Tesla Berlin Will Use the Big 4680 Battery and Structural Battery Pack

Elon Musk says Tesla Berlin will produce and use the new 4680 battery cell and the structural battery pack. Tesla Berlin should be producing cars in Q1 or Q2 of 2021.

Rob Mauer of Tesla Daily estimated how the 10 GWh pilot line and 100 GWh by 2022 mentioned at Battery Day could be used.

If Tesla was sandbagging then the production for Berlin might begin in the second half of 2021. Perhaps 5GWh from each of two lines in Berlin in 2021.

SOURCES- Elon Musk, Tesla Daily
Written by Brian Wang, Nextbigfuture.com (Brian owns shares of Tesla)

64 thoughts on “Tesla Berlin Will Use the Big 4680 Battery and Structural Battery Pack”

  1. You are correct, but your emphasis is a bit "off". Even if you would accept thick floors without battery packs, its not certain that the honeycomb would be lighter/cheaper than the normal floor.

    With the battery pack, the weight of the battery shells is "not counted". You get the can metal "for free", since it is counted in the "budget" of the cell weight. If you just made a honeycomb floor with empty battery cans, this floor might be heavier than the normal floor. And probably more expensive…

  2. Who gave you one down vote? Makes no sense at all to downvote such a common sense comment.

    Of course you are correct in what you are saying. But it's not just Tesla and Porsche who get lots of power from motors; 1100 HP Lucid air, 1000 HP next electric Hummer, Rivian 700 HP, to mention a few.

    But you are correct that not all EV makers provide lots of power, but it seems easy enough for many to do so.

    Perhaps it's a question of hybrids? Small battery only allows for low power and hence weak electric motors?

  3. Ah yes, they can't use an expanded honeycomb floor in a normal car, because they want the floor to be as thin as possible because everyone wants their car low to the ground with maximum ground clearance.
    The EV design forces people to accept a nice thick floor, so they take advantage of that to use a properly structured floor and get a good stiffness to weight ratio.

  4. I got the strong impression that they want the 'tubes to be used in the current production systems, and new ones that it makes possible, rather than making their own battery with a specific chemistry, altho they could certainly do both, esp as a product demo. I'm guessing that when you make a battery, you start with sheets of electrode stuff and go from there. They would do the process early in the build.

  5. Fun fact. The author of the article also interviews another battery researcher, and he believes that the NAWA electrode will be very expensive:

    "The issue in commercialization is the cost associated with producing aligned carbon nanotubes. My guess is the cost would be much more than x10."

    I guess I am not the only one that doubts the commercial viability of using VCNTs for batteries…

  6. …And now I have read the New Atlas article…

    I don't think I can trust very much in that article, because the author thinks that "silicon batteries" means that you replace the lithium with silicon [1]
    . "Silicon batteries" means that you replace part of or all of the graphite in the anode with silicon, since silicon can store lithium ions per atom compared to graphite. He states that you can

    If the author does not know or does not understand such a basic fact, what else has he gotten wrong?


  7. Yes, I'm sure, but my point is that whatever cans or pouches that will be used for their cells, it needs to be a part of a structural battery pack to compete with Tesla's cells.

    Since Tesla has patented the tables electrodes, I don't see NAWA making 4680-cells, but most likely they would have to use a smaller form factor. Also, if anyone other than Tesla would use their chemistry, this company would have to develop a structural pack technology.

    And how long do you think it would take a normal company to implement such a technology? A few years at the very least….

  8. Thanx for response! This is the sort of thing I have no special knowledge of, but with a general understanding of Physics, that I get a *smell* for that is often correct. But often incorrect, too. They don't claim anything about any chemistry at all, indeed promise to use existing chemistry as is, just this one part different, the actual conduction function of the "electrode", however that is defined! That may be the problem there, too general a word? The 'tubes are just a smaller and better conductor than powder. Perhaps they have separated the functions, so the chemistry and conductor are not in each other's way?

  9. You know Dan, the more I read about this battery the more convinced I am that they have not measured anything. Perhaps you do not know, but you need a chemical reaction at the cathode. At the anode, the lithium can be adsorbed, but the at the cathode it is different.

    So, even if you would use the VCNTs for scaffolding and conduction at the anode, what reactant is also present at the cathode? In this Wiki article [1], they give the half reaction for a cobalt oxide cathode. And how can you reduce the weight of the cathode by a factor of 10 given that you need these other chemicals as well?

    Is NAWA really claiming to reduce the weight of the cathode by a factor of 10? There is no iron phosphate, no lithium manganese…? They have invented new cathode chemistry, as well? Or in their cell you only need 10% of the original amount of FePO4, MnO or CoO2?

    No Dan, this doesn't smell right..

    I think that they have some laboratory result with storing lithium ions in the VCNTs, and from that they infer that a battery cell with their electrodes would be much lighter/better. But since they do not have any real cell results, they keep all statements extremely vague, so that no one will charge them with fraud while at the same time attracting more investors…



  10. I'm glad you asked, since this is one of my favourite technical details of Teslas battery plan.

    It works like this: they make a honeycomb sandwich [1] out of metal sheet (bottom), batteries (middle) and metal sheet (top). They pour epoxy between the cells and glue the top/bottom of the cells to the metal sheets. Now, the resulting structure is very stiff and strong.

    The pack weight here is the sum of the metal sheet (top and bottom), the epoxy and controlling electronics. This weight is less than the weight of the "belly" of the car that the battery pack replaces. Hence the "negative" pack weight.

    Now, how can the pack weight be less than the original "belly" sheet of the car? The honeycomb "borrows" strength from the cell cans as to make it a lot stronger than the thin top and bottom metal sheets would have been by themselves. Forces are transferred from the sheets to the cell casings.

    Note that if you added the cells without gluing them to the honeycomb structure they would still have cans for structural integrity, but here the cans have a dual use.

    Clever, no?


  11. Savings on ICV cars are, if I remember from another vague story I read a couple of years ago, currently running at something like $7/kg (for a standard economy hatchback I think).

    So Toyota would pay $70 dollars more for component that was 10 kg lighter on a corolla. This not only lets them save money in the engine/driveline/suspension/wheels/tyres/chassis etc. It also gives them better fuel economy (or CO2 emissions, same thing) which is worth everything from boasting rights in the ad campaign to reduced tax rates and avoiding various financial penalties, depending what market they are selling in.

    (Note that this will not be a smooth function. They have suspension arms X available, that will support a vehicle up to 1500 kg. They also have a cheaper, lighter suspension arm Y available that supports up to 1200 kg. If they drop the vehicle gross mass from 1350 to 1300, they still have to use suspension X, so no suspension savings there.)

    I'll also note, without being able to explain, that while Tesla and Porsche manage to get vast amounts of power and torque out of their EVs, many other makers such as Nissan with their Leaf are still pottering around with vehicles that are below average performance by even ICV standards.

  12. the pack structure is actually lighter than the car structure that it is replacing.

    I don't see how that could work. Surely you could use whatever marvellous materials and construction you used to make the pack, and also use that for the structure?

    Unless they mean that the pack, because it is such a thick floor, allows them to make it much stiffer (for a given weight) than a normal car design where the floor is usually as thin as possible. That would be possible.

  13. The 10x lighter electrodes lead to the 3x energy capacity. Other factors, such as recharge time and cycle number make it 10x better, IMHO. Otherwise, you have always been absolutely correct, as all readers can verify.

  14. Dear Dan, you should read more carefully. I'm claiming exactly what I claimed from the very beginning of our discussion.

    A comment about the "new atlas" source. I assume "10 times better" actually means "10 times higher power density". Is the wording "better" something that this Boulanger came up with? Or is it the reporter?

  15. "one is that CVD is highly unlikely to produce low price bulk material to where their battery could compete on price." Certainly true. They have a process of coating a surface with a structure that is superior to bulk materials. And you may certainly be right about the rest. Or wrong. Glad you have stopped making claims w/o knowledge.

  16. We may never have complete knowledge of the price of their process. They may run out of money and we will never know why they failed.

    At this point though, we do have sufficient knowledge to know a few things. And one is that CVD is highly unlikely to produce low price bulk material to where their battery could compete on price. We also know that Boulangers comments are so vague that they fall into the category of marketing cow-manure.

  17. The article is in recent New Atlas, to bring up a competeing blog. It is about electrodes, so both types, and yes 10x better electrodes is 3x overall battery improve.

  18. "three kilos of graphite with one kilo of graphene" the 3x overall density comes from a 10x factor, not the 3x factor you ignore. "Just to give you some numbers," continues Boulanger, "the cost for
    depositing anti-reflective coating inside a PV panel is a few cents per
    square meter. It's the same, we just deposit our material, because we've mastered the process." is pretty clear he IS stating a price range.

  19. That's the thing, though, the statements can be absolutely true and we still know almost nothing. The process may be the same, but is the cost the same? Boulanger never states anything about their costs.

  20. And then there is of course the question of the energy density. Does NAWA claim that a battery with a NAWA electrode has 3x the energy density per cell weight, or is it only the anode by itself that has 3x the energy density? The difference is of course huge…

    As a matter of fact, we have no idea of the NAWA system. Do they use a VCNT cathode as well, or is it just the anode? When they say 3x energy difference, is that a theoretical value, a measured value from a lab anode or a measurement from an actual prototype battery cell? It makes all the difference.

    A graphite anode make up about 25% of a standard lithium cell per weight, so if a NAWA anode would bring that down to a third, you would save 18% on the cell weight. Not that dramatic.

    Now if you save a factor of 3 on the complete cell weight, that is something else.

    But we just don't know, do we? If you do, please enlighten me and point me to the source.

  21. No, I have deliberately stopped before making the last step in the comparison where I compare three kilos of graphite with one kilo of graphene because I end up in a territory where the prices and gains are so uncertain. I stop at the point where I think that everything points to a very expensive NAWA electrode.

    The uncertainties are so big for both cost and performance as to make the comparison almost meaningless. Boulanger says that coating optics cost "a few cents per m2". So does this mean that it is eaqually cheap for his process? He does not state this, nor do we have any price points for their CNTs. There might be more steps involved in VCNTs compared to optical coating. Up until recently, growing VCNTs entailed laying down a (metal) catalyst layer on the substrate which would seed the CNTs. This implies at least one more vacuum deposition step, i.e. at least twice the process cost. Or perhaps no if the technology has moved beyond this step….

    The point is, that he does not state that the NAWA growth process costs a few cents per m2. A comparison with other industrial producers of lower quality CVD-grown CNTs shows that the end price is most likely much higher than "a few cents" per 100 microns, and hence a much higher volume price, i.e. 10-100x times more expensive. But we do not know the precise answer, since Boulanger is being extremely vague.

  22. "miraculously cheap CVD process" like perhaps "the cost for
    depositing anti-reflective coating inside a PV panel is a few cents per
    square meter. It's the same, we just deposit our material, because we've mastered the process." making cheap surface area, exactly what you need in the battery. It makes sense to me, somehow. Assuming they are giving true statements about tubes at all.

  23. Do you account for the fact that LESS of the CNT material is needed than of the powder? You keep going under the assumption that the masses will be equal, thus the cost comparison per battery will be just the difference in mfg per kilo, seems to me. For electrodes, the important thing is surface area, right? The 3x energy density is because of the 10x less electrode material, right? So, finding very thin (or perhaps fluffy and light) surface to work with gives the added benefits, right? Compared to the thick heavy powders, right? "ultra-lightweight nanotube scaffolding" has one cost per kilo, which you state/estimate, but then compare with cost per kilo of powder, not per surface area, thus weight, thus size, of battery, all better with tubes, right?

  24. A small note here. The new Tesla 4680-cell allows to make the pack weight "negative", i.e. the pack structure is actually lighter than the car structure that it is replacing. This means that a new battery cell would have to have a similar mechanical structure that allows for a "negative" pack material weight, or suffer the added conventional pack weight which may nullify any weight savings due to higher energy density.

    It is not obvious that a new battery technology will allow for this.

    To put this into numbers, we know that Tesla battery packs contains about 160 Wh/kg on a pack level and 250 Wh on a cell level. I.e., the pack ads 56% more weight to the cells. In your example, the pack would add about 150 kg. So, for the new battery cells to compete, it would have to be 150 kg lighter, i.e. about 500 Wh/kg, just to be on weight parity.

    I'm not saying a new battery excludes the possibility of structural packs, but I am claiming that it must be able to be a part of a structural pack to even compete on weight alone.

  25. Interesting. But I wonder how much is gained by reducing the weight on an electric car?

    There may be substantial differences. For instance, every HP in an ICE vehicle is quite expensive to add due to the high cost of the combustion engine, whereas the electric motor is very cheap per unit of power. So the cost savings on the motor and drive train may be completely different.

    The savings on suspension is probably similar for both types of vehicles…

  26. When I try to find any information about the NAWA lithium battery I find no concrete numbers, only a vague statement that their battery could have 3x the energy. Please correct me if I am wrong.

    So what we have is an unknown increase of energy density at an unknown price. If we assume a miraculously cheap CVD process compared to the other CVD producers, then price parity with conventional batteries may be within reach. If we assume that their production methods of CNTs is comparable to other volume producers of CNTs then their battery will be much more expensive. And price is all…

  27. OK Dan, let's work with those numbers. They keep the substrate in the furnace for one minute and they get 100 um nanotubes. And let us assume that they have a perfect load-lock (google it if you don't know the term) so that they loose zero time on loading and unloading samples. Let's say 5 cents per square meter. They then get 0,1 litres for 5 cents or 50 cents for a liter. Looks good so far. Of course, Boulanger never states that their process is this cheap…

    But then you look at the micrographs of their VCNTs and you see that the CNTs actually only make up a small fraction of the volume. And their "cartoon" shows that the lithium ion must be absorbed into the carbon nanotube. Not between them. This is their selling point; the CNT is so stable that it does not expand when the ion is absorbed.

    Taking this into account, the price per volume of CNT is probably a lot higher. At 5% volume ratio (5% CNTs in the film) you would get 10 USD per kg, which is still dirt cheap for CNT's. The market price is about 85 USD per kilo for the lowest quality CNTs [1], or about 110-120 USD per liter. Remember that you need "metalic" VCNTs to conduct the electricity in the battery cell, so if NAWA could produce their CNTs at 10 USD/kilo, it would be a revolution by itself. I highly doubt it.


  28. Yes, but the video of the battery day is not likely to be changed afterwords on account on such a minor detail, don't you think? So even if you rave every time the video is linked, it's not going to do much, is it? But the comment will get tedious quickly….

  29. At the beginning of the 20th century, the electric automobile market share in the US was 38%. However, what hasn't changed since then is that much of the electricity used to recharge EV batteries is produced by coal fired power plants, and EVs are still too expensive for the average American to purchase.

  30. I thought maybe it was the cart from Monty Python and the Holy Grail from the "Bring out your"– … actually, never mind, too soon. Also, that scene didn't have a horse, so I suppose it doesn't count. But the I'm pretty sure D's horse from Vampire Hunter D was powered by batteries. Either that, or the tears of the innocent.

    I should probably drink that coffee that I made at 5 AM, as it is currently 7:17 PM… >_>

  31. Wh/kg still helps a lot for cars. But it's a secondary consideration.

    If your total battery pack is lighter that saves you money in a host of ways:
    –You now need smaller structures to support that battery pack.
    — Smaller, lighter suspension, brakes, motors, driveline to cart that smaller pack around
    — Because the rest of the car is lighter, you can get away with a smaller battery pack for the same vehicle performance

    So there is a point at which a more expensive battery is worth it if it saves you enough kg. But you have to work out the numbers.

    I vaguely recall some claim, it may have been by famous car designer Gordon Murray, that on an EV it's worth about $30 to save 1 kg. (At the time and level of tech development when he said it, so maybe 2010?)

    So, if you have a 300 kg, 80 kW.h battery pack, and you can drop that to 250 kg, that's worth spending an extra 50 x $30 = $1500. Which means that improving energy density by 20% will let you increase the cost by nearly $20/kW.h.

    But given that they are talking about dropping the price by many times that, that outweighs any feasible weight drop.

  32. "Just to give you some numbers," continues Boulanger, "the cost for
    depositing anti-reflective coating inside a PV panel is a few cents per
    square meter. It's the same, we just deposit our material, because we've
    mastered the process. The growth rate for vertically aligned carbon
    nanotubes is known as being very, very fast. We can grow vertically
    aligned nanotubes up to, let's say, 100 microns per minute. It needs
    only one minute in the furnace. We've scaled this process on very large
    surfaces, and with a process that works at atmospheric pressure, at
    lower temperature, we can do it a little bit like making a newspaper.
    Not that fast, but almost the same idea."

  33. What Photon Plumber sez. "the volume almost inconsequential", thus the kilos, but not the important part, the surface area. Why is this so hard to get?

  34. Sure you are going to get more power per weight with a large area per weight. Sure you are also going to get more lithium into the carbon nanotube structure per weight compared to, say, standard graphite.

    And this changes absolute nothing about the cost per kWh. Every kilo of vertically aligned CNT's is going to be astronomically expensive. Why is this so hard to get?

  35. Because I get annoyed at marketing claims that assume history began only 10 years ago. My working life involves multiple disputes with marketing departments and their approach to reality, and so I'm sensitive to it.

  36. What we are dealing with here is a joke, or humorous comment, based on the fact that your metaphor of the horse and cart is actually subject relevant to transportation systems, the meta-subject of the entire post.

    My suggestion is to drink more of your morning coffee before responding to other people's comments that don't make sense at 5 am.

    Unless of course you are making a second level sarcastic joke and I'm the one who's not getting it, because I'm only halfway through my coffee at 7 am.

  37. Kind of a silly topic to cheerlead on. Up until now, batteries have been pretty generically and mass produced for a host of end uses–no one 'EV battery'. While the 4680 may have been completely designed and produced in the US, it'll likely be used in non-automotive applications, making it no different historically from other batteries.

  38. Can concur, the SiO2 AR coatings on our laser optics are quite thin, spread over a 20mm optic makes the volume almost inconsequential.

  39. How does that have anything to do with German battery talk when that plant has not been approved to make batteries? Red herring much?

  40. Agreed. I'm frustrated by people tallking about Wh/kg or Wh/L. That only matters for phones and airplanes. For cars what matters is Wh/$.

  41. hmm not sure but I think the Horse and cart are on the way out and the future is the Automobile, still time will tell.

  42. Tesla does not have approval from Germany nor the city to produce batteries in that plant. Putting the cart before the horse.

  43. I'm not really a battery fan, certainly not an expert, but these layers of nanotubes are far thinner than the bulk quantities "So when you calculate the *volume* price of vertically aligned CNT's the price is bound to be astronomical.". The fact that they are thinner and provide a *surface area* to interact with the *juice* whatever is used, is an advantage!

  44. But Dan, think a bit… Even if it's cheap to coate a window with a few hundred nanometers of anti-reflective coating, this layer is almost nothing in terms of volume.

    So when you calculate the *volume* price of vertically aligned CNT's the price is bound to be astronomical. I'm willing to bet quite a lot of money on that.

    Let's say you can coat 1 m2 of glass with anti-reflective coating for 1 USD. I would say that is dirt cheap. With 500 nm coating, that would translate to 0.0005 litres, or about 1000 USD per kg. Does that sound cheap to you? And that is assuming a ridiculously cheap coating process….

    Any production method that relies on bulk materials that are produced by CVD is doomed. It's a dead end. And so is this battery for mass adoption…

  45. the process we're using is the same process that's used for coating
    glasses with anti-reflective coatings, and for photovoltaics. It's
    already very cheap."

  46. Why is this important? You have posted about this a few times… Must have some great significance..?

  47. My conclusion is that Tesla is sandbagging the rate of 4680 battery adoption in the company.

    Tesla Berlin is supposed to start ramping their production in 2021, not in 2022. Now, they seem to be ahead of the plane, which means that ramping may start early 2021, which means that they would need those 4680 batteries in a few months.

    And if Tesla Berlin would produce even 200k model Y in 2021 – a low ball estimate – they would need at least 14 GWh for this model alone. And we are not even counting the Cybertruck or the Semi… 50 GWh in 2021? And 150 GWh in 2022? That's my guess…

  48. Jeezus wept…

    It's vertically aligned carbon nanotubes….

    A normal graphite anode costs about 8-20 USD per kg. What do you think the cost of vertically aligned CNT's will be, per kg?

    These naive guys talk about a growth rate of 100 microns per minute in CVD as if it would solve anything. Just think about it. If you have a giant CVD that has a chamber with 1 m2 surface, you would get about 6 mm in one hour, or about 6 litres, or about 12 kg. That's it.

    This is absolutely not for mass market batteries…

  49. Not important. Cost is all.

    I used to think like you, that the race was for more Wh per kg, and to some extent, it is for smaller applications (mobile phones, future airplanes etc), but the biggest market just wants low prices.

    If you have a sufficient number of cycles, say 4000, and the range of you car is sufficent, why would you care about the exact energy density of the battery cells in your car? Or in the power wall? Or in the utility power bank?

    At this point, it is just about the cost at a sufficient longevity/quality…

  50. Definitely the first one destined for mass production, and specifically designed for EVs rather than repurposed leisure batteries.

  51. I thought Semi was in Austin and at a Mw+ each would be the top user of 4680 with the strongest utility reason for the best power/weight. Semi is supposed to have priority over Roadster or even CT.

  52. 3x the energy density, 10x the power, vastly faster charging and battery lifespans up to five times as long.

Comments are closed.