3 ms·

I seem to remember being referred to as "that battery guy" here on HN, so here I am I guess. Unfortunately that is complete bullshit. (No offense intended to t

I seem to remember being referred to as "that battery guy" here on HN, so here I am I guess. Unfortunately that is complete bullshit. (No offense intended to the Bloomberg suit.) Energy densities have absolutely -not- tripled since 2010, in fact they have barely changed. Costs have gone down, yes, but not by 87%. Cost-wise: a simple counterexample is to look at any smartphone BOM from 2010. Phone batteries around 10-12 Wh cost around $5. That's somewhere around $416-$500 per kWh. And that's for a phone battery, AKA not exactly the cheapest batteries out there, and not at massive scale (phone batteries are tiny in capacity and almost all 2010 smartphones sold in the single-digit millions max.) Currently, as Bloomberg claims, cheap batteries are around $160/kWh. So that'd be a change of around 61% cheaper. Around 0.4x the 2010 price instead of the outlandish claim of 0.13x the price. Keep in mind that this I'm just basing this off mobile phone BOMs cause I can't be bothered to find anything cheaper - there certainly were cheaper batteries then. Energy density wise, it's even more bullshit. TRIPLED? No, energy density has sat around 200 Wh/kg for approximately forever. (Off the top of my head, since at least 2009.) 200 Wh/kg batteries have been available to the -consumer- at not-insane prices for over a decade. In 2013, anyone - business or consumer - could purchase Panasonic cells at 224 Wh/kg. Today, the highest energy density anyone claims is around 240 Wh/kg from Tesla's 2170. Let's take that at face value and assume it is 240.0 Wh/kg; that's a 7% increase from 2013, and at best somewhere around a 15% increase from 2010. The sad reality is that batteries are extremely hard, and few things have changed aside from price. Hell, even the form factor, the 18650, was invented about three decades ago! Pretty much the only thing that has changed recently is that volumes have gone up massively, driving down costs with at-scale manufacturing. But that has happened in the vast majority of industries, and was entirely predictable. Furthermore, the cost decrease for batteries specifically hasn't been as fast as nearly everyone predicted. All the other magical headlines, about energy densities skyrocketing, or fancy new chemistries, or swappable packs, have been pure bullshit.

owenversteeg 6y ago
I'm glad the above comment got some attention even buried here :) Another comment I'd like to make: Even with that increase, say 200 Wh/kg in 2010 to 240 in 2020, a lot of these higher-energy-density batteries are more expensive/easier to damage/have bigger heat management problems/etc. Keep in mind that is the raw cell weight. At 240 Wh/kg, you do not have any cooling, casing, wiring, BMS, safety systems - that is a plain cell in a vacuum. Once you've got all those goodies around the battery, even if you're best-in-class like Tesla your total energy density is around 150-160 Wh/kg. If you aren't making a luxury vehicle, you might end up around 80-90 Wh/kg (24kWh Nissan Leaf battery is in total 82 Wh/kg), and even if you do aim for something light and use fancy engineering you might end up around the 130 Wh/kg mark (2019 Audi e-Tron: 134 Wh/kg.) Batteries are hard.
yellowapple 6y ago
Do you happen to have sources / further reading on this?
- owenversteeg 6y ago
  I didn't bother to give sources because everything is pretty trivial to find, but you can Google "tesla 2170 bench test" or "Samsung INR21700-50E bench test" for details of the best cells today, "history of 18650" to see that the format is nearly 30 years old, and "(your favorite 2010 phone) BOM" for some example price data. (as an example: galaxy s4, $5 and 10Wh.) That said, I'm positive there were cheaper batteries at the time, given that cell phone batteries are typically more expensive to make even today and are on the order of 10,000 times smaller than car batteries. I wish I had some quality further reading to link, but unfortunately I don't. "battery Google" is dominated by 1) overpriced e-stores selling a few cells to vapers, 2) news about handmade cells in a lab reaching some ridiculous performance (yeah, and a handgun kills cancer in a lab), 3) complete unadulterated bullshit from the news/bloggers/social media about battery trends, 4) bullshit corporate PR of companies claiming they're a year away from 900 Wh/kg (anyone working with batteries would give their firstborn child for 900 Wh/kg no joke)... Sadly, it's a mess. The other problem is that an explanation is hard - very hard. It involves manufacturing (and scaling this), thermal properties, funding for research, an analysis of the players in the game, etc etc. I have considered writing one, but it even for me it would be quite a project and even posted here on HN I have a feeling it might fail to get any attention. Nobody wants to read a complex multifaceted failure analysis, everybody wants to read a Bloomberg headline saying we're racing into the future, even if it's total bullshit.
  - yellowapple 6y ago
    > I didn't bother to give sources because everything is pretty trivial to find If it was then I wouldn't be asking :) Best I was able to find was a chart (ostensibly from NASA) comparing different battery technologies, which is probably where the 3× claim originated, but doesn't say much about improvements over time (only by technology): https://www.epectec.com/batteries/cell-comparison.html https://www.epectec.com/batteries/cell-comparison.html To your point, though, it does indicate that lithium-ion specifically hasn't had such a 3× jump (but it does represent such a jump relative to older technologies like lead-acid or nickel-cadmium). > you can Google "tesla 2170 bench test" or "Samsung INR21700-50E bench test" for details of the best cells today That doesn't really say much about historical trends, but good to know. > "history of 18650" to see that the format is nearly 30 years old Is that relevant? AAs are even older than that, and yet seem to have improved quite a bit. > and "(your favorite 2010 phone) BOM" for some example price data. (as an example: galaxy s4, $5 and 10Wh.) That does line up with https://www.androidauthority.com/galaxy-s4-bill-of-materials-175146/ https://www.androidauthority.com/galaxy-s4-bill-of-materials... (which cites $5.60 for 2600mAh, or $4.90 for 2100mAh). Meanwhile, the Galaxy S9+ lists $4.90 for 3500mAh: https://technology.informa.com/601100/galaxy-s9-materials-cost-43-more-than-previous-versions-ihs-markit-teardown-shows https://technology.informa.com/601100/galaxy-s9-materials-co... Neither of these really say much for the mass, though, so it's hard to get a clear picture on energy density from this. > Nobody wants to read a complex multifaceted failure analysis I mean, I would, but I'm well aware of my insanity.

owenversteeg 6y ago

I'm glad the above comment got some attention even buried here :) Another comment I'd like to make: Even with that increase, say 200 Wh/kg in 2010 to 240 in 2020, a lot of these higher-energy-density batteries are more expensive/easier to damage/have bigger heat management problems/etc. Keep in mind that is the raw cell weight. At 240 Wh/kg, you do not have any cooling, casing, wiring, BMS, safety systems - that is a plain cell in a vacuum. Once you've got all those goodies around the battery, even if you're best-in-class like Tesla your total energy density is around 150-160 Wh/kg. If you aren't making a luxury vehicle, you might end up around 80-90 Wh/kg (24kWh Nissan Leaf battery is in total 82 Wh/kg), and even if you do aim for something light and use fancy engineering you might end up around the 130 Wh/kg mark (2019 Audi e-Tron: 134 Wh/kg.) Batteries are hard.

yellowapple 6y ago

Do you happen to have sources / further reading on this?

owenversteeg 6y ago
I didn't bother to give sources because everything is pretty trivial to find, but you can Google "tesla 2170 bench test" or "Samsung INR21700-50E bench test" for details of the best cells today, "history of 18650" to see that the format is nearly 30 years old, and "(your favorite 2010 phone) BOM" for some example price data. (as an example: galaxy s4, $5 and 10Wh.) That said, I'm positive there were cheaper batteries at the time, given that cell phone batteries are typically more expensive to make even today and are on the order of 10,000 times smaller than car batteries. I wish I had some quality further reading to link, but unfortunately I don't. "battery Google" is dominated by 1) overpriced e-stores selling a few cells to vapers, 2) news about handmade cells in a lab reaching some ridiculous performance (yeah, and a handgun kills cancer in a lab), 3) complete unadulterated bullshit from the news/bloggers/social media about battery trends, 4) bullshit corporate PR of companies claiming they're a year away from 900 Wh/kg (anyone working with batteries would give their firstborn child for 900 Wh/kg no joke)... Sadly, it's a mess. The other problem is that an explanation is hard - very hard. It involves manufacturing (and scaling this), thermal properties, funding for research, an analysis of the players in the game, etc etc. I have considered writing one, but it even for me it would be quite a project and even posted here on HN I have a feeling it might fail to get any attention. Nobody wants to read a complex multifaceted failure analysis, everybody wants to read a Bloomberg headline saying we're racing into the future, even if it's total bullshit.
- yellowapple 6y ago
  > I didn't bother to give sources because everything is pretty trivial to find If it was then I wouldn't be asking :) Best I was able to find was a chart (ostensibly from NASA) comparing different battery technologies, which is probably where the 3× claim originated, but doesn't say much about improvements over time (only by technology): https://www.epectec.com/batteries/cell-comparison.html https://www.epectec.com/batteries/cell-comparison.html To your point, though, it does indicate that lithium-ion specifically hasn't had such a 3× jump (but it does represent such a jump relative to older technologies like lead-acid or nickel-cadmium). > you can Google "tesla 2170 bench test" or "Samsung INR21700-50E bench test" for details of the best cells today That doesn't really say much about historical trends, but good to know. > "history of 18650" to see that the format is nearly 30 years old Is that relevant? AAs are even older than that, and yet seem to have improved quite a bit. > and "(your favorite 2010 phone) BOM" for some example price data. (as an example: galaxy s4, $5 and 10Wh.) That does line up with https://www.androidauthority.com/galaxy-s4-bill-of-materials-175146/ https://www.androidauthority.com/galaxy-s4-bill-of-materials... (which cites $5.60 for 2600mAh, or $4.90 for 2100mAh). Meanwhile, the Galaxy S9+ lists $4.90 for 3500mAh: https://technology.informa.com/601100/galaxy-s9-materials-cost-43-more-than-previous-versions-ihs-markit-teardown-shows https://technology.informa.com/601100/galaxy-s9-materials-co... Neither of these really say much for the mass, though, so it's hard to get a clear picture on energy density from this. > Nobody wants to read a complex multifaceted failure analysis I mean, I would, but I'm well aware of my insanity.