Tesla Model S 85 kwh Battery Pack Volumetric Efficiency Left Room For Improvement


2015 Tesla Model S 85

2015 Tesla Model S 85

We know that Tesla has squeezed more cells into the same physical size pack in their new P100D battery pack that was just announced August 23, 2016 in a conference call with journalists.

Tesla CEO Elon Musk & CTO JB Straubel talk all things battery and Gigafactory last month

Tesla CEO Elon Musk & CTO JB Straubel talk all things battery and Gigafactory last month

“The cell is the same – but the module and pack architecture has changed significantly in order to achieve adequate cooling of the cells in a more energy dense pack, and to make sure that we don’t have cell to cell combustion propagation” – Tesla CEO Elon Musk from the conference call on the new 100 kWh options.

So obviously there was room for improvement in the old 85 kwh pack design. If it was as good as it could be they never would have improved it –yes?

But how much improvement: Again we know the answer. One hundred kwh’s divided by 90 kwh is an 11% improvement.

Let’s put more numbers to it by looking at the old 85 kwh pack and applying some simple geometric calculations.

First, look at a photo taken of the old 85 kwh pack, and using power point let’s draw some circles around the cells. You can see the edge of the cells in the photo so one can draw the circles fairly accurately. More photos in this TMC forum post.

 Telsa 85 kwh battery pack with circles laid over the cells. Photo courtesy Tesla Motors Club Forum poster wk047.

Tessa 85 kWh battery pack with circles laid over the cells. Photo courtesy Tesla Motors Club Forum poster wk047.

Now eliminate the photo underneath and just leave the circles and it is easy to visualize how closely the cells are packed. We can draw a box around the cell group and calculate the ratio of the cells divided by the total area. The ratio is  64%.

Model S 85 kwh cell density

Model S 85 kwh cell density

You can also see that the gaps between the cells where the cooling snake goes is larger than when there is no cooling snake. The ratio of the area of the cooling snake goes to the total area is 5% as shown below.

Ratio of area for cooling snake to total area is 5%.

Ratio of area for cooling snake to total area is 5%.

So you can see from this simple geometric analysis why there definitely was room in Tesla’s old pack design for improvement….and improvements in packaging have been achieved in the new P100D pack.

….but what has Tesla actually done to increase packing density? A conceptual design is proposed in this article.

Categories: Battery Tech, Tesla


Leave a Reply

13 Comments on "Tesla Model S 85 kwh Battery Pack Volumetric Efficiency Left Room For Improvement"

newest oldest most voted

The next model is 100D with 350 mile range, based on the same %range increase of P90D to P100D.

In other words, cell density has not improved in ~five years. That’s a bit disappointing.

Typical Tesla bashing B.S. post from this short-selling FUDster.

Tesla went from a top battery capacity level of 85 kWh to 90 kWh in the same space (volume), just by improvements in cell energy density, in only about 3 years.

Good luck achieving a similar improvement in either gasoline or liquid hydrogen. 😉

It’s still much less than the 5% per year that Tesla spoke about.

85 * 1.05^3 = 98.4 – and that is where we ought to have been, really. Add in the 11% from improving the pack architecture itself and the one additional year, and we’re at 115.

Other makers are improving much faster. That’s understandable given the low base they established, but regardless the gap is shrinking fast.

I hope and expect Tesla to make bigger steps when they get going with cell production at the gigafactory.

My god, you are such a moron! I am surprised you don’t type in all caps or make more grammatical errors.

If I recall correctly, simple geometry shows that if you pack circles together as tightly as possible, you get 71% coverage of a given area. So if the original packing of the cells occupies 64%, then there doesn’t seem to be much room for tighter packing, assuming: (a) That Tesla will continue to use cells which are round in cross-section (b) That there needs to be room at the sides of the cells for the cooling “snake” tube. Someone suggested in a recent article here at InsideEVs that cooling/heating could be achieved by using just a plate against the bottom of the cells. While I doubt that is practical, because I suspect it would lead to too much heat buildup at the top of the cells, perhaps it is possible to adequately cool/heat the cells if you place cooling plates at both the top and bottom of the cells. That would indeed allow tighter packing. The problem with putting a cooling plate on the top of the cells is that Tesla needs to put a fuse on the top of each one. That would certainly complicate things, but perhaps it’s still do-able by putting holes in the top cooling plate.… Read more »

You can’t even work it out in your head, but have to rely on recollection?!? I thought you were a programmer.

A circle with a radius of one unit just fits in a square with sides two units long. The area of the square is four, and the circle’s area is Pi, Pu-Pu.

3.14/4 = 0.785.

So your recollection is not to be trusted.


“You can’t even work it out in your head, but have to rely on recollection?!? I thought you were a programmer.”

I wonder if you have even the slightest idea how obnoxious and arrogant your posts often come across to others.

No, I can’t do complex math in my head, just like most people. If everybody could, then we wouldn’t need pocket calculators, now would we?

I very specifically started my post with “If I recall correctly” to alert the reader that I was relying on memory, so could be mistaken on that point. I would have thought that didn’t need to be explained.

“So your recollection is not to be trusted.”

Neither is your understanding of basic geometry. At least I knew that circle packing is properly (that is, most densely) done in a hexagonal arrangement, not the circle-inside-a-square arrangement you mistakenly cite.

So for all your presumed “superiority” in math, your analysis was no better than my memory.

No need to argue because both answers are wrong and also because geometry is supposed to be fun. Pushmi-Pullyu did not remember correctly but he didn’t do the square based packing mistake that Terawatt did. A round in a square will not give the correct percentage of cell area because it just isn’t a square packing but actually an hexagonal packing. One cell is surrounded by 6 other cells. In that case in a drawing the fundamental unit is a triangle of a 2 cell radius side and an area of 2r time 2r time sin 60º and divided by 2. The space in between is what is left when we subtract the 3 sixth circles included in that triangle. So that is pi time r square and divided by two. Taking for example r=1 gives a triangle of 1,73205 and the 3 sixth circles together give a total of 1,57079 so the maximum percentage of circle area is 1,57079 / 1,73205 = 0,90689 which was the actual right answer. Note that since the hexagonal packing has to finish on the sides of the pack, there will be some gaps left over that will slightly diminish the maximum packing in… Read more »

This was interesting. Maybe Tesla needs to make 2 sizes of cells.

“A compact binary circle packing with the most similarly sized circles possible. It is also the densest possible packing of discs with this size ratio (ratio of 0.6375559772 with packing fraction (area density) of 0.910683).”



I’d be interested in what you could do with ovals/ellipses.

Thanks for linking to the Wikipedia “Circle packing” article, Kdawg.

I was wrong about the geometry; the actual density of circles packed as tightly as possible (in a hexagonal pattern) is 90.69% of the available space. So if the density of packing in current Tesla battery packs is only 64%, then there certainly is a lot of room left, making significantly tighter packing a very real possibility.

But with tighter packing, there will be a need for better active cooling, since the amount of passive cooling will be diminished by packing the cells closer together.

Elliptical cells seems enticing since it would allow you to choose how much area to utilize and thus how much room to leave for cool-aid (hehe). But cylindrical cells aren’t common because they leave gaps between cells, but because they are easier/faster to manufacture in volume, which makes them cheaper. I am not sure, but I doubt the same advantage would apply to elliptical cells. With sufficient scale I think pouch and prismatic cells are really the best formats. I don’t know though why they are always so large, but it seems to me that’s a good thing if you can get away with it (keep them cool). The idealization of mathematical lines as cell boundaries obviously fails to account for the space “wasted” by the packaging of each individual cell, so the larger the cells the more of the chemicals there is room for. Tesla went with 18650 not because it was technically better as a cell format for EVs, but because it was the best and most cost-effective cells available. That was so thanks to production advantages, but nobody was making batteries for EVs in volume back then. The fact that Tesla is sticking to cylindrical cells and… Read more »

Well, there is cell architecture and then there is pack architecture. I think that pouch cells may actually be cheaper to produce than cylindrical cells, but the pouch cells need a lot more support; they need to be installed in a rigid framework before being assembled into groups to be installed in the pack. Cylindrical cells, with their metal shell, need no such framework. Assembling them into groups and packs is much simpler. However, due to the larger number of smaller cells needed to achieve the same capacity, I don’t know which approach is more labor-intensive. I would guess it’s easier to automate assembly of the hard-shelled cylindrical cells into packs, because they’d be less likely to be damaged by pick-and-place machines than the soft-sided pouch cells.