Close

Project Log 32: Ok, maybe hydraulic legs ain't it.

A project log for DIY Mech/Exoskeleton suit.

Mechs are not viable, nor cheap, so I will try to design and build one alone anyway.

fulanodetailFulanoDetail 03/05/2023 at 18:230 Comments
Sunday, 15:04, 05/03/2023

Well... Sigh...

I wrote it as an edit on Project Log 31, but I'm writting a new Project Log just to "officialize" it, I guess.

Unless I made a big extra mistake on the equations/calculations on the subject, it seems my reallistic expectations on Project Log 30 for a reallistic mech were more feasible.

Again...


So, for context, I was recalculating the amount of muscles required for the mech, and one really important detail that I forgot throghout all these 30+ Project Logs is that the lower part of the body is 3 times stronger than the upper part of the body.

So, if you want to lift 100 kg, your legs must output 300kg. If you want to lift 1 ton, your legs must output 3 tons.

And this is also true for mechs, and even if I used more power on the legs instead of more muscles, the output would be the same.


Until now, the best bet would be Nitinol Shape Memory Alloy (there is other types of Nitinol, so this distiction is important), which, accordingly to the previous calculation I've made (that is more or less incorrect), I would need 3 kg of Nitinol for a mech/suit that can lift 100kg.

Nitinol SMA costs 500-700 dollars per kg, meaning I would need around 1500-2000 dollars (around 10,000 reais) to build a suit (without all the other issues related to control).

And McKibben muscles, on the other hand, besides being cheap and simple to build, they require a pump, and pumps are insanely expensive.

Just like the calculation I've made on Project Log 31, I would actually need around 2500 dollars with only the costs of the braided sleeves and hydraulic pumps alone.


So, it came to my mind that maybe it could be possible to make a homemade hydraulic pump with car/motorcycle engine parts, an hydraulic piston pump...

Until I looked at the equations, and I would require an even more expensive power plant to power it up. Around 40hp instead of the 1-2 hp.


So... Unless somehow a new kind of artificial muscle that is cheap and easy enough to build appears, I think I will be forced to make a threaded mech. 


And the tanks win over mechs on the real world again...


So, this is me making a tank-mech instead of a humanoid mech, everybody will probably have fun, but it is not the way I wanted it. >:(


Edit¹:

Ok, this time I actually tried to use as little actuators as possible.

So, no McKibben muscles for fingers, no Stewart platforms on the shoulders, and so on.

In the first attempt, it was:
+6 groups of actuators on each thigh with 3 muscles per group (Stewart platform would be strong enough to lift the body [I think]), and on each feet, giving 36 muscles in total.
Now, it would have more 9 muscles on the knee to go up and more 9 to go down, so 54 muscles in total.

The upper part would have +3 muscles on the shoulder, +2 muscles to go up and down, +2 muscles to move the forearm up and down.
No wrist, no finger etc.
14 groups in total with 3 in each group, 42 muscles in total.
96 muslces on the entire body.

You know the drill, third of a second, 380.16 liters per minute at 8 bars, which would consume more or less 0.8 hp (or 600 watts per hour), which could fit in a single 580 LPM pump.


Edit²:

It is not actually a pump that can output 580 liters per minute, it is just a pump that is called 580L/M

So all my calculations until now were useless.

I'm not okay.


Also, since this is all actuators fully activated at once, then it could be cut in half to 190 liters per minute.

If I connect opposite actuators, so when one unactuate, the fully pressurized fluid goes to the opposite side, I could cut it in half even more to 80 LPm.

... But I don't know if the force will be strong enough, unlike the hydraulic cylinders case, where the surfaces are different in size.


The second attempt was using even less muscles on the legs for just 3 groups on the hips and so on.

Trying to reduce as much muscle as possible McKibben:
+3 shoulders, +2 arms, +2 forearms
7
Muscles need three 10mm McKibbens
14x3 = 42
+3 on the hips, +2 on thighs, +2 on legs, +2 on the feet.
9
The bottom part needs to be 3 times stronger than the top part.
18x9 = 162
162+42 = 204
204 x 0.022 = 4.5 LPM
In a third of a second all of that would need 807.84 LPM, 400 if cut in half, as not all the muscles would be contracting (supposedly).


If I reduce the hydraulic cylinders as much as possible, I will have:

+3 shoulders, +1 biceps, +1 shoulder.
+3 thigh, +1 thigh, +1 leg +1 foot.
11 on one side = 22 total.

I need 0.141 liters in one third of a second.
558.36 in total.
If it is half 279.18 LPm
5 horsepower + pressure limiter on the top


Of course, these two options would be just like the Guardian XO exoskeleton, where you don't have hands nor complex body movement, it would move like a brick.

But it would move, and it would be feasible, unlike fully actuated body.

This would still consume more energy than any mech-tank or exo-tank suit (I think).


Discussions