Additive Manufacturing: State of the Art and Practical Design Notes

Transcription

1 Additive Manufacturing: State of the Art and Practical Design Notes Aaron M. Dollar John J. Lee Assistant Professor of Mechanical Engineering and Materials Science

2 SOA in Commercial Systems Same basic concepts as 1995, just a whole lot better (mostly) and cheaper!: Shooting lasers at beads (SLS) Shooting lasers at resins (SLA) Coiling melted plastic threads (FDM) Others Gluing together beads/powders (e.g. ZCorp) Inkjet-based deposition (e.g. lost wax forms)

3 Selective Laser Sintering (SLS) Basic concept: Vat of beads locally fused via laser heating Build plate descends through reservoir

4 Selective Laser Sintering (SLS) Pros: Different materials than other techniques Nylon, polystyrene Metals! No support material/waste Cons: Poor surface finish/resolution Machines are large/expensive Generally too expensive for Universities

5 Selective Laser Sintering (SLS) Commercial Systems: (many, with varying specific technologies)

6 Stereolithography (SLA) Basic concept: Liquid resin locally cured via ultraviolet laser Build plate descends through liquid reservoir OR Resin dispensed via nozzle, laser curing at exit

7 Stereolithography (SLA) Pros: High resolution No support material/waste Cons: Limited material choice (but expanding) Machines are large/expensive Generally too expensive for Universities Materials are unsafe/toxic

8 Stereolithography (SLA) Commercial Systems:

9 Fused Deposition Manufacturing (FDM) Basic concept: Thermoplastic is heated and extruded Coiled layer-by-layer to make the part

10 Fused Deposition Manufacturing (FDM) Pros: Inexpensive Safe/simple High part toughness Cons: Limited material choice (but getting larger) Requires support material/waste Limited resolution (~0.007in)

11 Fused Deposition Manufacturing (FDM) Commercial Systems: Stratasys! Acquired Makerbot, merged with Objet (polyjet) 3D Systems

12 Practical Notes Not just for (rapid) prototyping anymore! Can make: Fully functional mechanical systems Multi-material parts Commercial-grade components Commercially-viable manufacturing processes Small-batch fabrication Complex/lightweight geometries Continually gaining traction

13 Practical Notes Think before buying a machine Costs keep coming down, quality coming up Do you really need 0.001in resolution? X exotic material? YxYxZ build volume? Don t have X machine don t worry! Many easy-to-use, low-cost fab houses! E.g. Shapeways.com

14 Practical Notes Don t bother with Hobbyist machines (at least for now improving rapidly, though) Poor resolution/repeatability/mechanical props E.g. Makerbot Higher resolution = higher cost And also often lower mechanical strength You can do a lot with commercial FDM! Generally available in Universities Stratasys machines (e.g. uprint, Dimension)

15 Caveat Eng-tor (let the engineer beware) Myth: 3D printing makes mechanical design and fabrication easier Reality: 3D printing makes BAD design and low-quality fabrication easier Parts/systems must consider the mechanical requirements and system tolerances to work effectively Particularly for load-bearing and alignmentsensitive parts (e.g. motor/gear mounts)

16 Design for Additive Manufacturing Chapter in Additive Manufacturing Technologies Gibson, Rosen, and Stucker - Springer Much more to be done Moving target! Caveat Eng-tor (let the engineer beware)

17 Tutorial Overview Stretching commercial Additive Manufacturing technologies for functional parts/systems Multi-material Molding and Printing Complete Systems (Aaron) Fabricating high-quality/resolution components and systems using additive manufacturing (Matei Ciocarlie)

18 Multi-Material Molding and Fabricating Complete Systems

19 Yale OpenHand Project All our experimental hand projects now use similar rapidly prototyped designs Why not start releasing hands for free?

20 Yale OpenHand Project Inexpensive designs Off-the-shelf components Rapid Prototyping Promote fast, iterative experimentation with robot hardware Promote novel mechanisms Promote design sharing and reproducible experiments [Ma, Odhner, and Dollar, ICRA 2013]

21 Credit: Raymond Ma and Lael Odhner

22 OpenHand Model T Based on Harvard SDM Hand Four underactuated, interlaced fingers Optimized for power-grasping - Model T - SDM Hand

23 Inside the Model T Can be downloaded, printed, cast and assembled Complete IKEA instructions are available Videos made to clarify build process

24 Modularity

25 Model T.42 ( for two ) Independently actuated sides for two-finger tasks

26 Starting larger effort on Open Hardware Send us links to your Open Source projects!

27 Major Obstacle to Open Hardware: Easy, inexpensive fabrication Leverage and stretch the capabilities of common rapid fabrication technologies - Especially FDM and Lasercutting machines

28 Multi-material molding with FDM

29 Rubber pads and flexures Extend material selection for 3D printing Custom mold cavities, surface features

30 Basic steps 1. Print hard parts with thin-wall cavities for rubber components 2. Pour urethane into the printed cavities 3. Cut away cavity walls to finish part

31 Printing Pad Features Can customize surface features of fingers Pad ridge thickness ~ 1.5 mm

32 Printing Anchors Secure flexure joint and pad material to RP bodies Can embed anchors completely within urethanes

33 Sealing Printed Cavities Most adhesive tapes can be used to seal mold cavities Overflow/leakage is expected, not critical

34 Vacuum chamber & dessicator Polymerization produces outgassing To solve this, boil the mixed resin at room temperature in a low vacuum chamber

35 Why degas? Higher density Higher stiffness Mfg. repeatability Alternatively, use low-outgassing resins and follow directions carefully

36 Cavity Wall Removal Avoid cutting into flexure/pad material Can cut partially through walls and then snap the remnants off

37 Precision and Repeatability over Machines/Technologies

38 Print Accuracy & Calibration Stratasys uprint SE Model Material: ABS Support Material: SR-30 soluble Print resolution: mm Process: FDM Cost: USD 15,000 (a smaller one costs $10k)

39 Minimum feature sizes Nozzle width limits the smallest feature you can print; this shows up as an asymptote in measured feature width Test comb ~1.03 mm.05 mm

40 ad hoc calibration fixtures Press-fit pin test fixture Plastic-on-plastic test picture

41 Adjusting the OpenHand models Many CAD engines support adjustable parameters or equations We use a SolidWorks tool (parameter tables) to adjust the model to your printer

42 Part Orientation for Build/Print

43 In-plane and out-of-plane accuracy Part accuracy will vary with orientation!

44 Holes Top view Top view Side view Side view

45 Strength problems Out-of-plane In-plane Edges on each planar slice Reinforcing edge layers strengthen on-axis holes Parts tend to split between layers when printed off-axis RP failure