I'm not sure if there's an inherent advantage to the smaller modules on a clean sheet. I think they all use very similar high power transistors to manage the power, so I'd suspect the cost would scale fairly consistently.

Tesla derives two advantages from the modular approach for their specific case.

First, the modules they use in the Superchargers are the same as the onboard chargers they install in every car, so they presumably get big economy of scale cost benefits since they are building hundreds of thousands of them per year now.

The second benefit is granularity in charging. The electronics are connected to a pair of stalls, and by having 12 discrete modules, it has the ability to connect some to one outlet and some to the other, letting it max out the first car to plug in and give everything that's left to the second car, so they get twice as many plugs for a given input power and almost the same investment cost. Charging is slower for the second half of the cars if the first group of cars are low on charge, but it's much better than being left waiting for the other session to finish.

 

10 kW modules. That's the total Power, but no mention of open DC voltage or amperage curve.
10 x 40 VDC modules in series would have an output of 400 VDC.
4 x 125 VDC modules would be 500 VDC open terminal voltage.
The current is what is controlled to maintain the kW amount during the charge process.

I suspect it is very complicated to 'One-off' a home brew DCFC.

Besides, what home are you going to plug this 416 Amp / 240 VAC device into?
I'm pretty sure DCFC units require 3 phase power.

Edit: I got to take a look inside an Efacec 50 kW DCFC unit.
These are as big as a refrigerator, but there is lots of empty space on the inside.
Down at the bottom is a very big and heavy transformer. Big wires too!

 

Tesla's approach is the other way - 12 ~10-15 kW capable modules connected in parallel, each outputting up to 400 V DC as required.

 

I'm not really considering this for home use. This is more to inform myself about the logistics of building out high-speed public charging infrastructure. As saghost mentioned, Tesla benefits from an economy of scale, but to build out the Supercharger network as effectively as they have indicates that it is a viable, economic strategy.

 

multiple modules does seem to make a lot more sense since if one module fails it's unlikely to bring the whole system down where if one module were used, it would render that charging station dead. Another way to possibly look at it is how supercomputers are built. They don't use one super fast processor but rather dozens or hundreds all working together. From a final cost standpoint, it's not cheap for either, but from a redundancy standpoint a failure of one or two smaller components is unlikely to impact the entire system and cause a total failure.

 

I don't think the equipment cost is a big deal. The major cost is getting that much power to the charging station. After construction the big driver for operating costs is demand charges which you can mitigate with battery storage. Locating the chargers would be 1000X easier for electrical utilities.

I'm also under the impression that the CCS/CHAdeMO chargers have only one module.

 

If you really wanted, you could use the new Gen2 chargers from Tesla that produce 18kW each. Just a small stack of six will get you ~100kW, depending on your input voltage. Keep in mind, you would need ~500A on single phase to run it. If you take it to 208VAC three phase, it drops down to ~90kW, but you need only 200A service with a 480VAC transformer. Six chargers would split nicely with 3phase, as well with two chargers per phase.

You'll need a 000 gauge busbar to connect the DC side, and the water cooling can be done outside the garage. In the winter, you can even use it to heat the house.