Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Battery Format for Model 3 - 18650 commodity cells or large format batteries

What cell format will Tesla use for the Model 3?


  • Total voters
    6
  • Poll closed .
This site may earn commission on affiliate links.

Reeler

Decade of Pure EV Driving
Oct 14, 2015
1,766
1,318
Denver, CO
Thus far, the Roadster, Model S and Model X have used commodity 18650 format cells. Other manufacturers with their own factories have instead gone for large format cells. Tesla's decision was pragmatic in that economies of scale kept costs low, but that much wiring is complex and has other issues.

Now that Tesla will have their own battery factory, I don't personally see any reason not to follow what everyone else has done--large format batteries. I predict that this will happen with the Model 3 release and completion of the Gigafactory. What do you think?
 
  • Funny
Reactions: SW2Fiddler
This is really interesting stuff. Are we sure they're still going to use a similarly sized cylindrical cell, in general ? The Lithium Ion packs in my cellphones are not cylinders. The one in my old Motorola Razr (the rebooted one) was even highly soft and bendable.

What about the GM cells, are they cylindrical inside the "packs" ?

I watched a pile of Elon Musk interviews this weekend and one "a-ha" moment he shared was the team asking what the "ingredients" were in a battery, and what the commodity market costs were, to start with a base input cost. And then later, they.... :cool:
 
  • Love
Reactions: Snerruc
This is really interesting stuff. Are we sure they're still going to use a similarly sized cylindrical cell, in general ? The Lithium Ion packs in my cellphones are not cylinders. The one in my old Motorola Razr (the rebooted one) was even highly soft and bendable.

What about the GM cells, are they cylindrical inside the "packs" ?

I watched a pile of Elon Musk interviews this weekend and one "a-ha" moment he shared was the team asking what the "ingredients" were in a battery, and what the commodity market costs were, to start with a base input cost. And then later, they.... :cool:

GM uses LG Chem Prismatic (rectangular) cells for all of their EVs to date - a smaller number of much larger, much more powerful cells. - the original Volt has only 288 cells (a 96 group series string of parallel triplets,) and the new Volt and Spark are actually even fewer despite the larger capacity - 192 cells arranged in pairs.

This approach makes manufacturing simpler, but the failure of a single cell has significant consequences for the car as a whole, so the battery supplier has to have extreme quality control in place (and mostly seems to - the Volt has been putting up a pretty impressive reliability history.)

Tesla took a RAID approach instead - they have 74 cells in parallel in each of their 96 series groups (6 groups in each of the 16 modules,) each individually fused and isolated in intumescent goo (in the original 85 kWh packs, anyway - the others are similar in concept.) This increases the odds of failure, but it allows a cell to fail without having significant impact on the car as a whole - the fuse pops, but as long as the coolant and goo are enough to stop a thermal runaway, the event ends there - and the car lost 1.4% of its range, with the odds strongly against the next failure costing any additional range.

It'll be interesting to see which approach ends up dominent in the industry...
Walter
 
GM uses LG Chem Prismatic (rectangular) cells for all of their EVs to date - a smaller number of much larger, much more powerful cells. - the original Volt has only 288 cells (a 96 group series string of parallel triplets,) and the new Volt and Spark are actually even fewer despite the larger capacity - 192 cells arranged in pairs.

This approach makes manufacturing simpler, but the failure of a single cell has significant consequences for the car as a whole, so the battery supplier has to have extreme quality control in place (and mostly seems to - the Volt has been putting up a pretty impressive reliability history.)

Tesla took a RAID approach instead - they have 74 cells in parallel in each of their 96 series groups (6 groups in each of the 16 modules,) each individually fused and isolated in intumescent goo (in the original 85 kWh packs, anyway - the others are similar in concept.) This increases the odds of failure, but it allows a cell to fail without having significant impact on the car as a whole - the fuse pops, but as long as the coolant and goo are enough to stop a thermal runaway, the event ends there - and the car lost 1.4% of its range, with the odds strongly against the next failure costing any additional range.

It'll be interesting to see which approach ends up dominent in the industry...
Walter
Its too bad that GM can't sell great cars with those great batteries.
 
What is it about the "commodity" cells that you think is better? They are known to have lower energy density and a higher cost/kwH than the existing 18650.
The 2170 batteries will be 40% better on both metrics than the 18650s.

Pouch cells are used by legacy automakers in compliance cars, period.
Lucid and Faraday are not using them, because they are not designing compliance cars.
 
  • Like
Reactions: Red Sage
What is it about the "commodity" cells that you think is better? They are known to have lower energy density and a higher cost/kwH than the existing 18650.
The 2170 batteries will be 40% better on both metrics than the 18650s.

Pouch cells are used by legacy automakers in compliance cars, period.
Lucid and Faraday are not using them, because they are not designing compliance cars.
Me? I was being ficesous.
 
  • Like
Reactions: Red Sage
The "commodity" factor was useful when Tesla was a tiny/small/medium consumer of Li Ion cells.
In 2017 Tesla will consume upwards of 10 GWh worth of cells. That's over a billion cells in a year.
And of course they have their own factory not far from beginning their own cell manufacturing.
And whatever Tesla does, a few EV manufacturers will follow. Tesla is the leader.
The game has changed.
The only disadvantage of 2170 over 18650 is you get less total voltage (with the whole pack in series) cause large packs pack higher amps-hour but same voltage. Hence another doubt is will Tesla use 2170 cells for 75kWh packs ? If they moved 100kWh packs to 2170, would it reduce acceleration/top speed ?
 
The only disadvantage of 2170 over 18650 is you get less total voltage (with the whole pack in series) cause large packs pack higher amps-hour but same voltage.

No that's not true. Series connections make voltage additive, they can just connect the same number of the larger 2170 cells in series to have the same voltage. The larger cells means they can connect fewer cells in parallel to get the same total amp hour capacity, volts x amp hour = kWh.
 
  • Helpful
  • Like
Reactions: YBT and MP3Mike
JRP3, the assumption is if you migrate from 18650 to 2170 keeping the same total capacity, you would get less cells hence less voltage. If the current 100kWh pack is using 18650, migrating it to 2170 would generate less voltage.
Since Tesla increased performance everytime they increase the MS/MX pack, I would assume going to 2170 would rollback unless they make it a 125kWh pack or something. But I'm really asking.
 
JRP3, the assumption is if you migrate from 18650 to 2170 keeping the same total capacity, you would get less cells hence less voltage. If the current 100kWh pack is using 18650, migrating it to 2170 would generate less voltage.
Since Tesla increased performance everytime they increase the MS/MX pack, I would assume going to 2170 would rollback unless they make it a 125kWh pack or something. But I'm really asking.

You wouldn't have to get less voltage. You switch some cells from being in parallel to being in series to get to the voltage you want/need. (And that wouldn't change your overall kWh of the pack, it just puts more stress on the individual cells.)

In theory you could put 96 cells in series and give someone a ~400 volt pack. Of course it wouldn't last very long...
 
  • Like
Reactions: Red Sage
JRP3, the assumption is if you migrate from 18650 to 2170 keeping the same total capacity, you would get less cells hence less voltage. If the current 100kWh pack is using 18650, migrating it to 2170 would generate less voltage.
Since Tesla increased performance everytime they increase the MS/MX pack, I would assume going to 2170 would rollback unless they make it a 125kWh pack or something. But I'm really asking.

You seem to be thinking that every cell in a Tesla pack is connected in series. If that were true, you'd have a point.

However, all the recent packs except the 100 have had groups of 74 cells in parallel in each of the six sets hooked up in series in each of the 14 or 16 modules connected in series (thus a total of 96 groups of 74 cells in the 85 and 95, and 84 groups in series of 74 cells in parallel in the 70 and 75.)

I don't think we have an exact cell count for the 100 yet, but it has to have upwards of 80 cells in parallel in each module from the limited guidance we've been given.

Replacing the 18650 cells with higher capacity 2170 cells would allow you to reduce the number in parallel while keeping the same number of series strings and the same total voltage and capacity. (Or they could put the same number of cells in and get a bigger battery - assuming the new cooling system can make that fit.)
 
Saghost is correct. Simple example, I can make a 12V battery with 100 amp hour capacity for a total of 1.2kWh with different cell configurations and sizes:
10 parallel 3 volt 10 amp hour cells create a 3 volt 100 amp hour group, 4 of those groups in series makes a 12V 100ah battery
or
5 parallel 3 volt 20 amp hour cells also create a 3 volt 100 amp hour group and 4 of those groups in series makes a 12V 100ah battery. Both batteries have the same 1.2kWh capacity and potential peak power output.