How many kWh can they squeeze into the Model 3...?

[Rashomon]

You are both misinformed. The larger cell optimizes the packaging/content ratio, some 10% of density increase can be achieved. Also, the cells are longer and WILL fit in the same standing up orientation, that's another 7.7% per square meter of floor utilized.

See this thread on the cells: 21-70 cell information. We still know almost nothing on Panasonic 2170 performance. If we take Elon at his word, and these are the most energy dense cells available, they're likely to still only be a small incremental improvement on specific energy to the current 18650s. They will save money at the cell level and at the pack level by being larger, but they probably won't offer any dramatic breakthrough in pack weight or size for a given energy.

Every indication is that Tesla is trying to hit an aggressive price target on the M3 by high efficiency and minimizing battery capacity, giving just enough to hit range targets that are good (great compared to almost anything else) but not particularly record setting. I don't expect to see a pack bigger than 70-75 kWh offered in the first year or so of M3 production.

 

[Red Sage]

Hmmm... Napkin math alert !

1.4666666 × 1.3 = 1.90666658

Idunno about you guys, but from my vantage point ~191% is close enough on a per battery cell basis to qualify as '...about twice...' the energy density and stuff. The density throughout the entire battery pack might not be double, of course... But surely we can cut a guy a bit of slack for the sake of spirited debate. Yes? Good.

 

[stopcrazypp]

You are both misinformed. The larger cell optimizes the packaging/content ratio, some 10% of density increase can be achieved. Also, the cells are longer and WILL fit in the same standing up orientation, that's another 7.7% per square meter of floor utilized.

Have you actually seen the pictures of the 100kWh modules? Yes, some packing improvements can be had from the older 85/90kWh packs, but then the 100kWh packs already have the cells packed much more tightly:
8256 Cells vs 7104 cells or about 1.16x
Pics/Info: Inside the Tesla 100 kWh Battery Pack

Larger cylinders will have better casing to volume ratio, but that assumes the casing thickness remains the same. However, the battery presentation in the thread posted by @Rashomon suggests that 20700/21700 cells will have thicker casing than 18650 for safety. I'm presuming the casing thickness will scale proportionally to maintain structural integrity. Of course we don't know about the actual cells Tesla will be using (what the actual density is vs the current cells in the 100kWh pack).
21-70 cell information

As for the 7.7% improvement from taller cells, I already accounted for that. That advantage is exactly cancelled out by the smaller width and wheelbase of the Model 3.

 

[stopcrazypp]

Hmmm... Napkin math alert !

1.4666666 × 1.3 = 1.90666658

Idunno about you guys, but from my vantage point ~191% is close enough on a per battery cell basis to qualify as '...about twice...' the energy density and stuff. The density throughout the entire battery pack might not be double, of course... But surely we can cut a guy a bit of slack for the sake of spirited debate. Yes? Good.

The 1.466x is just the increase in volume of the cell, not the increase in density. Density is the energy per unit volume. If you increase volume and energy increases proportionally, the increase in density is zero.

As for the 1.3, I'm not sure where that comes from. Keep in mind that the reference point would be the cells in the 100kWh pack.

 

[Red Sage]

The 1.466x is just the increase in volume of the cell, not the increase in density. Density is the energy per unit volume. If you increase volume and energy increases proportionally, the increase in density is zero.

As for the 1.3, I'm not sure where that comes from. Keep in mind that the reference point would be the cells in the 100kWh pack.

The 1.3 refers to the 30% improvement in energy density for Generation III vehicles in 2017, compared to Generation II vehicles in 2012, as noted by JB Straubel several times since mid-2015.

 

[stopcrazypp]

The 1.3 refers to the 30% improvement in energy density for Generation III vehicles in 2017, compared to Generation II vehicles in 2012, as noted by JB Straubel several times since mid-2015.

That's what I suspected. But the 100kWh pack is already 1.176x that of the 2012 85kWh pack (actually slightly more if wk057's actual capacity numbers are accurate, given he measured the 100kWh pack at 102.4 kWh). That leaves an improvement of roughly 1.105x which given the 87.5 kWh I counted before, leaves 96.7 kWh. That extends my estimate to 95kWh nominal (using latest rounding convention), but still not quite 100kWh.

 

[Red Sage]

That's what I suspected. But the 100kWh pack is already 1.176x that of the 2012 85kWh pack (actually slightly more if wk057's actual capacity numbers are accurate, given he measured the 100kWh pack at 102.4 kWh). That leaves an improvement of roughly 1.105x which given the 87.5 kWh I counted before, leaves 96.7 kWh. That extends my estimate to 95kWh nominal (using latest rounding convention), but still not quite 100kWh.

OK. That's an interesting way to look at it. But again, I'm going by what JB Straubel said.

With a 30% improvement in energy density by mass/volume, that means the space needed to enclose 85 kWh in 2012 would be only 70% of that volume in 2017. Similarly, the volume required in 2012 for 100 kWh would be around 86% that amount in 2017. That fits projections of available volume under Model ≡ for a battery pack.

I expect there are efficiencies beyond the battery cells themselves within the battery packs. JB typically refers to the importance of 'battery management' and 'power electronics' systems as well as the software that controls them for charging and discharge. These will provide the additional support for Model ≡ to get more mileage from lower capacities than I currently expect. But it will by no means be the 'magical' key that unlocks a typical 150 Wh/mile to 200 Wh/mile efficiency as some seem to project.

 

[Cloxxki]

OK. That's an interesting way to look at it. But again, I'm going by what JB Straubel said.

With a 30% improvement in energy density by mass/volume, that means the space needed to enclose 85 kWh in 2012 would be only 70% of that volume in 2017. Similarly, the volume required in 2012 for 100 kWh would be around 86% that amount in 2017. That fits projections of available volume under Model ≡ for a battery pack.

No, just no, you mathemetical naysayer.
+30% makes the smaller number: 100%/1.3=76.9% of the larger number.

It's obviously vital with every cryptic oracle from the Tesla to track whether they were speaking about cells, packs, or by volume. Especially since the packs effectively gain 7.7% of volume by accomodating 2170's upright like the 18650's.
85kWh (really 80.7 remember) x 1.3 = just 104.91kWh. And the 100 pack is already at 102.4kWh. So, EXTREMELY unlikely this is how JB meant the 30%. And if he did, he pretty much lied intentionally, again about the long exposed overstated 85.

2 options here: the +30% statement has been overtaking by reality, or it was misinterpreted (likely as intended).

Let's say it was the density of the cells to increase by 30%, and on top of that 7.7% more utilization of the pack's outer size. And, as hinted and guessed and semi confirmed, similar utilization of floor surface to the 100 (102.4 kWh) pack.
102.4*1.3*1.077=143.3 kWh.

Perhaps the 7.7% was already taking into account, just not specified? Possible. Then, 133.3 kWh.
Which is my best bet based on my own little guestimate: 102.4 + 10% cell density from optimized 2170 format + 7.7% longer cells + 10% long overdue chemistry update = 133.44 kWh.

Also, it's called Model 3 now.

 

[stopcrazypp]

No, just no, you mathemetical naysayer.
+30% makes the smaller number: 100%/1.3=76.9% of the larger number.

It's obviously vital with every cryptic oracle from the Tesla to track whether they were speaking about cells, packs, or by volume. Especially since the packs effectively gain 7.7% of volume by accomodating 2170's upright like the 18650's.
85kWh (really 80.7 remember) x 1.3 = just 104.91kWh. And the 100 pack is already at 102.4kWh. So, EXTREMELY unlikely this is how JB meant the 30%. And if he did, he pretty much lied intentionally, again about the long exposed overstated 85.

2 options here: the +30% statement has been overtaking by reality, or it was misinterpreted (likely as intended).

Let's say it was the density of the cells to increase by 30%, and on top of that 7.7% more utilization of the pack's outer size. And, as hinted and guessed and semi confirmed, similar utilization of floor surface to the 100 (102.4 kWh) pack.
102.4*1.3*1.077=143.3 kWh.

Perhaps the 7.7% was already taking into account, just not specified? Possible. Then, 133.3 kWh.
Which is my best bet based on my own little guestimate: 102.4 + 10% cell density from optimized 2170 format + 7.7% longer cells + 10% long overdue chemistry update = 133.44 kWh.

Also, it's called Model 3 now.

Your analysis is predicated by assuming the 30% improvement is from current state of the art Model S to Model 3. That is not what JB was talking about.

Here is his direct quote:
"These batteries are steadily improving every single year – maybe around 5% improvement in their energy density — their ability to store energy in a given amount of mass. That’s probably one of the key metrics we worry about. And when we went from the Roadster to the Model S, they have improved by about 40% and when we were designing the Model 3, they were about another 30% better. That improvement just continues on every single year in the background."
Tesla Model 3 battery technology improves: JB Straubel cites 30% energy density increase [Video]

The original Roadster to Model S saw 40% improvement. The 30% is clearly talking about from original Model S (85kWh pack) to when they started designing the Model 3.

That means they are talking about 85kWh *1.3 = 110.5 kWh (or 81kWh*1.3=105.3 if you subscribe to wk057 numbers) for a Model S sized pack.

Then factor in that the Model 3 pack's 8 modules uses only the 14 module space out of the 16 modules (it doesn't use an extended double stacked section anymore), slightly longer cells, and smaller width*wheelbase.

85kWh * 1.3 * (14 modules/16 modules) * (70mm/65mm) * (8384.6 sq in / 9005.45 sq in) = 96.95 kWh.

or

81kWh * 1.3 * (14 modules/16 modules) * (70mm/65mm) * (8384.6 sq in / 9005.45 sq in) = 92.38 kWh.

Neither are at 100kWh. So again, Elon's claim that 100kWh doesn't fit in Model 3 is plausible for current tech and what we know so far about the Model 3 pack design.

 

[scaesare]

Incidentally, I don't assume Model 3 modules hold the same number of cells as a Model S As such, I'm suspect of using the number of modules in each pack as a ratio for predictive purposes...

 

[Cloxxki]

Your analysis is predicated by assuming the 30% improvement is from current state of the art Model S to Model 3. That is not what JB was talking about.

Here is his direct quote:
"These batteries are steadily improving every single year – maybe around 5% improvement in their energy density — their ability to store energy in a given amount of mass. That’s probably one of the key metrics we worry about. And when we went from the Roadster to the Model S, they have improved by about 40% and when we were designing the Model 3, they were about another 30% better. That improvement just continues on every single year in the background."
Tesla Model 3 battery technology improves: JB Straubel cites 30% energy density increase [Video]

The original Roadster to Model S saw 40% improvement. The 30% is clearly talking about from original Model S (85kWh pack) to when they started designing the Model 3.

That means they are talking about 85kWh *1.3 = 110.5 kWh (or 81kWh*1.3=105.3 if you subscribe to wk057 numbers) for a Model S sized pack.

Then factor in that the Model 3 pack's 8 modules uses only the 14 module space out of the 16 modules (it doesn't use an extended double stacked section anymore), slightly longer cells, and smaller width*wheelbase.

85kWh * 1.3 * (14 modules/16 modules) * (70mm/65mm) * (8384.6 sq in / 9005.45 sq in) = 96.95 kWh.

or

81kWh * 1.3 * (14 modules/16 modules) * (70mm/65mm) * (8384.6 sq in / 9005.45 sq in) = 92.38 kWh.

Neither are at 100kWh. So again, Elon's claim that 100kWh doesn't fit in Model 3 is plausible for current tech and what we know so far about the Model 3 pack design.

The ~105kWh based on a 30% CELL cell density improvement could well be. However, it obviously doesn't account for the longer cells, as then even without improved density we'd see 102.4 + 7.7% = 110kWh. So this 7.7% needs to be added still. And, really, do we know that the 16% (90->100kWh upgrade) that was crammed in extra was factored in? Not if JB was talking CELLS. 110 + 16% = 127,6kWh. If the 7.7% WAS factored in, it'd be 118.5kWh.

But, did JB really know a the time that the cell size would be adjusted (to get that ~10% improvement right there)?

And indeed, you can't just presume a module stays the same size for Model 3 in any way. Smaller car, right? They're alsogoing to have a different wiring due to the less numerous cells. Unless they accept a much lower voltage.

Shorter and narrower, that will really make the 3 pack smaller. The 2-stacked modules kept out also really hurts.

I'm sad. A 100kWh Model 3 would do some crazy EPA rated mileage. 400+.

 

[Red Sage]

Here is his direct quote:
"These batteries are steadily improving every single year – maybe around 5% improvement in their energy density — their ability to store energy in a given amount of mass. That’s probably one of the key metrics we worry about. And when we went from the Roadster to the Model S, they have improved by about 40% and when we were designing the Model 3, they were about another 30% better. That improvement just continues on every single year in the background."
Tesla Model 3 battery technology improves: JB Straubel cites 30% energy density increase [Video]

The original Roadster to Model S saw 40% improvement. The 30% is clearly talking about from original Model S (85kWh pack) to when they started designing the Model 3.

Thank you for the link! I'm not sure if I have watched this particular presentation before. But it seems as though he was covering points he gave at a Conference in Korea around mid-2015, then expanding upon them relative to Model ☰'s use of 2170 battery cells. Thanks again.

 

[stopcrazypp]

Incidentally, I don't assume Model 3 modules hold the same number of cells as a Model S As such, I'm suspect of using the number of modules in each pack as a ratio for predictive purposes...

Look a bit more closely at my numbers. I'm using the number of modules in Model S as a proxy to the differences in volume. Model S had 16 modules in total with 2 double stacked in a small extended section in front, so 14 of them are in the main rectangular section. The Model 3 uses 8 modules in the same general area that the Model S had those 14 modules (width*wheelbase differences aside). That means that the Model 3 will obviously have larger modules that can fit more cells per module.

Note that my numbers does not refer to 8 modules in Model 3 at all. The only assumption I am making is that the Model 3 will pack the cells proportionally close as whatever tech Straubel was referring to and that the Model 3 will use a rectangular pack with no extended section (as per graphics during Model 3 unveiling).

 

[Red Sage]

Look a bit more closely at my numbers. I'm using the number of modules in Model S as a proxy to the differences in volume. Model S had 16 modules in total with 2 double stacked in a small extended section in front, so 14 of them are in the main rectangular section. The Model 3 uses 8 modules in the same general area that the Model S had those 14 modules (width*wheelbase differences aside). That means that the Model 3 will obviously have larger modules that can fit more cells.

Possibly more battery cells per module, but the battery pack as a whole will have fewer cells. There is a discussion at the Tesla Forums where it has been calculated the Model ☰ might have a certain maximum number of cells. Per @Topher ...

"21mm cells gives: 5415 in a square arrangement and 6159 in a honeycomb arrangement."​

He explained how he came to those numbers in this post. Seems legit to me.

 

[stopcrazypp]

Possibly more battery cells per module, but the battery pack as a whole will have fewer cells. There is a discussion at the Tesla Forums where it has been calculated the Model ☰ might have a certain maximum number of cells. Per @Topher ...

"21mm cells gives: 5415 in a square arrangement and 6159 in a honeycomb arrangement."​

He explained how he came to those numbers in this post. Seems legit to me.

Right, I'm referring to cells per module, not total cells, I edited to make this a bit more clear. Model 3 has fewer modules (half as many as 100kWh Model S), but they are individually larger.

 

[scaesare]

Ah, OK.. you are solving for the percentage of the 85kWh pack energy attributable to the footprint occupied by 14 of the modules in that portion of your equation.

Cool.. makes sense.

Thanks for clarifying.

 

[Cloxxki]

8,256 cells for the "maximized" 100kWh pack, right?
So they didn't quite achieve that theoretical honeycomb number. But, close.

Also, are the 2170's really 21mm, or 20.3mm as I seem to remember being reported?
Using a 6,000 number for 2170's and the rumored/expected 20Wh/cell figure brings a somewhat disappointing theoretical maximum size for the 2170 S/X packs. Going by that, sub 100kWh for Model 3 is not so strange.