Hacker News new | past | comments | ask | show | jobs | submit login
GBatteries (YC W14) let you charge your car as quickly as visiting the pump (techcrunch.com)
65 points by mikek 63 days ago | hide | past | web | favorite | 92 comments



Super skeptical here, or at least the wording is misleading as hell.

> The product allows users to charge a 60kWh EV battery pack with 119 miles of range in 15 minutes as compared to 15 miles in 15 minutes today.

> The technology works with off-the-shelf lithium ion batteries and existing fast charge infrastructure by integrating via a patented self-contained adapter on a car charge port

First read says that they're delivering 60kWh in 15min, or they're pumping electrons at 240kW. This is just nonsense -- most deployed DC fast chargers are 50kW... the charger companies and the OEMs are experimenting with 200kW chargers, but those are liquid cooled. So physics says this is nonsense.

Okay, so second read: the 60 kWh is distracting technobabble, and what they're ACTUALLY doing is "adding 119 miles of range in 15 minutes."

The just-announced Nissan Leaf e+ has a 62 kWh battery with 226 miles of range. When you DC Fast Charge, going from 0-80% takes about as much time as going from 80-100% (EV roadtripping is about doing frequent small charging because of this phenomenon, not one-time top-offs like with gas).

The Nissan Leaf e+ is about a 60 kWh battery, and if they're adding "119 miles of range in 15 minutes", that's about half the Leaf's 226 mi range, so call it "30 kWh of charge in 15 min". That means they're charging at an average of 120kW. Okay, now we're back within the realm of physics.

Note the Leaf e+'s (yet unreleased hardware) still only accepts a maximum of 100kW DC fast charging. I think jaguar has experimented with 150kW charging on the ipace (but again, liquid cooled cables).

So they're saying they found a way to cycle fast charging and get about a 20% improvement in average charging rate from the current kinda-top-of-the-line tech, and only for the first 50% battery capacity.

Not nothing, but hardly "as quickly as visiting the pump", especially since whatever tricks they use likely won't continue for the next 50% of the battery.


Hi! I'm one of the founders at GBatteries, please let me try to clear that up.

We've demonstrated 5 min to 50% charge, 10 min to 100% - I’ll edit this to include a demo that we’ve done at CES 2019, and that we’re now exhibiting at the Detroit Auto show demonstrating.

[EDIT] Video here: https://youtu.be/kSLrqR4TfnU Note: Charging times can vary by ~1 minute here and there, because the algorithm is adaptive and characteristics of batteries change from moment to moment.

>First read says that they're delivering 60kWh in 15min, or they're pumping electrons at 240kW. This is just nonsense -- most deployed DC fast chargers are 50kW... the charger companies and the OEMs are experimenting with 200kW chargers, but those are liquid cooled. So physics says this is nonsense.

>Okay, so second read: the 60 kWh is distracting technobabble, and what they're ACTUALLY doing is "adding 119 miles of range in 15 minutes."

There’s a mistake in the article, in both cases it should be in 5 minutes; should be corrected shortly.

In this example, we’re talking about a 60kWh battery with 238 miles of range (Chevy Bolt). Right now the car charges 90miles in 30 minutes, or 15 miles in 5 minutes because the manufacturer limits the charging rate for the batteries in order to preserve their life. Our technology can enable these Li-ion batteries to charge in 5 min to 50%, or 119 miles. This example assumes that the car goes to a fast charging station that has enough power to charge the vehicle in this time; fast charge infrastructure that’s currently being put up by companies like Ionity (350kW), Electrify America, or Charge Point (500kW capability).

The problem with batteries today is not charging speed; it’s possible to charge any battery quickly, but the faster you charge a battery the faster it will degrade. Our technology is able to decrease the irreversible chemical reactions that happen during charging, so that the same batteries can be charged fast without compromising cycle life.


What are you doing differently? Unless you’re pulsing your charging or otherwise varying the current rapidly, then it seems that all you can really do is vary the current as a function of temperature, state of charge, and maybe some other variables. At the end of the day, if you’re going to deliver energy e in time t, you need average power e/t, and the fancier your curve, the more your peak power will exceed your average.

(And if you don’t want to overhead the cables or the connector, you care about current squared, giving you an added incentive to charge at near constant current or perhaps to charge some cells at a different rate than others.

What’s the trick here?


And to add to these questions - what are the consequences for increasing charging rates? Tesla has made some public statements regarding fast(er) charging (specifically targeting the 350kW Ionity chargers) to the effect that pushing that much power will degrade the batteries far faster for little gain.

Edit:

> Our technology is able to decrease the irreversible chemical reactions that happen during charging, so that the same batteries can be charged fast without compromising cycle life.

This seems to imply it's doing something to the battery chemistry? Maybe a brief pulse of high-rate discharge every now and then to help balance things out?


> This seems to imply it's doing something to the battery chemistry? Maybe a brief pulse of high-rate discharge every now and then to help balance things out?

That's usually within the domain of the battery management system. There's not much the charger itself can do (there is intelligent communication between EV and charging equipment, but not as granular as the view the BMS has).


You're correct - On a high level, our technology is based on pulses that are fully adaptive, and their parameters are adjusted by the AI controlling the charging.


This makes more sense....I posted another comment, but the long and short is that you would need to be able to bypass the manufacturers controller to do this, which it sounds like you are saying.


That's correct, we're not able to fast charge at these levels without approval from the OEM because of the limits they have built. We need to either be integrated inside of the vehicle, or on the charger but have the vehicles "approval" for charging at these higher rates, as at Level 3 and up it's DC to the pack.


Correct me if I'm mistaken, but this seems like a B2B/acquisition play masquerading as a B2C play; your value (if proven) is a superior battery management system (versus the OEM's or a supplier's) and charge controller for non-Tesla EVs (hinging on the battery pack instrumentation architecture).


You're absolutely right; our business model is to license the technology to manufacturers (OEM, Tier 1, etc) - not B2C to anyone with an EV.


Awesome, hope your team is successful.


Why do you say it's masquerading as B2C? It seems like straight up B2B to me...


The techcrunch article tone makes it sound like they're pitching charge rate increases directly to consumers. That doesn't seem to be the case based on the in-thread discussion, but how techcrunch made it sound.


The battery is the most expensive part of an electric car. If you have come up with something that MITMs the signaling between my car and the charger, causing the charger to deliver more power than my car says it can handle, that's a huge problem. First off, there's no way this won't immediately invalidate every warranty on my car, and secondly, why should I trust you over the company that manufactured my car and its battery?


Read the founder’s comment - sounds like they would only do this with explicit approval (and presumably license payment) from the manufacturer.


Thanks for chiming in! Batteries and charging are complicated as is, when translated into Journalist they get even more so :)

There was a lot of concern with the first generation EV's about battery life, a lot of which the data has shown to be misguided now that the first gen's are coming off-lease. How long is the process to convince the OEMs this won't degrade their battery packs?


Sounds like you know the struggle :)

In the battery industry there's a saying; liar liar battery supplier. That definitely doesn't help things move fast!

Automotive OEMs do move slowly, but once they see value things definitely speed up - A lot of it has to do with being in front of the right people.


Thank you for popping in and clearing that up.

It sounds like this would push the battery harder than they're designed for. Does using your charger void the warranty?


Desulfating process while charging?


Desulfating?


"119 miles of range in 15 minutes" is also nowhere near as fast as an IC car "visiting the pump". It does not take 15 minutes to fill the tank on a normal car.

Short of physically swapping out batteries, charging an EC will never be as fast as pouring gasoline into a tank. Transferring that much energy that quickly over wires is ... I don't want to stand anywhere near it. Fillup time is not an are where EC should try to compete.


I never finished school, so correct anything I've got wrong here.

Diesel has a specific energy of 48MJ/kg and weighs 0.832kg/L.

48MJ converts to 13.33kWh. My diesel tank will take 70L from nearly empty. I've never timed it, but it's definitely substantially quicker than 10 minutes to fill.

So I'm transferring about 776kWh in to my diesel tank in <10 mins.

70 * 0.832 * 13.33 = 776.33kWh

So it would take a 1MW charger three quarters of an hour to transfer the same amount of energy. <gasp>

Did I get any of that right?


Seems about correct by energy going into your tank, but not energy actually going to the road later. Divide your result for that by 4 or 5.


That's to say nothing of a caravan of trucks such as yours with full 70l tanks, barrelling down the freeway. If three such vehicles pass a given point per second, that's like 144 MW of power, I think!


> "119 miles of range in 15 minutes" is also nowhere near as fast as an IC car "visiting the pump".

The article was corrected to say 5 minutes. See also replies from timsher upthread.


Tesla superchargers also peak out at 120KW.


I wonder if you need to cool the battery pack charging at that rate you could imagine having a connector that hooked up a battery both to electricity and to a cooling circuit that pumped water round the battery pack as it charged


VAG is releasing a car shortly (March) that can charge at 150 Kw, and has deployed some charges in Europe and the US that can put out that much power. Last I heard, the iPace is listed as accepting 100 kW, but no one has seen over 80 yet.


476mi/h of charging isn’t gorundbreaking. Aston Martin, Tesla and Porsche already charge at those speeds or faster.

https://www.greencarreports.com/news/1118575_gm-working-on-u...

I believe their “secret sauce” is allowing existing packs which charge slower to get “up to speed”. If so, that’s a pretty limited market as it’s likely that within a few years, all EVs sold will charge faster than the listed speeds. They’ll likely have a maximum tota addressable market of a few hundred thousand cars with only a few thousand real consumers. Most people buying electric cars with slower chargers are doing so because those vehicles fit their lifestyle and wouldn’t have a massive incentive to buy a new charger.


Hi! I'm one of the founders at GBatteries - I’ve partially answered this in another comment, so will copy paste portions.

We've demonstrated 5 min to 50% charge, 10 min to 100% charge - below is a demo that we’ve done at CES 2019, and that we’re now exhibiting at the Detroit Auto show demonstrating this.

[Video here: https://youtu.be/kSLrqR4TfnU Note: Charging times can vary by ~1 minute here and there, because the algorithm is adaptive and characteristics of batteries change from moment to moment.]

The problem with batteries today is not charging speed; it’s possible to charge any battery quickly, but the faster you charge a battery the faster it will degrade. Our technology is able to decrease the irreversible chemical reactions that happen during charging, so that the same batteries can be charged fast without compromising cycle life.

We’re working on enabling existing packs to charge as fast as possible. We’re not able to fast charge at these levels without approval from the OEM as they have limits built-in. We need to either be integrated inside of the vehicle, or on the charger but have the vehicles "approval" for charging at these higher rates, as at Level 3 and up it's DC to the pack.


Thanks for the added clarity. Being able to charge from below 50% without damaging cycle life is huge. Would be great to see some cycle life charts showing GBatteries vs traditional chargers.

One of the other challenges you may face is the lack of consumer knowledge about how charging affects cycle life. Since consumers tend to charge batteries in ways that dramatically reduce cycles, this is potentially huge.

Figuring out how to create demand from consumers will be very important as OEMs don’t really have any motivation here. Cycle life isn’t published nor really considered by consumers when they buy devices with rechargeable batteries.


Should have a current clamp on the cable to show that you're not cheating when doing the demo.


That sounds like a cry to be bought by the highest bidder. Just of the shelf hardware and some pixie dust, err, "AI". Sure thing they'll find some car company that will sink some millions into their company.


Remember the Tesla insider that leaked a bunch of firmware and system details?

https://twitter.com/atomicthumbs/status/1032939652393521156

If it's to believed, the charge rate limit from a Supercharger is already maxed out -- they're pushing energy into the battery as fast as the charger can supply it (and taking some safety risks in the process).

This same company has been around for a few years with spurious claims about improving laptop battery life with their AI algorithms. So far nobody has seen evidence that it works, and nobody is lining up to buy them.


This is a straight up lie right?

My car (2017 Honda Accord) has a 500+ mile range on a tank of gas. It'd take over an hour to charge to 500 miles.

Even at 15 minutes that's a lie. It takes no more than 2-3 minutes of actual pumping to fill my 500+ mile tank of gas.


Came here to post exactly this. Misleading click baity headline; it doesn't take me much more than 2-3 minutes to fill up 55 litres of fuel to my 65 liter tank, for 700km+ range. (I'm not in the habit of running it dry!)


Can take 15 minutes+ in Oregon since we can't pump our own gas. But yeah, last road trip I got ~600 miles out of a full tank. This company has a loooong way to go if they want to compete with that.


Even in Oregon (or NJ where I'm originally from), it takes longer because of stuff outside of the actual pumping. So if for some reason you're not allowed to charge your own car, it'd just move from 15 minutes to 30 minutes.

Also in NJ it definitely doesn't take more than 5 extra minutes waiting for the attendant 95% of the time.


depending on the station, and how many attendants are working, you can be up around 30 minutes. I've seen some very efficient stations that do a lot of turnaround, and some very inefficient stations that have 10 lanes of 6-8 cars per lane waiting.

that said, even at a full tank, that's only 220 miles for me, so unless I'm leaving Oregon on a trip, a 15 minute charge sounds amazing.

the good news is that I've not seen a charging station that had an attendant.


The system uses AI to optimize the charging systems in electric cars.

Here's one thing that stinks about current/last gen electric cars. I drove with my wife from my apartment in Nob Hill to Santa Cruz, using up about 92% of my range. I was able to find a charger and get fully recharged for my trip back. All of that is fine. What was disturbing, was that while I was climbing those initial hills out of Santa Cruz, my range figure dipped way below my miles to go figure. As I left behind that initial climb, my range figure then went comfortably above my miles to go. This is one place where some judicious AI could help. (Along with integration with GPS/Maps and destination information. I should also update my electric car to technology ca. 2019.)


Telsas do that, FYI. They take elevation and temperature during a trip to estimate where on the energy usage curve you are at any point in your trip.

Here's an article from 2015 when they introduced that feature: https://www.teslarati.com/closer-look-tesla-firmware-6-1-tri...


Absolutely true and it works. For example, my Tesla S (2014) very accurately predicts mileage going from Los Altos to Tahoe (up hill, have to recharge on the way) and from Tahoe to Los Altos (down hill, can barely make it back without charging).


Chargers already pay attention to charging curves and temperature. Is there really much room for safely charging existing batteries faster?


The Porsche Taycan and the future Audi e-tron GT can make use of 350 kW chargers to get to 80% charge in 15 minutes. 350 kW chargers are being deployed on the Ionity and Electrify charger networks now:

https://insideevs.com/ionity-debuts-350-kw-ultra-fast-chargi...

Volkswagen probably will be the biggest producer of battery electrics within 3 years. They own 12 car brands (including Porsche and Audi https://www.volkswagenag.com/en/brands-and-models.html). VW also has a stake in Ionity and they own Electrify America and Electrify Canada.

If this charging technique really does work then I imagine VW will license it.


Reading between the lines, the constraint is that today’s chargers are always monotonically increasing state of charge with a roughly steady DC applied current (tapering slightly at the high end). Adding in something like pulsed charging or even some reverse-voltage segments could potentially change the charging dynamics enough to let the battery charge much faster without damage.


There is a fix for slow charging that is literally as fast as filling a tank but nobody wants to use it. (I assume it is the same reason that there are pages and pages of battery improvements that never make it to market.)

https://en.wikipedia.org/wiki/Flow_battery


Flow batteries, at least the kind that are commercially adoptable in the present/near-future (Vanadium redox), scale in capacity very easily. The downsides are twofold:

  * Poor power density
  * Relatively limited nominal power output
These compromises are fine for stationary applications, but for vehicles, they'd be a riskier bet as far as I'm aware.

(disclaimer: not a subject matter expert)


In the case of flow batteries, don't a lot of them have poor power to weight figures? Those liquids are heavy, but aren't nearly as power dense as gasoline.


The other fix would be swappable batteries. But it would require a huge engineering effort.


Why not fill up with loose cells? They are small, they roll. Couldn't they roll into a ribbed tray and then a rail on both sides clamps them in place and connects them? Or pulled in on a chain like a machine gun loading mechanism?


This is an original idea that I haven't seen elsewhere.

The big issue is that assembling loose cells into a battery pack is a hard enough operation. It must achieve good mechanical properties (the pack must be rigid enough), good electrical contact with the battery terminals (bad contact would ruin a pack, could start a fire as a bad contact could overheat) and good cooling (battery packs heat, so afaik all designs use liquid cooling, which needs very good thermal contact between the cells and the cooling assembly). Loose cells would essentially mean you're feeding the loose cells inside the pack and the pack has some robotic assembly arms that it uses to put the cells in the right places while it achieves the three objectives.

With some different, taylor made cell design you could make this self-assembly easier, maybe even make it mechanically passive - have the cells orient vertically and slide into their places, and then they're locked. But the current cell design is definitely not chosen for this.


tailor-made, unless we're talking golf, ha


>swappable batteries would require a huge engineering effort.

Worked just fine back in 2013.

The people's habits were the only obstacle: owners wanted to retain their own battery pack, instead of getting pre-owned one left by somebody else, in unknown condition. And/or drive over a weird contraption.

[1] (loud music warning) https://www.youtube.com/watch?v=H5V0vL3nnHY


I wonder why it bothers people about battery packs, but not propane tanks. They seem to have equal chances of exploding.


Propane technology has matured and stabilised; a 2019 propane tank is no better than a 2009 propane tank, it'll be the same size, supply the same amount of energy, and have the same connectors. A 2029 propane tank will, in all likelihood, be the same.

The same cannot be said of EV battery technology.

So why the Tesla demo video? According to [1] just doing the demo entitled them to $90 million of Californian "ZEV credits", and they never believed in or planned to roll out the technology.

[1] https://www.quora.com/Whatever-happened-to-Tesla-battery-swa... apologies for the Quora link


What if some jerk replaces half his cells with rocks then swaps the pack and you get it?

You have to trust the charging station that you’re getting an equivalent pack in trade. Are incentives aligned to ensure that?


Gogoro has been quite successful with this model.

The difference, of course, is that a scooter takes batteries small enough to swap by hand.


The Gogoro model is really cool and I would love to see it expand to cars. I’m just providing an example of why people might object to the idea.


This could definitely revolutionize the electric car market by removing one of the main complaints, slow charge time. The other pending issues for even greater mass adoption are availability of charging stations & range. If charge time decreases, the turnover of cars at stations could dramatically increase thereby improving the economics of charging stations. We could see more charging stations then and more quickly remove these barriers.

Has this AI approach been used in other charging applications? Curious to hear more details on what's going on behind the scenes.


If you could recharge while waiting in line for fast food, that would eliminate the need to ever go to a 'gas station' at all. These guys should partner with sonic or something, drive-in style.


UBeam et al have demonstrated that flat out ignoring physics can be an extremely reliable way to raise funding and score press visibility. Here’s hoping this won’t end up on the same list.


Do people in the US take 15m to visit the pump?


No, not really. The trade off I make for my Tesla is that I spend 0 time at the pump 359 days out of the year and about 45 min at the pump on 6 days when I take a long trips. Though, on those days, we are in the restaurant eating while it is charging, not just sitting there staring at it.


Hi, I'm one of the founders at GBatteries - sorry, the article had a mistake on publishing; should be 5 minutes for that example.


No, it's usually more like 5 minutes


When I'm on road trips with my friends, I can never get them out of a rest stop faster than 25 minutes. When I'm on the road with my wife, 15 minutes is about the right time.


That's the time it takes to visit the convenience store. Relevant in some circumstances, but if you're not already in need of a break you only visit the pump and even five minutes is oddly slow. And the article and company site specifically say the time to fill the tank, not to visit a gas station. So they're blatantly lying.


For my personal situation, I can't fill the tank in under 25 minutes on road trips with my friends. So it works for me.


Are you sitting at the pump the whole time?


In the case of my friends, I'm sitting at the pump or standing in the store texting my friends. In the case of my wife, I'm waiting for her to get out of the restroom.


there are two states (out of 50) that do not allow drivers to pump their own gas, so in that case, yes, it can take more than 15 minutes to visit the pump.


Have you seen the size of our vehicles? Of course! ;)


At a Canadian Costco to save 10%-15%, yes.


If this works as advertised, it’s super cool. However, my anecdotal experience suggests that things that charge faster heat up, and I know that my phone warms up significantly on with the “fast charger” it came with. Knowing that batteries are potentially very volatile and prone to exploding and leaking, how safe is this technology?


All modern EV’s except the Leaf have liquid-cooled battery packs with active cooling from the radiator and the A/C compressor.


Are there going to be any issues with peak demand charges from utilities making fast charging cost prohibitive? I have to imagine that pulling that much power from the grid will either be very expensive or will require some grid upgrades to prevent issues.


The obsession with fast charging is getting tiresome.

Here in Norway we have more electric cars per head than anywhere else and most people wake up to a fully charged car.

It is really rare to need to charge the car at a public charger unless you have nowhere off road to park your car. What is most needed in most places is somewhere to charge the car overnight or while you are at work. Such chargers need no new technology, no extra infrastructure beyond a simple socket in a post beside the road or in each parking space in car parks.


Fast charging is needed for road trips, and for people who don't have access to a charger at home. (Think about people who rent or who need to park on the street)


A lot of countries are much denser and don't have off road parking where you can run a cooker circuit 415v to


Why don't we just do removable batteries?


Company would have to own the batteries, users have no incentive to treat them well. Tragedy of the commons.


How does one treat a battery poorly?

I mean, I know in the abstract sense that it's ideal if they ever get fully drained and, but that's not all that catastrophic. I'd consider it normal use in most cases—if manufacturers don't want the charge to ever dip that low, they should make the battery stop working earlier.


I presume you mean don't get fully drained. They also should not get fully charged either. They could make the battery stop working earlier and limit them from being filled, but this is going to limit a lot of one time use cases when the trade off is worthwhile.


> They could make the battery stop working earlier and limit them from being filled, but this is going to limit a lot of one time use cases when the trade off is worthwhile.

Which brings me back to my original question: how does one treat a battery poorly? Or perhaps more accurately, how would people treat batteries they rent more poorly than ones they own?

In a world where consumers were highly cognizant of maintaining battery health, and only fully drained/charged their batteries when absolutely essential, I can see how rented batteries would be a problem. In practice, I doubt many consumers actually think about this, if they're even aware of it to begin with.

Ergo, I don't forsee rented batteries dying significantly sooner than owned batteries, because most consumers aren't going to change their behavior.


I expect most people would fill up to 100% when not swapping.


Tesla have investigated this a bit.

It appears that it's for some of the same reasons why people don't want to just swap their car - when you have a large asset that can be in a variable condition, you don't really want to be swapping it for another one randomly.


They should track the rating of your personal battery management better. This is just an information sharing problem.


Is the sort of assertion you can make, but need to test with customers. Tesla did test with customers and it didn't hold up.


> as compared to 15 miles in 15 minutes today.

My Tesla Model 3 gets 75 miles in 15 minutes at the supercharger.


Might be naive here, but isn't there some sort of limitation on the battery side as well?


I hope people realize this "article" is nothing more than a PR puff piece.


If a car's battery system is not set up for this kind of a charge rate (IE: it maxes at 7.2 amp 240 volt level 2), there is no way their system can just attach to your port and charge your car faster. There are on-board controllers etc. that are controlled (for good reason), by the car/manufacturer that would need to be bypassed.




Applications are open for YC Summer 2019

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

Search: