Tesla's 18650s versus larger format automotive cells

[JRP3]

There's simply no way any new battery technology will not end up in laptops and Teslas.

There are a lot of different battery chemistries out there that aren't used in laptops for various reasons.

Nor does the Tesla's cooling system
figure more than incidentally in its cost at this point or any likely point in the future.

It's not just the production costs, it's the energy cost as well that reduces the charge efficiency.

The Volt's cooling system is quite elaborate
as well, and it uses large format batteries.

I'm comparing to the LEAF system which is quite simple.

As to
the matter of 6800 cells, just look at the number of cells in the Volt's paltry 16 kWhr battery pack.

288, if scaled to Tesla size it's less than 1000, 1/6 the number that Tesla uses, that is still significant.

 

[JRP3]

NMC doubles the density with little additional cost. So it will be cheaper when mass produced.

That's good to know. That means as theBike says it all comes down to volume, if the volume of large format cells increases enough they might become cost competitive with the laptop cells when considering the lower number of parts to assemble. Simple is usually better.

 

[qwk]

I'm comparing to the LEAF system which is quite simple.

Nissan is going to have to add a battery heater to the LEAF if they want happy customers in the states that have a winter. They are already working on that.

Every battery technology has drawbacks, there is no perfect choice.

 

[JRP3]

Yes I expect some active heating would be necessary but that can be pretty simple. Most DIY builds intended for cold weather have insulated packs with resistance elements. Preheating from the wall is usually all that's necessary.

 

[qwk]

Yes I expect some active heating would be necessary but that can be pretty simple. Most DIY builds intended for cold weather have insulated packs with resistance elements. Preheating from the wall is usually all that's necessary.

Since incorporating some sort of battery temp control system is necessary for an all around car, why not just take Tesla's approach and add water cooling/heating?

I realize not all battery packs need cooling, but they all need some sort of heater for the winter. This to me seems like the easiest choice since the motor and electronics already require cooling via antifreeze, so you already have all of the components there.

 

[JRP3]

Adding water passages is more complex and takes up more space than simple resistance elements. The systems would have to be separate since you'd want to be able to preheat the pack while plugged in but you don't want to preheat the motor.

 

[qwk]

Adding water passages is more complex and takes up more space than simple resistance elements. The systems would have to be separate since you'd want to be able to preheat the pack while plugged in but you don't want to preheat the motor.

The temp could be controlled within one system for everything.

It's going to be interesting to see how much range the Leaf gets in very cold climates in relation to the California, Florida, Arizona etc. cars. I'm sure there will still be a penalty.

 

[JRP3]

Certainly there will be a penalty. However, if the reduced range is still within the range you need, there is no effective limitation. Charging and discharging the battery also generates some internal heat on it's own, and it's somewhat self regulating. Colder temps mean higher internal resistance, which causes more heating, as it heats up resistance drops again. Remember, unless the Tesla pack is plugged in it's range is also being affected by temperature, you can't take energy from the pack to heat the pack and not lose range.

 

[Doug_G]

Remember, unless the Tesla pack is plugged in it's range is also being affected by temperature, you can't take energy from the pack to heat the pack and not lose range.

The Roadster doesn't heat the pack unless you're plugged in and not in Storage mode. Self-heating gets the pack up to the ideal temperature range in 10-15 minutes of normal operation, i.e driving, even if the car was stone cold when you started it. The only slight nuissance is that it disables the regen when the battery pack is below freezing.

 

[doug]

I'd say it takes about 68 times as long to assemble 6831 cells as a 100, even for a robot.

It depends on the robot. Depending on the physical design and program it might assemble many cells in parallel. Again, don't confuse redundancy for complexity. I'm sure they considered the cost/benefit in the assembly of the packs. And I'd bet they can finish making a pack before the rest of the car is done, so again, I don't see it as a manufacturing bottleneck at least.

Those cells provide for amazing packaging freedom and versatility. Tesla has managed to come up with a fantastically flat and dense pack for the Model S.

 

[EVNow]

And as if on cue ...

http://www.reuters.com/article/idUSTRE70A4QY20110111

Toyota sees Tesla EV battery cost at 1/3

"It could be as low as one-third of the cost of batteries being developed by car makers, because (laptop) batteries are produced in massive volumes"
...
Toyota has said the RAV4 EV would target drivers traveling longer distances, while its own EV would be suitable for short distances.
...
Even if the RAV4 EV passes Toyota's rigorous durability tests and can be offered at relatively low prices, Uchiyamada said Toyota still believed hybrids and plug-in hybrid vehicles had the best chance of mass proliferation

Though we don't know what actual cost they are talking about here - the EV battery costs are quoted from anywhere near $1,500 to $500/kwh. Consumer Li batteries are about $200/kwh.

 

[JRP3]

It depends on the robot. Depending on the physical design and program it might assemble many cells in parallel. Again, don't confuse redundancy for complexity. I'm sure they considered the cost/benefit in the assembly of the packs. And I'd bet they can finish making a pack before the rest of the car is done, so again, I don't see it as a manufacturing bottleneck at least.

I agree that the cars probably aren't sitting around waiting for the packs to be assembled. I see your point that once the robot is built and programmed it doesn't matter what it has to do as long as it's completed in time. Certainly their method can't be beat for cost vs. density at this point, but that may change as larger cells are built in volume.

 

[JRP3]

And as if on cue ...

http://www.reuters.com/article/idUSTRE70A4QY20110111



Though we don't know what actual cost they are talking about here - the EV battery costs are quoted from anywhere near $1,500 to $500/kwh. Consumer Li batteries are about $200/kwh.

I think Nissan has quoted $375/kwh, real volume should bring that even lower. Without the cooling demands of Tesla's it's a more efficient pack.

 

[Norbert]

I think Nissan has quoted $375/kwh, real volume should bring that even lower. Without the cooling demands of Tesla's it's a more efficient pack.

Although Tesla is getting a custom battery from Panasonic, it is also reported that using the 18650 format will give it the advantage of mass-production cost benefits. Why shouldn't Tesla be able to get quite close to the $200/kwh figure? Would you know how to compare the effects of cooling/heating demands on cost of either side?

 

[Eberhard]

as i understand, only with the small form factor of the 18650 cells, the temperature controlled can be done more easily by liquid, compared the bigger cells, because of the heat which has to move from inside of the cells to the surface. Smaller cells means short ways and i do not think, that the cells material-composit has a high thermal conductivity.
second problem with big cells: if is there a little problem within the components of the big cell - the whole cells fail - big loss. if one of a hundred of smalls cells fail, 99 are still fine.

 

[JRP3]

The higher number of small cells also means more possible failure points, the cells themselves and extra connections. The volatile nature of the LiCo chemistry forces the use of aggressive liquid cooling where other chemistry used in large format cells does not.

 

[Eberhard]

why is water-glycol aggressive? there is no acid or similar inside. the liquid even prevents corrosion.

 

[kgb]

as i understand, only with the small form factor of the 18650 cells, the temperature controlled can be done more easily by liquid, compared the bigger cells, because of the heat which has to move from inside of the cells to the surface. Smaller cells means short ways and i do not think, that the cells material-composit has a high thermal conductivity.

My Chemical Engineering background left me thinking that it had more to do with surface area. Heat Flux (transfer) is increased by more surface area. If you take two batteries with the same amount of heat energy, but one is broken up into 100 smaller parts. The one that is broken into parts has more surface area over which to transfer its heat. So the small form factor will be able to transfer heat more quickly and therefore remain at a constant temperature. (So it is not the distance that the heat has to travel, but rather the amount of surface area over which it has to transfer)

One of the down-sides to the small form factor that I see is something I like to call the "Hamburger bun phenomenon." Would you rather have a single quarter-pound burger in a single bun, or a hundred small sliders making up the same weight? The problem with small form factor is too much bun! Or in the case of batteries, you waste too many materials on packaging. It would be more efficient if someday the small form factor batteries did not come individually wrapped.

 

[JRP3]

why is water-glycol aggressive? there is no acid or similar inside. the liquid even prevents corrosion.

I didn't mean aggressive as in corrosive, I mean aggressive as in the LiCo cells tend to explode if they get too hot so you need more aggressive, or active, cooling, which usually means liquid cooling, to remove excess heat. LiFePO4 and LiMn cells aren't as heat sensitive so can get by with air cooling.

 

[JRP3]

One of the down-sides to the small form factor that I see is something I like to call the "Hamburger bun phenomenon." Would you rather have a single quarter-pound burger in a single bun, or a hundred small sliders making up the same weight? The problem with small form factor is too much bun! Or in the case of batteries, you waste too many materials on packaging.

Right but even though there is wasted space and material with the small commodity cells the LiCo chemistry has higher density than LiFePO4 and LiMn so the Tesla pack still has better density. Theoretically I guess if you could make a LiCo cell in larger prismatic format you'd get even better density, but because of the aforementioned cooling issues with LiCo it may not be possible without some sort of chemistry breakthrough. Also C rates usually drop when going from cylindrical to prismatic construction so the LiCo prismatics might end up with lower C rates.