Tesla's 85 kWh rating needs an asterisk (up to 81 kWh, with up to ~77 kWh usable)

[Bangor Bob]

As previously stated Tesla probably designed the 85kWh pack first, using the numbers from Panasonic. When they then started designing the 60 they realized that the cells didn't quite make the specs from Panasonic in the real world. My guess is originally they fully intended the 60 pack to be a 65 pack.

Well... Except the 60 is closer to 65 than the 85 is to 85.

 

[NuclearPowered]

If these cells were in a car they would only have lost roughly 2.2% range after ~60,000 miles. Again, the cells are awesome cells and are perfect for their application. They just don't meet the advertised capacity spec.

Any clarity using your info into why the rated range for many owners indicates loss significantly greater than 2.2% after fewer miles?

 

[wk057]

Any clarity using your info into why the rated range for many owners indicates loss significantly greater than 2.2% after fewer miles?

Because the BMS is bad at capacity estimates if you short cycle a lot. For example, charging to 80% to 65% and recharging to 80% regularly is a surefire way to throw the BMS all out of whack. Eventually I guess Tesla might fix this, and some people will magically get some range back.

Then again, there are a lot of factors that could cause real degradation, like always charging to 100%, or leaving the car at high SoC in the heat and such. Even with thermal management it's not going to mitigate 100% of the accelerated degradation these things cause.

 

[dhanson865]

I understand your concern, but you can compare the advertised range between different battery packs and base your purchase on that. The range is correct, since it is tested under lab conditions by an independent organisation (EPA). I think most (if not all) buyers will only look at the advertised range and not so much to the advertised capacity.

I wish people would stop saying this. EPA doesn't test range on cars, and they don't test efficiency on most cars.

Download the test data from Test Car List Data Files | Cars and Light Trucks | US EPA and pay attention to the field called "Test Originator" and note how 99% of them say "MFR" and a few say "EPA". As in the EPA didn't test most of those cars and they have never tested Tesla cars. For example on the 2016 spreadsheet only 14 of 75 Toyota Prius configurations say "EPA" and the other 61 say "MFR". In the case of Tesla that is 12 "MFR" and 0 "EPA". I personally didn't download 2012-2015 and check them all manually so if you want to go through the trouble and tell us if I'm wrong I'll be happy to go back and download additional data files to confirm but only if you do the labor first and tell me which year the EPA bothered to test a Tesla vehicle if any at all.

I'm not sure why you assume or believe that the EPA tests these vehicles that they aren't testing. I can tell you most of the time they didn't.

I'll leave you with this blurb from Emission and Fuel Economy Test Data | Cars and Light Trucks | US EPA

Fuel Economy Test Data

All new cars and light trucks sold in the U.S. are required to have a economy label posted on the window sticker. The label contains the city and highway miles-per-gallon values, as well as other related information. To calculate these values, laboratory tests are performed on pre-production vehicles. Most testing is performed by manufacturers at their own testing facilities. EPA audits the data from this testing and performs its own testing on some of these vehicles to confirm the manufacturers' results. All of this test data is available for review and analysis, and is presented in two formats, One is a downloadable, detailed description of vehicle fuel economy tests, formatted for printing. The other is a downloadable, delimited vehicle description of data for importing into databases/spreadsheets Regardless of format, the data is presented on a model year basis.

 

[stopcrazypp]

It is incredible. Don't get me wrong, these cells are awesome and are performing extremely well in my testing as far as degradation goes. My point is that they just are not 11.965Wh cells in any test under any condition that I've been able to try.

I want to hold off on releasing a bunch of data from my tests until I can publicly release info on the devices I'm using, along with summaries of the mountains of data I'm generating. But, I'll answer your question. This particular test is showing incredibly low degradation. My car-like test has been running for about 6 months now and the equivalent of about 10,000 miles per month of wear (my fake driver must only sleep while the fake car charges hehe). Admittedly, I estimated the anti-brick buffer at 3.5 kWh back when I started the testing, so I even discharge them a little more at times than Tesla would when my function allows the car to be fully drained (few times per month). If these cells were in a car they would only have lost roughly 2.2% range after ~60,000 miles. Again, the cells are awesome cells and are perfect for their application. They just don't meet the advertised capacity spec.

Makes sense. Keep in mind that the Panasonic datasheets generally don't include Wh, only Ah. I think Wh is the more important number, personally.

I understand you will have concerns about releasing data given the special equipment you are using. I will wait until you get the permission to do your data dump (along with releasing test conditions) before commenting further. Doing this discussion in a Q&A format (rather than just looking at data with laid out test conditions, like for example in a datasheet) makes it seem very nit-picky, so I don't want to give that impression. Thank you for taking the time to address the questions raised.

- - - Updated - - -

Quick question for the age / degradation theorists...

How do you explain my 60kWh battery having more than 60kWh after 20 months and 12k miles?

All this theorising of how Tesla got to the 85 figure is fine, but doesn't explain how they were accurate on the 60.

For me the range / performance / specs are fine if consistent across the board. This is yet another case of stretching the gap between models via differing measurement standards. Clearly aimed at boosting ASP. I had thought Tesla were beyond these sorts of used car dealer tactics, but clearly they are not.

Fine while they have no competition, but I've personally lost trust to the point I'd consider an alternative manufacturer for my replacement car.

I have a simple non-malicious theory to explain the whole 60kWh vs 85kWh deal, but it keeps getting buried. It is slightly different than JRP3's. They announced capacities first, designed later (in a 85kWh -> 60kWh -> 40kWh progression rather than all at once), but ran into degradation/discharge/charge rate constrains in terms of choosing the number of cells to blank out (they ended up blanking out 10 cells per group, when they could have blanked out 14 to match same ratio as 85kWh).

Way back then my original theory was that 60kWh pack would either use different cells (slightly more power optimized) or would be the same exact modules but with fewer. Turns out neither was true: they instead chose to blank out cells per module (with two fewer modules). This has all the signs of a last minute design (that came well after they announced capacities).

I find it weird how people have to always assign malicious intentions but completely forget the history of what happened. Long story short: Battery capacities announced late-2011, the Model S was produced in June 2012 with 85kWh as the main pack, 60kWh came half a year later in January 2013 (40kWh was cancelled April 2013). The decisions made on what to do in the production version likely came far later than when they announced the 60kWh and other battery capacities in late 2011 (and it did not make sense to change the number).

I say this especially because they didn't even have the 40kWh developed yet up to the point they cancelled it and if you look at the 60kWh pack (which has cells blanked out), I think it is very obvious Tesla did not design the pack in a systemic way for all the capacities from the start.

My comment below:

 

[DavidB]

Because the BMS is bad at capacity estimates if you short cycle a lot. For example, charging to 80% to 65% and recharging to 80% regularly is a surefire way to throw the BMS all out of whack. Eventually I guess Tesla might fix this, and some people will magically get some range back.

In your opinion, wk, is there any harm to almost always only charging to 65%, other than the estimating software getting out of wack?

 

[Dbitter1]

So I just have to ask... is anyone taking bets on what a 70 or 90 battery pack really is?

 

[Quantum`]

Nothing in the owner's manual says anything anywhere about 90%.

The three things that shorten a LIon's life are
- Heat
- Repeated excessive discharge
- Number of charge/discharge cycles.

Unless the pack's BMS is so poorly-designed as to allow excessive heat buildup on full-charge trickle, there is no harm in charging to 100%.

I realize that this is unheard-of heresy, but it's also established fact.

 

[dhanson865]

Nothing in the owner's manual says anything anywhere about 90%.

The three things that shorten a LIon's life are
- Heat
- Repeated excessive discharge
- Number of charge/discharge cycles.

Unless the pack's BMS is so poorly-designed as to allow excessive heat buildup on full-charge trickle, there is no harm in charging to 100%.

I realize that this is unheard-of heresy, but it's also established fact.

Unheard of? It's the daily existence for Nissan Leaf owners in the southern US.

A pack with no active cooling and a BMS that allows it to bake itself while charging, check and check.

The only silver lining is leaf owners don't care about charge cycles or miles on the odometer because heat degradation swamps that variable into non relevance.

 

[Beltsbear]

Nothing in the owner's manual says anything anywhere about 90%.

The three things that shorten a LIon's life are
- Heat
- Repeated excessive discharge
- Number of charge/discharge cycles.

Unless the pack's BMS is so poorly-designed as to allow excessive heat buildup on full-charge trickle, there is no harm in charging to 100%.

I realize that this is unheard-of heresy, but it's also established fact.

the fourth item you left out is being in high and low states of charge. Over 90% and under 10% shorten life too.

 

[stopcrazypp]

Nothing in the owner's manual says anything anywhere about 90%.

The three things that shorten a LIon's life are
- Heat
- Repeated excessive discharge
- Number of charge/discharge cycles.

Unless the pack's BMS is so poorly-designed as to allow excessive heat buildup on full-charge trickle, there is no harm in charging to 100%.

I realize that this is unheard-of heresy, but it's also established fact.

Sitting at 100% SOC is a sure way to kill a battery quickly. Cycling to 100% and using it up quickly however is a lot better.

 

[smac]

I understand your concern, but you can compare the advertised range between different battery packs and base your purchase on that. The range is correct, since it is tested under lab conditions by an independent organisation (EPA). I think most (if not all) buyers will only look at the advertised range and not so much to the advertised capacity.

A small point of technical trivia EV manufacturers have an exception in the rules for independent testing. IOW Tesla just submit their own numbers.

In fact the official EPA submission by Tesla for the 70, understates range with a little * and a footnote saying as much.

 

[AntronX]

The three things that shorten a LIon's life are
- Heat
- Repeated excessive discharge
- Number of charge/discharge cycles....

According to Prof. Jeff Dahn of Dalhousie University, the biggest factor that shortens lithium battery life is time spent at high voltage. Here is an excerpt from his presentation. That entire presentation is worth watching for the depth of technical detail alone.

 

[jpet]

Because the BMS is bad at capacity estimates if you short cycle a lot. For example, charging to 80% to 65% and recharging to 80% regularly is a surefire way to throw the BMS all out of whack. Eventually I guess Tesla might fix this, and some people will magically get some range back.

And my car is a good example of the BMS freaking out for exactly this reason. I have registered ALL of my range charges in the MaxRange Tesla Battery Survey sheet (look up jpet) and now I can compare my "Tesla estimated" max range with 224 other cars.

I cycle between 30 and 70% because I know this is better to get more cycles out of the cells. I only charge above 70% when I need the range (and that is not often). The result of doing things a little differently than the majority is that I have an estimated range that is probably the worst from the group (see the purple dots).

So what am I going to do about it? Nothing! The range displayed in the car is nothing more than a best estimate and unless you torture the cells a lot, this estimate is going to be off by quite some margin. The majority will KISS and that's fine but at the end of the day (make that a significant number of years) degradation and cycle life will not be equal.

 

[OddB]

And my car is a good example of the BMS freaking out for exactly this reason. I have registered ALL of my range charges in the MaxRange Tesla Battery Survey sheet (look up jpet) and now I can compare my "Tesla estimated" max range with 224 other cars.

I cycle between 30 and 70% because I know this is better to get more cycles out of the cells. I only charge above 70% when I need the range (and that is not often). The result of doing things a little differently than the majority is that I have an estimated range that is probably the worst from the group (see the purple dots).

View attachment 110538

So what am I going to do about it? Nothing! The range displayed in the car is nothing more than a best estimate and unless you torture the cells a lot, this estimate is going to be off by quite some margin. The majority will KISS and that's fine but at the end of the day (make that a significant number of years) degradation and cycle life will not be equal.

Interesting !
Here are a recent survey among the norwegian owners. Only one data point pr car.
The apparent large differences can probably be explained by inaccurate readings, bad BMS estimates AND (not mentioned in this thread (?)) unbalanced cells.
As I understand BMS estimate errors doesnt hinder you in utilize the full capacity but bad balancing will ?

 

[Pajda]

At first I must say great thank you to wk057 for doing this tests and presenting results. But this discussion starts to be strange. So again I ask the fundamental question:

Is there anybody who can present another test results or graphs which shows that wk057 results are wrong?

Till now I don`t have lucky to get some Tesla cells for my own tests, but from my experiences with many other industrial grade 18650 cells I have no reason to believe that wk057 results are wrong.

 

[jpet]

As I understand BMS estimate errors doesnt hinder you in utilize the full capacity but bad balancing will ?

BMS estimate errors will indeed not hinder using the full capacity and that's probably why some people like to take the risk driving below the 0 range point.
When the pack is not fully balanced the cells with the lowest voltage will probably make the BMS shut the whole system down to protect them.
But that is not a concern I have since I cycle between 30 and 70% which leaves plenty of room for the cells to be unbalanced.
When I need to have the full capacity / range available, I first do a range charge which triggers the balancing so that things happen naturally.
You don't really have to think about the battery. It manages itself pretty well. If you like, you can adopt a charge habit that optimizes cell life but with an 8 year warranty there's no need to worry for most owners.

 

[CO2CLEAN]

I wish people would stop saying this. EPA doesn't test range on cars, and they don't test efficiency on most cars.

Download the test data from Test Car List Data Files | Cars and Light Trucks | US EPA and pay attention to the field called "Test Originator" and note how 99% of them say "MFR" and a few say "EPA". As in the EPA didn't test most of those cars and they have never tested Tesla cars. For example on the 2016 spreadsheet only 14 of 75 Toyota Prius configurations say "EPA" and the other 61 say "MFR". In the case of Tesla that is 12 "MFR" and 0 "EPA". I personally didn't download 2012-2015 and check them all manually so if you want to go through the trouble and tell us if I'm wrong I'll be happy to go back and download additional data files to confirm but only if you do the labor first and tell me which year the EPA bothered to test a Tesla vehicle if any at all.

I'm not sure why you assume or believe that the EPA tests these vehicles that they aren't testing. I can tell you most of the time they didn't.

I'll leave you with this blurb from Emission and Fuel Economy Test Data | Cars and Light Trucks | US EPA

I stand corrected, I did not know that the EPA does not test the cars itself (except for a few, as you mentioned above). I am from Europe, and the NEDC test (European EPA-like test) is done by an independent organisation, that is why I incorrectly assumed that the EPA did all tests themselves.

 

[jpet]

wk057 has more than deserved our respect and gratitude for sharing his research and findings with the community.
So without further ado, I would like to officially nominate him for the title of "The leader of the pack". :tongue:

 

[Andyw2100]

wk057 has more than deserved our respect and gratitude for sharing his research and findings with the community.
So without further ado, I would like to officially nominate him for the title of "The leader of the pack". :tongue:

Careful! With the kind of research and findings he's been coming up with, you don't want to give Tesla any ideas. (They do have OTA access to his brakes!)