What does the battery degradation curve look like?

[kevincwelch]

dsm363-

Thanks.

Wondering how the 90% fits into that degradation curve now. Either way, battery life sounds great!

 

[Zzzz...]

Excellent post, johnchamplinhall, thank you!

The NCA you were dealing with, what electrolyte were used there? Lithium hexaflurophosphate based with something like Ethylene/Diethyl carbonates? Or something more funky? Like lithium imide etc?

 

[pguerra]

I have to vote johnchamplinhall as having the best first post of all users here in the forum. Thank you for such an informative and timely post. Man, I hope you're right about the NCA in Model S. 20yrs is crazy long...

 

[Citizen-T]

At some point Elon said they expected the battery to be useful for most automotive applications for twice the warranty period. So that's 16 years and is pretty well in-line with the above model.

 

[stopcrazypp]

I've posted the cycle testing graph of the actual cells that the 85kWh Model S uses. After 500 cycles, the cells retain about 76.6% the original capacity. In miles, 500 cycles is about equal to 500 x 88% x your full range when the car is new. For example if you get the EPA rated 265 miles on a full charge, then 500 cycles ~500*.88*265 miles=116600 miles. The 88% is there as a linear approximation for the degradation over the 500 cycles.
http://www.teslamotorsclub.com/showthread.php/7747-Standard-Warranty-Revealed/page13?p=172284&viewfull=1#post172284

As I note in my comment, degradation slows down significantly after about 300 cycles (it follows #3 of the OP's suggested graphs). The rate is about ~1.35% loss per 100 cycles. So it might take another 488 cycles before it reaches 70% capacity (assuming there isn't a sharp drop at some point for other reasons). So likely past 200k miles before it reaches 70%.

I also have a link to satellite based calendar life testing (it seems that's the industry where calendar life is very important) and it seems to follow what johnchamplinhall posted (of course his data is more updated).

 

[Johan]

I've posted the cycle testing graph of the actual cells that the 85kWh Model S uses. After 500 cycles, the cells retain about 76.6% the original capacity. [...]

What I can't quite figure out from the above graphs, and the original document from Panasonic (http://industrial.panasonic.com/www-data/pdf2/ACA4000/ACA4000CE240.pdf), is if this is the equivalent to constantly range charging the Model S, then driving it so that the discharge rate is constant at 1C, emptying the battery to 0%, waiting 20 minutess, then range charging again etc. It says in the graph that the cells start off at 4.2 V (range charge equivalent?) and discharges at 1.0lt (equivalent of 1C?) and down to 2.5 V (equivalent of empty/0% on the Model S?) and doing this in a temperature managed state (at 25 deg. C or maybe at maximum 25 deg. C?).

If what I interpreted above is the case, than this is the worst case scenario since you are doing constant cycles of range charging! You would also need to take in to account that when accelerating the car forcefully you are discharging at burst of up to 3.5C, but while driving at steady speeds you will likely be discharging at well below 1C (0.5C has been said for constant speed highway driving at 60 mph). So driving style (track driving for example) is likely to have some sort of impact on battery degradation, but to what extent? Constant range charging will have the most detrimental effect IMO. Often driving the battery close to 0% will probably also have some detrimental effect, but less than frequent range charging. All of this also naturally is dependent on effective and reliable temperature management (which it seems Tesla has on the Model S, not so much on the Roadster with the PEM fan issues etc.).

 

[eledille]

What I can't quite figure out from the above graphs, and the original document from Panasonic (http://industrial.panasonic.com/www-data/pdf2/ACA4000/ACA4000CE240.pdf), is if this is the equivalent to constantly range charging the Model S, then driving it so that the discharge rate is constant at 1C, emptying the battery to 0%, waiting 20 minutess, then range charging again etc. It says in the graph that the cells start off at 4.2 V (range charge equivalent?) and discharges at 1.0lt (equivalent of 1C?) and down to 2.5 V (equivalent of empty/0% on the Model S?) and doing this in a temperature managed state (at 25 deg. C or maybe at maximum 25 deg. C?).

This is pretty aggressive cycling, and much worse than any car is likely to be subjected to. If I'm not mistaken, 4.2V is absolutely full and 2.5 V is completely empty. I would be surprised if Model S will charge above 4.1 V and discharge below 3.0 V. Also, charge rate for the 85 kWh battery would be 42.5 kW and discharge rate 85 kW, which is quite a bit.

What are the empty and full cell voltages for the Roadster? *edit* 4.15 and 3.0 volts, see here A Bit About Batteries | Blog | Tesla Motors. Standard mode charges to 4.10 V. This means that 4.2 V is much worse than range mode charging a Roadster.

This looks promising...

 

[Johan]

[...]
This looks promising...

Let's just not forget that the graphs and testing above does not take into account battery ageing over several years, which comes in addition to the charge/discharge degradation. We know for a fact that there are some Roadsters out there that have (sadly) been just sitting and gone just a few hundred miles since 2008/2009 (George Cloney's for example, recently auctioned out). Would be interesting to know how those batteries are doing? Also would be interesting to know if sitting at standard charge v.s. storage mode has a lot of impact? (Note: Model S doesn't come with a storage mode as far as I've gathered).

 

[eledille]

Let's just not forget that the graphs and testing above does not take into account battery ageing over several years, which comes in addition to the charge/discharge degradation.

As far as I understand it, johnchamplinhall's graph takes aging into account too.

johnchamplinhall: Great post, thanks! Why is the vertical axis in volts? I would have expected mAh or %?

 

[Johan]

As far as I understand it, johnchamplinhall's graph takes aging into account too.

Yes, true. I was referring to Panasonic's testing of the NCR18650 cells - this testing seems to have been done over a few days/weeks (however much time it takes to charge and discharge 500 times in a row).

 

[stopcrazypp]

What I can't quite figure out from the above graphs, and the original document from Panasonic (http://industrial.panasonic.com/www-data/pdf2/ACA4000/ACA4000CE240.pdf), is if this is the equivalent to constantly range charging the Model S, then driving it so that the discharge rate is constant at 1C, emptying the battery to 0%, waiting 20 minutess, then range charging again etc. It says in the graph that the cells start off at 4.2 V (range charge equivalent?) and discharges at 1.0lt (equivalent of 1C?) and down to 2.5 V (equivalent of empty/0% on the Model S?) and doing this in a temperature managed state (at 25 deg. C or maybe at maximum 25 deg. C?).

Yes, your interpretation is correct. It's basically cycling from absolute 100% SOC 4.2V and absolute 0% SOC for the cells. The Model S won't actually allow you to do that even if it indicates SOC is 100% (it'll be 90-something percent and a couple percent SOC left on the bottom). The charging rate is 0.5C, so it's not quite supercharging (it's half speed). So this is close to a worse case in terms of cycling. Ambient temperature is 25C (I don't think there's any active cooling involved, it's designed to simulate a typical environment instead).

This graph does not factor in calendar losses and manufacturers rarely have graphs of calendar testing. From looking at battery loss models used by the US NREL, the losses are not additive, it's whichever one is worse. So it's a race between the two. From the satellite 18650 test results I found, calendar losses are about 1% loss per year in the years after the first year (which is about 7%). So in most cases, I think the cycle losses will be worse unless you have the car sitting.
http://www.teslamotorsclub.com/show...evealed/page14?p=172328&viewfull=1#post172328

 

[gaswalla]

Error

 

[mhpr262]

I do recall an article about a Tesla Modes S batterypack being tested in a lab and after 165.000 simulated miles it was still at 85% capacity. Maybe they didn't use 100%-0% cycling but something closer to what the pack would normally undergo in daily use. Either way the result is pretty encouraging. Even 70% of 265 miles is a pretty useful range, and if it stays that way for a decade after ....

Edit: I think this is where I found the info:

Recent Articles Slamming Tesla for Battery Pack Issues | Forums | Tesla Motors

about 2/3 down:

bshortell | June 28, 2012
At the Newport Beach Fashion Island event, an engineer stated that they already had an 85kW battery pack with over 165,000 miles and it is at 85% capacity.

 

[fasferrari]

I received my Model S Performance 10 days ago. I am waitibg for my high speed charging devise to be delivered so I may install it at the garage in my house.
In the interim I am using my 240 Voltec charger I had for my Volt.

THe first tiome I charged my Model SP it read 248 miles. The second time 238 and the third time 228. I used the standard charge setting each time. At this rate in one week it will only charge up to 100 miles. IS THIS COMMON?
Also I went on a 130 mile round trip. At the start the car read 238 miles. When I returned home after doing 131 miles, the car had only 67 miles lft. In other words 41 miles were lost. I drove very conservatively and the weather was in the 50"s. Is this typical?

.... [Mod note: Personal e-mail removed]

I am concerned that I should have kept my Volt with the gasoline engine--which Volt only did 42 mile on a 4 hour battery charge in the summer without air conditioning and 20 miles on a winter day with terrible heating system.
Thank you very much.

 

[wraithnot]

I received my Model S Performance 10 days ago. I am waitibg for my high speed charging devise to be delivered so I may install it at the garage in my house.
In the interim I am using my 240 Voltec charger I had for my Volt.

THe first tiome I charged my Model SP it read 248 miles. The second time 238 and the third time 228. I used the standard charge setting each time. At this rate in one week it will only charge up to 100 miles. IS THIS COMMON?
Also I went on a 130 mile round trip. At the start the car read 238 miles. When I returned home after doing 131 miles, the car had only 67 miles lft. In other words 41 miles were lost. I drove very conservatively and the weather was in the 50"s. Is this typical?

.... [Mod note: Personal e-mail removed]

I am concerned that I should have kept my Volt with the gasoline engine--which Volt only did 42 mile on a 4 hour battery charge in the summer without air conditioning and 20 miles on a winter day with terrible heating system.
Thank you very much.

My relatively new (almost four weeks old and 2k miles) 85 kWh MS usually shows 242 miles immediately after a standard charge, but the vampire load will often reduce this to 238 or so by the morning. Did you observe your 228 value immediately after the smartphone app showed "charging complete" or was it the next morning? Some vampire load and a little cold whether might explain what you are seeing.

As far as the real world range, I tend to get about 80% of the EPA rated range when I drive on the freeway with the flow of traffic (75+ mph around here). This seems reasonably consistent with your numbers. If I understand the EPA rated range correctly, it is for a blend of city and highway driving and the highway driving is at the speed limit. You'd have to set the cruise control to something like 62 mph and not need much cabin heating or cooling to get the EPA rated range on the highway. From what I hear, rain or a headwind will also have an effect on range. Speeding up and slowing down a lot will also reduce range.

 

[Discoducky]

Hoping more owners can state their standard range charge shows right after completing the charging as well as stating how many miles they've driven.

Would be really interesting to hear from owners with over 10K miles as a comparison to what Roadster owners saw/see at that mileage.

 

[Zythryn]

Hoping more owners can state their standard range charge shows right after completing the charging as well as stating how many miles they've driven.

Would be really interesting to hear from owners with over 10K miles as a comparison to what Roadster owners saw/see at that mileage.

9100 miles, over 6 months, standard charge results in a 240-241 range.
Roadster over 27,000 miles and 27 months, range loss a bit under 3%.
YMMV

Please note, the range is noted generally 3-5 hours after charge completes. Software version 4.3.

 

[Tommy]

Hoping more owners can state their standard range charge shows right after completing the charging as well as stating how many miles they've driven.

Would be really interesting to hear from owners with over 10K miles as a comparison to what Roadster owners saw/see at that mileage.

Just competed 12.5k mile service this weekend. A standard charge yields 233 miles right after completing the charging. The charging is 240v/40A. Tesla tech confirmed the drop from my initial 242/240 miles after a standard charge to now 233 miles is nothing to be worried about.

 

[bhuwan]

Just competed 12.5k mile service this weekend. A standard charge yields 233 miles right after completing the charging. The charging is 240v/40A. Tesla tech confirmed the drop from my initial 242/240 miles after a standard charge to now 233 miles is nothing to be worried about.

3% degradation in less than a year? Did I calculate correctly?

 

[Discoducky]

3% degradation in less than a year? Did I calculate correctly?

Yeah, I just did the same calc. 70% life after my 125K miles in 6 years sounds tough to hit using that measuring stick.