Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

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

This site may earn commission on affiliate links.
Maybe we both read different graphs, but till now I just did not seen any graph where the Tesla cell discharge significantly more than 3100mAh.



Looks like about 3300mAh to me

Overdischarge after 10A run.png
 
Not everyone has the time or ability to read and keep up with the rapid pace of wk057's projects and tinkering, but he has described all the instrumentation and methods in various posts while dismantling a couple of tesla battery packs, designing and installing a huge solar array grid on his house, including wiring up all the connections, controllers and other equipment necessary to use the packs to store the solar energy and power-up his house and charge his teslas completely from that system.

And he didn't die of electrical shock or burn down his house or car, and he still has all his fingers and toes.

Oh yeah and in his free time he reverse engineered the CAN busses used by tesla so he could read the actual brick voltages, pack currents, motor torque and power, and all the diagnostic screen info, and described how all that was done in another thread.

Like i said it's difficult to keep up, but he has plenty of validity and credibility amoung the technical crowd. His knowledge, skills and abilities have been demonstrated numerous times--nobody practiced in the art of electrical engineering would doubt his results.
This. Anyone can throw a cell on a test bench, and come to a conclusion. Few are hacking into and diagnosing complex systems like wk. Voicing your opinion has little to with credibility.
 
Finally, on Flathill I agree with the other posters - tone it down bud. On the other hand, I understand his frustration. This thread has been a good example of a failure in this forum (not due to the mods) - We have posters like ... MukeBur who clearly have nothing substantial to add continuing to post - while wk has been driven from the thread, and he is one of the few who SHOULD continue to contribute.

Don't get me wrong - I know wk is very sharp and capable of understanding how to make the argument he's trying to make - it just wasn't complete and he got a little ahead of himself in terms of technical understanding - but I know he's sharp enough to get it. He is also a revolutionary / whistle-blower personality type - which I also respect. :)

- - - Updated - - -


This post is a perfect example of why the technical minded, wk, myself, other... will always be driven from further posting - and threads that have the potential to make sound technical conclusions will continue to be rare on this forum.

Thanks bud. I will calm down a bit. As an EE with a metrology background and over 20 years of experience working with EV's the one thing I can tell you is the smartest scientists (excepting physicists) and engineers are usually the worst at making accurate/precise measurements. The other point to make is NONE of us have access to virgin cells. I too am almost certain the cell is the BE, but not 100%. Even getting your hands on fresh grade A BE cells is a challenge. Even if some USED Model S cells don't measure 11.9Wh that still does not allow us to draw any conclusions on whether Tesla's spec is wrong.
 
Last edited:
Right. Even if okashira is right, if I guess at 25kW average highway cruising draw, then I get approximately 1A per cell, plus or minus depending on pack voltage, etc. So even his data says that under normal usage you'll never get 85kWh out of the pack.

...and that still doesn't explain the difference between 60 and 85. Although I don't believe EPA vs EPA is true at all for 70 vs 90.

I actually responded to a similar post by you. 1A draw (which is roughly C/3 for these cells) most definitely does not simulate the UDDS EPA test. Something like C/16 or C/17 or 0.2A is the correct draw to use because it would take 16+ hours to run down the battery on a 85kWh Model S on the UDDS (and this is only the usable part).

Even if you look at EPA HWFET (48.3 mph average speed), that is 6-7 hours (AKA C/6-C/7) which is closer to a 0.5A draw.
Look at test conditions. That is 1A discharge at the slowest one. You aren't going to get the nameplate capacity of any cell at 1A discharge. That test was what I referred to before, where he wasn't trying to get max capacity. To do that most testers use a 0.2A load (he suggested even lower up thread).

For reference, the EPA UDDS cycle is 19.6mph average speed and Model S gets 320+ miles on it, so 16+ hours of discharge, probably 17+ hours if looking at nameplate (AKA C/16 to C/17 discharge). That nicely works out roughly to 0.2A for the cells in question.

And for purposes of this discussion we are trying to determine nameplate cell capacity, which 1A draw will not give you (esp. for a cell that is not fresh off the assembly line).
 
Last edited:
So... I went looking for a the datasheet for teh NCR18650BE, after okashira pointed out the likelihood that those are the cells (or very similar) that Tesla is using. Despite their being readily available for other Panasonic NCR18650 variants, I cannot find an authoritative datasheet on the BE anywhere, whech may be telling in and of itself.

One thing I noted in looking at the NCR18650B: although Panasonic dos not list a power (in watt-hours) spec for their cells, they do list the following:


  • Gravimetric energy density (in Wh/kg)
  • The weight of the cell

For example:
18650Bsnip.PNG


Thus, for this cell, the energy density per gram is 243/1,000=.243 Wh/g. And given the cell weighs 48.5g, we can determine the energy would compute to 0.243*48.5=11.8 Wh. Thus, we can reasonably extrapolate that these cells have a manufacturer nameplate rating of 11.8Wh,

A Tesla 85 pack with 7,104 of these cells would have a manufacturers rating of 83,827Kw/h. That rounds to 84 for whole numbers, or 85 for the nearest "5".

Given that the NCR18650BE undoubtedly has slightly different chemistry from the NCR18650B, I don't think it's outside the realm of reason to suggest that Panasonic themselves rate the cells such that the pack logically rounds to 85KW/h.

Anybody find a datasheet for the BE??
 
Last edited:
One thing I noted in looking at the NCR18650B: although Panasonic dos not list a power (in watt-hours) spec for their cells [...]

Thus, for this cell, the power density per gram is 243/1,000=.243 Wh/g. And given the cell weighs 48.5g, we can determine the power would compute to 0.243*48.5=11.8 Wh.

Nice find! Just a quick comment though: Wh is a measure of energy, not power (that would be W) :wink:
 
Thus, for this cell, the power density per gram is 243/1,000=.243 Wh/g. And given the cell weighs 48.5g, we can determine the power would compute to 0.243*48.5=11.8 Wh. Thus, we can reasonably extrapolate that these cells have a manufacturer nameplate rating of 11.8Wh,
I would be careful on that. The weight specifies the max cell weight. And if you see the little (1), (2), (3) footnotes that the test conditions may not be the same on the different numbers.

Manufacturers will not post an official nameplate Wh. They will give nominal voltage and nameplate capacity in mAh. You have to do the math yourself. For the cell in question, the official capacity is "3350 mAh", but everywhere it will be advertised at "3400mAh".
 
From the linked review below, it looks like the BE cell was tested at about 10.5 Wh of total energy discharge:

Comparing Panasonic 18650 BD vs. Panasonic 18650 BE 18650 Battery | BATTERY BRO

- - - Updated - - -

I would be careful on that. The weight specifies the max cell weight. And if you see the little (1), (2), (3) footnotes that the test conditions may not be the same on the different numbers.

Yep, and from the tests I linked above it came in at about 47g (not 48.5)
 
Nice find! Just a quick comment though: Wh is a measure of energy, not power (that would be W) :wink:

Derp on my end... thanks. Corrected.

- - - Updated - - -

I would be careful on that. The weight specifies the max cell weight. And if you see the little (1), (2), (3) footnotes that the test conditions may not be the same on the different numbers.

Manufacturers will not post an official nameplate Wh. They will give nominal voltage and nameplate capacity in mAh. You have to do the math yourself. For the cell in question, the official capacity is "3350 mAh", but everywhere it will be advertised at "3400mAh".

My point is that the rating may very well be based of what the manufacturer provides. That same spec sheet for the B lists 3350mAh @3.6Vnom as well, which equates out to > 12Wh per cell... so it may be premature to assume that these cells would have an industry accepted rating substantially less than what Tesla specs them to be to comprise their pack.
 
Last edited:
From the linked review below, it looks like the BE cell was tested at about 10.5 Wh of total energy discharge:

Comparing Panasonic 18650 BD vs. Panasonic 18650 BE 18650 Battery | BATTERY BRO
Not a full capacity test. Test conditions:
Charge end at 4.16 V (vs 4.2V)
20 or 25 degrees Celcius ambient temperature (difference between 20C and 25C gave a 5% difference in the NCR18650B, although the claim is less than that for NCR18650BE)
.61 ampere discharge current (vs 0.2A or even 0.05A suggested by okashira)
2.5 volt cut-off point when the discharge ends
2.5A charge rate (Panasonic is rated 0.909A)

- - - Updated - - -

Derp on my end... thanks. Corrected.

My point is that the rating may very well be based of what the manufacturer provides. That same spec sheet for the B lists 3350mAh @3.6Vnom as well, which equates out to > 12Wh per cell... so it may be premature to assume that these cells are would have an industry accepted rating substantially less than what Tesla specs them to be to comprise their pack.
I agree with your point, I just say that you have be careful in extrapolating from energy density figures when the test data and conditions may not be the same between the two specs. Simple multiplication of nominal voltage and capacity is typically how "nameplate capacity" is figured.
 
I would be careful on that. The weight specifies the max cell weight. And if you see the little (1), (2), (3) footnotes that the test conditions may not be the same on the different numbers.

The footnote for energy density reads:
Energy density based on bare cell dimensions

That would seem to apply specifically to the volumetric spec, not the gravimetric one I used for the calculations.

- - - Updated - - -

I agree with your point, I just say that you have be careful in extrapolating from energy density figures when the test data and conditions may not be the same between the two specs. Simple multiplication of nominal voltage and capacity is typically how "nameplate capacity" is figured.

Yup, agreed... extrapolating is more or less of a sanity check.

Given that the numbers we DO have from manufacturer spec sheets on similar cells produce an energy rating in the correct ballpark, until which time we have an official spec for these cells, I think it's premature to claim foul. I think we agree there as well...
 
Looks like about 3300mAh to me

View attachment 111088

Ok we are now looking on the same graph. :smile:

You need to read the voltage axis! Okashira in this particular test discharged the cell to 0V, but the nominal capacity for the most NCA cells is measured to 2,5V and you lose the manufacturer warranty if you discharge the cell below this value! So the nominal capacity for this test setting is ~3050mAh to 2,5V.

The second important thing is that below 3V the voltage drops rapidly for the most Li-Ion cells. So it significantly affects the usable energy. For example in the range of 3V-2,5V you got only 5% or less of energy. And in the range 2,5V-0V it is less than 1% because the lack of voltage for integration.


By the way nominal capacity/energy for the most modern cells (see datasheets) is measured under these conditions:

Charge: 0.3-0.5C with cut-off current C/50
Discharge: 0.2C

and this can be translated as you don't get significantly better capacity/energy if you set even lower testing currents.
 
Okashira,

I appreciate you laying out the case that there is room for error in wk057's data or at least that there are variables based on his methodology which he has not yet presented here. I assume that you object the spot welder comment from MikeBur, because as far as I've seen you have not stated that you have tested cells and drawn 11.9Wh from them. Will you address that aspect of his post?

If in fact, wk057's testing methodology underestimates the battery capacity in a fashion that happens to be consistent with Tesla's own BMS does it then follow that if the true capacity of the pack is 85Wh, then his calculated capacity of 61Wh for the 60 pack should actually have been ~ 64Wh?

I do find it mildly amusing that, assuming you are correct, the answer to the 85 vs 81 debate will turn out to be essentially the same as the hp debate -- the batteries could theoretically produce that much energy, just only on a test bench -- never in the actual car.
 
...
Charge: 0.3-0.5C with cut-off current C/50
Discharge: 0.2C

and this can be translated as you don't get significantly better capacity/energy if you set even lower testing currents.
Maybe this is true for charge current, but definitely not for discharge, esp. when we are talking about single percentage differences being significant (we are disputing a 5% overall difference in this case). The IR losses at least partially are going to be what contributes to this (edit: mAh capacity also changes even more significantly).

For example, the test below of NCR18650B at 0.5A (~0.15C) measures 11.856Wh and at 0.2A (~0.06C) measures 12.142Wh. That is a 2.4% difference. Note that the test below cut off at 2.8V.
http://lygte-info.dk/review/batteries2012/Panasonic NCR18650B 3400mAh (Green) UK.html
 
Last edited:
Not a full capacity test. Test conditions:
Charge end at 4.16 V (vs 4.2V)
20 or 25 degrees Celcius ambient temperature (difference between 20C and 25C gave a 5% difference in the NCR18650B, although the claim is less than that for NCR18650BE)
.61 ampere discharge current (vs 0.2A or even 0.05A suggested by okashira)
2.5 volt cut-off point when the discharge ends
2.5A charge rate (Panasonic is rated 0.909A)

On top of that note their inconsistent methodology:

"I always like to run through three to five discharge cycles before using the data from the discharge curve for analysis. "

Which is it 3 or 5? That can make a huge difference with an NCA cell.

Also note the claimed BE spec sheet pictured in the batterybro article (which is incomplete) has footnote
2G23X0KYKU which is actually the 18650B datasheet so this was either a Chinese forgery (happens very often) or Panasonic made an error (early release). Judging from the fact the "Cell Type NCR18650BE" is not aligned properly I would say it much more likely a forgery.

They also don't give you the date of mfg, time it was tested, or the grade of the cell. I have found it almost impossible to buy fresh Grade A cells from China or in the USA. All USA panasonic resellers on the internet buy them from the Chinese grey market. There is only one source I know of in Germany that will sell factory fresh cells direct to consumers and they are 3-4X the "street" price.
 
Last edited:
Kelvin testing is 4-wire, not 2-wire.

A good time measuring device is also needed in addition to the current sense and voltage measurement.

The panasonic data sheet above would also correctly be interpreted as nominal 3200Ah x 3.6V = 11.52 Wh, times 7104 gives 81.84 kWh.

Yeah, so some older Panasonic data sheets use 3.7 v nominal and you can see if you search for NCR18650B's online... lots of sellers still quote the 3.7v number. That's always been voodoo to me as to why Panasonic sometimes uses 3.6 and sometimes uses 3.7 when you look at the discharge graphs of those cells, it doesn't look like there should be a difference in nominal voltage.
 
So... I went looking for a the datasheet for teh NCR18650BE, after okashira pointed out the likelihood that those are the cells (or very similar) that Tesla is using. Despite their being readily available for other Panasonic NCR18650 variants, I cannot find an authoritative datasheet on the BE anywhere, whech may be telling in and of itself.

One thing I noted in looking at the NCR18650B: although Panasonic dos not list a power (in watt-hours) spec for their cells, they do list the following:


  • Gravimetric energy density (in Wh/kg)
  • The weight of the cell

For example:
View attachment 111090

Thus, for this cell, the energy density per gram is 243/1,000=.243 Wh/g. And given the cell weighs 48.5g, we can determine the energy would compute to 0.243*48.5=11.8 Wh. Thus, we can reasonably extrapolate that these cells have a manufacturer nameplate rating of 11.8Wh,

A Tesla 85 pack with 7,104 of these cells would have a manufacturers rating of 83,827Kw/h. That rounds to 84 for whole numbers, or 85 for the nearest "5".

Given that the NCR18650BE undoubtedly has slightly different chemistry from the NCR18650B, I don't think it's outside the realm of reason to suggest that Panasonic themselves rate the cells such that the pack logically rounds to 85KW/h.

Anybody find a datasheet for the BE??

48.5g is the max weight. The typical weight is 47.5g. In any case the Model S pack is not a 18650B...

b.PNG
bf.PNG
 
Let's see how well real experts from The Electrochemical Society do compared to the "expert" here.

They measure two NCA cells (Panasonic NCR18650A and Panasonic NCR18650B)

"A testing methodology consisting of six stages of testing was proposedand conducted on five commercially available Li-ion cells. Size measurement found some cells to notably deviate from the format associated sizing. These variations have the potential to cause standardization and exchange ability issues and can further distort battery development trend analyses that use the format as a baseline for comparing energy densities. Initial capacity testing found that the cellsfrom well-known manufacturers (Panasonic and Sanyo) performed towithin 8% of their rated capacity."

expert.PNG


Quick short Panasonic stock :)

I still see problems with their comparison methodology (NO DATE OF MFG!!!, AVG OF 5 FOR FOR NCA DOWN TO ONLY 2.8V, 6TH CYCLE ONLY FOR NCA DOWN TO SPEC 2.5V MEANING NO FIRST CYCLE DOWN TO 2.5V DATA AT ALL FOR NCA CELLS) but note:

"coils removed and wires replaced in order to decrease resistance. Furthermore,four-terminal sensing was used to minimize the measuredvoltage drop due to applied currents. The voltage measurements weretaken directly from the surface of the cells, whereas the currents wereapplied via the upgraded holder wirescoils removed and wires replaced in order to decrease resistance. Furthermore,four-terminal sensing was used to minimize the measured voltage drop due to applied currents. The voltage measurements were taken directly from the surface of the cells, whereas the currents were applied via the upgraded holder wires"

. Despite the 50–130 cycles undergone representing only a relatively small fraction of the commonly assumed Li-ion cycle life of several thousand cycles, a general ageing trend is clearly visible. While the Panasonic and Sanyo cells show a capacity decrease of only 5–10%,

This inconsistent result suggests that the sample size of two cells per type and discharge cycle is insufficient for conclusive studies of the impact of slight cycle variations on ageing


http://jes.ecsdl.org/content/162/8/A1592.full.pdf

Now do you see where Tesla is coming from when it says:

"It is very difficult to replicate the exact discharge profile to extract the maximum available energy”
 
Last edited: