Well, it probably wasn't that much effort. When you're 3D printing you're going to end up printing everything 2-3 times anyway, so why not dial in the ratio while you're at it?
And you can't really declare your design is "high precision" and present yourself as someone others should take transmission design advice from if you aimed for a gear ratio of 8 and achieved "somewhere around 7.9 to 8.2"
It probably doesn't matter so much whether it's 7.913 or 8.186, but it would be important to know the exact value for kinematics. One way to do that is to build an object very accurately, the other is to build inaccurately and then measure the result after the fact.
It's also interesting because competing actuators with strain-wave, cycloidal, or planetary gearboxes will state exactly what the ratio is. The actual gear teeth may not be spaced out perfectly around the circumference, but the number of teeth is an integer with an infinite number of zeros.
> Also, I wonder how resistant this mechanism is to wear and fatigue.
He actually discussed this in an earlier video for his initial tests on the capstan drive. He ended up testing the rope he used for around 358 hours (two weeks) on continuous use in the drive itself with very low backlash
Because when a real engineer puts 2 and 2 in and gets 3.8 out, it vexes them and they want to at least know why they can’t get 4. He’s trying to make a machine that does what he told it to do, so that he understands what is actually happening.
I think it's about kinematics, the more precise your gears the better the model fits the real world.
That's why pro crews don't use gears and ropes. At high impulses deformations and elasticity throw the kinematics off what's actually happening. Modeling the deformations and the elasticity is a computational no no. Instead what you see is the motors right on the joints.
At least that was the case last time I had a look at robotics.
> That's why pro crews don't use gears and ropes. [...] Instead what you see is the motors right on the joints.
The answer here, as with so many things in robotics, is: It Depends.
UR10e robot arm that can lift a 4kg object with a reach of 1m and has sub-1mm repeatability? Strain wave gears in the base and shoulder joints, 100:1 ratio.
MIT Mini Cheetah robot dog that can do backflips? 6:1 planetary gearbox.
Shadow Hand with 20 degrees of freedom? Tendon driven, with the 20 motors in the forearm to keep the fingers slim.
Little dinky Huggingface SO-101? Servo motors, integrating 1:345 gearing with a series of 6 tiny brass gears.
Mid-price CNC milling machine, if you call that a robot? Really long ballscrews, driven by stepper motors.
> At high impulses deformations and elasticity throw the kinematics
Sure. Yet evolution has achieved astonishing kinematics with all manner of deformation and elasticity inherent to the materials, and also constantly changing physical properties, using low resolution data. We cannot build permanently lash free mechanical devices at reasonable cost and reasonable size/weight. Eventually, the answer must be pervasive real-time compensation throughout the kinematic model.
> Modeling the deformations and the elasticity is a computational no no.
Why? Nervous systems do this. That's why you can change your shoes and still walk upright.
In the video he mentions how the specific type of Dyneema cord he's using is well-suited for the application (compared to other kinds of rope/cord). It's particularly strong, light, and inelastic; a lot of climbing equipment uses a version of it for similar reasons.
> Modeling the deformations and the elasticity is a computational no no. Instead what you see is the motors right on the joints.
That sounds like “It’s not wrong, we just don’t do it”. There are some amazing examples of imprecise drive systems compensated for by excellent control systems all over the world, for millions of years.
I don't think the number of the gear ratio really matters, what matters is that you know what it actually is (since every IK calc depends on said ratio); 8:1 is probably arbitrary and/or looks nice & might simplify some stuff.
michaelt|7 months ago
And you can't really declare your design is "high precision" and present yourself as someone others should take transmission design advice from if you aimed for a gear ratio of 8 and achieved "somewhere around 7.9 to 8.2"
LeifCarrotson|7 months ago
It's also interesting because competing actuators with strain-wave, cycloidal, or planetary gearboxes will state exactly what the ratio is. The actual gear teeth may not be spaced out perfectly around the circumference, but the number of teeth is an integer with an infinite number of zeros.
ErigmolCt|7 months ago
jedimastert|7 months ago
He actually discussed this in an earlier video for his initial tests on the capstan drive. He ended up testing the rope he used for around 358 hours (two weeks) on continuous use in the drive itself with very low backlash
https://www.aaedmusa.com/projects/capstandrive
https://youtube.com/watch?v=MwIBTbumd1Q&t=10m
hinkley|7 months ago
throwawayffffas|7 months ago
That's why pro crews don't use gears and ropes. At high impulses deformations and elasticity throw the kinematics off what's actually happening. Modeling the deformations and the elasticity is a computational no no. Instead what you see is the motors right on the joints.
At least that was the case last time I had a look at robotics.
michaelt|7 months ago
The answer here, as with so many things in robotics, is: It Depends.
UR10e robot arm that can lift a 4kg object with a reach of 1m and has sub-1mm repeatability? Strain wave gears in the base and shoulder joints, 100:1 ratio.
MIT Mini Cheetah robot dog that can do backflips? 6:1 planetary gearbox.
Shadow Hand with 20 degrees of freedom? Tendon driven, with the 20 motors in the forearm to keep the fingers slim.
Little dinky Huggingface SO-101? Servo motors, integrating 1:345 gearing with a series of 6 tiny brass gears.
Mid-price CNC milling machine, if you call that a robot? Really long ballscrews, driven by stepper motors.
topspin|7 months ago
Sure. Yet evolution has achieved astonishing kinematics with all manner of deformation and elasticity inherent to the materials, and also constantly changing physical properties, using low resolution data. We cannot build permanently lash free mechanical devices at reasonable cost and reasonable size/weight. Eventually, the answer must be pervasive real-time compensation throughout the kinematic model.
> Modeling the deformations and the elasticity is a computational no no.
Why? Nervous systems do this. That's why you can change your shoes and still walk upright.
BHSPitMonkey|7 months ago
PaulDavisThe1st|7 months ago
the motors were so sloppy the company wasted a ton of money [0] having me write heuristics to tackle the errors they accumulated over several hours.
one of his whole points is that by using dyneema (rope), there's almost no elasticity at all in the capstans.
[0] relative to the cost of better motors
mlhpdx|7 months ago
That sounds like “It’s not wrong, we just don’t do it”. There are some amazing examples of imprecise drive systems compensated for by excellent control systems all over the world, for millions of years.
skeeter2020|7 months ago
Isn't having more decimal places the exact definition of precision (vs accuracy)?
munchler|7 months ago
sabareesh|7 months ago
dvt|7 months ago
ErigmolCt|7 months ago