Design of a Linear Actuator for motorizing the legs

I’m working on a design for a motorized leg. My idea is to integrate a linear actuator into the tibia to manage the ankle movement. The concept from James Bruton shown in his video seems to be feasible to integrate. Further, the current ankle design will probably not manage the dynamic load of InMoov if he starts walking. So the design of the knee looks very robust. 

I try to adapt the knee joint to the ankle to get a strong pivot point (D) and add a hub flange (S) as shown in the picture. The pivot point will be strong enough for the weight of InMoov and the hub flange is the access point for the linear actuator.

First Prototype

This is the first prototype of the heel with a very early version of my linear actuator. It’s not ready yet, but it’s a start. I'm testing different motors with different gears right now.

Next thing is a modified and combined version of the original ankle parts. The clamp for the 20x20 profile of the linear actuator goes to the ankle axis to reduce weight from the printed parts. 

Second Prototype

I checked my first prototype inside the lower leg and saw that the distance between the 20x20 profile and the lead screw and to the ball button actuator was too large to fit into the tibia-parts.
Therefor I changed my actuator design to be more compact. After seeing a video from Yens for a small linear actuator ( using two square tubes and a threaded bar gave me the right direction. 

My local hardware store offers square tubes with 20x20 mm and 16x16 mm. With a wall thickness of 1.5 mm, the first tube has inner dimensions of 17x17 mm. This should be an easy fit for the inner tube. If tests show a loose fitting, I think about covering the inner tube with strips of Teflon tape, as Yens already suggested.

While the outer tube is aluminum, the inner tube with the required dimensions is only available in steel. Despite of the weight, this is good, because the M8 hexagon nut can be welded into the square tube for a steady fit. My goal is to keep the inner tube as short as possible. I go for an experimental approach to determine the necessary length for the actuator. 

The whole design adds a real skeleton to InMoov. The lower 20x20 square tube is supposed to start on the knee joint and goes directly to the ankle joint to unburden the weight of InMoov from the printed parts. 

The left part is the motor flange for a 37D mm gear motor with encoder. The middle part is the holder for the outer tube of the linear actuator which can be secured from slipping by an M6 screw thru the top. The round fitting is for an M8x22 mm ball bearing for the M8 threaded bar. The right part is the holder for the other end of the outer tube of the actuator.  All three parts have another opening and a 6.2 mm bore so that they can be screwed with an M6x40 mm screw to the 20x20 square tube which acts as a bone from knee joint to ankle joint.

Third Prototype

The second actuator with the threaded bar was much to slow. With a pitch of 1.25mm per turn, it needed forever to extract the inner tube.

That learned, doing some calculation helped. Since the motor has an encoder with 64 pulses per revolution, a ball screw with a 5mm pitch has accuracy of 0.08mm which is more than enought. 

The 3rd design uses a cheap T8x2 Lead Screw I had in stock. My local hardware store offers a round aluminium tube with 10mm and a wall thickness of 1mm. Perfect for the Lead Screw which is 8mm thick!
The printed parts are even more compact with the advantage, that the block for the lead screw nut lies on top of the 20x20 square tube, so that it cannot twist.
The round tube is pushed out of the right holder, when the motor is turning.

more to come...