DIY Electric Car Forums banner

A look at the new Tesla cells

11915 Views 124 Replies 14 Participants Last post by  Kevin Sharpe
2170 cells replaces 18650 standard at the Gigafactory.

Doubtful they will be available to DIY community, though. The big boys will probably contract full production.
61 - 80 of 125 Posts
fwiw I do think that a clutchless 2 speed shifting is gonna be a thing at some point. With hi resolution control over motor speed or even gear tooth position (yes, possibly the return of sliding mesh, it is simpler mechanically and more efficient) it will be able to snick it up or down no problem. but it will be for lower powered commuters, or Bonneville.

I really don't see where it helps looking in the graph in post #30, you drop the ratio by 10% the rpm drops to a point with %10 more motor torque (no change in wheel torque), and the power is basically flat anyway., and the performance is more than adequate as is for a road car.
We made a pretty sweet clutchless two speed for the Nikola truck, but they never got the damn thing running so we never got to see it in action.
Im not a motor technology guy, but there was a comment on the Tesla forum suggesting that the profile of the Torque plot on that graph , ...in particular the flat constant value for the first several thousand rpm....suggests that the maximum torque available is being limited by the control systems (current limiting ?).
Implying that even more torque is potentially available via software and or harware alterations...One of which may well be a higher power capacity battery.
Major, ?..Brian ?...anyone,...any view on this ?
I'd have thought the flat top for the first few thousand is exactly what you'd expect from a PMSM motor see here esp. fig 8

Using this as a model predicts the S100DL performance quite accurately with published numbers, but as the model shows, more torque low in the rev range will not help as the limiting factor is tyre grip.
Text Line Font Parallel Technology


This model also lets you add an ideal two-speed box, but playing with that shows it doesn't seem to buy much on a road car, given the wide rev range of a Tesla motor, as dcb has said.
See less See more
I'd have thought the flat top for the first few thousand is exactly what you'd expect from a PMSM motor.......
Yes, but the question is what limits it to any specific peak torque level ?....
....current, control hardware, software, ...or motor design/hardware ?
So far , Tesla have raised the peak torque simply by increasing the amps, but is there still room for more that way ?
Im sure it would be possible to squeze a little more than 7C from those cells for a couple of seconds,....but could the controller and motor convert that into more torque ??
The performance model is very good, but not 100% accurate, as it seems to predict a best 0-60mph of 2.6 secs, whilst There are reliable reports of sub 2.3 sec times on record.....Probably down to tyre and track condition.
Note the 2.6s number is for 0-100kmh. 0-60mph is 2.5, which I believe is the 'standard' Ludicrous number, which the parameters of this vehicle profile aim to model. Of course, there is a limit to how accurate a simple model such as this, with only a few dozen parameters, can model an entire vehicle performance!

But the beauty of having a model is that you can change the model parameters to see what it would require to get a model S to 60 in 2.3s, and thus discover exactly what Tesla are doing.

Note that I've had to assume quite high friction values (mu=1.25) even to get to 2.5s number.

So assuming the tyres haven't changed and we are wheel slip limited in the 0-30 range, it would imply Tesla are raising the base speed somehow to extend the flat torque range higher to improve the 30-60 number.

Of course this another way of saying, they are increasing the motor peak max power. The 2.3s number could be achieved by upping the total motor power from 600kW to 700kW, but this pushes the max current from 1600A to ~2000A. This sounds astronomical, but given some of JB Straubels comments regarding their inverter, may not be so far fetched...
The power rating increases are largely enabled by better and better power electronics – transistors or IGBTs which is what we are using. As those improve, we can continue pushing up the current rating and continue increasing performance. If we can make the inverter just half a percent more efficient, that’s half a percent less battery pack that we have to put in the car or half a percent more range that the customer gets to have so it a very strong and virtuous feedback cycle with high efficiency.
And, this could be done within the 10C discharge envelope of the pack.

OTOH if the tyres can achieve mu=1.6, that would pretty much do it...
See less See more
I'd have thought the flat top for the first few thousand is exactly what you'd expect from a PMSM motor.......
.. I forgot to mention...
...you do know the Tesla is an Induction motor , not a PMSM. ?

( sory but i had a hitch with the forum and could not get a login for a while :confused::mad:)
I'd have thought the flat top for the first few thousand is exactly what you'd expect from a PMSM motor...
I agree - basically constant torque up to a threshold point, then constant or dropping power (so dropping torque... just a matter of how steeply) is normal. That is, normal for AC permanent magnet motors...

...you do know the Tesla is an Induction motor , not a PMSM. ?
... and that behaviour is also normal for induction motors (although the slip speed must be controlled appropriately to get peak torque for any given shaft speed), as major's graph shows in post #58 - just follow the peaks of the slip/torque curves for each speed.


And although real Tesla fans all probably knew this long ago, I learned from another discussion that the Model 3 is using a permanent magnet motor, so just "Tesla" is not going to be specific enough for these discussions.
Im not a motor technology guy, but there was a comment on the Tesla forum suggesting that the profile of the Torque plot on that graph , ...in particular the flat constant value for the first several thousand rpm....suggests that the maximum torque available is being limited by the control systems (current limiting ?).
Implying that even more torque is potentially available via software and or harware alterations...One of which may well be a higher power capacity battery.
Major, ?..Brian ?...anyone,...any view on this ?
The flat-torque region can be limited by current. I'm certainly no Tesla expert, but it seems like some of the performance improvements have been the result of more aggressive (less restrictive) current limiting as they have gained experience with what the battery and electronics can handle - no physical change in components involved. Of course there are limits to that, both in voltage and physically in both the motor and the battery.
For anyone that missed it, the "44160" cell format rumour has been neatly killed and explained as a miscommunication during a Tesla presentation where some tech guy stated that the Model 3 would use 4416 cells......
.....he litterally meant a quantity of 4,416 cells in the pack !:D
The cells are , of course, the new 21700 , 5.0 Ah capacity Tesla cell.
Which puts the M3 "long range" pack right at 83-85 kWh max capacity.
.. I forgot to mention...
...you do know the Tesla is an Induction motor , not a PMSM. ?

( sory but i had a hitch with the forum and could not get a login for a while :confused::mad:)

Not anymore :), but as Brian says, same applies to induction motor.
Yes , the M3 has gone to PM, but the Mod S that was being discussed will remain as n induction motor.....at least for now !
And some more details on the M3 pack
https://electrek.co/2017/08/24/tesla-model-3-exclusive-battery-pack-architecture/
Yes , the M3 has gone to PM, but the Mod S that was being discussed will remain as n induction motor.....at least for now !
And some more details on the M3 pack
https://electrek.co/2017/08/24/tesla-model-3-exclusive-battery-pack-architecture/
Yes saw the M3 pack details, included it in my M3 simulator

Had to assume 4.75Ah for the 2170 cells, as published by Samsung (cf Panasonic??)

But with that assumption, the short range version comes in ~10% under the 220 mile number, so presumably they are relying on a process improvement down the road.

Attachments

See less See more
Why use the Samsung spec ?
Tesla had to declare details of the pack to the EPA for approvals.
Those official documents declare the "long range" pack to be 230 Ah from a 46p pack....so 5.0 Ah per cell.
Presumeably Tesla are "selling" the pack as a lower capacity (4.4 Ah) to allow for a conservative operating voltage range (3.0-4.1 v ?). To ensure long working life and give some "headroom" for capacity loss compensation.
So , maybe you should use the "Tesla rating" of 4.4 Ah for the simulator ?
Couldn't find a definitive 20700 cell capacity (if you have please post) but from the pack details you posted the cell capacity must be ~4.75Ah.

(4.4Ah gives a pack density of 70kW and range of 288miles)
Tesla are highly guarded of their actual cell specs, so until someone gets hold of a few cells and tests them ( give it a few months) we have to rely on information such as that supplied by Tesla to the EPA for certification
https://electrek.co/2017/08/07/tesla-model-3-new-details-revealed/
230Ah, 350v etc
Added to the fact that we also know its a 4,416 cell count, in a 96s, 46p pack configuration .
The rest is just maths.
Infact Tesla are being even more cagey than normal with information regarding these latest cells, and of course they will continue to develop, refine, and improve the cell chemistry , so things will likely change again in the coming months and years.
Any reliable updates on the 21700 Tesla cell specs?
Any reliable updates on the 21700 Tesla cell specs?
As others have pointed out, things don't pencil out for the new Semi and R2 to be using today's battery technology. It seems like the batteries in these vehicles must have energy densities improved by a factor of ~ 2.
As others have pointed out, things don't pencil out for the new Semi and R2 to be using today's battery technology. It seems like the batteries in these vehicles must have energy densities improved by a factor of ~ 2.
Tesla could use some "real" good news. I would think even if they had improved energy density by 20%, it would be huge news. I am afraid we may have hit a hard ceiling. The 3.4Ah came out 3 or 4 years ago and there has not been any density improvement by Panasonic (or Tesla) since then.
As others have pointed out, things don't pencil out for the new Semi and R2 to be using today's battery technology.
Yeh, I get that but was more interested in the Model 3 battery and how we might use the salvage :D
The 3.4Ah came out 3 or 4 years ago and there has not been any density improvement by Panasonic (or Tesla) since then.
I thought the 21700 in the Model 3 was 'confirmed' as ~4.6Ah given we 'know' the long range version has 4416 cells and a capacity of either 74kWh (here) or 80kWh (here) :confused:

Attachments

See less See more
As others have pointed out, things don't pencil out for the new Semi and R2 to be using today's battery technology. It seems like the batteries in these vehicles must have energy densities improved by a factor of ~ 2.
There is so little hard data for the Semi that this seems like a difficult conclusion to support. A battery technology improvement over the Model S/X may be necessary, but a factor of two? :confused:

Tesla has been claiming a high content of Model 3 components in the Semi; with the gearboxes being different, and none of the cab, structure, steering, brakes, suspension, axles, wheels, or tires being shared, it seems necessary (and reasonable) that all of these components would all be straight from the Model 3:
  • battery modules (from this description, four in a Model 3 and about four dozen in the Semi)
  • inverters (one per motor)
  • motors (four of them in the Semi, one to two in a Model 3)
  • BMS
Yeh, I get that but was more interested in the Model 3 battery and how we might use the salvage :D
Since Model 3 modules built of the 2170 cells are presumably superior to the 18650's in current production models, my guess is that modules which you salvage from Model 3's after they go into actual production will be the same as what will be used in the Semi (and presumably Roadster 2). Or at least the cells in them will be the same.

The Semi's requirements may have driven the shift to larger modules, so that there is not such an impractically large number of them in the 1 MWh Semi pack... or perhaps it was just simplification for production cost efficiency. Either way, those big modules are going to be a challenge for packaging in many DIY projects.
See less See more
61 - 80 of 125 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top