I'm suspicious this is actually the case as well. And part of the reason, I suspect GENIII is another major leap ahead - to be back-annotated to SnX
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Model S Battery Pack - Cost Per kWh Estimate
- Thread starter CapitalistOppressor
- Start date
-
- Tags
- TSLA
I'm suspicious this is actually the case as well. And part of the reason, I suspect GENIII is another major leap ahead - to be back-annotated to SnX
stopcrazypp
Well-Known Member
Well, Gen III is projected at 2016-2017 (production more likely 2017 and beyond). The 4000mAh was announced in 2009 and will start production some time this year. There's still a relatively huge gap in time until Gen III's production start, so I find it unlikely Tesla will be using current cells.
I've been using a 2 year technology gap as a conservative estimate (judging from Model S history: the first 3100mAh cell came out of the factory in 4/2010, first Model S in 6/2012). So that pegs Tesla using 2015 cells for the Gen III. This will be technology unannounced by Panasonic and already in the lab already (or Tesla can switch to other manufacturers too given the use of commodity cells).
Yes, I see this as a key reason to expect Tesla to achieve 25% gross margins, irregardless of regulatory credits.
- - - Updated - - -
My own estimates are that Tesla needs a minimum of 2-3 years to properly engineer a vehicle around the battery. The idea that they can just switch batteries willy-nilly, just seems like nonsense, and is the biggest reason why I am confident they did not utilize 3400mAh cells. There are major engineering complexities involved with even relatively minor changes to vehicles. Manufacturers need to have time to thoroughly engineer and test their solutions after picking a battery.
smorgasbord
Active Member
Optional equipment, like 3rd row seats, has a weight difference as well, yet I'll bet Tesla didn't have to do either crash nor EPA testing with and without the 3rd row seats. What's the limit on weight changes to trigger retesting?
If Tesla went from 8000 cells to 7000 cells, at about 47 grams/cell that would be about 100 pounds in weight difference.
For that matter, does anyone know the curb weight difference between identically configured 60 kWh and 85 kWh cars?
Right, and why Musk is confident GenIII will meet their cost projections, because the "hard" part of battery cost optimization has already been achieved, or clearly will be shortly with expected improvements and iterations.
I guess one shouldn't be surprised as they've been doing battery and drivetrain research for over 10 years, but it's still amazing how they have taken the unconventional route. If I can summarize your research as I understand it in a narrative it's as follows: "Let's not reinvent the wheel in cell production - because other companies can do that better/cheaper and adapt/optimize faster. Let's keep our overhead and exposure low, because moreover the mass unit (18650) is ideal in many ways for safety and pack optimization. And yet we WILL intervene with the manufacturer and tell them how to simplify which we can do because of our volume ordering. Instead we will engineer the safety features at a pack level, thereby further reducing costs while being confident about the safety protocols we've designed in ourselves."
adiggs
Well-Known Member
I think of Tesla's approach to battery packs, as the Hadoop of energy storage (many many, commodity units of storage that fail, with failure handled at a higher layer in the solution stack - battery pack in Tesla's case).
I wouldn't quite put it as building a car around the battery, since Tesla has a black box battery design with an active cooling system. I'd suggest they'd be trying to build a battery for the car that meets the physical characteristics they need and meets or exceeds the performance and durability characteristics they'd need. Development can happen in parallel.
stopcrazypp
Well-Known Member
Within 250lbs, crash tests are still considered comparable. But I don't know what is the retest trigger though and whether a base model weight difference would be more significant than a optional equipment difference.
http://www.safercar.gov/FAQ
There will be a 10% difference in cell count (3400mAh vs 3100mAh), plus some overhead in terms of modules removed.
Curb weigh difference is 183lbs.
http://www.teslamotorsclub.com/show...s-85-khw/page5?p=261350&viewfull=1#post261350
True, willy-nilly seems like nonsense - that's a good point - safety testing, charge rates etc. I agree too many just think of it as a block replacement which is not accurate. In addition, though on the other side, I think the Tesla experience to date, the now larger market drive position of Tesla, etc. will produce a 2 year max time frame rather than min time frame. That gives them 1-2 years from now to commit. By that time the iteration AFTER the current testing soon to be mfg will take the stage for GENIII. In fact, I believe this is exactly what is driving the time schedule of the GENIII vs a more aggressive release. I think Elon wants to have several years head start on competition for GENIII (as competition for that sweet spot will be fierce). Tis also conforms to the 7 years major redesign time for current model S. They'll launch a new Model S at that time with both the new battery chemistry and the corresponding S design mods to leverage same
kevin99
Member
Great research! And good timing too! I would be a bit concerned reading Barron's article without getting educated by this post. It would be interesting to see hoe the market respond tomorrow. It seems obvious the Barron guys get mad and decide to badmouth tesla, even though Elon told them significant reduction and they choose to doubt him.
kevin99
Member
CapitalisrOppressor, one thing I can help to further this research is to help you verify or get more details on the battery pricing from the Chinese sellers. I can call them up. Just let me let know any particular info you need and I can try, although I am just getting myself educated in this space. Pm me if needed.
I have been attempting to contact them. Based on user forums you need to be a verified buyer with verified deep pockets to get their best prices. Until today, they had ignored my inquiries. However, I finally broke down and made a "creative" profile and put out a general request to purchase a relatively small number (100,000) of batteries. In context, Tesla probably uses 3,000,000 per week.
The one reply I've gotten so far was for around $350/kWh. The problem is that even 100,000 batteries is a tiny number for these guys. They will just box em up and ship them in one shipment, and they'll throw in PCB's for free. They also promised to let me inspect the shipment before it went out (presumably to protect against counterfeit goods).
Here is an example of a PCB, which they are just throwing in for free -
Tenergy 32004 PCB for 7.2V / 7.4V Li-ion 18650 / 18500 7.2V Battery Packs
I don't know what PCB's are costing them, but there is obviously a lot of fat there. To get the best contracts probably requires in person negotiation and a commitment to huge quantities over an extended period. So I don't know how to move the research much further, other than to root it in the IEK report (which is authoritative, and references prices from $120-$200/kWh) and the ads. Given that, $2/cell still seems very safe, and there are competitive reasons to think that the $1 prices might be real.
stopcrazypp
Well-Known Member
$2/cell is pretty safe ($2.5 is safer). But $1/cell might be tough even with negotiations, because I suspect that might be undercutting material costs. From everything I have read, the costs of the cells are already very close to material cost (economy of scale has been exhausted) and cell costs have not decreased significantly in the last couple of years ($/kWh is primarily being driven by energy density improvements).
And somewhere up thread, you use Ultrafire prices as a reference for $1/cell and I would advise against that. Ultrafire is well known in the industry, but not in a good way. They are well known for overstating their capacity significantly (by 2x). And these are the "real" ones. The fake ones are just recycled laptop batteries wrapped in a new wrapper!
http://www.candlepowerforums.com/vb/showthread.php?316999-Ultrafire-18650-4000-mah
http://www.candlepowerforums.com/vb/showthread.php?316888-Fake-Ultrafire-18650-battery-warning
I've mentioned them here (the "real" versions) and why they can't be used for comparison.
http://www.teslamotorsclub.com/show...w-battery-tech?p=111988&viewfull=1#post111988
I should add that I had the option of having the PCB's attached, or unattached.
- - - Updated - - -
That makes sense. That would push their effective per kWh up to ~$138, and the cheapest of the cheap would be ~$120, which anchors us back into the I.E.K. report. Of course, that's assuming they were making an ad-hoc "adjustment" like this.
The link from Aweite for $1.8-3.2 for a Panasonic 3400mAh cell was an unusually firm looking price in comparison to the normal $1-5 or $2-7 (or even their $1-3 for the same cell). But I am very much inclined to stick to the range specified by the I.E.K. and $147/kWh for a NCR18650B seems to be getting very close to implausible.
Anyways, I wrote up a story for GCR (dunno if they will publish it yet), and while I mentioned the Ultrafire phenomenon I was careful not to tie my estimate to it. Instead I just took the $2/cell price and gave Tesla a ten cent per cell discount for eliminating the stock cap (so $1.90/cell), which resulted in a per cell cost of ~$170/kWh. Frankly, I'm still just as comfortable with the $160/kWh estimate in the OP, but I went for conservatism in the estimate, while leaving the spicy possibilities open for debate.
P.S. I found a new link to a study in 2009 showing 18650 costs at the time of $200-$250/kWh -
http://www1.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2009/energy_storage/es_02_barnett.pdf
- - - Updated - - -
That makes sense. That would push their effective per kWh up to ~$138, and the cheapest of the cheap would be ~$120, which anchors us back into the I.E.K. report. Of course, that's assuming they were making an ad-hoc "adjustment" like this.
The link from Aweite for $1.8-3.2 for a Panasonic 3400mAh cell was an unusually firm looking price in comparison to the normal $1-5 or $2-7 (or even their $1-3 for the same cell). But I am very much inclined to stick to the range specified by the I.E.K. and $147/kWh for a NCR18650B seems to be getting very close to implausible.
Anyways, I wrote up a story for GCR (dunno if they will publish it yet), and while I mentioned the Ultrafire phenomenon I was careful not to tie my estimate to it. Instead I just took the $2/cell price and gave Tesla a ten cent per cell discount for eliminating the stock cap (so $1.90/cell), which resulted in a per cell cost of ~$170/kWh. Frankly, I'm still just as comfortable with the $160/kWh estimate in the OP, but I went for conservatism in the estimate, while leaving the spicy possibilities open for debate.
P.S. I found a new link to a study in 2009 showing 18650 costs at the time of $200-$250/kWh -
http://www1.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2009/energy_storage/es_02_barnett.pdf
JRP3
Hyperactive Member
"Ultrafires" were also used by some college team for their solar vehicle, which caught fire and almost killed the driver who barely got out in time with minor burns. Possibly the worst 18650 cell on the planet.
- - - Updated - - -
- - - Updated - - -
I guess that explains why the 60 is only 180lbs lighter than the 85, when it should have been lighter than that.
I thought for sure an estimate of the Gen III battery pack had been done by someone, but not finding any through a search, I went ahead with my own back-of-envelope calculations based on the conversations of this thread. Please troubleshoot and/or improve upon these numbers as my number-crunching skills are not anywhere near CapOp's.
My estimate is fairly straightforward, simply taking 85% of the Model S battery packs due to the quoted 20% size reduction of Gen III and allowing that some elements like power electronics, chargers, airbags, etc. don’t scale linearly but are closer to constant. This may be conservative from the standpoint of Gen III being a more modest car with less range expectation, but I don’t think Tesla wants to skimp on the range (they dumped the 40kWh Model S after all), and the 200mi range quoted for Gen III is only slightly less than the 208 mi EPA certified rating of the Model S 60kWh pack.
Model S Battery Packs (estimated)
Gen III Battery Packs (extrapolated from 85% size of Model S as per 20% size/weight reduction)
I started with a direct 80% weight reduction (48kwH/68kWh) but moved to 85% to be more conservative. I wonder though if it can’t in fact reach 80% because one, the Model S may have some luxury features which add weight that Gen III will forgo; two, the engineers may find more weight savings through everything they’re learning from the S/X platform; and three, most significantly the weight savings of the batteries will be greater if the more powerful 3400 and 4000mAh cells do not weigh proportionally more than the 3100mAh cells they will replace.
It does suggest that even if the cost of cells remains the same, the combined gains of next generation cells with weight savings of a smaller platform start to have significant effects on battery size, and in turn total cost.
My estimate is fairly straightforward, simply taking 85% of the Model S battery packs due to the quoted 20% size reduction of Gen III and allowing that some elements like power electronics, chargers, airbags, etc. don’t scale linearly but are closer to constant. This may be conservative from the standpoint of Gen III being a more modest car with less range expectation, but I don’t think Tesla wants to skimp on the range (they dumped the 40kWh Model S after all), and the 200mi range quoted for Gen III is only slightly less than the 208 mi EPA certified rating of the Model S 60kWh pack.
Model S Battery Packs (estimated)
| Rated Size | Size w/buffer (1.05X) | Range (est./cert.) | Cell Type | #Cells |
| 60 kWh | 64 kWh | 230/208 | 3100 mAh @3.6V | 5735 |
| 85 kWh | 89.25 kWh | 300/265 | 3100 mAh @3.6V | 7997 |
| Rated Size (Model S x.85) | Size w/buffer (1.05X) | Range (est./cert.) | Cell Type | #Cells |
| 51 kWh | 53.55 kWh | 230/208 | 3400 mAh @3.6V | 4375 |
| 4000 mAh @3.6V | 3719 | |||
| 72.25 kWh | 75.9 kWh | 300/265 | 3400 mAh @3.6V | 6197 |
| 4000 mAh @3.6V | 5268 |
It does suggest that even if the cost of cells remains the same, the combined gains of next generation cells with weight savings of a smaller platform start to have significant effects on battery size, and in turn total cost.
Last edited:
You got this size buffer calculation wrong, since it should be at 53.55 and not 50.4, and it probably threw off your other calculations too.
Yes, thanks - that gives slightly higher cell totals of 4375 and 3719 for the 51kWh pack - I corrected the table in my first post.
Hi there,
First, I think that the Tesla Model S is truly an astonishing car and a great achievement.
Still, I think that it remains to be seen if Tesla is able to make money at all. These numbers regarding battery cost are way to optimistic in my opinion.
Some numbers: Based on Tesla's 3rd quarter report the cost to build and manufacturer (that is including every cost except R&D and sales) one Model S is at app. 95.000$ with an average selling price of app. 112.000$ worldwide.
By comparison: Ford has a cost of app. 20.400$ for every car produced. That is including R&D but without sales. Of course their margin is a lot lower than Tesla's, but they are selling their cars directly to the dealers and not to the customer.
If Tesla wants to maintain their current strategy of selling cars directly to customer, they will have substantially higher sales cost over time. If you include R&D and sales, Tesla is still not making any money at present. Yet they will have to expand their point of sales massively to get anywhere close to the volumes they want to achieve. In Germany they have 3 shops at present. I say shops not dealerships because they are really tiny. The one in Munich has not more than 2 sales people.
Back to costs. I believe that even the hand-made Porsche Panamera is a lot cheaper to produce than the Tesla Model S. Porsche is making money on every car sold including the base model Boxster, starting at 50.000$ at the dealer. The Porsche Panamera is much more expensive, yet production costs are only slighly higher. Even a full sized Turbo S will not be more than app. 60.000$.
In 2012 during an Investor presentation Tesla gave a very short insight into the Model S' battery cost. They said, they expected it to be around 400$/kwh. That is including material, manufacturing and all cooling requirements related to the battery. If I take a look at the Q3 I believe this number is a lot more valid than the fantastic sub 200$ calculations.
Car manufacturing is a very complex, difficult process. Such a big battery has a very rich demand profile. It has to bee cooled, heated, it has to absorb shocks, accidents, standing water etc.......yet it also has to cope with incredble power and charging requirements. Under full load this battery will be empty within minutes. That's just the most demanding environment for a battery system you can think of.
First, I think that the Tesla Model S is truly an astonishing car and a great achievement.
Still, I think that it remains to be seen if Tesla is able to make money at all. These numbers regarding battery cost are way to optimistic in my opinion.
Some numbers: Based on Tesla's 3rd quarter report the cost to build and manufacturer (that is including every cost except R&D and sales) one Model S is at app. 95.000$ with an average selling price of app. 112.000$ worldwide.
By comparison: Ford has a cost of app. 20.400$ for every car produced. That is including R&D but without sales. Of course their margin is a lot lower than Tesla's, but they are selling their cars directly to the dealers and not to the customer.
If Tesla wants to maintain their current strategy of selling cars directly to customer, they will have substantially higher sales cost over time. If you include R&D and sales, Tesla is still not making any money at present. Yet they will have to expand their point of sales massively to get anywhere close to the volumes they want to achieve. In Germany they have 3 shops at present. I say shops not dealerships because they are really tiny. The one in Munich has not more than 2 sales people.
Back to costs. I believe that even the hand-made Porsche Panamera is a lot cheaper to produce than the Tesla Model S. Porsche is making money on every car sold including the base model Boxster, starting at 50.000$ at the dealer. The Porsche Panamera is much more expensive, yet production costs are only slighly higher. Even a full sized Turbo S will not be more than app. 60.000$.
In 2012 during an Investor presentation Tesla gave a very short insight into the Model S' battery cost. They said, they expected it to be around 400$/kwh. That is including material, manufacturing and all cooling requirements related to the battery. If I take a look at the Q3 I believe this number is a lot more valid than the fantastic sub 200$ calculations.
Car manufacturing is a very complex, difficult process. Such a big battery has a very rich demand profile. It has to bee cooled, heated, it has to absorb shocks, accidents, standing water etc.......yet it also has to cope with incredble power and charging requirements. Under full load this battery will be empty within minutes. That's just the most demanding environment for a battery system you can think of.
Similar threads
