6 ms·

Drives like an electric, refuels like a liquid fuel car (it uses a compressor, not electric charge).

bryanlarsen 5y ago
If by "electric" you mean "Prius", you're right. Tesla's and similar cars have great acceleration because they support massive currents at high voltages. They need this to support fast charging, the incremental cost of adding a high torque motor is small. Hydrogen cars have tiny batteries that don't need to support fast charging. Like a Prius.
- hwillis 5y ago
  > Hydrogen cars have tiny batteries that don't need to support fast charging. Because that's what they think consumers want, not because of technological limitations. Size is not a requirement for high power batteries. The Porsche 918 has a 6.8 kWh battery that can do 210 kW. That's ~31 kW per kWh, compared to a top-end Model S at ~6 kW/kWh.
  - bryanlarsen 5y ago
    Yes, a $800K Hydrogen car would probably have a small but high power battery. But Nexo's and Mirai's don't.
    - hwillis 5y ago
      High power batteries aren't like high power engines. They're built on the same production lines. High power engines prioritize power over cost; high power batteries just prioritize it over capacity. It's not challenging to make a cheap, small battery that can provide 200 kW. It just happens that it hasn't been prioritized. A provable example: 1,826 Sony VTC4 18650 cells will push 200 kW continuously. That's only 14 kWh of storage and weighs <100 kg. It's a 28.5"/72 cm square 3"/7.5 cm thick. I could personally build that battery for <10k USD. A full scale automaker should be able to do it for half that. That would almost triple (136 -> 336) the Mirai's power and put it (2050 kg @curb) in spitting distance of the 550i (390 kW, 2020 kg @curb). The BMW's MSRP starts 27k USD higher. Of course you would need to change the suspension and geometry completely, make the interior far nicer, and add even more power to be able to actually compete with the 550i, but the point stands. A small, powerful battery is cheap and easy to add to a FCV.
      - bryanlarsen 5y ago
        If you are wiring all 1826 cells in parallel, that means you need 1826 battery controllers to achieve reasonable life and reliability. Plus the hassle and cost of using bus bars instead of wires. And it means that the battery module output is 3.6V at an insane amperage and needs a massive transformer for a reasonable voltage.
        hwillis 5y ago
        > If you are wiring all 1826 cells in parallel, that means you need 1826 battery controllers to achieve reasonable life and reliability. NB that I'm an electrical engineer and have built many batteries before, including from 18650s. First off, it is absolutely not necessary to have so many controllers; cell-level controllers are only used in the cheapest batteries. High-quality batteries will match cells individually and switch them in groups, which reduces the losses from running a bunch of parallel circuitry. This is not really different from simply using larger cells, as long as you can match them sufficiently well. Cell-level BMS is very cheap as a result, for example this single cell chip that costs 7 cents each: https://lcsc.com/product-detail/PMIC-Battery-Management_HY2213-BB3A_C113632.html https://lcsc.com/product-detail/PMIC-Battery-Management_HY22... > Plus the hassle and cost of using bus bars instead of wires. Only if you're using 18650s; if you are using prismatic cells you just solder or crimp the tabs to the HV supply. If you're using small-format cells, you would typically used stamped sheet current carriers rather than busbars. > And it means that the battery module output is 3.6V at an insane amperage and needs a massive transformer for a reasonable voltage. It certainly does not, because batteries are DC and you do not use transformers to convert their voltage. Transformers require you to convert to AC and back, but you can use an inductor ~100x smaller by using one of the many "boost" topologies. The basic idea is that you only need to convert to AC, and you don't need to rectify back to DC. You chop the high-current DC up very finely, then store 1-3 cycles in a very small inductor. If the inductor is sized just right, it will drain out current that is very close to DC. Naturally though you would really want to operate between the 245 volts of the Mirai's battery or the 650 volts of the Mirai's fuel stack. That would be something between a 68s27p pack and a 180s10p pack. That's far more efficient. You can still have cell-level BMS for every single cell- that IC I linked above is daisy-chainable, meaning it does not need a real 0v ground reference. It's perfectly happy working at whatever voltages on the two ends of the particular cell its connected to. In this case I'd opt for something closer to 68s- the more cells you have in series, the more susceptible your design is to degraded performance over time.

If by "electric" you mean "Prius", you're right. Tesla's and similar cars have great acceleration because they support massive currents at high voltages. They need this to support fast charging, the incremental cost of adding a high torque motor is small. Hydrogen cars have tiny batteries that don't need to support fast charging. Like a Prius.

hwillis 5y ago
> Hydrogen cars have tiny batteries that don't need to support fast charging. Because that's what they think consumers want, not because of technological limitations. Size is not a requirement for high power batteries. The Porsche 918 has a 6.8 kWh battery that can do 210 kW. That's ~31 kW per kWh, compared to a top-end Model S at ~6 kW/kWh.
- bryanlarsen 5y ago
  Yes, a $800K Hydrogen car would probably have a small but high power battery. But Nexo's and Mirai's don't.
  - hwillis 5y ago
    High power batteries aren't like high power engines. They're built on the same production lines. High power engines prioritize power over cost; high power batteries just prioritize it over capacity. It's not challenging to make a cheap, small battery that can provide 200 kW. It just happens that it hasn't been prioritized. A provable example: 1,826 Sony VTC4 18650 cells will push 200 kW continuously. That's only 14 kWh of storage and weighs <100 kg. It's a 28.5"/72 cm square 3"/7.5 cm thick. I could personally build that battery for <10k USD. A full scale automaker should be able to do it for half that. That would almost triple (136 -> 336) the Mirai's power and put it (2050 kg @curb) in spitting distance of the 550i (390 kW, 2020 kg @curb). The BMW's MSRP starts 27k USD higher. Of course you would need to change the suspension and geometry completely, make the interior far nicer, and add even more power to be able to actually compete with the 550i, but the point stands. A small, powerful battery is cheap and easy to add to a FCV.
    - bryanlarsen 5y ago
      If you are wiring all 1826 cells in parallel, that means you need 1826 battery controllers to achieve reasonable life and reliability. Plus the hassle and cost of using bus bars instead of wires. And it means that the battery module output is 3.6V at an insane amperage and needs a massive transformer for a reasonable voltage.
      - hwillis 5y ago
        > If you are wiring all 1826 cells in parallel, that means you need 1826 battery controllers to achieve reasonable life and reliability. NB that I'm an electrical engineer and have built many batteries before, including from 18650s. First off, it is absolutely not necessary to have so many controllers; cell-level controllers are only used in the cheapest batteries. High-quality batteries will match cells individually and switch them in groups, which reduces the losses from running a bunch of parallel circuitry. This is not really different from simply using larger cells, as long as you can match them sufficiently well. Cell-level BMS is very cheap as a result, for example this single cell chip that costs 7 cents each: https://lcsc.com/product-detail/PMIC-Battery-Management_HY2213-BB3A_C113632.html https://lcsc.com/product-detail/PMIC-Battery-Management_HY22... > Plus the hassle and cost of using bus bars instead of wires. Only if you're using 18650s; if you are using prismatic cells you just solder or crimp the tabs to the HV supply. If you're using small-format cells, you would typically used stamped sheet current carriers rather than busbars. > And it means that the battery module output is 3.6V at an insane amperage and needs a massive transformer for a reasonable voltage. It certainly does not, because batteries are DC and you do not use transformers to convert their voltage. Transformers require you to convert to AC and back, but you can use an inductor ~100x smaller by using one of the many "boost" topologies. The basic idea is that you only need to convert to AC, and you don't need to rectify back to DC. You chop the high-current DC up very finely, then store 1-3 cycles in a very small inductor. If the inductor is sized just right, it will drain out current that is very close to DC. Naturally though you would really want to operate between the 245 volts of the Mirai's battery or the 650 volts of the Mirai's fuel stack. That would be something between a 68s27p pack and a 180s10p pack. That's far more efficient. You can still have cell-level BMS for every single cell- that IC I linked above is daisy-chainable, meaning it does not need a real 0v ground reference. It's perfectly happy working at whatever voltages on the two ends of the particular cell its connected to. In this case I'd opt for something closer to 68s- the more cells you have in series, the more susceptible your design is to degraded performance over time.