm not seeing the CAC imbalance on my app. Am I missing something?
View attachment 1012931View attachment 1012932

The question was asked, so it's not dumb
CAC does not show for the older MS. It's also broken now on the newer MS with the holiday update.
Cell imbalance is still a good tell for a pack in good condition. Anything less than 10mv resting is good.

 

[kirkusss]

@Koike On Android you need to request the beta version from the store. On iOS CAC is populated in the regular version.

 

[aerodyne]

@Koike On Android you need to request the beta version from the store. On iOS CAC is populated in the regular version.

I have both versions on Android, the CAC is broken on beta still. Ranges from 25 to 800 A/H.

 

[Koike]

Ah thank you for the clarification! Really appreciate it

 

[jensk2]

Update: I went for a drive and the CAC values populated in the Android beta app.

View attachment 952352 View attachment 952353

Wonder how BMS calculates Nominal Capacity?

My Model S 70D Oct 2015 with 93.000 km has CAC 220,7 - 225,0 AH

I think my Nominal Capacity should be around 3,65V x 220,7Ah x 6 Bricks/Module x 14 Modules == 67.67 kWh
But BMS/ScanMyTesla claims 65.1 kWh.

If I reverse your numbers you have 77,9 kWh and 221.2 Ah Min CAC. That means that your capacity is calculated from this voltage:
76900 Wh / 221,2 Ah / 16 Modules / 6 Bricks/Module == 3,62 V. So half you energy is below 3,62 V.
My car has:
65100 Wh / 220,7 Ah / 14 Modules / 6 Bricks/Module == 3,51 V. But that looks wrong, here is approx Voltage Versus SoC:

At 50% SoC, my voltage is above 3,73 V (Voltages may be too high due to not waiting long enough for resting voltage, I will take a new sample after resting 24 hours!)


Using you rnumbers I should have Nominal:
220,7 Ah * 3.62 V x 6 Bricks/Module * 14 Modules = 67.11 kWh. Even 'correcting' for ScanMyTesla reporting Buffer = 4 kWh instead of the original 2.4 kWh on the 14 modules pack does not help (makes it worse). Perhaps Tesla wants to make sure that they CAN replace my Pack with a refurb instead of a new pack.

Should I assume that I am Voltage Capped, because I have NOT charged enough for BMS to calculate all Voltages at sufficient different SoC?

BTW: Teslamate calculates my Available as 65.17 kWh based on Ideal Range 343 km and 0,190 kWh/km.

 

[jensk2]

Wonder how BMS calculates Nominal Capacity?

My Model S 70D Oct 2015 with 93.000 km has CAC 220,7 - 225,0 AH
View attachment 1016141
I think my Nominal Capacity should be around 3,65V x 220,7Ah x 6 Bricks/Module x 14 Modules == 67.67 kWh
But BMS/ScanMyTesla claims 65.1 kWh.

If I reverse your numbers you have 77,9 kWh and 221.2 Ah Min CAC. That means that your capacity is calculated from this voltage:
76900 Wh / 221,2 Ah / 16 Modules / 6 Bricks/Module == 3,62 V. So half you energy is below 3,62 V.
My car has:
65100 Wh / 220,7 Ah / 14 Modules / 6 Bricks/Module == 3,51 V. But that looks wrong, here is approx Voltage Versus SoC:
View attachment 1016140
At 50% SoC, my voltage is above 3,73 V (Voltages may be too high due to not waiting long enough for resting voltage, I will take a new sample after resting 24 hours!)


Using you rnumbers I should have Nominal:
220,7 Ah * 3.62 V x 6 Bricks/Module * 14 Modules = 67.11 kWh. Even 'correcting' for ScanMyTesla reporting Buffer = 4 kWh instead of the original 2.4 kWh on the 14 modules pack does not help (makes it worse). Perhaps Tesla wants to make sure that they CAN replace my Pack with a refurb instead of a new pack.

Should I assume that I am Voltage Capped, because I have NOT charged enough for BMS to calculate all Voltages at sufficient different SoC?

BTW: Teslamate calculates my Available as 65.17 kWh based on Ideal Range 343 km and 0,190 kWh/km.

PS: I do have two Voltage readings at higher SoC, but they were taken BEFORE disconnecting 11 kW charger, so are elevated. But they do show that 92,5% is at 4,099 V when charging, which argues that 100% is close to 4.2 V, so no capping :-(

 

[jensk2]

PS: I do have two Voltage readings at higher SoC, but they were taken BEFORE disconnecting 11 kW charger, so are elevated. But they do show that 92,5% is at 4,099 V when charging, which argues that 100% is close to 4.2 V, so no capping :-(
View attachment 1016143

My car has been parked for 12 hours and now sits at 52.6% SoC (and 51% Usable (Battery pack is at 2.5 Celsius)) and Min Cell Voltage is 3.722 V. So BMS surely thinks that 50% of energy is below 3,722 V meaning that 50% Voltage is approx 3,722 * 50 / 51 == 3.65V, which is not far from the previously calculated (Model S 90D) 3,62V. But then my Nominal should be:
220,7 Ah * 3.65 V * 6 Bricks * 14 Modules == 67.67 kWh.
BMS now reports Nominal of only 64.5 kWh, so appears to be temperature calculated.

I can think of two possibilities:
A: ScanMyTesla uses Buffer incorrect (and why is it 4 kWh and not 2.4 kWh for my 350V Battery?)
B: Nominal could be wrong because of Cold Scandinavian Temperatures

BTW: For what it is worth TeslaMate calculates Usable as 65.5 kWh which is higher than ScanMyTesla Nominal

 

[ajadan]

I’m fairly confident in the SMT buffer of 4kWh based on how far down below 0% I’ve driven my 90D on the old dying battery. I’ve gone 11 miles past 0 at ~310wH/mi = 3.4kWh. And I wasn’t dead yet, just too scared to keep going.

Not sure how temp plays into the calculations.

 

[islandbayy]

I’m fairly confident in the SMT buffer of 4kWh based on how far down below 0% I’ve driven my 90D on the old dying battery. I’ve gone 11 miles past 0 at ~310wH/mi = 3.4kWh. And I wasn’t dead yet, just too scared to keep going.

Not sure how temp plays into the calculations.

The buffer is capacity that is not accessible to the driver. Your BMS's capacity calculation was just off. Once the car shuts down, that 4kWh buffer, is what is still left in the pack AFTER shutdown. This is to prevent the battery from being bricked if not immediately charged after shutdown.
IF the vehicle is left to sit, and self discharge drains that buffer, the vehicle will not be able to be charged by the owner. You will then have two options. Option 1, Tow the vehicle to tesla, or 2, have mobile service visit.
Once that buffer is drained, A service technician will need to connect to the vehicle. The vehicle will need to preform a full battery diagnostic before charging begins. This can take some time, as well as charge rate will be extremely slow, until that 4kWh buffer is filled. It does this to make sure there is no damage to the cells. If it fails the test, you will need a replacement battery (per tesla's policies, not going to comment on third party repair).

So again, what you were able to drive below 0 rated range, was not the buffer. It was just the BMS calculation being inaccurate. A couple discharges below 10% SOC then all the way to 100% SOC without interruption will help the BMS recalculate out useable pack capacity.

Shutdown will happen regardless of the calculated usable capacity, as once the lowest voltage cell hits shutdown voltage, that's it, you are done.

 

[brainhouston]

Anyone know what voltage is the true zero cut off with 4kwh buffer left?

 

[islandbayy]

Anyone know what voltage is the true zero cut off with 4kwh buffer left?

It's changed a few times over the years, Not sure what the current shutdown voltage is, however, I'm seeing approximately 2.5v as a hard limit under load. I'd guess when it cannot sustain 2.5v per cell under any load/circumstance, thats when it will shut down.
Without doing it myself, which I have in the past to see these numbers, though it's been a really long time, 2014-ish, and I didn't have BMS access back then. I'd need someone like @wk057 or @Recell to comment on shutdown voltages.

 

[ajadan]

The buffer is capacity that is not accessible to the driver. Your BMS's capacity calculation was just off. Once the car shuts down, that 4kWh buffer, is what is still left in the pack AFTER shutdown. This is to prevent the battery from being bricked if not immediately charged after shutdown.
IF the vehicle is left to sit, and self discharge drains that buffer, the vehicle will not be able to be charged by the owner. You will then have two options. Option 1, Tow the vehicle to tesla, or 2, have mobile service visit.
Once that buffer is drained, A service technician will need to connect to the vehicle. The vehicle will need to preform a full battery diagnostic before charging begins. This can take some time, as well as charge rate will be extremely slow, until that 4kWh buffer is filled. It does this to make sure there is no damage to the cells. If it fails the test, you will need a replacement battery (per tesla's policies, not going to comment on third party repair).

So again, what you were able to drive below 0 rated range, was not the buffer. It was just the BMS calculation being inaccurate. A couple discharges below 10% SOC then all the way to 100% SOC without interruption will help the BMS recalculate out useable pack capacity.

Shutdown will happen regardless of the calculated usable capacity, as once the lowest voltage cell hits shutdown voltage, that's it, you are done.

I’ve been fed this info many times, and heard lots of conflicting info on whether the buffer is accessible or not. My drive well below 0% was after recalibrating my BMS many times (more than a dozen) as recommended by Tesla and others on here as one of the bricks in my battery was dying. High charge, low charge, sit at high charge, leave pack balancing above 93%, sit at low charge, DC charge, level 1 charge, none of it made any difference on the calibration.

Long prior to that, I had the car shut down due to “no energy remaining” with 3% showing, that was bad BMS calibration, but I was able to recharge no problem after getting to a power source, so clearly didn’t get to the brick point. I’m not confident I didn’t drive into the buffer portion of the battery when I went well past 0%. Obviously many threads on here debating this, but everything from the EPA-derived range (driving until they can’t drive further, not 0%) and my own experiences do not fit the inaccessible buffer theory.

 

[ran349]

My car has been parked for 12 hours and now sits at 52.6% SoC (and 51% Usable (Battery pack is at 2.5 Celsius)) and Min Cell Voltage is 3.722 V. So BMS surely thinks that 50% of energy is below 3,722 V meaning that 50% Voltage is approx 3,722 * 50 / 51 == 3.65V, which is not far from the previously calculated (Model S 90D) 3,62V. But then my Nominal should be:
220,7 Ah * 3.65 V * 6 Bricks * 14 Modules == 67.67 kWh.
BMS now reports Nominal of only 64.5 kWh, so appears to be temperature calculated.

I can think of two possibilities:
A: ScanMyTesla uses Buffer incorrect (and why is it 4 kWh and not 2.4 kWh for my 350V Battery?)
B: Nominal could be wrong because of Cold Scandinavian Temperatures

BTW: For what it is worth TeslaMate calculates Usable as 65.5 kWh which is higher than ScanMyTesla Nominal
View attachment 1016148

New nominal full pack for an S 70 was 68.88 kWh. So what is being labelled as Usable (new) and Usable (now) in you chart is incorrect.
Those chart values are the same as SMT nominal full pack values.
The actual usable does not include the buffer so those readings should be 4.0 kWh less than what the chart says, or 64.8 (new), and 61.6 (now).

The good news for you is your degradation is only about 5%!

 

[ajadan]

New nominal full pack for an S 70 was 68.88 kWh. So what is being labelled as Usable (new) and Usable (now) in you chart is incorrect.
Those chart values are the same as SMT nominal full pack values.
The actual usable does not include the buffer so those readings should be 4.0 kWh less than what the chart says, or 64.8 (new), and 61.6 (now).

The good news for you is your degradation is only about 5%!

Probably lots of factors at play in why SMT and other apps' nominal capacity don't match the straight V * A * bricks * modules math, the main ones being:

- Voltage drop is not linear from 100%-->0%, therefore estimates of 50% voltage are just that, estimates. Though estimating from 51% should be darn close.
- CAC is also just that, calculated, presumably backwards from nominal capacity and voltages, but I don't know that for sure. It takes some time for CAC to show up, and I have to assume some variables we can't say play into the estimation of min, max, and avg CAC. In addition, using the average may or may not be a good estimate depending on if means or medians are being used by the BMS. Average to me says mean, not median, but who knows.
- Buffer may or may not be accessible (or some combo of both) and may or may not be 4kWh (shown) vs 2.4kWh.
- Temp may be a factor as mentioned above.

A 70 battery should not be voltage capped, it's just a 14/16 module version of the larger batteries of it's generation (85 & 90). The very original 40kWh batteries were voltage capped as they could be software upgraded to a 60.

 

[ran349]

Probably lots of factors at play in why SMT and other apps' nominal capacity don't match the straight V * A * bricks * modules math, the main ones being:

- Voltage drop is not linear from 100%-->0%, therefore estimates of 50% voltage are just that, estimates. Though estimating from 51% should be darn close.
- CAC is also just that, calculated, presumably backwards from nominal capacity and voltages, but I don't know that for sure. It takes some time for CAC to show up, and I have to assume some variables we can't say play into the estimation of min, max, and avg CAC. In addition, using the average may or may not be a good estimate depending on if means or medians are being used by the BMS. Average to me says mean, not median, but who knows.
- Buffer may or may not be accessible (or some combo of both) and may or may not be 4kWh (shown) vs 2.4kWh.
- Temp may be a factor as mentioned above.

A 70 battery should not be voltage capped, it's just a 14/16 module version of the larger batteries of it's generation (85 & 90). The very original 40kWh batteries were voltage capped as they could be software upgraded to a 60.

I think the bottom line is that CAC is useful for comparing cell to cell, but not the best method for determining nominal capacity for reasons like you mentioned.

 

[NV Ray]

I am the founder/sole admisistrator of BMS_u29/018 Facebook. Scan My Tesla is heavily used by our 2000+ members as an owner's tool to either check their health of a current good pack or failed pack. The Group is fortunate to have the developer, Amund from Norway as a member of the Group. Recently he shared some info that might be useful to this conversation:

"I am realizing there is a lot of interest for Scan My Tesla in this group. I appreciate the support, and want to fix some issues I have been made aware of. I will roll out some updates to improve the situation for you. The backstory is that I added the CAC values when I stumbled upon them, but was never sure they were neither useful nor correct. I did not see any interest from any users, until I was invited to this group, so it was not added to the classic Android app properly (I believe it's there for Model 3, but not Model S).
1. Add Model S/X CAC to the 'classic' Android app. Until the fix is out - use the beta version
2. Add CAC to the BMS tab. Some users struggle with getting CAC to show in the Battery tab. I believe these have slower bluetooth adapters, where the slower CAC signals drown in other data, and might never reach the app. The BMS tab only recieves 1 slow signal, so CAC should make it just fine along with the cell voltages. Workaround is to use the graphical CAC/Cell balance dashboard
3. Fix for Model 3/Y/Plaid CAC signals. These broke in a firmware update a while ago, but I have found the new way to interpret them after a lot of hard work, probably the most time I've ever spent on 3 signals. Not directly relevant to u029/u018, but now that you know the value of the CAC readings, the newer cars will also need it as they age.
PS. The IOS app will not have any of these problems, but it will still inherit the addition of CAC to the BMS tab. I also have a re(tro)-designed BMS tab in the pipeline, so the new apps will get a similar heatmap of cell voltages as the classic app."

 

[coryjfowler]

I am the founder/sole admisistrator of BMS_u29/018 Facebook. Scan My Tesla is heavily used by our 2000+ members as an owner's tool to either check their health of a current good pack or failed pack. The Group is fortunate to have the developer, Amund from Norway as a member of the Group. Recently he shared some info that might be useful to this conversation:

"I am realizing there is a lot of interest for Scan My Tesla in this group. I appreciate the support, and want to fix some issues I have been made aware of. I will roll out some updates to improve the situation for you. The backstory is that I added the CAC values when I stumbled upon them, but was never sure they were neither useful nor correct. I did not see any interest from any users, until I was invited to this group, so it was not added to the classic Android app properly (I believe it's there for Model 3, but not Model S).
1. Add Model S/X CAC to the 'classic' Android app. Until the fix is out - use the beta version
2. Add CAC to the BMS tab. Some users struggle with getting CAC to show in the Battery tab. I believe these have slower bluetooth adapters, where the slower CAC signals drown in other data, and might never reach the app. The BMS tab only recieves 1 slow signal, so CAC should make it just fine along with the cell voltages. Workaround is to use the graphical CAC/Cell balance dashboard
3. Fix for Model 3/Y/Plaid CAC signals. These broke in a firmware update a while ago, but I have found the new way to interpret them after a lot of hard work, probably the most time I've ever spent on 3 signals. Not directly relevant to u029/u018, but now that you know the value of the CAC readings, the newer cars will also need it as they age.
PS. The IOS app will not have any of these problems, but it will still inherit the addition of CAC to the BMS tab. I also have a re(tro)-designed BMS tab in the pipeline, so the new apps will get a similar heatmap of cell voltages as the classic app."

That certainly explains why I do not see those values on my vehicle. And Facebook has yet to show me that information.

 

[ArgusIII]

Of course I made a service center appointment immediately, and when I arrived I was quickly informed I'd be getting a new HV battery under warranty! So that definitely made my day. They'll have the car for a couple weeks. I don't have an invoice to note what part# I'm getting, but the service center thought it would be new, not reman. Not sure if reman 90 packs exist (I hope for the sake of their original crappiness that they don't).

What replacement battery did you end up receiving? Can you provide the invoice part number? Was it new or remanufactured?

 

[jensk2]

New nominal full pack for an S 70 was 68.88 kWh. So what is being labelled as Usable (new) and Usable (now) in you chart is incorrect.
Those chart values are the same as SMT nominal full pack values.
The actual usable does not include the buffer so those readings should be 4.0 kWh less than what the chart says, or 64.8 (new), and 61.6 (now).

The good news for you is your degradation is only about 5%!

Hmmm, According to Jason, a Model S 70 2015 had 71.2 kWh Brutto and 68.8 kWh Netto - that is why I entered 68.8 kWh as Initial

TeslaMate Battery Health page allows one to specify initial capacity. If one does not do that, the Page will use whatever maximum Capacity is hidden in the Teslamate historic data. That will report an Initial Capacity of 66.1 kWh because I didn't start using Teslamate untill summer 2021 at which time the best charge revealed Projected Available as 66.1 kWh, So it estimates that my degradation has almost stopped (I cycle between 20% SoC and 65% SoC mostly) and at least the Ideal Range has been almost constant at 340 km'ish (Although with a strange recovery 0m 2022-07-25 after the car had been unlogged for almost a week, due to Tesla forcing Password/Token update)

Using only registered data eliminates possible errors in assumed initial capacity (a problem that Tessie has with the many different Model Y battery packs, with very differing initial capacity), but ofcourse means, that one needs to enter 'assumed' initial capacity in order to get estimates, when the App/Logger was not running since car was new.

 

[jensk2]

Hmmm, According to Jason, a Model S 70 2015 had 71.2 kWh Brutto and 68.8 kWh Netto - that is why I entered 68.8 kWh as Initial

I looked up the Jason reported numbers and they were:

	Total	Usable
Original 60	61​	58,5​
85/P85/85D/P85D	81,5​	77,5​
90D/P90D	85,8​	81,8​
Original 70	71,2​	68,8​
75/75D	75​	72,6​
Software limited 60/60D	75​	62,4​
Software limited 70/70D	75​	65,9​

Interesting that I now have approx same capacity as a Software limited 70D, so Tesla made a good choice on setting the Software limit lower on the newer S 70D's as they knew they would degrade slower, when users could only charge to 88-92% .