Graded Materials Graded Materials Very Common Nature Few Examples of Functionally Graded Materials in Man-Made Assemblies.

Graded Materials

• Graded Materials Very Common Nature

• Few Examples of Functionally Graded Materials in Man-Made Assemblies

Graded Materials

• SDM Allows Variability in Compliance and Damping throughout a Candidate Design

Shape Deposition Manufacturing

• Cycle of Material Deposition and Removal• Complex 3D Geometry, Multi-materials

Part

Support

Deposit (part)

Shape

DESIGN

PROCESSING

FABRICATION

DESIGN

Original Design SDM Re-Design

•Rocker Pin Joints Replaced With Flexural Regions to Introduce Compliance and Damping

Benefits of Graded Materials

•SDM with Graded Materials Allow Control of Material Location and Properties in 3D geometry•With Graded Materials, Flexure Size Can Be Increased To Increase Strength of Part

Graded Materials• Un-Actuated Five-Bar Leg Mechanism Illustrates

Benefits of Heterogeneous Material Properties• Flexure Joints Replace Pin-Joints to Add Compliance

and Damping

Graded Materials• Desired Performance of Structural and Flexural

Regions Very Different• Fabricating With Single Material Would Result In

Compliant Structural Regions or Brittle, Failure Prone Flexures

• Ideal Solution Requires Varying Material Properties Between Different Regions of the Part

Graded Materials• Graded Interface Increases Surface Area, Resulting in

Increased Bonding• Mixing in Arbitrary Ratios Not Possible• Function Needs to Be Applied To Discretize the Graded

Regions Based Upon a Specified Tolerance Parameter

BondingWet-Wet Bonds:•Excellent Bond•Difficult To Control•Avoid Over-mixing

Wet-Dry Bonds:•Excellent Control of Surface•Increase Surface Area•Roughen Surface•Keep Clean - machine as last step before new pour if possible

Wet-Dry BondWet-Wet Bond

Material SelectionPolyurethane (part):•Hardness: Machinable/Cast Only•Bond Issues - cure time, shrinkage (#’s)

IE-90A

IE-70DC

IE-65D

(Cast Only, Very Soft)

Material SelectionWax (support):•Machinability•Shrinkage•Melting Point

Red Wax

Blue Wax

Green Wax

Palette Blocks; Excellent Machinability

Castable; Medium Machinability

Water Soluble; Poor Machinability

Design Constraints• 2.5D/3D

• Tool Size Constraints

• Ordering

• Materials

Example of 2.5D/3D GeometryWhite Regions (Soft Material) in 2.5DClear Regions (Hard Material) in 3D

Ordering Defined By Urethane Hardness. Processed to Minimize

Machining on Soft Surfaces

Flexures

2.5D/3D

Increased Width

Maximize Bond Area

Materials

Minimize Sharp Edges

Plastic Fails in Tension

PROCESSING & FABRICATION

Pouring Maintenance & Technique

• Clean Technique:• Clear Division Between Clean and Unclean

Objects• Attire• Dry Gas

Under Gloves

Outer Gloves

Mask

Eye Protection

Toe-Covering Shoes

Lab Coat

Good Cures

• Aesthetic and Practical Importance• Pre-vacuum• Over Pour• Cure Times

Tolerance Issues

• Over Pour/Under Machine• Excess Material: Part Deformation and HAAS

Tolerance

Processing

• Combine Differing Grades of Material Roughly in Order of Hardness

• Bond Surfaces Should be Freshly Machined

Ideal Machining Parameters

• Please be Conservative (tool speeds very approximate)

• Cooling: always provide direct cooling from compressed air

Wet-Dry Bond

• Cutting Depths: Roughly Half Tool Diameter

• Feed Rates:

1. 500-750mmpm 1/8” or larger

2. 200-400mmpm 1/16”

Application of Graded Materials

• Berkeley 1 DOF Walking Machine• Four-Bar Linkages Represent Practical Application

Well Suited to Use of Graded Materials

Compliance for 1 DOF Machine

Original Design SDM Re-Design

• Reduce Assembly Complexity, Increase Robustness• Four-Bar Mechanism Utilizes Two Rotary Joints and Two Rocker Joints•Rocker Pin Joints Replaced With Flexural Regions to Introduce Compliance and Damping

Compliance for 1DOF Machine

• New Design Features With SDM

• Geometry: Constant Ground Contact

• Replaced Pin Joints With Flexural Region: Introduced Compliance & Damping

• Leg Preflexes Defines by Build Orientation

• Future Analysis and Experiments to Tune Compliance to

Locomotion

Acknowledgements

Thanks to Prof. Mark Cutkosky, Prof. Fritz Prinz, Sanjay Rajagopalan, Jorge Cham, Yanjie Sun, Jianpeng Dong and Sangkyun Kang and the

other students and staff of RPL and CDR for their help in generating the results described in this presentation and Dr. Noe Lozano and the SURF

program for supporting this research.