If the cell can't leave the factory with a a certain capacity you can't label it with that capacity. Imagine if they did that to anything else. Like say gasoline "Oh yeah it was full gallon in the underground tank, but by the time it made it to your car, 5% was gone, sorry we didn't mean it"
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
If the cell can't leave the factory with a a certain capacity you can't label it with that capacity. Imagine if they did that to anything else. Like say gasoline "Oh yeah it was full gallon in the underground tank, but by the time it made it to your car, 5% was gone, sorry we didn't mean it"
So Tesla engineers knew that were size constrained, knew that the cells that they were going to get from Panasonic were already not going to give them the max capacity listed on the label when they got them, but still listed the battery size as 85 kWh. Then instead of correcting this with the 90 kWh battery, they decided to then list it as a 90 kWh battery instead of an 85 kWh battery (to make more money).
- - - Updated - - -
Yet, Tesla keeps wanting to be NOT like the other EV companies. So why do it like everybody else does it?
JRP3
Hyperactive Member
Actually a small amount does evaporate from the nozzle as you pump, but yeah.
They may not have known that when the designed the first 85 packs.
We don't know that to be true, at least I'm not aware of anyone testing a 90 kWh pack outside of the car.
There is a difference between being "honest" and stupid. When everyone is forced to state the usable capacity based on some technical standard only then will Tesla do it.
You frankly don't have to. There is clearly only 83kWh (or less) usable. This thread has basically taken "We give Tesla the benefit of a doubt" to "Here's proof the capacity was never there". Any reason we should extend that from the 85 to the 90 now? Credibility is the thing in question.
Chickenlittle
Banned
They do not recommend routine greater than 90 % charging but occasional fine.In short there is none.
The 100% best source for a guidance would be Tesla, and they offer none explicitly other than keep it under 90%.
FWIW I charge mine to 80% in general, but then I fully intend handing it back and getting another.
One thing I will say is I set mine to a full 100% trip, and I forgot to unset it there once, next day I went out to my car and found it sat fully charged. At that point I had a minor moment of "regret" for the pack. But hey it will be going back to them long before range degradation is my problem....
... besides I don't feel too bad, it's their choice if they wont make it clear what the optimum is and leave charging to some customer determined slider, and won't force it to reset after a "trip button", just to avoid range averaging on the spec sheet like the Leaf
Tesla is NOT "pre-cycling" their cells. They are QC'd at Panasonic with one cycle (battery of measurements) and then a quick self-discharge test, and then QC'd at Tesla once with once cycle (battery of measurement) and then a longer term self-discharge test. If Panasonic was SAFT all the testing would be done at the cell mfg, but Panasonic is not that type of cell supplier. Some mil-spec QC tests done at SAFT last up to 4 months, before the cells are given the OK.
Interesting you found the the rapid decline in capacity (3+% !) for the first 5 cycles with your NCR18650A cell which is a NNP (NEW NICKEL PLATFORM, A.K.A NCA) chem. OMG CALL THE COPS AND HIRE A LAWYER AND SELL ALL YOUR PANASONIC STOCK AND SWEAR TO NEVER BUY ANOTHER PANASONIC PRODUCT AGAIN
You will find if you take a brand new cell and cycle it twice and then leave it full charged for a few weeks (simulating a long-term self-discharge test), when you measure the capacity the 3rd time, it may be closer to the capacity of cell that was not stored fully charged for a few weeks and cycled maybe 5-8 times depending on which NNP cell you have exactly (NCR-18650,18650F,A,B,E,BF)
The cell can be "broken-in" either by cycling or simply increasing the calendar aging rate artificially, but like I said Tesla's goal is not to pre-cycle or "break-in" the cells, it is to screen for duds (aka out-of-family outliers) which abnormally high self discharge.
In any case, I don't think any of the off-the-shelf NNP cells are Tesla cells. I even have one of the newer NCR18650BF cells which are similar to the cells used in the 90 pack (High density NNP Cathode and Silicon monoxide Anode). It does appear to "break-in" slower which could explain why some 90 packs appear to lose capacity faster than new than the 85 pack.
Again none of the NNP cells are Tesla cells. One of the first tests I requested okashira do when he got his hands on some Model S cells a couple years ago is a zero-volt storage tested where a resistor is used to short out the cell and then the cell is stored. I am happy to report after a year stored at 0V the cell recoved to 94% full capacity!
The reason why I requested this test is because I heard a rumor Quallion was involved in the Model S cell design. Now this does not mean for sure the cells have Quallion "Zero-Volt" tech but all signs are pointing to yes.... If true this means the Model S pack in cannot be bricked (unless there is a imbalance that cause some cells to pull others negative).
Note okashira sells both cells, modules, and custom packs made using Model S cells (which custom cell holders that pack the cells tighter than off the shelf ones). He pioneered a method to remove the cells from the holders without a scratch by using solvent. One of the first things I had him check is to tear out the can seal and soak it in the solvent. I wanted to make sure the solvent would not compromise the can seal. After a week in the solvent there was not softening, swelling, shrinkage, or cracking of the seal.
On a side note one of the reasons the cells are glued in place is because these consumer type 18650's don't have a hermetic seal. Gluing them in place stops vibration that can break the bond wire fuse (main purpose), but the other benefit is it serves the better seal the cell against humidity.
I suspect the gigafactory cell will use a mil-spec glass-metal hermetic seal which will allow them to drop a hundred pounds in adhesive. The new Model 3 pack uses flex PCB interconnects (positive and negative connection made on one side of cell) and the other "terminal" side is base plate cooled. Cooling thru one end of the cell is much more effective than even a full wrap around the can (which is impractical anyways). Again flex PCB means the pack is much simpler to make and there are no bond wire fuses that quickly break with vibration, and most importantly tighter packing (but still a bit or airspace to prevent runaway if that one dud you didn't catch vents fire). The LG Bolt soft pouch prismatic pack is a disaster waiting to happen.
Last edited:
Added this note to the first post:
Hint on upcoming project: My instrument cluster is looking pretty epic.
2016-02-11: Just adding an important note pointing out that there is a LOT of misinformation getting thrown around and perpetuated in the replies here on this thread. I have no control over what other people post/reply, and I can't waste time battling random people on the Internet every time someone posts something that's completely false, misleading, or otherwise twists the facts into something they're not. I've posted the data that I can and my conclusions based on the data I have collected. I hope, among other things, my reputation here speaks for itself when you come across replies here with outrageous claims, unsubstantiated claims of knowledge of Tesla's inner workings, and other attempts to discredit me or my findings. I'm always fine with someone presenting data that opposes my own, but no one here has done so (nor in my opinion is really able to legitimately do so given that the data pretty much speaks for itself).
In an effort to spare my sanity and make much better use of my limited free time I'm unsubscribed from this thread and will not be following up on this matter further. Keep an eye out for more fun projects.
Hint on upcoming project: My instrument cluster is looking pretty epic.
From your first post:
Tapping out?
One "word": lol.
Maybe next time you should keep your "conclusions" to yourself given their implications....
Tapping out?
One "word": lol.
Maybe next time you should keep your "conclusions" to yourself given their implications....
¯\_(ツ)_/¯
Last edited:
The cells are NCR18650BE, I said this a year ago...
Yes, you can get 11.9Wh out of a cell when new, even close to new.
They are rated at 3200 mAh.
I don't have a problem with Tesla's 85kWh rating
I don't know what Tesla cells you tested, but NCR18650B's are pretty poor performers. NCR18650BE will have much lower impedance because the B has a PTC, which increases resistance.
All modern cells no longer use a PTC.
500mA is too fast drain to measure Wh capacity. You need to use a quasi-static discharge... barring that... 50mA might be slow enough to get a proper Wh reading.
Also make sure you're charging to 4.2V with a 50mA charge cut. Kelvin connections are also a necessity.
Yes, you can get 11.9Wh out of a cell when new, even close to new.
They are rated at 3200 mAh.
I don't have a problem with Tesla's 85kWh rating
I don't know what Tesla cells you tested, but NCR18650B's are pretty poor performers. NCR18650BE will have much lower impedance because the B has a PTC, which increases resistance.
All modern cells no longer use a PTC.
500mA is too fast drain to measure Wh capacity. You need to use a quasi-static discharge... barring that... 50mA might be slow enough to get a proper Wh reading.
Also make sure you're charging to 4.2V with a 50mA charge cut. Kelvin connections are also a necessity.
Seems the forum doesn't want me to unsubscribe to this thread. It keeps popping back up in my list, so, guess I'll try later.
No, just done having to defend my data and conclusions against people with no credibility. Private discussions elsewhere are much more appealing and useful at this point. My absence in this thread is in no way "tapping out" since that would imply that I somehow have lost the battle, which I clearly have not. It's a fight not worth fighting. Again, the data speaks for itself. I've logs from over 20 cars, from brand new to a signature VIN S00XXX. Not a single one shows pack capacity at or even close to 85 kWh as reported by the car itself. Data which matches my own testing of cells outside the car, and the tests of others.
I don't care what Tesla does with the cells prior to delivering them to customers, or not. If the product I received doesn't meet spec, then it doesn't meet spec. Plain and simple. People are welcome to verify this for themselves, and I invite them to do so. I've outlined the relevant BMS CAN messages here.
First, let's actually quote things with context.
This was obviously my attempt to show that the capacity difference was significant, using a monetary value derived from Tesla's own pricing, for folks who may or may not have understood the delta.
I could care less about the implications of what the data tells. The data says that customers never received a pack that Tesla labeled "85 kWh" that could actually be charged to 85 kWh in or out of the vehicle. That's a fact based on mountains of data, including admission of the cars themselves.
If the data said that the pack was say an 86 kWh pack (it does not) we wouldn't be having this conversation, and people would be praising Tesla for under promising on the spec and everyone would be happy. Instead it's data that shows that the spec doesn't match the advertised spec, however little the delta, and for whatever reason that triggers half of the forum to jump on the dog pile with the messenger at the bottom.
Personally, I'm just curious as to what the blog post that Tesla will eventually post in response to this issue, trying to explain it away like the P85D range issue and P85D horsepower issue, will read. The range issue blog I'm reasonably certain I personally triggered with my side by side testing of the P85 and P85D... either that or it was a complete coincidence that Tesla posted their blog post exactly as I was doing my pre-announced and publicized side by side testing.
- - - Updated - - -
Actual final post here, since I respect okashira on this, and if desired he's welcome to continue this discussion in private.
I'm not sure how you came to the conclusion that they're NCR18650BE. We can't both have the same cells, awesome test equipment, and come to drastically different conclusions. Some variable is different.
I'm well aware of how to test the cells. Actually, the latest iteration of the equipment being developed for this that I've been field testing for the company designing it includes complete thermal controls capable of holding the cell anywhere from -20C to 75C and capacity calculations accurate to a ridiculous +/- 5 mWh. It's literally accurate enough to measure degradation from cycle to cycle, and I really wish I could share more information on the equipment because it's absolutely awesome. I know quite a bit about the cells and how to test them, and I had even made my own cycler for them prior to hooking up with the company I'm assisting. Using my lower end equipment I never got close to 11.9 Wh. I thought I did a few times, but it turned out to be an issue with the hardware (because the cells definitely don't hold 13+ Wh). Using the new equipment we've tried many different methods of charging and discharging the cell in an attempt to recover 11.9 Wh, or even close to it, and have not been able to. Thermal controlled, uncontrolled, constant current, constant power, varied power, varied current, etc etc etc. No dice on 11.9 Wh.
Fun that we have conflicting data, though. However, Tesla's BMS does in fact agree with mine pretty much exactly... so, not sure where that leaves us. Edit: Added data point being that I collect mountains of charge/discharge data from my ~16000 cells in my solar setup, and they also match nicely with my other data extrapolated out to the large scale full pack. I've not discharged my solar pack down to nothing, but I have charged it to 100% several time, and have discharged it on a few occasions below the voltage Tesla considers 0%... and the data still matches up with my single cell testing.
No, just done having to defend my data and conclusions against people with no credibility. Private discussions elsewhere are much more appealing and useful at this point. My absence in this thread is in no way "tapping out" since that would imply that I somehow have lost the battle, which I clearly have not. It's a fight not worth fighting. Again, the data speaks for itself. I've logs from over 20 cars, from brand new to a signature VIN S00XXX. Not a single one shows pack capacity at or even close to 85 kWh as reported by the car itself. Data which matches my own testing of cells outside the car, and the tests of others.
I don't care what Tesla does with the cells prior to delivering them to customers, or not. If the product I received doesn't meet spec, then it doesn't meet spec. Plain and simple. People are welcome to verify this for themselves, and I invite them to do so. I've outlined the relevant BMS CAN messages here.
First, let's actually quote things with context.
This was obviously my attempt to show that the capacity difference was significant, using a monetary value derived from Tesla's own pricing, for folks who may or may not have understood the delta.
I could care less about the implications of what the data tells. The data says that customers never received a pack that Tesla labeled "85 kWh" that could actually be charged to 85 kWh in or out of the vehicle. That's a fact based on mountains of data, including admission of the cars themselves.
If the data said that the pack was say an 86 kWh pack (it does not) we wouldn't be having this conversation, and people would be praising Tesla for under promising on the spec and everyone would be happy. Instead it's data that shows that the spec doesn't match the advertised spec, however little the delta, and for whatever reason that triggers half of the forum to jump on the dog pile with the messenger at the bottom.
Personally, I'm just curious as to what the blog post that Tesla will eventually post in response to this issue, trying to explain it away like the P85D range issue and P85D horsepower issue, will read. The range issue blog I'm reasonably certain I personally triggered with my side by side testing of the P85 and P85D... either that or it was a complete coincidence that Tesla posted their blog post exactly as I was doing my pre-announced and publicized side by side testing.
- - - Updated - - -
Actual final post here, since I respect okashira on this, and if desired he's welcome to continue this discussion in private.
I'm not sure how you came to the conclusion that they're NCR18650BE. We can't both have the same cells, awesome test equipment, and come to drastically different conclusions. Some variable is different.
I'm well aware of how to test the cells. Actually, the latest iteration of the equipment being developed for this that I've been field testing for the company designing it includes complete thermal controls capable of holding the cell anywhere from -20C to 75C and capacity calculations accurate to a ridiculous +/- 5 mWh. It's literally accurate enough to measure degradation from cycle to cycle, and I really wish I could share more information on the equipment because it's absolutely awesome. I know quite a bit about the cells and how to test them, and I had even made my own cycler for them prior to hooking up with the company I'm assisting. Using my lower end equipment I never got close to 11.9 Wh. I thought I did a few times, but it turned out to be an issue with the hardware (because the cells definitely don't hold 13+ Wh). Using the new equipment we've tried many different methods of charging and discharging the cell in an attempt to recover 11.9 Wh, or even close to it, and have not been able to. Thermal controlled, uncontrolled, constant current, constant power, varied power, varied current, etc etc etc. No dice on 11.9 Wh.
Fun that we have conflicting data, though. However, Tesla's BMS does in fact agree with mine pretty much exactly... so, not sure where that leaves us. Edit: Added data point being that I collect mountains of charge/discharge data from my ~16000 cells in my solar setup, and they also match nicely with my other data extrapolated out to the large scale full pack. I've not discharged my solar pack down to nothing, but I have charged it to 100% several time, and have discharged it on a few occasions below the voltage Tesla considers 0%... and the data still matches up with my single cell testing.
Last edited:
hockeythug
Active Member
Let me get a new cell in my tester, I am sure I will get 11.9Wh out of one. ;-)
My test equipment is only accurate to 0.002Ah or so. :-/
Yes, they are NCR18650BE. If it quacks like a duck, looks like a duck....
Seriously, I am suprised you would debate this and it was settled.
-They both have the same DCIR.
-They both look the same (and this is important - the Model S cell and the NCR18650BE are the ONLY two cells in the world in existence that have the exact same triangular top.) NO OTHER PANASONIC CELL EVER MADE HAS THIS TRIANGLE TOP!
-They both have the same capacity
-They both have the same discharge curve profile
-They have matching weight to +/- 0.01g
-They cycle the same
-They have the same temperature response to DCIR.
To get +/- 5 mWh is not really that special? Any decent DC load can do that. It would only require about 0.002mV accuracy and about 0.1% accuracy on current. Any decent multimeter over $300 can read that....
My cheap 8-channel chinese cycle tester will consistently hold 0.001V across the range of 2.5-4.2V, calibrated with a benchtop fluke of course.
Btw: Model S modules make for a great power supply for a spot welder:)
I have pulled 3,900 amps welding straight to the bus bar. Good thing for only 9ms.
The scope is there - just finished measuring current of the spot welder.
The battery on the bottom is a 20 series 7 parallel pack made from Model S cells for an E-bike.
The cells were pulled from a 2014 85 with about 5k miles.
I have a 2015 70D in the driveway that we're about to tear into. I am hoping we get lucky and they are the new high cap cells - likely NCR18650GA.
- - - Updated - - -
Yes, I left two Model S cells discharged to 0V for a year. I kept them shorted out with a 400 ohm resister.
After about 10 months, I removed the resister and let them sit for another month. They slowly picked up about 0.4V.
I charged the cell to 4.2V and the cell still held around 94% capacity.
Endless-sphere.com View topic - Tesla Model S 18650 Cell Test Data
It did pick up some impedance, however.
My test equipment is only accurate to 0.002Ah or so. :-/
Yes, they are NCR18650BE. If it quacks like a duck, looks like a duck....
Seriously, I am suprised you would debate this and it was settled.
-They both have the same DCIR.
-They both look the same (and this is important - the Model S cell and the NCR18650BE are the ONLY two cells in the world in existence that have the exact same triangular top.) NO OTHER PANASONIC CELL EVER MADE HAS THIS TRIANGLE TOP!
-They both have the same capacity
-They both have the same discharge curve profile
-They have matching weight to +/- 0.01g
-They cycle the same
-They have the same temperature response to DCIR.
To get +/- 5 mWh is not really that special? Any decent DC load can do that. It would only require about 0.002mV accuracy and about 0.1% accuracy on current. Any decent multimeter over $300 can read that....
My cheap 8-channel chinese cycle tester will consistently hold 0.001V across the range of 2.5-4.2V, calibrated with a benchtop fluke of course.
Btw: Model S modules make for a great power supply for a spot welder:
)I have pulled 3,900 amps welding straight to the bus bar. Good thing for only 9ms.
The scope is there - just finished measuring current of the spot welder.
The battery on the bottom is a 20 series 7 parallel pack made from Model S cells for an E-bike.
The cells were pulled from a 2014 85 with about 5k miles.
I have a 2015 70D in the driveway that we're about to tear into. I am hoping we get lucky and they are the new high cap cells - likely NCR18650GA.
- - - Updated - - -
Yes, I left two Model S cells discharged to 0V for a year. I kept them shorted out with a 400 ohm resister.
After about 10 months, I removed the resister and let them sit for another month. They slowly picked up about 0.4V.
I charged the cell to 4.2V and the cell still held around 94% capacity.
Endless-sphere.com View topic - Tesla Model S 18650 Cell Test Data
It did pick up some impedance, however.
Last edited:
Andyw2100
Well-Known Member
Well, then it must be true. I mean even though there is lots of data supporting what wk057 has reported in the opening post, we should definitely just assume it is all wrong. That makes a lot of sense.
A vendetta, huh? Is that why he has said over and over again that the lack of degradation the Tesla pack demonstrates is amazing? I don't know about the vendettas you are used to, but the ones I'm familiar with don't generally involve the people waging them having both positive and negative things to say about the target of the vendetta.
Most don't find the mods here oppressive. In fact, I'm going to guess just a handful do. The mods are volunteers, do a very hard job, and do it well. They deserve our gratitude and respect. They do not deserve to be insulted with comments like the above.
And very noticeably missing from Tesla's official reply was any statement about the specifications of the battery packs being accurate. If Tesla is confident that the specs are accurate, why not say that?
Yes I do not trust the self proclaimed battery expert who has not shared much details about his instrumentation or methods used to test individual cells before thrashing Tesla let alone raw data
now all over the internet we have headlines such as "battery expert shows blah blah blah"
Here is a rational post from the reddit thread:
"TL;DR: Numerous plausible justifications of cell capacity disparity still remain in within ambiguity of disclosed tests.
I certainly don't think the "expert" (clearly he's experienced but maybe not being a total pro [frosty + exhaustively thorough] about it prior to throwing bombshells?) Maybe it's not warranted, but I can't shake the feeling that he (wk057) has a chip on his shoulder from the HP debates (irrespective of the merits of that debate) as he's openly professed "ranting". The following took minimal effort to find, but he didn't seem to address them thoroughly in his thread (I only read the one linked and briefly his other old threads.) I'm taking his results with a grain of salt until there's more detail as I don't want to get sucked into any ostensible (tame) vendetta. Maybe my doubt is misplaced I dunno. =?"
now all over the internet we have headlines such as "battery expert shows blah blah blah"
Here is a rational post from the reddit thread:
"TL;DR: Numerous plausible justifications of cell capacity disparity still remain in within ambiguity of disclosed tests.
I certainly don't think the "expert" (clearly he's experienced but maybe not being a total pro [frosty + exhaustively thorough] about it prior to throwing bombshells?) Maybe it's not warranted, but I can't shake the feeling that he (wk057) has a chip on his shoulder from the HP debates (irrespective of the merits of that debate) as he's openly professed "ranting". The following took minimal effort to find, but he didn't seem to address them thoroughly in his thread (I only read the one linked and briefly his other old threads.) I'm taking his results with a grain of salt until there's more detail as I don't want to get sucked into any ostensible (tame) vendetta. Maybe my doubt is misplaced I dunno. =?"
Was there another set of test data than what you posted here last year? This looks more like 3Ah x 3.6V = 10.8 Wh
LOL. So let's do a quick hand integration and arrive at Wh by subtraction. Take the two limits, 4V and 3Ah, that is 12Wh. Since we have a curve we know it MUST be a value less than that. OK, now each square is .2V*.2Ah=0.04Wh. Now count all the full squares between those lines and the top curve, estimating on the semi full ones, and subtract from 12. I got 10.96Wh. Or 77.8kWh for an 85 pack with 7104 cells.
Hmmmmm
Similar threads
- Question
