User:Brianwcallaghan/myWork

Project Goal
Build a working example of a Theo Jansen StrandBeest type walking robot, powered possibly by wind.

My First Task
Draw a 3D model of StrandBeest in 123D.

Summary of actual work over first weekend
I began by attempting to learn AutoDesk 123D. Initially, I made some rough drawings and started to get some knowledge under my belt. But, as I tried to grasp the geometry of the mechanism, I felt that I needed something physical and abandoned the 123D for the time and constructed a cardboard mock-up of the leg mechanism instead. I found a great video of a similar prototype someone made from wood. Check that out here. I found lots of other examples of uses for the leg mechanism. If we had more time and I knew how to weld, I would love to try to build one of these

Week1 Narrative
Began work in AutoDesk 123D. My first impression was that it seemed to be an update of Inventor. My knowledge of AutoCAD may actually have been a hindrance. I kept trying to relate the functionality of the program to the way things work in CAD and found the two programs to be quite different.

The drawing tools were similar to AutoCAD, but the layout is different. There is no command prompt which means clicking on a lot of menus and buttons. I prefer the efficiency of typed commands, but I can see the usefulness in some situations. It took me a while to discover the units in the lower right hand corner of the screen. After finding that, I was able to change from cm to inches, etc. Found that drawing on a 2D plane is more effective. Then you can extrude later.

After 4 or 5 hours of experimenting in 123D, I began construction of a working cardboard prototype of on leg pair mechanism. I found geometry for the parts online and cut pieces from a piece of cardstock. I assembled the parts with staples that I cut in half to make pins. This unit did not function very well, so I decided to make another out of better materials.

For my second attempt I used a thicker piece of cardboard, and fixed the whole mechanism to a cardboard back plate. In addition, I used small screws instead of staples to make the joints. However, the screw used to hold the connecting arms to the center wheel/handle for the crank was too small to properly grip and turn. In addition, the screw in the center wheel itself would back out of the back plate after a few turns. Lastly, the screw hole at one joint on the left leg, made the cardboard weak and it would bend forward while turning.

I addressed the issues by replacing the center screw with a flat headed thumb tack that I soldered on the other side of the back plate as a permanent shaft, replacing the crank handle/connector screw with a roofing nail. I put a small dab of solder on the outside of the nail to hole the arms in place and put a cardboard bracket across the entire top section to keep it from bending. The final product works great.

Finally, I tested the walking curve of the leg mechanism. To do this, I tacked my prototype to a corkboard and tacked a piece of plain white paper behind the feet. I then marked with a pen, the point where the foot rested. I turned the crank approximately 15 degrees and marked the new position of the foot. I repeated this process until completing 1 full turn of the wheel. The resulting curve was not the optimal walking curve. It would probably achieve movement, but very small, jerky, and inefficient steps. Learned later from team mate Matthew Wnuk, that the geometry I found was not the same as the measurements on Jansen's own website. Info on Matt's research can be found here; Mwnuk.

My Second Task
For my next task I will complete 2D drawings of the parts needed to assemble our Beest. We will use 12 individual legs/6 pair. These drawings are to be imported into EnRoute software at work (Maryland Sound) and cut from wood on the CNC router. I will attempt to get wood donated to our cause from my employer. If not, we will have to ask the school to procure wood for us.

Summary of actual work over second weekend
I did pretty much what was planned. I drafted the parts for our StrandBeest in AutoCAD 2012. I also learned a little about what I need to do to get the drawings into EnRoute.

Week2 Narrative
I Began by taking the measurements of the parts from Theo Jansen's website. I then derived the angles of the triangles with a triangle solver app on my iPhone. After drawing the parts, I arranged them into a 5 foot by 8 foot rectangle to fit the sheet size for the CNC router.

Some online research and a phone call to a co-worker taught me that I need to save the file as a DXF, which can I did in AutoCAD. I now need to open the .DXF in EnRoute at work and set the toolpath. The MultiCAM router will then cut the parts.

My Third task
I will cut out the parts on the MultiCAM CNC router at work. I will then try to assemble a working set of legs to show the walking curve in class.

Summary of actual work over third weekend
I was able to accomplish everything I planned.

Week3 Narrative
I made the final tweaks to the files for the CNC router to cut. Spent several hours rearranging parts to fit the wood. Once everything was set and the toolpaths were added, I made a test cut with the router. This showed that I needed to add small bridges to hold everything together on the table. With bridges added, I began the cutting of the parts. The bridges were cut by the machine on the second pass, so they would be thin and easy to remove.

I assembled one leg linkage as an example. The geometry seems off. The inner connectors bump into each other during movement.

My Fourth task
Finalize leg mechanism.

Summary of actual work over fourth weekend
Plan was to finalize the leg mech. I think I came close.

Week4 Narrative
Made some changes to turning radius measurement. A slight reduction improved the movement drastically. Made a note of this measurement on all the connector parts.

Complete Project Page
Follow the "project done" format.