Hello there all. I work as a CAD engineer in the UK. I've carried an odd little exercise for myself using screenshots of the video Tesla put on Twitter.
Due to the amount of corroborating information on that picture, the fact that the battery modules are quite obvious as containing one cell height, the Tesla Model 3's battery cell details, and of wk057's work & product information and other details (listed below), I thought it might not be actually be that difficult to determine the geometric maxima.
Obviously not trying to pursue accuracy here, but this is the next best thing.
How many 21x70mm cells could be hexagonally arranged in there? I inserted cropped screenshots of the Semi at different stages of the video -please excuse the chair- and scaling them to satisfy the (mostly much larger) knowable dimensions> such as:
-
An assumptive width of 2.55m. I think this is the average Euro limit, US is 2.6m if i recall correctly. But obviously not more so than the standard container of 2.44m (8 ft). It is obvious the vehicle smoothly lines up with it's trailer section more so than a traditional semi.
-
Matching up axle pitches from both planes of screenshots.
-
Battery box cover Z dimension lineup The modules are obviously shorter, see centre gap, mind for perspective-induced fuzzy edge on front one)
-
Tyre diameter (which appears to correspond well 42" if a slight bit large - suggests 2.5 metre width?) Left sketch overlaps alone due to slight perspective apparent.
-
Elon Musk's height. Go ahead, laugh... Best person to make comparisons with due to his known height of 188cm as a celebrity and seen clearly specifically in the unveil video, stepping out of the vehicle with his height almost perfecting matching the body/window boundary on the truck, and from some camera distance as well. Also, tyre and number plate can be compared.
Door corresponds to somewhat 170cm (>6ft) height from cab floor, please let me know if this sounds incorrect.
Battery modules, as described by others earlier in this thread, in the Tesla Twitter video, appear to be composed of four (bent metal?) "module carriers" with 3 modules each inside of them. I have judged the outside edges of the battery compartments to be 80cm +/- 1 cm along the lengthwards direction (minor floor plane) and around 220cm in the widthwards direction (major floor plane) due to the rectilinear C- shaped "hem" visible in the red highlights of the battery compartment from above in the Twitter video.
To make it clear, I am not aiming for technical accuracy, but rather the absolute geometric limits of what number of 2170s could possibly fit in the space of the modules.
Using this reprehensible methodology and creating a battery module footprint of 80x222cm and assuming a hexagonal pattern of 21mm diameter circles with a 23mm pitch (2mm spacing representing aluminium cooling fins... please let me know more about this) I have found that the largest number of circles is basically
3,726 to 3,774 depending on the direction of orientation.
For those who are concerned that my sketching may not be accurate, due to packing inefficiencies, it turns out that changing the floor dimensions of the module by as much as +/- 1.5 cm actually makes no or little difference, whereas changing the orientation, or pitch by +/- 1mm, will have fairly dramatic results in comparison. Based on this property I propose that the maximum possible cell capacity is predictable to determine even with >5 pixels in perceptive error.
See the screenshots below.
Judging by earlier information elsewhere on the Tesla Model 3, with the Standard Range having 2,976 cells and the Long Range having 4,416 cells with a maximum of 80,500 Wh, we can be sure that the Wh per 2170 cell is 18.23~18.24. Dividing that number by 1.46, given that the 2170 is said to have 46% more volume than the 18650, gives the result of roughly 12.5Wh, which is basically what one gets if tapping in numbers from wk057's breakdown of the Tesla 100 kWh battery pack.
That story over, we can now determine the absolute capacity of one of these modules:
3,726 cells x 18.23 Wh => 67,924.98 Wh
3,774 cells x 18.23 Wh => 68,800.02 Wh
So let's call them 67.9kWh and 68.8kWh respectively.
The image below is a revealing screenshot from Tesla's Twitter video, if you haven't seen it, please do so.
It's a bit dark, but you can reasonably tell several things:
1. There are
four "module holders" or possibly welded frames.
2 There are
three modules or module "layers" in each "module holder" making up a total of 12.
3. The "hems" or "frame edges" of the "module holders" continue vertically downwards and do not appear around the central "bends"
4. The divisions at the top produce a number that is too numerous to be separate modules.
5. The modules are narrow enough to have one height of cell (Upon sketching they appear to be exactly the height of S and 3 modules...)
6. There is a pattern on top that goes upwards judging by the light pattern on the motors, and is therefore probably related to module holder engineering, rather than a supposed hidden number of smaller modules.
Therefore, we can take the numbers of kWhs earlier, and multiply them by 12 to obtain the absolute capacity of the total.
The total capacity is therefore:
67.9 kWh x 12 = 814.8 kWh
68.8 kWh x 12 = 825.6 kWh
I think this strongly suggests a usable of
800kWh, with a megawatt hour being impossible with the same chemistry as the Model 3.
I suspect that the real answer is in between, and that the "module holders", are designed to have a usable capacity of 200kWh to be multiplied as necessary, given the non enclosing appearance of the body underframe, for different regimes (but not for swapping).
Since each one has what I safely assume to be three modules, those would need a usable capacity of 66.7kWh per module.
What about efficiency and charging?
Quite obviously, it's going to be
1.6 kWh per mile or less. If we treat any "500 mile range" with the same degree of trustworthiness as the ranges of the Tesla Model S, then the capacity could still be 750 kWh, but no less as 1.5 is stretching things as it is.
The Cummins AEOS, Daimler FUSO both appear to have an efficiency close to 1.4 for lower average speeds and weights, while the Freightliner Supertruck seems to be equivalent to 1.7-1.8 when I account for the diesel being exactly 50% efficient. Given that the Tesla has better aero, retracting skirts and regenerative power downhill, therefore 1.6 makes sense to me.
...That leaves the "400 miles in 30 minutes" figure.
I'm going to make a controversial assumption. I think what it really means is "
to 80% of range from a minimum reserve that the driver can see in his tank/battery", because, let's face it, almost no-one refuels or charges from flat, and certainly not from a battery vehicle under a great deal of expectation and skepticism.
Size of this "minimum reserve in mind is probably anything from 10% upwards. Say it's 10% of range. So for a usable of 800 kWh, 10% to 80% is going to be from 80kWh to 640kWh, a difference of 480kWh, which in terms of charge-rate considerations, is 59% or 58% of 815 or 825kWh absolute capacity, and would require only 960 kW average, with a higher (tapering) input to begin with, I guess about 1.2-1.4 MW, hence the "Megacharger".
So a charge rate of 30 minutes here, is not anywhere further outside the bounds of current battery technology performance.
Let me know if I have anything disastrously wrong here, it's not going to be completely accurate, but it is a potential maximum and I am having quite a bit difficulty seeing how it could have an absolute capacity more than 840 kWh.