Welcome back everyone, I'm Jordan Geisigee and this is The Limiting Factor. In the last video I showed the tear down of the Generation 2 4680, or Cyber Cell, that I received from Monroe and Associates. One thing I didn't show in that video was that the battery cell had a voltage of less than half a volt, which meant that it was likely damaged. Due to that damage, doing electronic tests to calculate its energy density would have provided inaccurate results, so we didn't proceed with those tests. That won't stop us of course from doing materials testing on the cell, and I'll release all that data in the next video, but it does mean that we have to look to another source for the energy density of the 4680.
欢迎大家回来,我是Jordan Geisigee,这里是"The Limiting Factor"频道。在上一个视频中,我展示了Monroe and Associates给我的第二代4680电池(也叫Cyber Cell)的拆解过程。但有件事我没在视频中提到,那就是这个电池的电压小于半伏,这意味着它很可能已经损坏。由于这个损坏,如果我们进行电子测试来计算其能量密度,结果会不准确,所以我们没有进行这些测试。当然,这并不会阻止我们对这个电池进行材料测试,我会在下一个视频中发布所有测试数据。不过,这也意味着我们需要从其他来源获取4680电池的能量密度数据。
Luckily, Monroe and Associates shared their own results for the energy density of the Cyber Cell on YouTube, which was 272Wh per kilogram. I was able to contact the person who ran the energy density tests and confirmed that the testing procedure that was used was nearly identical to the procedure that UC San Diego used for the Generation 1 4680 cell, which returned an energy density of 244Wh per kilogram. That means the Generation 2 4680 has an energy density that's 11.5% higher.
The question is, how does that compare to other battery cells on the market? I've decided to devote an entire video to the topic because I found that, as you'll see, the EPA data that much of the Tesla community has been relying on for energy density appears to be highly inconsistent and often inaccurate. After correcting for that, the conclusion I've come to is that, in terms of energy density, the Cyber Cell is significantly better than the average 2170 battery cell, and highly competitive with the cells that Tesla buys from Panasonic, even with generous assumptions. That's not only at the cell level, but also the pack level. Let's get into the video and I'll walk you through exactly how I came to that conclusion.
Before we begin, a special thanks to my Patreon supporters, YouTube members and Twitter subscribers, as well as RebellionAir.com. They specialize in helping investors manage concentrated positions. RebellionAir can help with covered calls, risk management, and creating a money master plan from your financial first principles.
Much of the data that I've used in the past to benchmark the energy density of Tesla's battery cells was derived from certificates that are publicly available on the EPA website. I'd never questioned that data in the past and assumed it was well vetted, but for my analysis of the Cyber Cell, I decided to double check it to make sure my benchmarks were rock solid.
That analysis turned up some surprises. To understand that analysis, let's start by doing a quick walkthrough of how I source the EPA data so that if you'd like, you can have a look for yourself and cross-check my analysis. Battery pack data for EVs is contained in the EPA's Transportation and Air Quality Document Index system, or DIS. The compliance document type that it's contained in is the Certificate Summary information, and the manufacturer is Tesla. After hitting the search button, it returns the documents that we're looking for. If we open any of these documents, there are two places where we can find information about the energy capacity of Tesla's battery packs.
The first is by looking at the end state of charge information, which provides the exact pack energy in kilowatt hours down to the third decimal based on discharge tests. The second, which I became aware of since my last tear down video thanks to Troy Tesla Nikon X, involves multiplying the total voltage of battery packs by the battery energy capacity. That returns an exact figure in watt hours. Then it's just a matter of dividing by 1000 to get kilowatt hours. This page also provides the battery-specific energy, which is in watt hours per kilogram at the pack level. With the EPA data in hand, it's just a matter of combining it with information from tear downs and other public information, such as cell weights and pack weights, to work out most of the key stats of Tesla's battery packs.
第一种方法是通过查看充电状态的最终信息,该信息提供了基于放电测试结果精确到小数点后三位的电池组能量(以千瓦时为单位)。第二种方法是通过观看我的上一期拆解视频后得知的,感谢Troy Tesla Nikon X,他讲解了将电池组的总电压乘以电池的能量容量。这会得到一个精确的瓦时数值,然后只需要除以1000就能得到千瓦时。本页面还提供了电池的比能量,即每公斤电池组的瓦时数。手中有了EPA数据,就可以将这些数据与拆解视频和其它公开信息(例如电池单体重量和电池组重量)结合起来,计算出特斯拉电池组的大多数关键参数。
Let's now take a look at my analysis of the accuracy of the EPA data. In order to interrogate the data, I pulled up the Certificate Summary information for all of Tesla's vehicles and trims, loaded it into a spreadsheet, and began noticing some discrepancies and quirks, which are shown on screen. First, the calculated energy capacity rarely aligns with the end state of charge energy capacity. That doesn't surprise me because my assumption is that the end state of charge capacity is usable energy capacity, or the energy available to the customer. Whereas the calculated energy capacity is the nominal energy capacity, or the total energy including any safety buffers for both overcharging and over discharging. If that's the case, it appears the fields aren't being used correctly, and at the very least aren't being used consistently.
That's because I found at least two instances where the calculated energy capacity is the same or lower than the end state of charge capacity. If calculated energy capacity is meant to show nominal or total energy capacity, that figure should always be higher than the end state of charge or usable energy capacity figure. Second, the pack level energy density figures are inconsistent. For example, as far as I'm aware, the 2021 Model Y and Model 3 long-range vehicles use the same battery pack. That's so Tesla could use the packs interchangeably and maximize economies of scale. However, the Model Y pack shows 180Wh per kilogram, and the Model 3 pack shows 165Wh per kilogram, which doesn't make sense if it's the same pack. I think both of those numbers are wrong and I'll provide my own estimate later in the video.
Third, even if the end state of charge figure is meant to show usable energy capacity, it doesn't seem to align with other data sources. For example, for the 2021 US-made long-range Model 3 that uses Panasonic battery cells, which I have, the TESSI app shows that the average original capacity is 77.8KWh of usable energy capacity. Note the 81KWh showing in the 2021 EPA documents. That TESSI data is backed up by several charging tests that can be found online, including this one by Out of Spec Reviews, which shows 76KWh of usable energy capacity for the 2021 US-made long-range Model Y, which had the same battery pack as the Model 3. Although 76KWh is lower than what's showing in the TESSI app by 2KWh, the vehicle has had 13,000 miles on it, which means it would have seen about 2 to 3KWh of initial capacity loss.
第三,即使终止充电状态的数字是为了显示可用的能量容量,它似乎与其他数据来源不一致。例如,我的2021年美国生产的长续航版Model 3使用松下电池,TESSI应用程序显示其平均原始容量为77.8千瓦时的可用能量容量。注意,2021年EPA文件中显示的是81千瓦时。TESSI的数据得到了几个在线充电测试的支持,包括Out of Spec Reviews的这个测试,显示2021年美国生产的长续航版Model Y的可用能量容量为76千瓦时,Model Y和Model 3使用了相同的电池组。尽管76千瓦时比TESSI应用程序中显示的少了2千瓦时,但这辆车已经行驶了13,000英里,这意味着它可能已经损失了约2到3千瓦时的初始容量。
With all that in mind, it leaves me with two takeaways. First, we can't blindly rely on the EPA data. If we do use a figure from the EPA documents, it needs to be carefully selected and with an explanation of why it was used. Second, the only way to get a true measure of a battery cell or pack is from sources like Monroe and Associates or UC San Diego, where we can guarantee an Apple's-to-Apple's comparison by controlling for testing conditions. Now that we have an understanding of the EPA data, let's take a look at my energy density table. For the generation 1 4680 cell, all of the data was from testing by UC San Diego, so it's a solid benchmark. For the generation 2 4680 cell, the nominal energy density of 272Wh per kilogram was from Monroe and Associates.
考虑到这一切,我有两个结论。首先,我们不能盲目依赖EPA(美国环境保护署)的数据。如果我们确实要使用EPA文件中的某个数据,就必须仔细选择,并解释为什么使用它。其次,要真正测量电池单元或电池组的性能,只能依赖像Monroe and Associates或加州大学圣迭戈分校这样的来源,因为他们能够通过控制测试条件来确保苹果对苹果的对比。现在我们对EPA数据有了一定的了解,让我们来看一下我的能量密度表。对于第一代的4680电池单元,所有数据都来自加州大学圣迭戈分校的测试,因此是一个可靠的基准。对于第二代的4680电池单元,272瓦时每千克的标称能量密度来自Monroe and Associates。
For the usable energy density, I used the 122.4KWh calculated pack energy from the EPA documents, divided it by 1,344 battery cells in the Cybertruck pack, and then divided it again by the cell weight, which is 350 grams or 0.35 kilograms. The end result was 260Wh per kilogram. In this instance, why did I use the 122.4KWh figure from the EPA document? Because real-world charge tests of the Cybertruck showed that from 0 to 100%, the total charge replenished came to 122KWh, so it matches the EPA documentation and provides double verification. Next, let's skip to the LG 2170 battery cell, which has been used in China-made, long-range Tesla Model 3s and WISE. Once again, all of the data shown here was from UC San Diego, so it's a reliable benchmark.
That finally brings us to the Panasonic 2170 cell, which many people consider the gold standard for energy density when it comes to Tesla vehicles. The nominal energy density of this cell is often said to be around 280Wh per kilogram, but I couldn't find any reliable data that backed that up. Let's take a look at my calculation. For the usable energy capacity, as I showed earlier, according to Tessie, the average original capacity of a long-range Model 3 battery pack using Panasonic cells is 77.8KWh. That pack contains 4,416 cells, and each cell weighs at minimum around 68 grams each. The cells are actually probably more like 69 grams each. That means, aspirationally, the usable cell-level energy density of the Panasonic 2170 cells is 259Wh per kilogram, which is surprisingly a tad lower than the cybercell's usable energy density of 260Wh per kilogram.
For the nominal energy capacity across multiple sources, the most common figure I've seen claimed for the long-range Model 3 and why is 82KWh, which would mean at the cell level, the total energy density would be 273Wh per kilogram. Again, I think that's generous because the cell is probably slightly heavier than I've listed here. What all this means is that the second-generation 4680 cell is right up there with what many people consider the best high-nickel battery cell that's used in Tesla vehicles. It's 272Wh per kilogram compared to 273Wh per kilogram for Panasonic, which as far as I'm aware is so far untested and may actually be lower.
As a side note, yes, there are other cells on the market with higher energy density, but at least for the time being, the vast majority won't meet Tesla's requirements in terms of other factors, like cost, cycle life, charging speed, form factor, or availability. Moving along, cell-level energy density is one thing, but what really matters is pack-level energy density. The EPA documentation states that the Cybertruck battery pack is 170Wh per kilogram. That information appears to be reliable because it's backed up by battery passport information, which includes a weight figure. A 721 kilogram battery pack with an energy density of 170Wh per kilogram would have a total pack energy of 122.6 kilowatt hours, which is right in line with the 122.4 kilowatt hours from the EPA documentation.
It also aligns with the 122 kilowatt hours we've seen in real-world charge tests, which means that 170Wh per kilogram appears to be usable energy density. That means the nominal or total energy density would actually be higher. At the cell level, if the ratio of usable to nominal energy density is 95.5%, that means the nominal energy density of the Cybertruck pack is more like 178Wh per kilogram. How does that compare to Tesla's other battery packs? As I showed earlier, for the 2021 Model 3 and Y that used the Panasonic 2170 cell, despite presumably sharing a common pack, one has an energy density listed of 165Wh per kilogram and the other 180Wh per kilogram.
So what's the reality? BatteryDesign.net, which I find to be reliable, calculated the energy density to be 171Wh per kilogram. That was based off a total capacity of 82KWh, which matches my assumptions and a pack weight of 481 kilograms. If that's all correct, then the energy density of the Cybertruck battery pack is higher than the 2170 battery packs that use Panasonic battery cells and by a solid 4% margin. What about the Model S, which according to the EPA data is 186Wh per kilogram? Most of the plaid Model S show that the battery pack can accept at least 96KWh, but it's probably closer to the 96.9KWh listed in the battery management system. The EPA documents show 97.8KWh, which is, as I often saw in these documents, on the high side by about a kilowatt hour, or 1%. I suspect the same is true for the calculated energy capacity of 104.5KWh. It translates to a pack-level energy density of 184Wh per kilogram rather than 186Wh per kilogram.
But it could be even lower. Monroe and Associates claim the plaid battery pack has an energy density of 181.5Wh per kilogram. However, I don't know how Monroe calculated that figure, so I'll be using 184Wh per kilogram instead to give the plaid battery pack the benefit of the doubt. That means the plaid battery pack only has about a 3% advantage over the Cybertruck battery pack. Considering that the plaid pack has been through several revisions over the past decade to reach this point and Tesla's structural battery pack is only on its second revision in two years, Tesla's making excellent progress. That's especially true if we take into consideration other factors beyond pack-level energy density.
What do I mean by that? First, the 4680 pack is cheaper and easier to assemble than Tesla's other high-nickel battery packs. That's because, among other things, it uses battery cells that hold about 5 times more energy, which means 80% fewer parts, which in turn means fewer welds and faster manufacturing speeds. Second, the plaid Model S uses a bulky structural floor pan above the lid of the battery pack, which is redundant and adds weight. For the Cybertruck, the top of the battery pack is the floor, which cuts a lot of weight. That means despite the slightly lower energy density of the structural battery pack in the Cybertruck, it reduces weight at the vehicle level by replacing part of the vehicle structure.
我的意思是什么?首先,4680电池组比特斯拉其他高镍电池组更便宜且更容易组装。原因之一是它使用的电池单元储能量大约是普通电池的5倍,这意味着零件数量减少了80%,从而减少了焊接点并加快了制造速度。其次,Plaid Model S使用一个笨重的结构地板,位于电池组的顶盖上方,这既多余又增加了重量。而在Cybertruck中,电池组的顶部就是车辆的地板,这大大减轻了重量。这意味着尽管Cybertruck的结构电池组能量密度稍低,但它通过替代部分车辆结构,在整车层面上减轻了重量。
Third, because the Cybertruck bolts the seats directly to the battery pack, it makes the vehicle faster and easier to manufacture, because the seats can just be lifted into the cabin on top of the battery pack. That's as opposed to the Model S, where the seats have to be awkwardly shoehorned in as a separate manufacturing step. As a side note, my guess is that the reason why the measured energy density figures in the EPA documents often come in 1-2kWh over what customers are finding is that for the EPA and state of charge, they could be draining the battery packs past 0% until they die. However, if that were true, it should also be true for the Cybertruck, but it's not, because the EPA and customer data matches. Once again, there seems to be some inconsistency, either in how the testing is being done or how the EPA documents are being completed.
In summary, at the cell level, Tesla's Generation 2 4680, or Cybert cell, is on par with the Panasonic 2170, which is the highest energy density battery cell that Tesla currently sources from their suppliers. At the pack level, my view is that the battery pack used in the Cybertruck is now the best high energy density battery pack in Tesla's fleet. That's because although it doesn't have the highest energy density at the pack level, it saves weight at the vehicle level and will be cheaper and easier to manufacture thanks to 80% fewer cells. Beyond that, given that this is only the second iteration of the 4680 battery pack, there's likely still more room for optimization. Overall, the cell and pack level energy density results were a nice surprise, because I had assumed that the Generation 2 4680 cell and pack were still about 6 to 9% lower than the high nickel cells and packs used in Tesla's other vehicles. But a closer interrogation of the EPA data shows that those assumptions were likely incorrect. As always, if I'm incorrect here and I've made an error in my calculations, let me know in the comments below.
With all that in mind, the next question is, how did Tesla achieve such high energy density with their 4680 cell in terms of both design and chemistry and what are the implications? That's what I'll cover in the next video in two weeks, so stay tuned. If you enjoyed this video, please consider supporting the channel by using the links in the description. Also consider following me on X. I often use X as a testbed for sharing ideas, and X subscribers like my Patreon supporters generally get access to my videos a week early. On that note, a special thanks to my YouTube members, X subscribers, and all the other patrons listed in the credits. I appreciate all of your support, and thanks for tuning in.