Leaf Spring Mounting Cad and Actuator Concept

Progress 10/15/2014

Series Elastic and Parallel Leaf Spring CAD

Figure 1: Pass-through design for series elastic spring (knee)

I based the series elastic spring designs on the CAD files that Nitish Thatte (in our sponsor Professor Geyer's lab) provided for his research prototype. The design insight that I took from his work was to pass the series elastic springs through empty space in the joint (Figure 1) so that length of the spring can be chosen independently from the length of the moment arm.

Actuation Concept

Based on the CAD models of the leg with structural, encoder, series and parallel elastic, and joint components included, I developed a concept for the actuation of each joint using a motor, gearbox, and ballscrew (Figure 2). The respective motor attaches proximally on each limb component, and the output of the motor is rotationally coupled through the gearbox to the ballscrew. The ball nut is affixed to a hollow tube that serves to house the ballscrew when the actuator is fully contracted. This hollow tube attaches to the dorsal end of the series elastic spring via bearings. At the proximal end of the actuator system, the motor is attached to the limb through a compliant connection (such as a hinge or a flexible member) that allows rotation through a small angle. Finally, while waiting for delivery of the structural and joint components, I continued fabricating parts using the two Makerbot Replicator 2X printers in lab. I fabricated two of every part necessary to assemble the knee and ankle joints with encoders attached.

Figure 2: Actuator system conceptual design

Challenges

Limited Thigh Length

The length of the thigh component is limited by the length of the amputation stump. When the added length of the socket, padding, and prosthetic leg coupling are taken into account, the thigh may need to be as short as 10 cm from knee joint to beginning of amputation stump. With such a short thigh length, attaching a sufficiently capable actuation system presents many difficulties. To begin to shorten the distance required, the knee actuator will have the motor and ball screw mounted in a reverse fashion as shown in Figure [fig:Actuator-system-conceptual]. The extra pair of gears involved will have to be carefully chosen/designed to avoid introducing additional backlash into the system.

In addition, although the thigh mount for the knee actuation system must be located close to the knee, the motor and proximal end of the ballscrew may lie parallel to and overlapping with the prosthetic stump as long as they do not physically interfere. While this design relieves the space restriction somewhat, rigidity of the actuation chain may be harder to maintain.

RP Part Tolerance

To avoid the extra cost and time required to fabricate parts in aluminum, I have been relying heavily on 3D-printed parts. These parts perform sufficiently well except in terms of clearances for shafts and bearings. These dimensions demand precision. Currently, I use a sanding drum on a Dremel tool to remove material until the bearing or shaft fits into the part. However, if I continue to rely on RP parts, I will need to determine the error in the printing process and alter the CAD model appropriately to produce parts with more accurate dimensions.

Source: https://sites.google.com/site/mrsdproject2...