I'm sure you guys are bored with this topic over & over again but I found myself doing all sorts of calculations about the new 2170 cells in the Model S this morning and wanted to share my thoughts with you guys. First, the cell. - 2170 will have 45% larger volume than the 18650. - Chemistry has been improving at an average rate of 7% a year. Proof? Besides statements from Tesla execs, 2008 Roadster cells were 2100mAh, 2012 Model S cells were "rounded up" to 3100mAh and 2015 cells were 3300mAh. So a 2017-2018 18650 should have been ~3700mAh. Add to that the larger volume and we have a 5300mAh, 19,5Wh, our saviour 2170 cell. Next, the fitting. Fitting circles in a rectangle is a real challenge depending on configuration. Thanks to this website we can enter our specs and get the physical limits of fitting the cells in our modules. We can find out about the size of a Tesla Model S module here at this eBay listing. When we plug the data into the calculator, being generous and entering 18,5mm for cell diameter and 0,8mm for cells being apart(listing states 0,8mm tubing, that might have decreased with the 100 pack but we'll ignore it for now) , we get 552 cells possible to fit in a triangular pattern. Next, my P100D cell count estimation. I am "close to sure" on the new packs having 87 cells per group (13 more than before). How do numbers check out? - 87p96s = 8352 cells. We know Tesla uses the same cells and pack voltage has to be the same (so we're sure on the 96s part). With the 2015 3300mAh, 12,2Wh cells we get ~101,8kWh nominal capacity. - Pack's Ah rating is 287Ah. Multiplying that with Tesla's 6,14 c rate used for ludicrous accelaration we get 1762 Amps peak draw equating to about 540kW of power. Very consistent with findings, I'm convinced. Getting back to the circle fitting problem we see that Tesla has fitted 522 cells per module into a physically limited 552 cell area. They have stated this is the theoretical limit for now too. If it weren't for any chips or housing or glue or connectors or chips it could get 552 cells but it is getting 522 cells. So a utilization rate of 94,5%! Impressive. Off to 2170 calculations. I'll be generous again as I was with 2170 capacity estimations and assume a less efficient, 90% utilization on the pack as a larger cell could need more cooling although unlikely. I'm doing this so the results will be 'at least' that amount. So our calculator shows a possible 420, 2170 cells fitted into the same old Model S module. - 90% utilization would give 380 cells per module. - 380 * 19,5 = 7,4kWh per module. - Keeping the 16 module setup is a Model S P120D with 380 mile range and probably even faster accelaration. And remember this calculation is very modest on developments. So I wouldn't be surprised to even see a 130kWh pack in 2019. (A nominal capacity rating of 20Wh and 93-94% util. rate for the pack would easily make this happen) With the new cells even the 14 module current 75 setup is 100 kwh. So my guess for the Tesla battery lineup in early 2019; Model S & X; - 100 (62p, 84s) - 120kWh. (62p, 96s) Model 3; - 55kWh (37p, 84s) (~54kWh usable) - 70kWh. (37p, 96s) (~66kWh usable) Although I'm fiddling around with the idea that the Model S and Model 3 should be sharing modules as long as production streamlining and costs are involved but I can't come up with the math. They have to keep the 96s part same for voltage and change the parallel part also there's the size difference. Unless they drastically redesign their packs and use much more modules and reduce in-pack series configuration, they should be different modules unlike VW, Faraday Future's modular setups. Anyway, my .2c. Will be interesting to watch. Even though there are diminishing returns above the 100kWh battery I think Tesla will go for it because 1) they have to keep the high top margins and 2) why not? What do you guys think?