-- MatthewNorcia - 25 Jun 2009

Week of August 3

This week, I finished construction of the new test plane, then did some test flights and further development. The plane flies well, and is quite stable. The two propellers effectively reduce torquing on takeoff, and the plane is able to fly straight off the wall. We can also control the aelerons independently, so they can serve both as roll control and as the thrust vectoring that previously required additional nose flaps. The plane also uses actuated landing gear to push the plane off the wall and tilt it back. Overall, the new plane is far more controllable than the old one during takeoff. I also began developing a spine release mechanism using Nitinol wires for actuation. This system uses my arduino board (for now) and a power transistor to source current through the wires, which pull on a cam that disengages the spines. It is looking promising, and should be far lighter and more compact than the previous servo-actuated system. This will only be used if we decide to use a tail spine and roll the plane backwards, rather than use the power to aerodynamically maneuver the plane off the wall. The rolling method would probably be more difficult, but might allow us to rely on less thrust for takeoff.

Week of July 27

Monday: Made a new plane for testing. I used the dimensions of the Flatana, and made the plane out of depron sheet and carbon fiber. This is much lighter and more durable than the foam-core planes. The plane has all control surfaces, and re-enforced areas for mounting the landing gear.

Tuesday: Made a new version of the claw detachers and mounted it to the test plane. It works as it should.

Wednesday Morning: Tested out the plane -- it flies pretty well, but it went out of radio range and crashed. I then repaired the plane and began testing take-offs. I made new flaps on the front of the plane, right behind the propeller that use the wind from the top to push the front of the plane away from the wall. Initial testing of this method worked quite well, and I got a few successful takeoffs. The plane pulls off the wall and turns upside down. This was all without using a tail hook, however, so it relies on thrust/weight >=1. It is able to do it without hitting the prop on the wall though because of the new control surfaces.

Thursday: Further testing of takeoff. Took off quite successfully, right into Dharma's arm, so much of day spent seeking stitches.

Friday: Began construction of new double propeller test plane for takeoffs. We were having problems with the single prop model torquing as it came off the wall, so the new one has two fuselages, and two counter-rotating props, and is more durable than the old one.

Week of July 20

Monday and Tuesday, wednesday: Basically just manufacturing stickybot parts, and dealing with Pro-E problems. In down-time from stickybot operations, began design for sweet bdml gecko sdm glowing usb drives for the lab.

Thursday and Friday: While the Haas was down, I began working on Alexei's plane project. I made a prototype design for releasing the front claws while simultaneously engaging a rear claw. This method seems to work pretty well.

Week of July 13 Monday Created a full stickybot model in solidworks. This model includes a new body design, an updated version of our prototype legs, and both updated vitaflex force sensors and spring based sensors.

Tuesday finished the stickybot model and made the manufacturing models for the necessary machining steps. What a time consuming process...

Wednesday Machined molds for test parts for Noe and Allen, began a solidworks design for the new stickybot tail, created machining operation for carbon fiber reenforcements for stickybot. Spent hours getting petty cash reembursement.

Thursday: finished making the test pieces for Allen and Noe, attended physics SRC seminar on energy choices for the future, began machining new stickybot, but broke the endmill so must start over. Also, finished a design for new stickybot tail.

Friday: More CAD and CAM work on stickybot parts, and began manufacturing of stickybot parts

Week of July 6

Monday and Teusday: Finished fabrication and assembly of test leg. We are having trouble with the vitaflex 20. It does not cure fully, then splits and delaminates. It is also very goo-ey rather than springy. This may have to do with the tint that we added. Then, started experimenting with new materials to make stickybot out of. Candidates include permutations of urethane, carbon fiber, pulltruded carbon rods, plywood and balsa wood. We are making a standard test geometry with which we can quantitatively test the stiffness of the materials. In addition, I have begun designing force sensors that use preloaded springs instead of stretchy material. This should reduce on-center slop in the sensors, and improve the stiffness of the limbs under light load.

Wednesday: I made some stalks and a new top cap out of vitaflex with Noe, then worked on making solidworks models of the adjustable preloadable force senors, and modified the original test leg cad file to use shoulder screws rather than through bolts for the joints. We also laid up the test spars that use task 9. We had some issues getting it to cure again. I don't think that the task 9 is an appropriate material for the cores.

Thursday: Continued work on the solidworks files for spring based force sensors, and updated the scalloping and width of our polymer based force sensor to reduce delamination. Dharma and I went on a mission on search of food, composites and polyester. We had great success with the food, thanks to Dittmer's, but were less successful on the other goals. I then slow-cooked the pig legs over night.

Friday: BBQ in morning, then group meeting, then redid the scalloping CAD model to fix some referencing problems. I now have the Solidworks models ready to create a new test leg at the beginning of next week.

Week of June 28:

What a painfully inefficient week... We spent far more time trying to find a working solid-works license and struggling with Pro-E than actually designing anything, but so it goes I suppose. However it is unclear whether this has been resolved, as the solid works license expired on Friday. Also, two people working on one solid works is torturously inefficient. Anyway, this week we designed a test leg with two additional degrees of freedom over the previous generation sticky-bot. The new leg incorporates pull-in capabilities with force sensing, as well as rotatable feet, also with force sensing. Design wise, this is pretty simple -- just a reworked, reconfigured version of the old legs to incorporate the added sensors and actuators. I suppose the main goal of all this was for Dharma and I to get experience with the Solid-Works to Pro-E to Haas chain. We will surely redesign the leg next week when we find out that we overlooked some limitation of this system, but hopefully this will allow us to do so in less that a week.

Week of June 21:

First week in Lab -- basically got introduced to the lab and my summer project, then began designing, prototyping and testing potential foot and toe designs for the Gecko.

Administratively, I met with Mark and Sal to discuss plans for the next generation StickyBot, attended several kickoff events, and got set up on the computer systems.

For practical familiarization with the lab, Aaron showed me how to cast and assemble the hierarchical adhesives used on the Gecko, and we made new patches of the adhesives for testing our prototype fingers and hands.

Dharma and I began to explore foot designs that take advantage of internal forces to provide normal forces without requiring shear forces. We made several prototypes, first out of foam core, then out of laser cut plywood and telescoping metal tubing. The results were mixed. Our full-hand prototype failed to have good surface contact, but simple two-finger prototypes proved more successful. Using a simple prototype with two patches each of slightly under one square inch, sprin loaded to provide opposing force, we were able to achieve tangential forces of just over one Stickybot (about 400g), and pure normal forces of around 150g. This could be a useful improvement for a gecko that can climb in arbitrary directions, as it reduces the dependence on the tail. We also experimented with finger designs that utilized a layer of angled foam to provide an extra layer of suspension to the hierarchical adhesives. We explored the effects of varying several parameters, including foam stiffness, angle and shape, and found a setup that seemed to work quite well. It proved to be far more forgiving than the toes with only the hierarchical adhesives and no foam. We encountered issues with stability in some of the designs, but by only covering part of the finger pad with foam, we found that we could achieve both good stability and compliance.

I will be/ was out of town on Friday, June 25th, and plan to come in on Sunday the 27th to take advantage of unused Solidworks.

This site is powered by the TWiki collaboration platformCopyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback