Breakfast Bytes Blogs

On September 9th, Tesla held a "Battery Day." Elon Musk also gave an update on the status of the models, the factories, the software, and so on. I embedded a video of the whole day at the end of this post. I think the whole presentation is worth a watch. If you are just interested in the battery part then skip to 34 minutes.

One comment from the earlier part of the presentation is about scaling manufacturing. This is also applicable to batteries (and semiconductor fabs, for that matter):

As you've heard me say before, the hardest thing is scaling production, especially of a new technology. It's insanely difficult. Making a prototype is relatively easy. The real achievement of Tesla wasn't making exciting prototypes. Tesla was the first company in about a century in the US to reach volume production and be sustainably profitable. This has not happened in 100 years. That's the actual super-hard part.

The difficulty of designing the machine that makes the machine is vastly harder than the machine itself. Making a model 3 or a model Y prototype is really quite trivial compared to designing the factory that makes it. Especially if it is new technology and you want to use new manufacturing methods. It's at least ten to a hundred times harder to do the factory than the prototype. That's why you see a lot of companies out there who will bring out a prototype but they can't get it over the hump for manufacturing, because manufacturing is the hardest thing by far. 

When Tesla started manufacturing in Fremont, to everyone's surprise (mine, anyway) they seemed to have difficulties getting cars out of the door. I assumed manufacturing cars was a solved problem, since there are automotive factories all over the world that manage it, and thousands of professional engineers who know how automotive factories work. But it turned out to be really hard, not least because nobody had really manufactured electric cars at scale before. Here's another thing that surprised me, in a different industry. Walmart is known for having some of the best logistics of anybody. How hard could it be for them to create a successful online business with fulfillment using their logistics expertise? It turned out that Amazon learned logistics faster than Walmart learned e-commerce.

Batteries are important in all sorts of markets, such as smartphones and laptops, but the most important of all are electric vehicles (EVs) and storage for intermittent renewables like solar and wind. As I'm sure you know, there is a shortage of automotive semiconductors. See my post Automotive Supply Chains for more details. In the Q&A at the end of battery day, Elon was asked what were the limiting factors on Tesla's production:

In the short term, semiconductors. In the long term, batteries.

Battery Technoiogy

That's how important batteries are. I'm sure if they are the limiting factor for Tesla, it is going to be the same for all manufacturers of electric vehicles, and that is basically all manufacturers. The big challenge with automotive batteries used to be to get the mileage per charge up to an acceptable level so that EVs were more than a toy for rich people, like the original Tesla Roadster. That has pretty much been done. The big challenge going forward is to get the cost of batteries down so that an EV costs about the same as an internal combustion engine (ICE) vehicle. That was the main focus of battery day.

Tesla uses two kinds of batteries today. LFP stands for Lithium Iron (Fe) Phosphate and is the main battery technology used in Chinese EVs and in most Teslas manufactured in China. It has some advantages over nickel based batteries, not least that iron is cheaper and more abundant than nickel. But the energy density is not quite as high. To some extent that is compensated by the fact that they can be charged faster. One big disadvantage is that they are not manufactured at scale in the US. Tesla is apparently importing batteries from China for its low-end models, at least partly because it is out of capacity for its own batteries in the US.

One way to get the cost of batteries down is to build bigger ones. Battery sizes have a number xxyy which means a diameter of xx mm an a length of yy mm. The three main sizes ar:

• 1865 (slightly larger than an AA battery) used in the Models S and X. Outside of Tesla this is called 18650 but since nobody could explain the zero, Tesla decided to delete it for consistency.
• 2170 used in models 3 and Y.
• the new 4680, the main focus of the battery day presentation.

According to Panasonic, who manufacture the battery the 4680 will store five times as much energy and cost half as much to build as the 2170 lithium-ion cells Tesla currently uses in the Model 3 and Y, which are also provided by Panasonic.

Battery Construction

A quick primer on batteries. They consist of an anode, a cathode, and a separator. These are all thin sheets (think of aluminum foil) that are then tightly rolled up into a cylinder. At some point during the manufacture, the rolling has to stop so that tabs can be added to form electrical connections to the anode and cathode (see in the picture above). This is a huge cost, not so much because the tabs themselves are costly but because it significantly slows down manufacture. Another issue is that with a single tab, all the electric current has to run all around the whole spiral to get to the tab.

The new 4680 are "tabless". The connections are made directly to the top and bottom of the roll. through "dozens of connections". As a result, they can be manufactured much faster and cost a lot less. This tabless approach also avoids thermal problems, which is one of the main challenges of making batteries bigger, since the electrical paths are 5 times shorter (not having to go around the spiral).

At scale, the requirement is to build a lot more battery (gigawatt-hours) per factory, otherwise Tesla's needs will be measured in hundreds of factories.

It's not just money, if we had a trillion dollars it's not that easy. You need to have a lot of people, build the machines, build the machines that make the machines so it's incredibly important to have that effort yield the most number of batteries.

We don't yet have a truly affordable car, but we will in the future, but in order to do that we've got to get the cost of batteries down.

The result is that the new battery holds 5X the energy, delivers 6X the power, and just from this change in form factor, there is a nice-to-have range increase of 16%. Production is ramping in a 10GWh pilot plant "just around the corner" although it will take about a year to ramp up to full capacity. The actual production plants will be more like 200GWh, so twenty times the size. The plans for Tesla internal battery production (on top of the batteries that it buys from Panasonic, LG, and others) is for 100GWh in 2022, going to 3TWh by 2030 or earlier.

Those are not the only changes. If you watch the whole video you'll see that they get an additional 20% range using silicon anodes (which is amazing—I know that nobody has been able to use silicon before since it expands so dramatically when the battery is charged). 

We have a three-tiered approach to batteries, starting with Iron (Fe) which is for medium range, nickel-manganese as medium plus, and pure nickel for long-range applications like the cyber-truck. For a semi-truck it is very important to have high energy density to get long range. Iron is good, not as good as nickel which is 40-50% better, but still pretty good. Great for stationary storage and for medium-range applications where energy density is not paramount.

Long term, batteries will be recycled (there is so much higher density of things like lithium or nickel in a battery versus ore). As Elon said:

Long term, once we reach a steady state, new batteries will come from old batteries

Another change is making the battery pack structural, so that it is an integral part of the chassis. This has all sorts of advantages in terms of stiffness, 10% weight reduction, moving weight to the center of the vehicle, 370 few parts...and another 14% increase in range.

The Summary of Battery Day

All of these improvements (and some more I didn't mention) provide:

• 54% increase in range
• 56% decrease in $/KWh
• 69% reduction in cost per GWh to build the factories

Other Manufacturers

Obviously, Tesla didn't talk about other manufacturers at their battery day, but it is worth pointing out a few things.  First, other manufacturers do not use cylindrical cells like Tesla does. They use rectangular cells. I don't know what the tradeoffs are between the two approaches.

Another development is solid-state batteries. The separation layer in most batteries is a liquid electrolyte. This has all sorts of disadvantages, in particular, if the battery leaks then the separator stops being effective and is subject to fire. So the dream is a solid-state battery with no liquid. Solid-state batteries can be charged faster (of the order of 10 minutes, so no much longer than filling the tank of an ICE vehicle with gasoline). They have good range, good lifetime in terms of number of charge-discharge cycles, and don't require any attention to heating and cooling. But solid-state batteries have been in development for 40 years or more, Toyota, which does a lot of its own battery development and manufacturing internally, has a major investment in solid-state batteries. Toyota has said that they will use these batteries initially in hybrids, where absolute maximum performance is less of an issue (since there is an ICE too).

There are two challenges with solid-state batteries. One is how good the basic technology is when the battery packs are charged and recharged many times over many years, and whether they develop issues. And the other is scaling production to TWh levels, which is non-trivial as Elon Musk discussed at the start of battery day and I quoted near the start of this post. But many manufacturers, not just Toyota, are working on the technology. Nissan and Volkswagen, to name two.

Watch Tesla Battery Day

By the way, it is often hard to write about Tesla without coming across as Tesla fanboy. For the record, I don't own a Tesla, and I've never even driven one. I have a Mini convertible, but mine has sensible tail lights, not ones that look like bits of the Union Jack (British) flag. If you want a laugh, take a look at Congratulations, Mini, You Made The Stupidest Turn Signals Ever.

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.

.