What are Issues of Transforming Robots or Robots that Combine into Bigger Mechs?

[Johnny1A.2]

Originally Posted by The Colonel

There's also issues like the requirement for massive engineering redundancy in any single transforming vehicle as mechanical loads change between forms. This goes up by another exponent if you then assemble the vehicle into a load bearing part of something else.
Also, the most important part of most vehicles is the pilot - if you assemble five vehicles into one, you've got four guys sat their like drunken monkeys whilst the fifth does all the work.

In the case of Combatra from the old comic that I cited, IIRC the pilot was in the head, as is usual in humanoid mecha (though that, too, makes little logical sense in itself, the chest would be more logical), and when the mecha split up into five, the pilot flew the head-aircraft and AI systems few the other machines under the general direction of the pilot, but there was provision for living pilots if they were available.

As I noted, using that set of assumptions, I could imagine that the pilot might link the vehicles up physically when not in battle or just in transit from point A to point B, just to simplify operating the whole thing. But why does the assembly need to be humanoid?

Originally Posted by Disliker of the mary sue

Hm interesting, The only justification for human like giant robots I found I liked is that in systems like in pacfic rim where you control the robot with your body/mind, that it easier for the human mind to grasp the idea to walk like a human rather then having to learn an entirely different method of walking. But brain controlled machinery seems to be a thing that won't be a thing before at least tl 9 to above if it even a thing that is possible.

Oh, it's almost surely possible. Difficult to implement, but there's no reason to think it can't be done, and it might even eventually be useful.

But here's another subtle point about giant mecha that is often overlooked. The 'human shape controlled by a human brain' idea is often used as a justification, but physics says no. It's the square-cube law again, that same old rule that messes the giant ants from the 50s SF movies.

Any object, when you increase its size by a factor of 'x', but keep the shape the same, increases its area by a factor of x squared, but it's volume by a factor of x cubed. If you have a cubical box 1 foot on a side, its surface area is 6 square feet and its volume is 1 cubic foot.

If we double it to 2 feet on a side, now its surface area is 24 square feet, and its volume is eight cubic feet. If we go to 3 feet on a side, now the surface area is 54 square feet and the volume is 27 cubic feet. At 4 feet on a side, the area is 96 square feet and the volume is up to 64 cubic feet.

See how the ratio changes? We go from 6-1, to 24-8, to 54-27, to 96-64.

At 5 on a side, now the ratio is 150-125, a 6 on a side it's 216-216, at 7 it's 294-343...

What that means is that mass is going to rise way faster than linear size. A humanoid shape twice as tall masses 8 times as much, so it needs a power source 8 times as large to get the same basic performance. But its ratio of surface-area to volume is only a fraction of that difference, so it has a harder time getting rid of excess heat.*

Structural strength rises as the square of the area of a supporting member, but mass rises as the cube, so a leg twice as big as a normal man's is only 4 times stronger, but supports a mass 8 times as great. A leg 3 times the size of a normal human leg is 9 times as strong (for a given material) but now has to support 27 times the weight.

What all this means is that a human-shaped mecha 30 feet high, or 50 feet high, won't move like a human, even if it's shaped like one. A human brain trying to apply human-scale motive instincts to a machine 50 feet tall will get...weird...results, at best, because it's as if he suddenly weighs 578 times as much but is only 40 times as strong. Leverage will be different, response time will be different, too.

My arm is a tad over 2 feet long. If I swing it in a round circle, my fingertip describes a circle roughly 14 feet in circumference. Let's say I take 1 second to do that, which is easily done. That means my fingertip is moving at about 14 feet per second or roughly 10 miles/hour.

Now let's say I'm piloting a 50 foot humanoid mecha and execute that same maneuver. My mecha-arm is now almost 18 feet long, it describes a circle not quite 56 feet in circumference. So in that 1-second circle, my mecha-fingertip is moving at 38 miles/hour instead of ten. So the kinetic energy is out of scale with my human form, the reaction time is off, the leverage is different because my 'body' is relatively massive compared to my strength, etc.

A human would have to relearn every motion to operate a 50 foot mecha. Nothing would come easily or naturally, even if the shape was exactly proportional.

*Remember that excess heat from the power plant? The bigger the human-shaped mecha, the worse that issue becomes, because that energy has to be radiated from the surface of the mecha, but every time the size doubles, the area increases by 4 and the volume (and heat) by 8, so the surface temperature to get rid of that heat has rise and rise and rise...

[RyanW]

The short version: each form of a transforming vehicle is going to have different requirements, so it's always going to have to be burdened with equipment that is useless in its current form.

Sometimes you hear the argument that higher technology will relieve this problem. The same argument against mechs (versus tanks) applies: Nearly anything that improves the inferior design could be applied instead to the already superior one. If A>B, A+N>B+N.

[Anthony]

Originally Posted by Johnny1A.2

But here's another subtle point about giant mecha that is often overlooked. The 'human shape controlled by a human brain' idea is often used as a justification, but physics says no. It's the square-cube law again, that same old rule that messes the giant ants from the 50s SF movies.

Actually, it's not the square-cube law -- the square-cube law is only a barrier if your artificial muscles and bones have the same performance as real muscles and bones. If your artificial materials have ten times the strength of human tissues, you can achieve ten times the size (thus, after applying the square/cube law, the 10x size version is 100x stronger because of increased cross-section, 10x stronger because stronger materials, for a total of 1,000x stronger, which makes up for being 1,000x heavier).

However, the human shape controlled by a human brain model still doesn't work, because the speeds are wrong. A human at a brisk walk takes about 0.5s per step, and each step is a bit under a yard. A 60' giant would take about 1.6s per step and each step would be ten yards. Reflexes that are suited to 0.5s steps aren't really very good for 1.6s steps.

[Johnny1A.2]

Originally Posted by Anthony

Actually, it's not the square-cube law -- the square-cube law is only a barrier if your artificial muscles and bones have the same performance as real muscles and bones. If your artificial materials have ten times the strength of human tissues, you can achieve ten times the size (thus, after applying the square/cube law, the 10x size version is 100x stronger because of increased cross-section, 10x stronger because stronger materials, for a total of 1,000x stronger, which makes up for being 1,000x heavier).

However, the human shape controlled by a human brain model still doesn't work, because the speeds are wrong. A human at a brisk walk takes about 0.5s per step, and each step is a bit under a yard. A 60' giant would take about 1.6s per step and each step would be ten yards. Reflexes that are suited to 0.5s steps aren't really very good for 1.6s steps.

I mentioned the speed too, though I used arm motion as an example.

But the square-cube still applies, it's a difference of degree. Yeah, stronger materials and power sources and power trains can make up for it, to a point, but the same technologies can make non-human-shaped vehicles and machines more effective too. The conceit of humanoid mecha is that you can scale up human interaction to giant machines, which doesn't generally work.