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

[vgrinshpun]

OK, what do you believe S85 battery total and usable capacity is?

 

[llavalle]

OK, what do you believe S85 battery total and usable capacity is?

I believe that the usable capacity on a "85" Tesla is around 77 or 78kWh.

I'm basing this info on the following (using miles for simplicity)

1. The in-car value "VAPI_ratedWattHourPerMile", on a 85D is 290 wh/mi and it's rated range is 270miles. 270 * 290 = 78.3kWh
2. Driving around, you'll end up with "estimated range" = "rated range" on the main display when the avg consumption is 290wh/mi. Brand new, my car was showing 271miles rated when full. 271*290 = 78.5kWh
3. The in-car value "BMS_nominalFullPackEnergyRemaining" of my car has always been around 77-78. 77.6 right now. Was closer to 78.3 when new. Also unlike the "VAPI_idealEnergyRemaining" value, this one does change with SOC. nominalFullPackEnergy always reports the full, usable pack, calculated by the BMS.
4. When driving around with my car, every time I lose 10% of SOC (with 1 decimal point accuracy using the CAN values, not the in-dash values), the trip meter shows ~7.76kWh used. 7.76*10 = 77.6kWh
5. I once ran out of juice and had to be towed to the nearest supercharger. The on-board charging screen showed 78kWh was put back into the battery once it hit 100% (took more than 1.5h!) No CAN logs of that one.. and no decimal point accuracy.
6. Decided to re-do #5... went a week of commuting without charging then ran in circles around my block until I hit 0% and had the "charge now" message.. Then plugged it in for a full charge Measured the consumption from AC (eGauge with 0.5% certificate using revenue grade CT - this thing also checks voltage) and compared with data on the screen. I'd have to dig the data out, but from memory, apparent charging efficiency was 0.92% (gen 2 chargers) and I pulled around 84-85kWh from the wall (240V@80amps) while the car showed the same 78kWh as on the supercharger.
7. I also happen to own a module from a 85 pack (~15000miles care). It's 1/16th of a full battery pack.. and when charging it from 3.11V up to 4.19V (what the car uses as 0 and 100% respectively), my charger calculated 4866Wh. IR was negligible since I was charging at a very low rate (<500W). 4866*16 = 77.85kWh
8. Draining the same module with a very accurate 800W resistance, I can run the heater for about 6hours 6 * 800W = 4800. 4800*16 = 76800kWh <-- this value is slightly low because I could not accurately account for the voltage sag of the 800W load and when the load was removed, SOC was slighly above the 0% voltage.. so I guess I could have run it a couple of minutes more.

With all of this in mind I'm pretty sure that 77 or 78 value is solid. Lot of people can do the tests I did with similar results.


For the total capacity, the only thing I can use is my usable capacity plus a static value from the BMS acting as the buffer. In the case of a "85" car, the buffer is reported as 4kWh by the BMS.

Using my 77-78 usable value and adding 4 gives me 81-82kWh total.

If someone can find the low-voltage cutoff used for testing the full capacity (what's the SOC voltage when the buffer is empty), I could do test #7 with my module.

So are we missing 4.something kWh. Probably not. But we're surely missing at least 3kWh.... and I'm not even mentioning that I strongly believe that usable capacity should be advertised... not total. "Hey I just bought this Tesla, it has a 85kWh pack but wait.. but I can only use 78kWh".

 

[llavalle]

Yes, you can make that argument. But I think there is still significant difference between Tesla providing a 85kWh (and then it wears down to 82 kWh shortly) vs Tesla providing a pack that was 82kWh from the start.

I agree with that. It's just that in both case, I would find it misleading.

Yeah, I would take the round numbers with a grain of salt. That is likely just a manufacturer provided number. The test itself only measures usable energy (specifically what is required for car to keep up with drive cycle; so if car enters a limp mode, that part is not counted, even though technically the car can still drive some distance).

That's my thought too. Manufacturer provided number.

I've yet to find an actual measured Ah value (not even kWh) in any EPA documents for the Model S.

 

[llavalle]

Life is so weird... I just received a notification that one of the channel I'm following on Youture put out a new video...

"Why You Can't Win an Internet Argument - If you're still in an internet argument that started weeks ago, you're probably better off just letting it go."

Perfect timing

 

[vgrinshpun]

I believe that the usable capacity on a "85" Tesla is around 77 or 78kWh.

I'm basing this info on the following (using miles for simplicity)

1. The in-car value "VAPI_ratedWattHourPerMile", on a 85D is 290 wh/mi and it's rated range is 270miles. 270 * 290 = 78.3kWh
2. Driving around, you'll end up with "estimated range" = "rated range" on the main display when the avg consumption is 290wh/mi. Brand new, my car was showing 271miles rated when full. 271*290 = 78.5kWh
3. The in-car value "BMS_nominalFullPackEnergyRemaining" of my car has always been around 77-78. 77.6 right now. Was closer to 78.3 when new. Also unlike the "VAPI_idealEnergyRemaining" value, this one does change with SOC. nominalFullPackEnergy always reports the full, usable pack, calculated by the BMS.
4. When driving around with my car, every time I lose 10% of SOC (with 1 decimal point accuracy using the CAN values, not the in-dash values), the trip meter shows ~7.76kWh used. 7.76*10 = 77.6kWh
5. I once ran out of juice and had to be towed to the nearest supercharger. The on-board charging screen showed 78kWh was put back into the battery once it hit 100% (took more than 1.5h!) No CAN logs of that one.. and no decimal point accuracy.
6. Decided to re-do #5... went a week of commuting without charging then ran in circles around my block until I hit 0% and had the "charge now" message.. Then plugged it in for a full charge Measured the consumption from AC (eGauge with 0.5% certificate using revenue grade CT - this thing also checks voltage) and compared with data on the screen. I'd have to dig the data out, but from memory, apparent charging efficiency was 0.92% (gen 2 chargers) and I pulled around 84-85kWh from the wall (240V@80amps) while the car showed the same 78kWh as on the supercharger.
7. I also happen to own a module from a 85 pack (~15000miles care). It's 1/16th of a full battery pack.. and when charging it from 3.11V up to 4.19V (what the car uses as 0 and 100% respectively), my charger calculated 4866Wh. IR was negligible since I was charging at a very low rate (<500W). 4866*16 = 77.85kWh
8. Draining the same module with a very accurate 800W resistance, I can run the heater for about 6hours 6 * 800W = 4800. 4800*16 = 76800kWh <-- this value is slightly low because I could not accurately account for the voltage sag of the 800W load and when the load was removed, SOC was slighly above the 0% voltage.. so I guess I could have run it a couple of minutes more.

With all of this in mind I'm pretty sure that 77 or 78 value is solid. Lot of people can do the tests I did with similar results.


For the total capacity, the only thing I can use is my usable capacity plus a static value from the BMS acting as the buffer. In the case of a "85" car, the buffer is reported as 4kWh by the BMS.

Using my 77-78 usable value and adding 4 gives me 81-82kWh total.

If someone can find the low-voltage cutoff used for testing the full capacity (what's the SOC voltage when the buffer is empty), I could do test #7 with my module.

So are we missing 4.something kWh. Probably not. But we're surely missing at least 3kWh.... and I'm not even mentioning that I strongly believe that usable capacity should be advertised... not total. "Hey I just bought this Tesla, it has a 85kWh pack but wait.. but I can only use 78kWh".

Thank you for the detailed reply. I will do the same later today or tomorrow.

Couple more questions, so I really understand where are you coming from. First, what accuracy would you assign to the battery capacity numbers in your post? Second, what ambient conditions and prior history parameters and to what degree do you think can affect your battery capacity numbers?

 

[digicool]

Not to beat a dead horse, but my personal observations of my Mx 90D led me to conclude that the usable capacity on my 90D MX is around 78kWh as well.

I havent had a chance to do a non-stop run from 100% at a relatively steady and low highway speed with just me in the car on a somewhat normal outdoor temperature (80F), no rain, no cargo, CC turned off to empty in-order to convince myself that my car indeed has a 90KWH battery. And that is what I was told is how to get the rated miles from the car.

But what are we going to do about it?

 

[llavalle]

Thank you for the detailed reply. I will do the same later today or tomorrow.

Couple more questions, so I really understand where are you coming from. First, what accuracy would you assign to the battery capacity numbers in your post?

That's a good question.
-The values coming from the car, I have no idea. You would assume the car's own BMS would be within 0.5% or so. (no clues really.. guesstimate)

-The values coming from my AC readings are within 0.5%. I'd have to check the certification & calibration paper and cross reference the actual CT port it was connected on but most of the ports are actually within 0.1% on my unit. The CT itself is within 0.1% since it was used at ~80% of its capacity. Datasheet here.

-I'll have to get back to you with my charger's spec on the module charging test. I do have a shunt that has around 1% error rate and while I don't have the value it recorded on hand, it was very close to the charger's value so my guesstimate would be within 1% at worst.

-As for the module discharging test. I'd say this one is pretty inaccurate. The 800W load is very precise (<0.1W sway) but the timing was a bit botched and I did not take the time to compute the cutoff voltage to account for the load itself...

Second, what ambient conditions and prior history parameters and to what degree do you think can affect your battery capacity numbers?

Not sure I'm understanding this one 100%. Most of my tests were done with ambient air at around 20C. The supercharger test was done at colder temp, It was around 12 or 13C outside IIRC. I did a few tests at VERY low temps (-15C outside temp) but it's really hard to get a good grasp of the actual DC usage (from AC) during charging because the battery heater is in use.

As for the prior history, most of my in-car testing was done during my first ~4 weeks of ownership. That means March / April of 2015. The 10% to SOC drop was actually done on my 3rd day of ownership.. car was only charged once before - car was at 63% SOC when I got it from service center... and my HPWC was defective when I got home - had to drive to my cottage so I stopped at a CHAdeMO and was a bit puzzled by the SOC% VS what the CHAdeMO station reported as being pushed into the battery (around 7.7kWh per 10%).

After that first charge.. I drove a good 380+km (2 straight shots of 190km without stoppig) and noted the SOC vs used energy was also around 7.8... so I started investigating further and further.

 

[llavalle]

Not to beat a dead horse, but my personal observations of my Mx 90D led me to conclude that the usable capacity on my 90D MX is around 78kWh as well.

I havent had a chance to do a non-stop run from 100% at a relatively steady and low highway speed with just me in the car on a somewhat normal outdoor temperature (80F), no rain, no cargo, CC turned off to empty in-order to convince myself that my car indeed has a 90KWH battery. And that is what I was told is how to get the rated miles from the car.

But what are we going to do about it?

You don't need to do a ride with AC off and all. Just charge to 100%.. and drive at around 65mph. The idea is to try to avoid pulling more than 50-60kW as you'll lose a lot of power to the internal resistance of the pack (waste heat). Make sure you reset your trip meter and do it in one go... the instrument cluster doesn't count electrons used while the car is off.

Highway is just easier because you can maintain a set speed for long.

You should have around 80kWh usable on your 90 pack.. maybe less if you have a lot of mileage. As mentioned earlier, degradation characteristics of the cells in the 90 pack seems to be worst than the 85-type cell.

I'm going off topic right now but it seems the 85-type cell suffer a "huge" drop (2%) in the first 20k miles then it becomes a very shallow curve at around 0.5%/25k miles or something like that (don't quote me on that, pulling that out of my head from memory).

The 90-type cell is a bit nastier.. You get a shallower beginning... 1% or so.. but it stays that way.... and continues on.

I need to dig out the info I found on cell testing, was very interesting.

 

[supratachophobia]

So I'd like my 6kwh back now, or the premium I paid to go to 90 from 85 when I ordered....

 

[llavalle]

So I'd like my 6kwh back now, or the premium I paid to go to 90 from 85 when I ordered....

Yeah.. you guys with 90 packs aren't lucky at all. Worst than 85 degradation. weird throttling at superchargers, etc.

 

[apacheguy]

I'm going off topic right now but it seems the 85-type cell suffer a "huge" drop (2%) in the first 20k miles then it becomes a very shallow curve at around 0.5%/25k miles or something like that (don't quote me on that, pulling that out of my head from memory).

Yeah, my S85 suffered a 3% drop in the first 10 K miles. It has stayed right there from then on (now have 40 K miles on the pack).

 

[wk057]

Ran some fun stats on the data dump I have.

There are just under 87k cars with "85" packs in this data out of about 270k cars.

The highest capacity pack in the database is showing at 84.6 kWh (80.6 + 4kWh bottom buffer). I think this one is a fluke, though, since the next highest is 80.1 usable, and the vast majority below 80.

For cars with an odometer between 500 and 1000 miles, "85" packs have an average of 78.62 kWh usable, or 82.62 kWh of actual capacity (212 vehicles matching).

I think cars under 500 miles (basically less than 3 cycles on the pack) are not representative of actual capacity, but, average capacity of 85 packs showing under 500 miles on the odo is still only 78.8, and only 82 cars reporting mileage this low, about 1/3rd of which are the "new" 85 packs that made it to Europe recently with an unknown configuration.

And probably even less useful of a stat, average usable capacity for all "85" packs in the database regardless of age or mileage: 76.7 kWh. Average odometer: 37k miles.

Fun stats "85" pack stats: About 6.66 GWh of usable capacity from all "85" packs in the fleet combined. A total of roughly 616 million 18650 cells.


For "90" packs things are even worse. There are about 3500 cars with 90 packs showing < 500 miles with an average usable capacity of 83.2 kWh (or 87.2 kWh total with the 4kWh buffer). Just over 1000 "90" pack cars reporting between 500 and 1000 miles on the odo with an average usable capacity of 82.8 kWh.

Average usable capacity for all ~65k "90" packs in the data: 81.3 kWh with an average odometer of only 11k miles.... that's an average drop from new of almost the same capacity on the 90 pack in 11k miles vs the 85 packs in 37k miles. Pretty crappy.

Edit/Update (per request):

"75" pack average 72.9 kWh usable (76.9 kWh total) under 500 miles (over 10,000 data points!)
"75" pack average 72.4 kWh usable (76.4 kWh total) between 500-1000 miles (~1500 data points)
"75" pack full average usable capacity of 71.3 kWh (75.3 kWh total) with 61k data points and average odometer of only about 6k miles.



Total usable capacity in the fleet: ~20.5 GWh.


This is data from a data dump provided by a Tesla insider. I've personally correlated data from this dump with known information from my own small "fleet" of Tesla vehicles (over 100 now) and it is definitely genuine data. This is raw data from Tesla's own fleet data. Certainly more useful than any EPA report or whatnot.

Since this probably wont even be enough to convince some of the die hard defenders and TSLA investors here, if I get more time, I'll make some capacity histograms and just put this to rest once and for all especially now that I have a verifiable snapshot of data for this from the entire fleet. (No, I won't be outing my source, and no I won't share the full data. If anyone wants to independently confirm its authenticity, and owns a Model S/X, PM me and if I have time I'll prove it to you under the condition that you publicly post here that you've been shown that the data is genuine).


Bottom line: Tesla lied about pack capacities. End of story. Do something about it, or get over it. Denying it simply is not an option in the face of the actual data.

 

[vgrinshpun]

So, once again, @wk057, what exact parameter are you using to state “highest capacity”? There are more than one parameter dealing with capacity.

Apologies for asking the same more than once, but my question was not previously answered.

Also, 78.62kWH average for cars with 500-1000 miles seem to be different than ~77kWh claimed in title of this thread. Could you explain the difference?

 

[wk057]

So, once again, @wk057, what exact parameter are you using to state “highest capacity”? There are more than one parameter dealing with capacity.

Apologies for asking the same more than once, but my question was not previously answered.

Also, 78.62kWH average for cars with 500-1000 miles seem to be different than ~77kWh claimed in title of this thread. Could you explain the difference?

Pretty sure I explained all of this pretty well, as have many others, but, sure... I'll put a few more bullets into this horse...

The capacities are the usable capacities as reported by Tesla's own battery management system from unique and individual vehicles. There is an unusable/locked bottom capacity cap of 4 kWh on all higher capacity packs (85 kWh and up). So, you take the usable capacity and add 4 kWh to it to get the full raw capacity of the batteries themselves. The is measured and calculated by Tesla's own system... and pretty much perfectly matches external testing of the batteries with external non-Tesla hardware (note reported capacity, dismantle battery pack, test battery modules... easily within +/-0.2 kWh extrapolated to the full pack level).

The ~77 kWh usable number came from the data I had available nearly a year ago, which was already pretty extensive, IMO, otherwise I wouldn't have bothered posting about it in the first place. Considering that going from a few dozen manually acquired data points up to tens of thousands of data points has only managed to reveal a 2% margin of error in my original values, I'd say it's pretty damn valid.

 

[supratachophobia]

Bottom line: Tesla lied about pack capacities. End of story. Do something about it, or get over it. Denying it simply is not an option in the face of the actual data.

As a 90kwh battery owner who saw their "upgraded" capacity rapidly dwindle away in the first couple months, what can we do about this? Or better yet, what would be a reasonable request of Tesla? Personally, that lost capacity means the difference of skipping superchargers on long trips and thereby saving hours potentially. Or not being able to make there-and-back day trips to places where destination charging isn't an option. And wonky supercharging speeds really chaps my bumper. If it comes to light that my randomly capped 45kw-65kw charging speeds are related to something goofy in the 90 packs, that's literally hours I've wasted unnecessarily. I never saw such weird behavior on my 85 pack. So yeah, considering the added cost at purchase, this is a pretty big deal.

 

[islandbayy]

Going to jump in on this conversation as well. My 90 pack will be hitting 50,000 miles next week. I'm behind in compiling and posting my videos on my YouTube channel for my range.

I sample my range (And CAN Bus data/Capacity) every 1000 miles. I can honestly say, compared to my old MS60 that was traded in with over 80,000 miles, my 90 pack is absolute garbage in terms of longevity. I've had the car just a hair over 1 1/2 years at this point.

Not to thread crap, but here is the latest that I have up.
My range update videos show the current range, followed by my Compiled data chart since a hour after taking delivery of my car. The Chart is missing some information that I collect now, mainly because I did not have BMS/CAN Bus data back when I got this car.
I've talked with Tesla Service. Reason "Nothing will be done", as they state, is my battery capacity is similar to the rest of the 90's in the Fleet. Well that doesn't say much.
If my 90D was really a 90kWh battery, factoring in usable capacity, so 90-4(reserve)=86kWh usable. My original starting range would be around what, 310?

So here's the deal, original Rated Range 294. Current rated range 1 1/2 years later and less then 50,000 miles is 271. In comparison, a friends 2013 P85 CPO, with about 48,000 miles, original battery, still gets 267 on a 100% charge. And to top it off, no Supercharger Speed Limiting. I'm also limited to 94kW charge rate MAX on superchargers.

So anyways, heres the video with my chart. Video is as of 46k miles. I have the 47, 48, and 49k videos which need editing and the charts compiled. Don't worry, it doesn't get better.

 

[Tforme]

Doesn’t the 100 battery use the same chemistry as the 90? I haven’t seen it reported that the 100’s are experiencing the rapid degradation of the 90’s. Maybe the modified cooling system in the 100s are making them perform better? Or maybe they are going to see the same rapid drop in capacity as the 90s.

My 4 year old S85 with 45k miles has shown only about 3.8% degradation. I’m considering trading for a new 100 - but I’d be unhappy to see heavy early degradation.

 

[thimel]

@wk057, does your big data dump have information that can be used to determine degradation vs mileage or calendar time? If so, would be interesting to see some scatter plots and fits.

 

[wk057]

As a 90kwh battery owner who saw their "upgraded" capacity rapidly dwindle away in the first couple months, what can we do about this? Or better yet, what would be a reasonable request of Tesla? Personally, that lost capacity means the difference of skipping superchargers on long trips and thereby saving hours potentially. Or not being able to make there-and-back day trips to places where destination charging isn't an option. And wonky supercharging speeds really chaps my bumper. If it comes to light that my randomly capped 45kw-65kw charging speeds are related to something goofy in the 90 packs, that's literally hours I've wasted unnecessarily. I never saw such weird behavior on my 85 pack. So yeah, considering the added cost at purchase, this is a pretty big deal.

I'm honestly not sure what's best to do about it. I think Tesla hoodwinked everyone here, especially those with the 90-type chemistry.

I have a feeling that, given Tesla's track record thus far on such things, it'll take a big lawsuit to get anything done. All Tesla has done thus far is try to mitigate the issue by dropping current/power limits on charge/discharge and tightening thermal limits (thus lowering overall efficiency) in order to make the issue not as obvious to people.

Going to jump in on this conversation as well. My 90 pack will be hitting 50,000 miles next week. I'm behind in compiling and posting my videos on my YouTube channel for my range.

I sample my range (And CAN Bus data/Capacity) every 1000 miles. I can honestly say, compared to my old MS60 that was traded in with over 80,000 miles, my 90 pack is absolute garbage in terms of longevity. I've had the car just a hair over 1 1/2 years at this point.

Not to thread crap, but here is the latest that I have up.
My range update videos show the current range, followed by my Compiled data chart since a hour after taking delivery of my car. The Chart is missing some information that I collect now, mainly because I did not have BMS/CAN Bus data back when I got this car.
I've talked with Tesla Service. Reason "Nothing will be done", as they state, is my battery capacity is similar to the rest of the 90's in the Fleet. Well that doesn't say much.
If my 90D was really a 90kWh battery, factoring in usable capacity, so 90-4(reserve)=86kWh usable. My original starting range would be around what, 310?

So here's the deal, original Rated Range 294. Current rated range 1 1/2 years later and less then 50,000 miles is 271. In comparison, a friends 2013 P85 CPO, with about 48,000 miles, original battery, still gets 267 on a 100% charge. And to top it off, no Supercharger Speed Limiting. I'm also limited to 94kW charge rate MAX on superchargers.

So anyways, heres the video with my chart. Video is as of 46k miles. I have the 47, 48, and 49k videos which need editing and the charts compiled. Don't worry, it doesn't get better.

Yep, the 90 chemistry is garbage. There is no doubt about it. My limited testing at the cell level proved it to me ages ago, but with the data trove bestowed on me a few months ago I can absolutely confirm it.

The supercharge speed issue is the worst, too, because it means an older 85/60 or a newer 100 can supercharge the same range much faster. Even worse is that the BMS for the 90 packs is less tolerant of cell imbalances and will quickly taper charge speed to mitigate a bigger imbalance during a charge.

Doesn’t the 100 battery use the same chemistry as the 90? I haven’t seen it reported that the 100’s are experiencing the rapid degradation of the 90’s. Maybe the modified cooling system in the 100s are making them perform better? Or maybe they are going to see the same rapid drop in capacity as the 90s.

My 4 year old S85 with 45k miles has shown only about 3.8% degradation. I’m considering trading for a new 100 - but I’d be unhappy to see heavy early degradation.

From what I can tell the 100 pack uses yet another chemistry. The degradation appears to be similar to the original 85s.

@wk057, does your big data dump have information that can be used to determine degradation vs mileage or calendar time? If so, would be interesting to see some scatter plots and fits.

Calendar degradation will be tough, because there are a lot of cars with non-original packs. I have the "birthday" for the car, but not the pack. I do have, however, "kWh charged" for every pack... which would be a good axis for a scatter plot (includes regen charging, so probably better overall too). Guess I'll do some later.

Edit: So.... I just did 85 and 90 pack scatter plots.... and I'm not even sure I should post them. Definitely going to open a can of problem worms for Tesla... ugh.

 

[dennis_d]

Doesn’t the 100 battery use the same chemistry as the 90? I haven’t seen it reported that the 100’s are experiencing the rapid degradation of the 90’s. Maybe the modified cooling system in the 100s are making them perform better? Or maybe they are going to see the same rapid drop in capacity as the 90s.

My 4 year old S85 with 45k miles has shown only about 3.8% degradation. I’m considering trading for a new 100 - but I’d be unhappy to see heavy early degradation.

100 packs use different chemistry , newer 75 assumed to have the same chemistry as 100 packs