Hacker News Reader

Comment by nine_k | Hacker News Reader

OK! Here comes a car with, say, 80 kWh battery. It's not entirely empty, so to reach 98%, it can take 60 kWh. It does so in 5 minutes.

This means that it gulps 60 kWh in 5 minutes; since 1 hour = 12 * 5 minutes, the power transmitted is 60 * 12 = 720 kW.

This is a scale of a neighborhood; a typical detached house is allocated 20 kW. And this is just one car.

Now let's imagine optimistically that the efficiency of the process is 99%, and a mere 1% is dissipated as heat. 7.2 kW take a serious cooling system to dissipate.

So no, I don't think it's fit for cars. For phones and drones, maybe, using a special actively cooled charging station.

View on HN

First, ultra fast charging is really relevant only for intermediate stops on long-distance travel. There you have dedicated charging stations which are directly connected to high voltage lines which can deliver 10-100 MW. Modern charging stations already deliver up to 350kw, doubling that power wouldn't be sci-fi.

But the peak charging rate is only one parameter and it isn't necessary to reach that in day to day usage. The biggest problem with current batteries is, that the maximum charge rate is good already, but only available while the battery is below 50% charge. So it isn't necessary to raise the maximum charge speed much, if this speed can be maintained for longer time. A battery which can take like 200-300kW till over 80% charge would be a winner and mostly "solve" recharging during trips.

Most batteries also degrade faster when charged at high rates. So at home, when you have the luxury of charging overnight, it’s better to charge slower.

Right. And if you have batteries, which could take 700kW peak, but are only charged with 350kW peak, it is no longer "fast charging" in the sense of impact onto the battery.

But the big future is anyway to have the car plugged into a modest outlet when parking during the day so they can take up any excess solar power available on the grid and possibly even partially power your home over night.

> Modern charging stations already deliver up to 350kw

Peak under optimal conditions ? Seeing how most stations fail to deliver on max charging claims I'd be sceptical we can realistically go so high without "lab like" conditions - available in 2 places in the world at 2 times so they can claim the number.

Well, I was referring to the charging infrastructure and all current Tesla superchargers installed can deliver up to 250kW and Ionity here in Germany always installs 350kW chargers - that the battery of current cars cannot accept that for long or at all is exactly the topic why this new battery technology might be a big improvement.

> ”Modern charging stations already deliver up to 350kw, doubling that power wouldn't be sci-fi.”

Not at all. In fact, MCS (Megawatt charging system) is already in development for charging heavy commercial vehicles. It’s designed to scale to charging rates as high as 3.75 MW!

https://en.m.wikipedia.org/wiki/Megawatt_Charging_System

The average power consumed by an electric car doing average miles is a few hundred watts. It's well within the capacity of the grid to supply. The issue you're highlighting is that fast charging makes the power usage very bursty with high peaks.

So why not smooth it out? Put another battery in the charger which slowly replenishes. Then the grid doesn't see the bursty load you're talking about.

But then we need to wait for the charging station to charge. How many of those batteries would a charging station need to ensure continuous service?

Likely we'll see such a system with electric motorcycles (and similar categories) first, which is like a "small scale" version due to the smaller battery capacity & size.

E.g. Honda is working on it: http://www.honda-mpp.com/in/

It doesnt need to completely disconnect from the grid, just smooth the load as much as is economicallly useful. A load that is already statistically smoothed by multiple cars charging at once, arriving and leaving at different, but statistically predictable, times.

And people are already attaching batteries to the grid to earn money by smootging supply and demand, it just makes sense to locate those near to demand so they can share cables. Then to the grid the car charger station looks like a consistent energy user that can flex a little to help the grid (possibly even feeding back during peaks), rather than a random one.

Another battery is another 10-15% loss of efficiency, which makes charging 10-15% more expensive. Far better in most situations to distribute the spike it to the grid.

> So why not smooth it out? Put another battery in the charger which slowly replenishes.

There are 1.4B vehicles one earth, the plan apparently is to replace them all with battery powered vehicles

I don't think people understand the scale of this and how many batteries we can produce and recycle given the limited resources we have.

> 1.4B vehicles one earth

Most of which spend approximately 90% of their time stationary and unoccupied.

I've never seen an empty gas station during the day in my city.

You can't have batteries to charge batteries, then you'd need batteries to charge the batteries charging the batteries. All of that because we simultaneously decided that nuclear was bad and all we need is windmills, which are like the worst thing to power a fully electric world with high demand peaks, and to move from gas to electric cars

Adding batteries won't solve any problem. Gas allowed us to live unsustainably and made it easy to do so. Wanting to replicate that with electricity/batteries is plain stupid.

If everyone needs three tonnes of steel and their own body weight in rare earth material to cover 30km per day on average we're totally fucked from the get go

Locally, the same 90%. It's no use sharing a car with someone who needs to commute somewhere else at the same time, you need to share it with non-conflicting use.

Nobody needs to charge their car at home in five minutes. You charge your car at home over night. The typical driver drives less than 100km a day, so needs less than 20kWh per day. The car is parked over night for around ten hours, so it takes about as much power as a kettle to charge it. Fast charging is for long highway trips where dedicated infrastructure can be constructed.

You math it a little confusing when it come to the cooling required. You scale up the power draw to an hour, but then take your cooling based on that. But each individual battery still only needs to dissipate the heat produced during its charge - so in your example, with 60kWh and 99% efficiency, the energy to dissipate is 0.6kWh per charge, over the duration of 5 minutes. Which sounds much more manageable than the 7.2kWh you end up with (which would be for 12 charges over the course of an hour).

He scales the kW/h hour down to five minutes to make kW/5min.

0.6kWh / 5min * 60min/1h is still 7.2kW. It only needs to handle those 7.2kW of heat loss for five minutes, not continuously, but the power doesn't change.

The power doesn’t change, but you get 12 systems that each need to dissipate 600Wh instead of a single system that needs to dissipate 7200Wh. You can for example partially buffer the heat if you know that you only need to handle that load for a few minutes. It’s much harder to sustain the same energy dissipation for an hour.

According to [1] petrol cars' fan-cooled radiators can dissipate 50kW

Dissipating a mere 7.2kW sounds achievable, by automotive standards?

[1] https://file.scirp.org/Html/6-1520083_67628.htm

> 7.2 kW take a serious cooling system to dissipate.

FWIW, gasoline car radiators can expend upwards of 100kW of heat. After all, internal combustion engines turn most of the energy in gasoline to heat.

Yeah, given that at least some battery packs are already liquid cooled, 7.2kW doesn't seem too bad.

Apples and oranges. At the temperature difference between the coolant of combustion engine to the environment heat dissipation is much faster than at the temperature difference between the coolant that is acceptable for batteries to the environment.

And the surface area of a battery pack is orders of magnitude larger than that of an ICE, which greatly helps the thermal transfer.

For a different comparison, 7.2kW is roughly equivalent to 20x 3090 GPUs. The area of a battery pack is (much) larger than the size of 20x GA102 dies. We've got a lot lot of experience extracting large amounts of heat from very tiny spaces. It's an engineering challenge, sure, but it's very well within our means.

I don't see the charge rate as a problem. [The new MCS standard for trucks will charge at around 3 MW](https://en.wikipedia.org/wiki/Megawatt_Charging_System), so clearly it is entirely possible to deal with even larger cooling needs than a 720kW system would entail.

There are already superchargers with multiple chargers that can in total handle way more than 720 kW.

The biggest issue when there are many super fast chargers operating at 1000v will be handling voltage spikes in the power grid.

I think they were suggesting a potential difficulty in cooling the cells, not the connector. Trucks may have 3MW but divided across many more cells.

The key here is the charge curve. We currently have 350kW chargers but most cars can only take advantage of that for ~10 minutes before dropping. This is why most EVs take an hour to charge to 100%. By the time they get to that 98% value the charge rate has slowed to 10s of kW.

If you could do 98% of the charge at 350kW, that'd be ~15 minutes to hit 98% at 350kW. Today, it takes around 30 minutes for most vehicles to go from 0->80%.

And the same tradeoffs would still apply, you'd still probably want to time your stops to charge from 10-20% to 70-80% when on a long trip to minimise total travel time and slow charge to 100 before you leave and after you arrive.

It would just be even faster than today. Which is why I'm suspicious of any announcement that talks about the unrealistic 0-100% use case. If that is the metric they are chasing they're probably not going to help the average road tripper much compared with more focused engineering for the real use cases.

Something as simple as multiple cables that mean cars can charge without waiting for the others to finish and their owners return, or smarter navs, or cheaper/bigger batteries, building more chargers etc. is likely to have more impact than a magic new battery chemistry, though every different aspect helps.

Wouldn't be possible to have very large capacitors that act like a buffer? Maybe charge the capacitor for an hour and when a fast charging car arrives, simply use the stored energy in the capacitor?

This doesn't solve heat dissipation.

Supercaps aren't really anywhere near high-end batteries in specific energy or density. Additionally, capacitors have a linear voltage curve which makes it impractical to use their full stored energy. You could try and work around a lot of the problems by jacking up the voltage very high, but supercaps break down at a relatively low voltage, and have poor self-discharge. It seems like you would need an absurdly large capacitor bank to store enough energy to make it worthwhile, and it would be incredibly inefficient.

You could use batteries as the buffer, though. Most batteries support higher discharge rates than charge rates, which is nice for longevity. You would need a battery pack the same capacity as the car's in order for it to really make sense. You would also probably use lower density cells too, since they don't need to move. They wouldn't fit into the user terminal, so you'd probably locate them in a nearby shack, and run very thick cables. If you had more than one charger, you would probably have an even larger shack. Before you get to millions of battery cells to manage, you might step back and realize that something like a giant flywheel would make more sense, and then you're only one or two steps away from building a power plant.

Hmm, kind of sounds like the battery replacing solution but instead of physically replacing the batteries, you simply move the energy from the buffer batteries to the consumer batteries.

You can build battery towers or underground battery storages depending on the physical constraints maybe? If its feasible, it can also be beneficial from renewable energy standpoint as they also need the storage.

One thought I have is that current battery technology depends on one nasty, rare chemical or another. It's difficult to extract from the Earth, possibly at the expense of human rights in some parts of the world. It's difficult to recycle also, to the point that it seems like a lot of companies are probably just quietly stacking up piles in warehouses to possibly but probably never deal with.

There are obvious benefits for convenient portable consumer devices, though. The benefits seem to outweigh the downsides, and all things considered, there's not much raw materials per-capita.

Electric vehicles are cool and fun to drive, but the benefits seem less obvious. Gas powered cars work very well, and if you look at the entire carbon cycle, it seems like they actually aren't all that bad relative to EVs. A lot of raw nasty materials to build EVs, by many orders of magnitude per-capita.

Going to batteries for wide-scale energy storage, for a vast EV charging network, just seems like a massive waste of resources. I'd rather those precious raw materials be used to build smart phones, sonic toothbrushes, and revolutionary new stuff like delivery drones, for an indefinite amount of time until a better battery technology can be developed.

We already have 350kW fast chargers, so this is just another factor of 2.

> So no, I don't think it's fit for cars.

Why? If it enables a flat charging curve at the existing 350 kW and 400 kW EV chargers then that's still a big win.

> ”Now let's imagine optimistically that the efficiency of the process is 99%, and a mere 1% is dissipated as heat. 7.2 kW take a serious cooling system to dissipate.”

Much of that heat is dissipated by the charger unit, not the car. The charger has massive heat sinks and fans etc, so not really a big deal!

The rest is just warming up the battery, which is a big heavy block which can store a lot of heat. This can be dissipated over time, which is actually very useful in winter! (heat pump scavenges battery heat to warm cabin, etc)

Are there any other reasons to use this battery tech beyond rapid charging time? Does it have more charge cycles than other Li batteries when charged at a more reasonable rate? Is it denser or does it have better discharge characteristics(i.e. lower internal resistance)?

Looks like they used a pretty standard lithium ion chemistry with high silicon anodes.

This is where the industry is currently going with state of the art lithium batteries. They've just been slower about increasing the silicon due to fears of what other tradeoffs will arise.

It's supposed to have better safety characteristics:

https://www.enovix.com/BrakeFlow/

Upon further reflection, I think the answer to this problem is to simply coordinate charging 'slots' on a sub-grid(not sure what you call a local delivery circuit with a certain capacity). It should be possible to request charging time slots from a local power company server and negotiate when your vehicle can charge with millisecond accuracy. Doing so also might help the power company supply that power more effectively/efficiently. Sure, this means in peak times you may not get the charging rate you want, because you have to wait for the network to be available, but in off-peak times you get the fast charging rate you're capable of.

It's chicken and egg. It will come, our grids are not ready, but perhaps these batteries can also discharge fast and you can charge your home battery during the day, charge your car when you come home, charge it through induction as your drive over a special track of highway. Many things are possible in the future, but we have to grow into it. In any case, this super fast charging is mostly relevant (as said in this thread) road side, where you can have special equipment. At home you can mostly charge over night.

The latest tesla supercharger has a liquid cooled cable and this sort of fast charging is for on the road. At home you charge at a max of 10kw~

What if the car battery is removed from the car and put into a charging station with cooling?

If external cooling becomes the solution (which makes a lot of sense) then I think it's more likely that a car will be provided with a water supply during charging - a water connection near (but not too near!) the electric plug would let the charger pump water through the battery pack and suck a bunch of heat from the car, without a complex battery removal workflow, and without the heavy cooling mechanism having to drive around with the car.

If this were to happen (and I don't think it will), then you'd want that water connector to be part of the charging cable to avoid a two step hookup. That requires a new charging cable standard (which is why I don't think it'll happen).

Also, water kinda sucks to deal with for something like this. What if, for example, the water flowing in isn't pure? What if it's hard water? Your battery cooling tubes could need flushing and might start growing nasty fungus and such.

You could do a heat exchanger - no mixing of the water then but added complexity.

You can easily dissipate 7.2 kW outdoors, just blow a bunch of air through.

ICE cars have to dissipate more than 100 kW of thermal energy and do so quite fine. The 7.2 kW are easy to cool in comparison. Of course that requires substantial fans as the cars are not moving while charging, but that is why they can get quite loud while charging :)

Reply on news.ycombinator.com