TWiki > Rise Web>ClimbingRobot>RiSELegDesign (19 Oct 2004, MatthewMessana? )
RiSE Leg Design Considerations

Introduction

RiSE is a platform of climbing robots. The research group at Stanford is primarily focused on leg and foot design while Boston Dynamics is responsible for the mechatronics.

While several different foot designs are being pursued, the legs are in need of a new design. The following is picture of the current leg.

Most of the work on the current leg has been integrating the electronics. This includes, but is not limited to calibrating and troubleshooting the strain gauges (used for feedback). The current leg is very bulky and its design can be significantly improved with materials readily available in the lab. These materials are advanced fibers (i.e. Kevlar), Urethane via SDM and conventional graphite-epoxy composites.

This page will begin by outlining some of the design considerations for the three main parts of the leg and continue to track the progress of prototypes to be built.

Spars

Current Design Issues

The “Spars” are the aluminum top and bottom portions of the leg, which are connected together by two springs. Note that the top portion is actually supplied by Boston Dynamics, so our design will only consider the bottom spar and how to interface with the top spar.

The basic goal for the new design of the spars is to lighten them. Also, it would be nice to make the spars easy to manufacture. The current design is all aluminum and requires a relatively large amount of machining for how small it is.

The spar contains several mounting holes that interface with the “ankle,” which connects to the foot. The screws for these mounting holes add a significant amount of weight and it would be nice to eliminate them.

New Design Considerations

Materials

Graphite-epoxy composites seem to be the most promising replacement for the aluminum spars. With a proper design, they can be lighter and stiffer than the aluminum. Fiber reinforced composites are difficult to work with and will require a significant amount of consideration to produce a part that is simple to manufacture. Depending on the epoxy resin used, an autoclave may be needed to cure the part. For testing purposes, a room temperature cure epoxy will be suitable. This will allow for easier manufacturing. When a final design has been determined, more advanced epoxies that require an autoclave can be implemented.

Another possible material would incorporate SDM. Some of the urethanes used (such as the IE-72DC from www.innovative-polymers.com) could be stiff enough for this application. The urethanes could be further stiffened via embedded fibers (little work has been done pertaining to this). SDM holds the promise of molding the part all at once.

Design

The overall shape of the part will be very similar to the way it is now. Ideally the mounting holes will be eliminated and replaced with adhesives. This will help reduce the overall weight of the part by eliminate the relatively “heavy” screws. However it will eliminate the versatility that we currently have, where we can easily interchange ankles and other attached parts.

If graphite epoxy is to be used, a rectangular tube will form the main part of the spar. Since holes in composites significantly weaken them, mounting holes will either be eliminated or metal collars embedded for holes.

If the part is made from urethane it will likely be a solid just like the aluminum. Urethane is not as brittle and is thus less affected by holes. The ankle could either be cast at the same time as the spar, or attached the same way as it is now.

Manufacturing Difficulties

(comment on the difficulties of carbon fiber and SDM manufacturing here)

Springs

Current Design Issues

The current springs have a couple of problems. The curved spring (yields in the y-direction) does not strain linearly with applied force. We think that the spring kinks at a certain point, changing the geometry and thus the spring constant. To fix this we can either change the type of spring or the geometry of this one.

Also, the current y-direction spring is not stiff enough according to JohnathanAmory? of Boston Dynamics. He recommends replacing it with a straight (coil?) spring to correct this.

The spring that provides flexion in the x-direction is adequate right now.

One advantage of the current geometry is that the strain sensors are nearly decoupled. To maintain this, the new design must have hinges in the same places as the current design.

New Design Considerations

The first thing that needs to be done is create some design requirements. The desired deflection and stiffness of the springs has not been nailed down as far as I can tell. Once that happens, some simple calculations can be performed to determine the type and geometry of the springs for the new design.

The springs could be either a coil, a curved beam or a straight beam. Ideally the geometry will not be changed significantly (i.e. a curved beam will provide y-direction stiffness and a straight beam will provide x-direction stiffness). Thus the interface with the strain gauges and the electronics will remain the same.

Materials

Spring stiffness is proportional to the Young’s modulus of the material. Steel has one of the highest modulus to density ratios, which makes it ideal for this situation. The problem with steel is that it would be difficult to integrate directly to the composite or urethane parts.

A spring could also be formed from urethane using SDM. This would be very easy to implement into an SDM spar. Urethane however is not very stiff and would require more material. It is likely that a urethane spring would be much heavier than a steel spring with similar stiffness.

One other material to explore is a composite spring. Composites are typically very brittle, however, using the right matrix material and fibers COULD result in a very light spring. This will require much more research and design.

Shape

The current springs need to be longer to eliminate kinking.

Hinges

Current Design Issues

The current hinges are bulky and take up too much of the spring length. This is very heavy and causes problems with kinking of the curved spring.

New Design Considerations

New hinges could either be made out of urethane (via SDM) or fabric (Kevlar?).

Materials

New hinges could either be made out of urethane (via SDM) or fabric (Kevlar?).

Prototypes

Composite Spars, Metal Springs, Kevlar Hinges

As I see it, this is the most optimum design as of now. It is NOT very versatile however. If spring constants need to be changed, a whole new leg must be built. Also, ankles will be glued on to the spar and be difficult to change. I have several ideas on how to manufacture this, which will become clearer as I complete some CAD drawings.

Spars

The spars will be made out a rectangular graphite-epoxy tube.

Springs

The springs will be the same as the current design. The top will interface to the main robot by bending the end of the spring into a small loop that a pin can pass through to form a hinge. See CAD drawings when they are available.

See hinge section for information about the bottom of the spring.

Hinges

The hinge for the curved spring will basically be two small pieces of Kevlar fibers. The spring will be sandwiched between them. The Kevlar will be stitched (as in “sewn”) onto the springs. The hinge will be coated with epoxy and cured with the spar, forming a single spar-hinge-spring.

CAD Drawings

(soon to come) 5.2 All SDM Leg (some thoughts to come later)

-- MatthewMessana? - 19 Oct 2004

 
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