To better understand the solutions, here the characteristic angles of human:
Frontal rotation:
Pronation: 20°
Supination: 30°
Sagittal rotation:
Flexion: 40°
Dosiflexion: 20°
Motor + spring:
I keeps the motor for sagittal rotation and a passive I creates rotation with a spring for the frontal rotation. I don’t choice a spring because I have two configuration: Rigid ankle or pliant ankle. I will make many tests to choose the best configuration
The blue and yellow parts form a spherical joint. The blue one is fixed to the MX-28 motor while the yellow one is linked to the foot (green) though a load beam.
Zoom on the passive spherical link
The two springs in the sagittal plane should have a really high gain to avoid undesired motion while the two springs in the frontal plan should have the adapted gain to produce the desired reaction (to determine).
Also there is 2 solutions for the spherical joint depending on which part is convex/concave.
Limitations
I have 2 main doubts about this design:
How will react a 3D printed spherical joint ? Especially I have concern about the friction between the two parts.
How we can measure the actual rotation position of the foot ?
I have some ideas about that.
First, can we make it work with only one “smooth” part?
If so, we can just print the spherical joint and glue half of a ping-pong ball on one part.
Or we can also simply use a marble!
For the measure of the position, it is more complicated…
Indeed, we can maybe imagine a solution with rubber-like or ceramic material.
Here is a proposition using a composite revolute joint:
The load is supported by the white parts (polyamide or ceramic), the shape of these parts only allow a revolution joint. Two rubber element are inserted between the two parts to add damping and stability. Then 2 springs are added to provide the desired behavior. I think a whole structure like this can be really compact.
Rubber material is way too soft to be placed in the sagittal plane. You will lose the force transmission of the ankle motor.
Also can you redo your design with dimension closer to an integrable solution ?
Your solution doesn’t allow the frontal rotation, max 5° maybe so we need 30°. Else, the rubber matérial is a good matrial but it’s impossible to control the properties (temperature, humidy…).
In my solution the rubber matérial acts as a damper.
I modified the foot ankle with the new foot design.
I tested two differents springs and I add 2 springs per foot for better performance during rotation.
I put this solution on Poppy with shoes. Poppy can stand up and move slowly
would be great if you replace the spring by a 2 filament print, one of the filament being the elastic type.
This allow to get very smart shapes with the possibilty to have spring effect on several axis.
Basdically it is like to printend tendons and muscles.
and the best thing is you can print the part in one shot with all the hard and soft part together.
so we could get something looking as nice/slick as this
while inside complex as much as this
with the bottom textured for grip like this
my opinion is a flat base on foot is creating problems, because if the foot is not touching flat immediately it cause the assembly to move, hence the needed complexity with many axis.
this can be greatly reduced if the bottom surface is rounded like regular foot and made from soft material.
basically it is to compare the cow hoof and the tiger paw
Your idea is very good and you are right for the problems
When I created the foot, I have the same idea to add ruber in the bottom of the foot, but for now I didn’t implement this option. I will try to put the rubber in the second version and the new sensors (like this => http://rubber3dprinting.com/) But for the moment I don’t have 3D printer for testing my new prototype.
If you would contribute to ameliorate this foot, it is wonderful for me and the Poppy community
