NOTE / PLEASE READ FIRST:
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.
--- (Original Post) ---
Alright, so this discussion has been going on in other threads and I keep chiming in with bits and pieces of details. I guess it's time to make the sure to be heated thread. Please read this entire post before commenting. It's likely I've answered your question or countered an argument you'll make already. I already know I'm going to catch hell for this one.
As many of you know, I'm pretty much the local self-proclaimed expert on Tesla's batteries. I'm using 15,984 cells from Tesla battery packs to power my house with my off-grid solar setup. I've disassembled two full Tesla battery packs, assisted with the disassembly of another's, and have disassembled a couple of modules into individual cells for various torture tests. I've also acquired roughly a dozen cells from a "60" pack with < 2500 miles to compare to my "85" pack cells.
I presently have had roughly 20 pairs of cells from Tesla's "85" kWh pack running 24/7 doing various cycle tests in various conditions. (Eventually I'll be posting detailed data from these tests, but I want to give them significant run times.) These cells were from a pack that had less than 1000 miles (or less than 5 charge cycles) on it, and arrived to me charged to roughly 50% (perfect for storage/shipment).
I've also done capacity testing on full modules.
More recently I have been investigating other aspects of the Model S via my bench setup. Many details on that in my other thread, but here I'll focus on battery related items.
I'll point out that I have no first hand or otherwise verifiable data on the 70 or 90 packs thus far.
Let's run through some quick facts.
Now, right away there is a problem and we haven't even gotten to any testing or other data yet. If 7104 cells total 85,000 Wh, then each cell would contain 11.965 Wh of energy. If 5376 cells total 60,000 Wh, then each cell would contain 11.161 Wh of energy. Hmmmm. These numbers don't add up. No big deal, maybe they used different cells in the "60" pack you might argue. Now for some test results.
OK, so, they've definitely the same exact cells. How can Tesla possibly be rating one pack as 60 and one pack as 85 then knowing the above? And we haven't even gotten very far here yet and this is already a red flag.
Now, OK, let's actually test these cells and see what happens. There has to be a real number. Turns out the cells are very consistent in capacity testing. VERY consistent.
In 6 hour charge/discharge cycle testing (1/6C charge and discharge) over a one month period the average capacity of the cells came out to 11.36 Wh per cell. The maximum capacity measured was 11.42 Wh/cell.
Extrapolate this out: 7104 * 11.36 = 80,701 Wh (80.7 kWh) for the "85" pack cell count, and 5376 * 11.36 = 61,071 Wh (~61.1 kWh) for the "60" pack. This means the "85" rating is short by at least 5%, and the "60" rating is under-rated by at least 2%.
Sounds small, but here's some more data.
This means that the "85" pack is "missing" about 4.3 kWh of capacity. That's about 14 miles of range.
"Oh, but it's a buffer... the anti-brick thing, wk!"
Glad you brought that up. Since I did my cell testing outside of the Tesla BMS and anti-brick stuff, I was able to discharge them well beyond what Tesla's setup would normally allow. So, no anti-brick buffer screwing with my tests.
Further, data gathered from my hacking efforts revealed that the the anti-brick buffer on the 85 pack is a static 4.00kWh (NO! THIS IS NOT USABLE CAPACITY BELOW 0 MILES). So that comes out to about ~77 kWh usable, max, on a brand new Model S with an "85" pack. Hmm... oddly matches pretty much everything we've ever seen. Applying to the "60" pack gives us about 57 kWh usable, although I believe the anti-brick buffer is slightly different on the 60 pack (unconfirmed). The car's BMS also reports the usable capacity as around 76.5 kWh with a 4 kWh buffer on my own car. See my CAN deciphering document for my CAN deciphering document which includes how to decode this data. The fact that the car reports these values that match perfectly (within 1%) of my cell testing means that Tesla is well aware that these are not actually 85 kWh packs.
So I can kind of see where the 60 number came from. It's 61, round down to an even 60. OK, good enough. But why "85"? This is not an 85 kWh pack. It's an 81 kWh pack at best, brand new, at super low discharge rates. I don't know about you, but when I went to school 81 would round to 80, not 85.
It might not sound like a big deal, but it kind of is. Today, Tesla has a 90 kWh pack upgrade (which is likely not actually 90 kWh either, but unconfirmed) that costs $3000 for an additional 5 kWh. Well, the "85" pack is missing 4 kWh to begin with, so extrapolating that out everyone with an 85 pack overpaid by about $2,400. As the buyer of a total of three 85 version Model S, that's $7,200 worth of capacity that I never received. That's significant.
Further, these cells are almost certainly NOT Panasonic NCR18650B cells, like many have assumed. Close, but not identical. I've tested actual Panasonic NCR18650B cells on the same equipment as the Tesla cells, and the retail Panasonic cells always perform better and have slightly lower internal-resistance. So, the NCR18650B specs are not quite relevant.
Many people keep bringing up hard drive capacity ratings as a parallel, but that's totally different. Sure, you could say that the usable capacity is analogous to the formatted usable capacity of a hard drive... I can go with that. But here's the thing. Prior to formatting a 1TB hard drive, there are 1,000,000,000,000 bytes available to the system to use, or exactly the definition of 1TB. So, the capacity is there on the physical device. In the case of Tesla's "85" kWh pack, the initial capacity isn't there to begin with (81 kWh total from completely full to completely dead). So, the analogy falls short. Let's not try and use it anymore.
Additionally, back to my solar project, I closely monitor the power output of my battery bank. The capacity measurement there matches my other testing and information gained from my hacking efforts to within 1%.
Essentially, there I can find no evidence that the "85" pack is actually 85 kWh, but multiple pieces of evidence that confirm that it is not in fact 85 kWh.
I've no idea what to really do with this info aside from share it with the community. I've already lost all faith in Tesla's published specs, personally. (See here and here for those rants...) I feel that at the very least owners of these cars and future buyers have the right to know the real specifications of the car they own or are buying.
"Let's get it on!"
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.
--- (Original Post) ---
Alright, so this discussion has been going on in other threads and I keep chiming in with bits and pieces of details. I guess it's time to make the sure to be heated thread. Please read this entire post before commenting. It's likely I've answered your question or countered an argument you'll make already. I already know I'm going to catch hell for this one.
As many of you know, I'm pretty much the local self-proclaimed expert on Tesla's batteries. I'm using 15,984 cells from Tesla battery packs to power my house with my off-grid solar setup. I've disassembled two full Tesla battery packs, assisted with the disassembly of another's, and have disassembled a couple of modules into individual cells for various torture tests. I've also acquired roughly a dozen cells from a "60" pack with < 2500 miles to compare to my "85" pack cells.
I presently have had roughly 20 pairs of cells from Tesla's "85" kWh pack running 24/7 doing various cycle tests in various conditions. (Eventually I'll be posting detailed data from these tests, but I want to give them significant run times.) These cells were from a pack that had less than 1000 miles (or less than 5 charge cycles) on it, and arrived to me charged to roughly 50% (perfect for storage/shipment).
I've also done capacity testing on full modules.
More recently I have been investigating other aspects of the Model S via my bench setup. Many details on that in my other thread, but here I'll focus on battery related items.
I'll point out that I have no first hand or otherwise verifiable data on the 70 or 90 packs thus far.
Let's run through some quick facts.
- Tesla's "85" kWh pack consists of 16 modules of 444 cells for 7,104 total cells.
- Tesla's "60" kWh pack consists of 14 modules of 384 cells for 5,376 total cells.
Now, right away there is a problem and we haven't even gotten to any testing or other data yet. If 7104 cells total 85,000 Wh, then each cell would contain 11.965 Wh of energy. If 5376 cells total 60,000 Wh, then each cell would contain 11.161 Wh of energy. Hmmmm. These numbers don't add up. No big deal, maybe they used different cells in the "60" pack you might argue. Now for some test results.
- Individual cell testing of capacities of the cells from the "85" pack and the "60" pack match to within 0.5%.
OK, so, they've definitely the same exact cells. How can Tesla possibly be rating one pack as 60 and one pack as 85 then knowing the above? And we haven't even gotten very far here yet and this is already a red flag.
Now, OK, let's actually test these cells and see what happens. There has to be a real number. Turns out the cells are very consistent in capacity testing. VERY consistent.
In 6 hour charge/discharge cycle testing (1/6C charge and discharge) over a one month period the average capacity of the cells came out to 11.36 Wh per cell. The maximum capacity measured was 11.42 Wh/cell.
Extrapolate this out: 7104 * 11.36 = 80,701 Wh (80.7 kWh) for the "85" pack cell count, and 5376 * 11.36 = 61,071 Wh (~61.1 kWh) for the "60" pack. This means the "85" rating is short by at least 5%, and the "60" rating is under-rated by at least 2%.
Sounds small, but here's some more data.
This means that the "85" pack is "missing" about 4.3 kWh of capacity. That's about 14 miles of range.
"Oh, but it's a buffer... the anti-brick thing, wk!"
Glad you brought that up. Since I did my cell testing outside of the Tesla BMS and anti-brick stuff, I was able to discharge them well beyond what Tesla's setup would normally allow. So, no anti-brick buffer screwing with my tests.
Further, data gathered from my hacking efforts revealed that the the anti-brick buffer on the 85 pack is a static 4.00kWh (NO! THIS IS NOT USABLE CAPACITY BELOW 0 MILES). So that comes out to about ~77 kWh usable, max, on a brand new Model S with an "85" pack. Hmm... oddly matches pretty much everything we've ever seen. Applying to the "60" pack gives us about 57 kWh usable, although I believe the anti-brick buffer is slightly different on the 60 pack (unconfirmed). The car's BMS also reports the usable capacity as around 76.5 kWh with a 4 kWh buffer on my own car. See my CAN deciphering document for my CAN deciphering document which includes how to decode this data. The fact that the car reports these values that match perfectly (within 1%) of my cell testing means that Tesla is well aware that these are not actually 85 kWh packs.
So I can kind of see where the 60 number came from. It's 61, round down to an even 60. OK, good enough. But why "85"? This is not an 85 kWh pack. It's an 81 kWh pack at best, brand new, at super low discharge rates. I don't know about you, but when I went to school 81 would round to 80, not 85.
It might not sound like a big deal, but it kind of is. Today, Tesla has a 90 kWh pack upgrade (which is likely not actually 90 kWh either, but unconfirmed) that costs $3000 for an additional 5 kWh. Well, the "85" pack is missing 4 kWh to begin with, so extrapolating that out everyone with an 85 pack overpaid by about $2,400. As the buyer of a total of three 85 version Model S, that's $7,200 worth of capacity that I never received. That's significant.
Further, these cells are almost certainly NOT Panasonic NCR18650B cells, like many have assumed. Close, but not identical. I've tested actual Panasonic NCR18650B cells on the same equipment as the Tesla cells, and the retail Panasonic cells always perform better and have slightly lower internal-resistance. So, the NCR18650B specs are not quite relevant.
Many people keep bringing up hard drive capacity ratings as a parallel, but that's totally different. Sure, you could say that the usable capacity is analogous to the formatted usable capacity of a hard drive... I can go with that. But here's the thing. Prior to formatting a 1TB hard drive, there are 1,000,000,000,000 bytes available to the system to use, or exactly the definition of 1TB. So, the capacity is there on the physical device. In the case of Tesla's "85" kWh pack, the initial capacity isn't there to begin with (81 kWh total from completely full to completely dead). So, the analogy falls short. Let's not try and use it anymore.
Additionally, back to my solar project, I closely monitor the power output of my battery bank. The capacity measurement there matches my other testing and information gained from my hacking efforts to within 1%.
Essentially, there I can find no evidence that the "85" pack is actually 85 kWh, but multiple pieces of evidence that confirm that it is not in fact 85 kWh.
I've no idea what to really do with this info aside from share it with the community. I've already lost all faith in Tesla's published specs, personally. (See here and here for those rants...) I feel that at the very least owners of these cars and future buyers have the right to know the real specifications of the car they own or are buying.
"Let's get it on!"
Last edited: