Hacker News new | past | comments | ask | show | jobs | submit login

People seriously underestimate both the current economic efficacy of existing lithium ion batteries, as well as the continuous and seemingly inexorable improvement in costs. This sort of improvement often doesn't get much news, but they have been happening at a very consistent rate across manufacturers.

I just hope that other technologies that are better for longer duration storage don't get overlooked for investment and development on industrial scales.




Longer duration battery storage is going to be very hard to make work financially since the capital costs are spread across far fewer charge/discharge cycles. At this point I feel it's more likely that it'll be cheaper to massively overbuild solar generation capacity to provide enough power in the winter than build seasonal battery storage.


Why do you say that? Researcher Jeff Dahn, who is a co-inventor of the lithium ion battery, has released what he termed a "million mile" electric vehicle battery. That is a pretty high level of charge / discharge cycles and seems like it would make stationary storage batteries, such as Tesla Energy's 3MWh Megapack.

https://cleantechnica.com/2019/09/09/jeff-dahn-claims-new-po...


I read 'duration' as the charge/discharge cycle time rather than the battery's lifetime in number of cycles. Clearly the more cycles a battery can handle the more cost effective it will be (more cycles over which to amortize the capital cost) but you still have to pay interest so even if the battery can handle infinite cycles if you only cycle it once a year (to balance seasonal differences in generation capacity) it will probably not be cost effective.

I think stationary storage batteries are great, but only for short duration storage (1 day) rather than long duration storage (months).


Yes, that's precisely what I meant for the term; storage that is economical even with few discharges cycles per month or maybe even per year.

I am somewhat skeptical that it can work, but flow batteries are still in their industrial infancy, and with the right sort of cheap stuff to add to water this could potentially scale very well for some sources.

I also strongly agree that we will likely build renewable sources to match their output to the seasonal minimum needs, rather than have many TWh of storage, it I have been so wrong in the past that I no longer want to cut off potentially promising avenues.

I see our future as being one of electrical energy abundance at nearly all times, and only a few times a year will there be a crunch to limit consumption; however as we get used to that abundance, conservation could get more difficult, and also people will get far more clever with time arbitrage of energy demand. And once there's a huge market for arbitrage, people will get more clever to serve that economic need, in ways that we have not yet anticipated.


Gigawatt-hours are coming online in the next few months:

https://www.greentechmedia.com/articles/read/say-hello-to-th...


This project is a (very large) lithium-ion battery which will presumably be used daily. It's not designed to store power between seasons.


I don't think you want to use batteries to store power between seasons, since things like Hydrogen or Methane are cheaper.


> Longer duration battery storage is going to be very hard to make work financially since the capital costs are spread across far fewer charge/discharge cycles.

The same argument could be applied to solar panels (25 year warranty) or fibre optic cables. But as so many solar panels are required they will be producing them for decades before the market is saturated. Of course after decades the old ones have indeed started dying, so it isn't an issue.


Lithium ion battery got what internal combustion engine had in beginning of the last century. Wide spread usage. Once money is steadily coming from end user, all the Innovation will be directed towards minor efficiency improvement, which will in long term, develop the tech that seems impossible as the biggininng. I can see by end of next decade, lithium ion battery will surpass both energy and power density of gasoline, which seems impossible even from physics point of view.


> lithium ion battery will surpass both energy and power density of gasoline, which seems impossible even from physics point of view.

Sure, I can do the math on that.

Gasoline has about 45 MJ/kg[1]. 1 kg of lithium contains 868x10^23 atoms, or 13.9 megacoulombs if each atom donates a single electron. Since a joule is one volt-coulomb, every electron must average out (nominal) to 3.24 volts. Todays li-ion batteries have a nominal voltage of ~3.65 V, and the nominal voltage of li-air batteries are 2.91 V.

The actual voltage of li-air is ~90% lower than the required voltage to reach raw parity with gasoline, which makes sense given that the theoretical limit of li-air batteries is 40.1 MJ/kg (.9 * 45 = 40.5).

Reaching the same power density as gasoline in vehicle engines (=<9 MJ/kg) isn't too bad (very close to 10x best current consumer batteries, closer to weirder chemistries), but parity with the real power of gasoline is a very hard ask. Donating more than one electron per lithium would require some serious re-evaluation of what we know about chemistry. If you only have one electron per atom, you have almost no weight left over for other materials to use to develop a voltage difference with. A battery that exceeds gasoline's full heat energy would need some very exotic materials- either lighter atoms[2], different ways of storing electrons[3], or ways to store energy besides electrons[4].

[1] divide by 5 for a realistic engine efficiency. At that rate, a gallon of gas provides 9.4 kWh- enough to get a Tesla model S over 35 miles.

[2] there are none, besides hydrogen, which itself isn't really helpful. It's not very good for electrochemistry.

[3] preferably, in a different state than me. Electrons hate each other. A 3 Ah battery has 11,000 hateful coulombs in it, each electron within a less than a nanometer of the others around it. That's a 10^9 on top and a 10^-9 on the bottom. Not positive on the math on that but I'm reasonably sure it's in the thousands of megatons. It is rapidly trying to change chemical and then geographical state.

[4] preferably on another planet from me. If you're storing energy with a different force than the electromagnetic force, you're pretty much guaranteed to be doing some real fuckery to subatomic particles, and they are not going to be happy.


Lol. Fabulous answer to the math, with nice humour! Unless the Gas industry learns how to make new Gas on the fly, Lion will continue to shine also because of the amazingly different ways that it's energy losses can be replenished (wind, water, sun, vibration...), even while it's doing its thing.

So, GP's comment of "develop the tech that seems impossible as the biggininng" (typo included) is the important part. It might be that in the future, because the motors and other moving parts become increasigly efficient (& are already much better than a combustion engine) then less energy is required for the same resulting work. Even more so, the ingenious ways that energy can be 'reharvested' to flow back into the system could in fact allow us to develop something that seems impossible...simply because we're harvesting energy and pouring it into our Lion dependant system better than Gas ever can.


> because the motors and other moving parts become increasigly efficient (& are already much better than a combustion engine) then less energy is required for the same resulting work

The resulting work is more important than the efficiency. Teslas have exceptionally low air drag, which is great for driving range.

However when you go from 80 to 90% system efficiency, you only increase range by 12%. Once youre at 99%, you really cant improve any more. 99% to 99.5% means the motor outputs half as much heat but goes hardly any farther.


You could make gasoline from the air. A bit inefficient, but if we somehow get to the fabled "free fusion" power plants, it could be a neat trick to have synthetic gas in pipelines flowing from fusion CO2 atmospheric scrubbers.


You don't need fusion. PV works too; you just need a lot of it. Making liquid fuel from air is currently the only viable solution for carbon-neutral aircraft, which is why the military likes the idea.


What about the weight of the entire drive system? Still not the same thing, as the weight of batteries doesn't change as you drive. Not sure, as a quick search doesn't give me the right data, but it seem like the ICE setup requires more engine/exhaust etc. weight, so it seems likely the amount of efficiency required for batteries is less than gas at the current tech level.


Last time I tried to estimate that it seemed with a passenger car you get some win there that offsets the battery weight. Something around 300lbs. My guess is if batteries got 30% more power dense the weight penalty would disappear. Part of that is as the weight of battery drops you can make the rest of the car lighter as well.


Thanks for the numbers, which I was aware of, but not exactly. The constant supply of money in Research can do impossible magic. If you look at it from social perspective, all the semiconductor progress in last 20-30 years happened just because people want to send their photo to another people. Finger crossed, but I'm hopeful.


No chemical battery that doesn't breathe air will ever be as energy-dense as gasoline.


> People seriously underestimate both the current economic efficacy of existing lithium ion batteries, as well as the continuous and seemingly inexorable improvement in costs.

Maybe because the cost of devices have skyrocketed and the battery dont last any longer... And in addition to that, you can no longer replace batteries.


How are people STILL repeating this? Does anyone actually remember what phones used to be like? Because I can stream videos for hours and listen to music all day for multiple days before I run out, and I still remember my iphone 3 dying in 6 hours playing music that was already on the device.

The iphone 11 has a 3110 mAh battery. The iphone 3 had 1220 mAh. I don't even remember the last time my phone died on me unexpectedly.


So far with every upgrade they have added features like more powerful processors and power hungry higher density screens. So most people haven't noticed the improvements in battery tech. Now I think phones have reached a point where I don't see many power hungry improvements going forward so we might start seeing longer lasting phones.


> power hungry higher density screens.

That is not exactly true. OLED can be very efficient. Not just in Dark Mode, but you need to be aware of those pitfalls and optimise for it. Since in overall package the Display is roughly the same power but you get much better quality.

Not to mention there are at least 3 - 5 years roadmap of technology in the pipeline that makes it even more energy efficient and thinner. ( Thinner equals potential for larger battery )


The higher density screen might need to require less power but the power to display on them used by the graphics etc is a lot higher. we were on 320x640 screen now we are at 3840 x 2160. Even if the individual pixel takes less power the number of pixel has gone up a huge amount.


I used to charge once a day (10 years ago), and I still do because my phone can't last 2 full days on a full charge... but it does last about 1.5 days ... it is not an iphone 11 but it is still an $800 phone


The number of hours you spend and what you do on your phone today compared to 10 years ago is also not the same.


True, but the hardware also uses a lot less power then it used to.


That's not true in the slightest. Energy per instruction varies a lot but has not decreased that much on recent nodes. The resistance growth just outpaces the shrinking capacitance of the gates.

On top of that processors are faster, applications and sites are hugely more intensive to render, and the hardware (mainly antennas) suck up much more energy.

The ARM11 in the iphone 3 used <1 watt. The A13 in the iphone 11 uses 6 watts.


Back when I had a basic phone I charged it roughly once per week.

But of course modern smartphones do a lot more, and I spend lots more time with my smartphone as well.


You're right and wrong. People buy iphones despite the cost and lack of user serviceable battery. They are a luxury item and a way to signal status. If people stopped buying iphones due to battery life then Apple may change. But Apple has a strong brand and the status of the item is important to people so they act against their own interests and pay lots of cash rather than use a $100 dollar android that's 90% as good.

Batteries aren't a feature to most people.

Also there's lots to be said about the scope of what most 'phones' do these days. The batteries are powering serious computers rather than just phones.


Yes computers are more powerful but they are also more efficient


This comment has a pet-peeve of mine, which is using the word "efficient" where it's unclear what the cost and benefit are. Efficiency is just cost per benefit, so it's not a very meaningful word if you don't know what the cost and benefit are.

I don't think you're wrong, I just am not entirely clear on what claim you're making. I'll venture based on context that the cost you're talking about is either electrical charge or time, but it's very unclear what benefit you're talking about.


Not the OP, but they compute things both faster and using less energy, so that’s an increase in efficiency on both counts.


Maybe? Tesla's working on a million mile battery suggests to me that battery lifespan will start to dramatically improves.


If they or their competitors can deliver, sure.

One could equally argue that the fact we don't have a million mile battery already suggests that there may be intractible problems to solve.


The issue with a million mile battery is one of testing. If your car has 500 miles of range, a million miles means 2000 charges and discharges. At a standard C/5 rate, that's 2.3 years to finish testing. Even at a fairly brisk 1 C that's 83 days. It makes iterating on a design -which can mean changing component amounts by fractions of a percent- incredibly time consuming.

Dahn's research is so well respected because he was incredibly good and thorough in measuring tiny amounts of heat and voltage changes in batteries so that you could trace their degradation over a much smaller number of cycles, then try them out at longer scales. Tesla has been funding much of his lab. There are very few people who can do this kind of research outside his lab, and IMO he's a wizard who will end up with a Nobel if he makes a few more big contributions.

The bottom line is that the only way to really tell if a battery will last a year or a decade is to watch it for a year or a decade. No amount of technology lets you forecast the future with perfect accuracy, and battery longevity testing takes time. That's the only reason we don't have a million mile battery.


>One could equally argue that the fact we don't have a million mile battery already suggests that there may be intractible problems to solve.

That would be a good argument if batteries had been stuck for decades at a much lower limit. But the fact is their lifetime has been steadily increasing, and will reach a million miles within a few years at most.


I'm confused by this. A million mile battery would make ICEs obsolete today. It would be more cost effective to spend $50k on a Tesla Model 3 than to buy 5 Hyundai i10 and driving those to dust.

A 1 million mile battery is kinda like fusion. Once it happens there is no going back so I don't think that failing to solve this problem is a blow to EVs.


A million mile battery does not mean the rest of the car lasts a million miles


Musk says modern Tesla's are designed to last a million miles because he wants them all to become taxi's. Full self driving taxi's that return more then they cost to the owner after Tesla takes it's take from the taxi earnings.

I don't know whether to believe this, but he has the motive, opportunity, and is definitely present at the crime scene.


Will Teslas become cost effective at that point?


I just ran numbers for Tesla Cybertruck. Depending on gas vs electricity prices, a "Tri FSD" CT, built to million-mile specifications, will pay for itself on gas savings by 0.5-1.5M miles.

I think "free truck" operated to design specifications constitutes cost effective.


You have to factor in the time value of money though. Car purchase price is an immediate cost while gas and electricity are costs over the life of the vehicle. Your average personal cars only gets driven about 15,000 miles a year and thus those gas and electric costs would be spread out over decades to reach a million miles. Probably still cost effective considering the average truck isn't that much cheaper though.


Too bad you can't buy one...


*yet


You can buy nuclear cars yet either...


No one is actively trying to build nuclear cars however. Additionally, no one has a prototype nuclear car, because that is a stupid and very dangerous idea.

There is a prototype cybertruck, and Tesla is about to announce the location for the factory to build them. It will be either Tulsa, OK or just outside of Austin, TX.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: