Construction of Physical Prototype

Construction of 1^st prototype:

The first prototype was designed to be a bare-bones representation of the final system, with the main goal being to simply demonstrate the gait path with the Theo Jansen mechanism. 

Figure 1: Incomplete model of 1st prototype

The above photo is a solidworks representation of the first prototype. The links were designed in solid works with the distance between the holes specified by the link lengths that Sung’s kinematic analysis solved. We planned to use quarter inch bolts as joints so the holes for the links were sized to be a little bigger than a quarter inch to allow for minimal friction about each joint. The two supports are identical to each other and have slots at the bottom to allow for precision positioning with respect to the base (not shown). The two Theo Jansen mechanisms are attached to a crank that offsets the relative motion of each mechanism by 180 degrees with respect to the other mechanism.

Once the solidworks design was complete, the various pieces where organized into files to be cut out of quarter inch acrylic by the maker’s space laser cutter and then assembled by the team.

Figure 2:  Team assembling 1st Prototype

Figure 3: Assembling the Crank 

Once assembled, there were some evident issues with the functioning of the prototype. One of the major issues was that the back support of the mechanism was two bolts connected by a nut. Since the connection was not perfectly rigid, the back axle sagged which added friction to the overall mechanism and contributed to binding. Also, the crank was assembled using only the acrylic pieces, bolts and nuts with the frictional contact between all three types of pieces being the only thing keeping each piece of the crank from rotating relative to another piece of the crank. So, sometimes when rotating the crank the torque caused by the hand rotation would be greater than the frictional torque keeping the crank rigid, thus causing the two mechanisms to move out of phase with each other. Another byproduct of the poor crank design was that sometimes torque would only be transferred to one of the mechanisms. The final major issue was that the joints would sometimes bind up due to the rotation causing the nuts on each joint to screw along the bolt until they came in contact with the acrylic links. However, despite these issues, the mechanism would function smoothly a majority of the time.

Construction of 2^st prototype:

 For the second prototype we wanted to fix the issues we had with the first prototype and also add features that would further distinguish our device from the original Theo Jansen mechanism on which we based our design.               

Figure 4: Final Prototype SolidWorks model

As shown in the solidworks model, there were two the major additions to the device. One was the gear train that attaches to the crank and is powered by a motor and the model of a person which is used to show how a human will move inside of the device. The gear train is a 9 to 1 reduction that was designed using solidworks in a process that was outlined in a YouTube video (link in "shared Links" section).  The person was designed by in a similar fashion to the Theo Jansen mechanism in that the thigh and shank link lengths were determined by the kinematic analysis performed in matlab and the rest of the body parameters were chosen so the body would look somewhat proportional.  The leg is attached to the Theo Jansen mechanism which causes the leg to move in a manner consistent with the foot trajectory traced by the mechanism. The link that connects the knee toe the arm is used to move the arm in tandem with the legs.

The final prototype was assembled using the same links for the Theo Jansen mechanisms and the supports but the base, the gears, gear supports and person links were all new. The gears, and gear supports were cut from quarter inch acrylic in the laser cutter and the person was cut out of wood using the laser cutter. The gears were attached to their respective shafts using nuts and the pinion was attached directly to the motor using an interference fit hole cut into the pinion itself.The pin joints were created using bolts as done in the previous prototype, however, hitch clips were used to secure the links instead of bolts which eliminated a lot of the binding issues that existed in the previous prototype. The other major improvement was the use of a single rod as the back axle of the device which helped counteract the weight effects of the mechanism and further reduced the likely hood of binding issues. 

 The major challenge in assembling the final prototype was making sure that each gear was properly aligned with its mate. This problem was caused by the fact that the way material is placed in the laser cutter causes the edges of the cut parts to be at a noticeable angle. This artifact of the laser cutting process caused all the holes in the gears and gear supports to be at a slight angle making perfect teeth alignment between gears impossible. This problem was rectified by essentially forcing the middle gear supports into the desired orientation and then using a zip tie to lock the supports in place, thus ensuring proper gear alignment and near optimum power transmission. Another issue involved the person model and the links that induced arm movement. This issue was the result of a toggle point that the knee joint experiences at a certain point in its motion. This toggle point rendered the joint incapable of transmitting motion to the arm connecting link and the combination of the toggle point and the connecting link caused the knee motion to deviate from the expected knee motion calculated in the simulation. Because of this issue, the group decided to remove the arm connecting link and just have the arms be static components of the device. Once these challenges were solved, the device ran smoothly. 

Figure 5: Final Prototype with Impedance controller