To get a torque output from it you need a proportional contact force to whatever you attach to it. Eventually you’re just twisting the contact area.
You’re limited by the strength of the thing the cube fits into. It’s like a 1/4" socket. It can handle a lot of torque, but eventually the steel gives way. That’s the thing you need to maximize.
There are only four in existence, so that’s the limit of how many you can line up. The torque is still limited by whatever you use to connect the cubes to each other.
If you have access to the material the cubes are made from, it’s a relatively easy problem.
But the generator only needs a finite torque to move and we can make things that handle that. The axles is turning at 10 RPM. Period. Always. No reason that has to break anything.
So what if the axle loses no rotation force when attached to something?
If you attached a 500 HP engine to a house fan and it was geared so the fan only received 10 RPM to the fan blades why would that break?
500 HP would not be transmitted to the blades in that case.
If you plunged the fan into something much more viscous than air, like pitch, then it might actually require significant horsepower. Except that the blades or axle or something would obviously break.
What power are you expecting to get from this? What’s the equation for power from rotation? Divide by the given RPM and you have the torque to generate that power. Instead of a cube, assume it’s already a shaft. Calculate the contact stress.
You can back out the maximum power from the limiting contact stress of whatever material you want.
Maybe if you have some electromagnetic coupling rather than physical you can do something there.
Am I missing something here? I’d say that it is turning quickly.
The generator needs a finite number, materials can provide a finite number, so what’s the problem? The problem is that there’s more than one finite number. Eventually, the finite number the generators need is more than the finite number real-world materials can supply.
That would be a very big, high-power generator at that point. You could do an awful lot, with just one Axle. But still only a finite amount.
I have a tool that turns at 10,000 RPM. Compared to that, 9 RPM is, in fact, very, very slow.
By my calculations, a single corner moves in a circle at about 0.11mph. That doesn’t seem very quick to me.
Why not 150 x 75mm? I mean, I appreciate that the contact patch nearest the axis of rotation won’t contribute as much as to the effort but may as well use everything you have.
You haven’t grasped the essence of the problem. The problem isn’t the power it’s the shaft.
If you have a 500hp engine with only a 1/8 inch (3.2mm) output shaft, what is the maximum size fan you can attach to it and turn at 10rpm? Clearly the limiting factor is not the engine power but the shaft. You can attach an enormous fan that would in theory take 500hp to turn at 10rpm, but it will do you no good because long before the engine is putting out maximum power, the shaft will break.
This question is asking the same thing - you have an infinitely powerful “engine” but only a 6 inch by six inch square cross-section output shaft - how much power can you take off that shaft before whatever connection you make to the shaft breaks?
Yeah, I could have done that, but the inner part contributes significantly less less in a non-linear way and I didn’t want to do the math 
. The 70% derating is also wrong, since it actually varies from 0 degrees (at the center of each face) to 45 degrees (at the edges). Like any Fermi problem, I tried to keep things simple and threw out the low order terms.
Yes, I thought of Roadside Picnic as well. I expect PTerry and Reynolds both read this story.
Would you describe the New Scotland Yard sign as turning slowly?
The DW Axle makes a complete turn in about the same time the NSY sign turns one face.
That would be slow for a blender.
That would be fast for a clock.
My apologies. Recalculating the cube speed, a point on the outer surface is moving at roughly 0.22 mph.
Assuming one face of the NSY sign is six feet wide (and I’m being conservative), a point on that face is moving at roughly 1.69 mph. Considerably faster.
But you could attach a wheel to the axle, and it would still turn once every 6.9 seconds.
Imagine a merry-go-round turning at that speed, one complete rotation every 6.9 seconds. Would you think it’s slow or very fast?
Naturally, the other part is how to anchor the other cube stationary to get power out.
The calculations for the “socket” are the same, but then you need to build a connection that doesn’t rip out of the ground when 70MW torque is applied. So you have a pure diamond pyramid socket - how many bolts of what diameter need to anchor it to what sort of base or wall to ensure that the support structure does not rip out of the ground or shred?
IANAEngineer, but I’m guessing the answer has to do with shear resistance of diamond vs. steel?
It’s kind of slow for a merry-go-round. Kind of fast for luggage carousel; about right for a windmill; way too slow for a helicopter; way too fast for a radio telescope.
Fast and slow are not absolute terms.
The smart thing to do would be to give it a clutch that would slip to protect the axle interface, rather than letting the axle destroy whatever you use to contain it. You could also make both sides of the axle rotate, and extract energy in opposite directions. Then, if one side is overloaded you’ll just slow down (speed up? I’m not a mechanical engineer) the other side rather than destroy the entire assembly.
Something like that is probably a practical idea to avoid accidental damage to the equipment, but it won’t change what the ultimate limit is.