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

4680 cell design, chassis integration & factory discussion for investors

This site may earn commission on affiliate links.
There are 8 gigacast machines inbound for Berlin.

1 each for:

MY front
MY rear
M3 front
MY rear
M3 SR+ floorpan
M3 LR / P3D floorpan
MY LR floorpan
MY Perf floorpan

Are the floorpans (below the battery) going to be gigacast? Can they be common across some of the above?
 
  • Informative
Reactions: EVCollies
Gigacast confirmed as a requirement for Plaid S presumably as Elon has confirmed structural 4680s. Will this be used for non Plaid? Anyone think the front and rear will be cast also?

While we don't know for sure. I am sure there was a comment about the structural pack needing front and rear castings.

If they can need to make front & rear casting for Plaid that might apply to all Model S/X as IMO they only want one version of the structural pack.

We also discussed in the main thread using the structural pack in a non-structural way to simply replace packs in existing older Model S/X.
The consensus was that even when used non-structurally the structural pack would weigh no more than the pack it was replacing.

Some of the advantages of Plaid obviously come from new motors and wiring, putting a 4680 pack into older cars isn't going toe reap the same benefits...

Eventually standardizing of 4680, structural packs , front and rear casting is all new cars will reap substantial benefits, unless there are good reasons why it can't be done for a particular model, that is what I expect to see. Of course it will take time to migrate..
 
  • Like
Reactions: Buckminster
There are 8 gigacast machines inbound for Berlin.

1 each for:

MY front
MY rear
M3 front
MY rear
M3 SR+ floorpan
M3 LR / P3D floorpan
MY LR floorpan
MY Perf floorpan

Are the floorpans (below the battery) going to be gigacast? Can they be common across some of the above?

I thought Elon said that the structural pack would be steel sheets top/bottom which means that it won't be a cast assembly. (Though it could be stamped.)

Yep, found it:
Elon Musk @elonmusk ·
Oct 7

Replying to @ChananBos @raytech247 and @WholeMarsBlog
Battery pack will be a bonded structure with cells providing shear transfer between steel upper & lower face sheets, eliminating most of the center body parts while providing better torsional rigidity & improved polar moment or inertia. This is a *major* breakthrough.

So it seems like casting will only be used for the front/rear sections of the frame.
 
Last edited:
  • Like
Reactions: MC3OZ and Fred42
I thought Elon said that the structural pack would be steel sheets top/bottom which means that it won't be a cast assembly. (Though it could be stamped.)

Yep, found it:


So it seems like casting will only be used for the front/rear sections of the frame.

This picture clearly shows channels below:
Tesla Battery Day Slides 2020/09/22
Also reduced number of channels on the top although this could be busbar space?

Metallurgy team could have cracked steel casting?
 
Munro has mocked up the battery pack. He is assuming 4 module (although physically they are all in the pack together). I am assuming Tesla find a way not to have the modules and extra wiring.

Gasket approach looks good but would it be okay as a structural pack? I was assuming that the cells would alternate orientation.
 
I watched the Battery Day presentation yesterday. it was a worthwhile exercise, it is very comprehensive and well explained.
One statement from Elon stood out as significant, front and rear castings are needed for a structural battery pack.
Given that Tesla use a special alloy for casting, other car makers seem like years away from being able to do structural packs.

The reason I re-watched is related to Plaid Model S battery pack.

We had some Reddit posts here around 24 hours ago from someone claiming to work on the Model S/X battery pack line.
My understanding was he was implying 18650s will be used in a new pack, and possibly for the Plaid Model S structural pack.

Th Reddit user has since deleted all of their posts and I think their user id, this tends to make them more credible IMO.
But there is no certainty their claims are correct, that they have the full picture, or that I have correctly understood their claims.

Elon confirmed a structural pack and possibly 4680 for Plaid Model S in a twitter exchange along these lines..

Omar::Will Plaid Model S have a structural pack and I assume the power density from 4680 helps.
Elon: Yep.

The Reddit user claimed that posit was misunderstood, Elon definitely confirmed that Plaid Model S has a structural pack and that the power density from 4680 helps. He didn't necessarily confirm the Plaid Model S pack will be built with 4680s, but that is a reasonable conclusion, and the vast majority of people would form that conclusion.

The other evidence for Plaid Model S having 4680s is:-
1) it was included in the Battery Day presentation, however I do not think they detailed what cells it was using.
2) Model S/X were listed in the vehicles specifically as using Nickel-Manganese cells.

However, I still have doubts about whether Model S/X is using Nickel-Manganese cells from Kato Rd Q1 2021.
It is possible Kato Rd is making Nickel-Manganese cells, but they are reserved for supporting the Berlin Model Y ramp.

We know Model Y at Fremont was initially built with a 2-peice rear casting and a non-structural 2170 battery pack.
So that path is possible for Model S/X with a later upgrade to a structural 4680 pack with front and rear castings.

However, the Reddit user claimed to down time was for a new pack architecture..
This may still used with a 2-peice rear casting, but could be a 18650 pack that is capable of structural deployment with front and rear casting. That was my reading of the Reddit user's comments.

However this does raise the question of whether the 18650 cells would need to be tab-less, I think tab-less 18650 and 2170 cells are possible, and that upgrade is worthwhile if these cells have a long term future.

However, I also seem to remember the Rivian pack is built with 2170s and end cooling, I doubt those were tab-less 2170 cells.
Again the Reddit user said something cryptic here about assumptions around cooling not being correct.

If an 18650 structural pack has been built, that may just be an interim step, until Model S/X can migrate to 4680.
However, we can be confident 18650s are available Q1 2021, provided they don't need to be modified for tab-less design. or that production line change can be done.

In this context, when I re-watched the Battery Day, I was looking for specific advantages 4680 cells might deliver over and above cost savings. Higher energy density, lower weight, and faster charging were 3 likely advantages..

Overall I'm reasonably confident about my interpretation of the Reddit users comments, and I think they have part of the picture.
If a Model S/X platform can be adapted to work with a structural 4680 pack, it can certainly use some kind of 18650 pack as an interim solution.
 
Last edited:
  • Informative
Reactions: Mo City and AZRI11
I.....

The reason I re-watched is related to Plaid Model S battery pack.

We had some Reddit posts here around 24 hours ago from someone claiming to work on the Model S/X battery pack line.
My understanding was he was implying 18650s will be used in a new pack, and possibly for the Plaid Model S structural pack.

Th Reddit user has since deleted all of their posts and I think their user id, this tends to make them more credible IMO.
But there is no certainty their claims are correct, that they have the full picture, or that I have correctly understood their claims.......

Overall I'm reasonably confident about my interpretation of the Reddit users comments, and I think they have part of the picture.
If a Model S/X platform can be adapted to work with a structural 4680 pack, it can certainly use some kind of 18650 pack as an interim solution.

I suspect they have several solutions running in parallel. Some are the lowest risk but least attractive (e.g. 18650 non-structural). Others are highest risk and highest apparent reward (4680 structural + full-castings). In between may be a 18650 structural module that can either be fitted into a cast frame, or used in non-structural adaptations of existing unibodies for development testing purposes. As they pass the switch points in the various projects they reset the base-case pathway of the overall programme to include different combinations of the individual project.

My personal suspicion is that Kato Rd 4680 production is going to have to support Berlin initial Y production. As the days pass it seems less likely that they will be able to get a Berlin 4680 cell facility built in time for initial Y production there. I could of course be wrong, especially as I do not know what forms the critical path, but it would seem that there is some value of delay, if only to get Kato Rd iteration #n absolutely nailed (mchy, solvent, water, yield, land grab) before committing to the Berlin 4680 cell build. By the time they get to the corresponding point in the Austin build the pathway should be fully understood so that there is no hesitation.

That in turn means that a mid-path fallback plan is required, and I think what you are suggesting as a 18650 structural module may form the core of that plan. This would enable the S/X refresh in Jan-21 which can then go to market using any combination of 18650 or 4680 in a way that is invisible to the client, i.e. the buyer would not know what is inside the pack. However in order to achieve that the 18650 structural module needs to be functionally near-identical to the 4680 structural module, and the form-near-identical, and the price (design/build penalty) of that modularity needs to be very low. If they can achieve that for the S/X in Fremont, then they can also achieve that for the Y launch in Berlin, and they can also drop it in to the 3/Y-lines at other factories at some point (Fremont, Shanghai).

That near-identicality of structural module needs to be sufficiently close that one set of homologations (certifications) will enable release into the relevant markets. So watching the (mooted) initial S/X refresh, and the Berlin Y homologation, may be the leading indicators.

Getting that design/build penalty down is very difficult indeed. Think, for example, of the relatively simple matter of how & where to create the coolant interfaces. There is always a cost to modularity, most especially in physical product.

Then one commits to build/launch of Plaid and Roadster and Semi when 4680 are fully solved and that modularity penalty can be eliminated, i.e. later in 2021.

It is fascinating to watch. Without knowing the overall programme, what the switch points are in it, and how progress is being achieved daily, it is very understandable that the Reddit leaker (if kosher) might themselves be somewhat confused. It would also be very understandable that they went into full-delete mode, as by reviewing what they have said and the mistakes they are making (or not making) one could probably nail who they are quite precisely.

regards, dspp/pb
 
Then one commits to build/launch of Plaid and Roadster and Semi when 4680 are fully solved and that modularity penalty can be eliminated, i.e. later in 2021.

Yes, we are in complete agreement especially on this point.

The Reddit poster probably realized they were working in a small team and easily identified, or management advise the poster/team to remove the posts.

Tesla management are aware employees leak information, so mostly employee know when they need to know and mostly only know what they need to know.

If we consider the improvements that could be made to the LR and Performance Model S/X with a hypothetical 18605 structural pack, thjey stack up:-
  • Less weight from structural pack and castings.
  • Weight moved to the center of the vehicle for better handling.
  • Faster charging via new cables/components
  • Simplified wiring harness - less weight
  • Heat pump / Octovalve
  • Interior and exterior styling
Not only would this be a considerable improvement on the existing versions, the cars should be easier to build via a simplified streamlined process.

Reflecting on the total picture I now think castings for Fremont Model S/X would be a higher priority than castings for Fremont Model Y.

The pack would not need to increase from 100 kWh to deliver more range, faster charging, better handling and performance.

The cells will keep improving .. so keeping in mind Tesla will choose a strategy that maximizes the number of cars they can build, there is no compelling need for a short term move from 18650 for LR and Performance models.

However for Plaid at the Roaster, I can make a case for 4680 especially high nickel 4680.

For Plaid models the pack will be bigger, they want to keep the weight centered for handling and they want maximum power and energy density.

The Roadster may be able to share a pack design with the Plaid models and avoid a complicated double pack configuration.

The new models need homologations (certifications) anyway, there is no additional penalty.

Whether or not the speculation is accurate we don't know. it all depends on the credibility of the Reddit poster
 
  • Like
Reactions: petit_bateau
I watched the Battery Day presentation yesterday. it was a worthwhile exercise, it is very comprehensive and well explained.
One statement from Elon stood out as significant, front and rear castings are needed for a structural battery pack.
Given that Tesla use a special alloy for casting, other car makers seem like years away from being able to do structural packs.

The reason I re-watched is related to Plaid Model S battery pack.

We had some Reddit posts here around 24 hours ago from someone claiming to work on the Model S/X battery pack line.
My understanding was he was implying 18650s will be used in a new pack, and possibly for the Plaid Model S structural pack.

Th Reddit user has since deleted all of their posts and I think their user id, this tends to make them more credible IMO.
But there is no certainty their claims are correct, that they have the full picture, or that I have correctly understood their claims.

Elon confirmed a structural pack and possibly 4680 for Plaid Model S in a twitter exchange along these lines..

Omar::Will Plaid Model S have a structural pack and I assume the power density from 4680 helps.
Elon: Yep.

The Reddit user claimed that posit was misunderstood, Elon definitely confirmed that Plaid Model S has a structural pack and that the power density from 4680 helps. He didn't necessarily confirm the Plaid Model S pack will be built with 4680s, but that is a reasonable conclusion, and the vast majority of people would form that conclusion.

The other evidence for Plaid Model S having 4680s is:-
1) it was included in the Battery Day presentation, however I do not think they detailed what cells it was using.
2) Model S/X were listed in the vehicles specifically as using Nickel-Manganese cells.

However, I still have doubts about whether Model S/X is using Nickel-Manganese cells from Kato Rd Q1 2021.
It is possible Kato Rd is making Nickel-Manganese cells, but they are reserved for supporting the Berlin Model Y ramp.

We know Model Y at Fremont was initially built with a 2-peice rear casting and a non-structural 2170 battery pack.
So that path is possible for Model S/X with a later upgrade to a structural 4680 pack with front and rear castings.

However, the Reddit user claimed to down time was for a new pack architecture..
This may still used with a 2-peice rear casting, but could be a 18650 pack that is capable of structural deployment with front and rear casting. That was my reading of the Reddit user's comments.

However this does raise the question of whether the 18650 cells would need to be tab-less, I think tab-less 18650 and 2170 cells are possible, and that upgrade is worthwhile if these cells have a long term future.

However, I also seem to remember the Rivian pack is built with 2170s and end cooling, I doubt those were tab-less 2170 cells.
Again the Reddit user said something cryptic here about assumptions around cooling not being correct.

If an 18650 structural pack has been built, that may just be an interim step, until Model S/X can migrate to 4680.
However, we can be confident 18650s are available Q1 2021, provided they don't need to be modified for tab-less design. or that production line change can be done.

In this context, when I re-watched the Battery Day, I was looking for specific advantages 4680 cells might deliver over and above cost savings. Higher energy density, lower weight, and faster charging were 3 likely advantages..

Overall I'm reasonably confident about my interpretation of the Reddit users comments, and I think they have part of the picture.
If a Model S/X platform can be adapted to work with a structural 4680 pack, it can certainly use some kind of 18650 pack as an interim solution.

Wouldn’t smaller diameter cells than 4680 have less heat problems leading to power density advantages? So more expensive, not as good energy density, but better power density might be possible with smaller cells.

‘My understanding as far as power goes is that tabless giveth and larger diameter taketh away. Remember the Supercharging graph vs. cell diameter?
 
Wouldn’t smaller diameter cells than 4680 have less heat problems leading to power density advantages? So more expensive, not as good energy density, but better power density might be possible with smaller cells.

‘My understanding as far as power goes is that tabless giveth and larger diameter taketh away. Remember the Supercharging graph vs. cell diameter?

They said Tabless results is a shorter path overall so less heat, but I think they were comparing to a 2170.

My hunch is Maxwell DBE requires a large diameter (cracking on smaller diameters was problem at one time).
And even if Maxwell DBE didn't require 4680, 4680 is the lowest cost option...

Fast charging a 4680 definitely needs tabless, the presentation started that, a cell with tabs would generate more heat

What they didn't say was, a structural pack needs 4680 cells or tabless.

Like you, I have a hunch that multiple smaller cells sometimes gives a power density advantage.

It also seems likely that for Model S/X\ performance and handling are the key targets, cost and range are lower priorities, especially for Plaid Model S/X.
 
Getting that design/build penalty down is very difficult indeed. Think, for example, of the relatively simple matter of how & where to create the coolant interfaces. There is always a cost to modularity, most especially in physical product.

Which might be why that isn't how they usually do things that I've seen. Of course, if they look at making something modular and find it easy to do, that would be a different matter since "usually" doesn't matter if something is easy.