Machines that Make machines

Hod Lipson

Mechanical & Aerospace EngineeringComputing & Information Science

Cornell University

Computational Synthesis Labhttp://ccsl.mae.cornell.edu

Cornell UniversityCollege of Engineering

The two meta-challenges of Engineering:

1. Design a machines that can design other machines

2. Make a machine that can make other machines

Machines that Design Machines

Lipson & Pollack, Nature 406, 2000

Need more design space

FabLab in a box

• Fablabers are distinguished by disciplinary desegregation

• Lots of machines can make parts of other machines

• Is there a universal fabricator?– Top down approaches– Bottom up approaches

Printable Machines

The Universal Fabricator

On a single machine

• Make arbitrary shapes / structure– preassembled mechanisms and parts

• Make arbitrary circuits– Sensing, processing, power and actuation

• Achieve large range of functionalities– Use large range of materials

• Increase design space– Afforded by co-fabrication

Analog vs. Digital

LinearMotor

ThreadedRod

SyringeBarrel Plunger

Deposition via Syringe Extruder Tool

>250um

MaterialFluid

Reservoir

PIEZO-ACTUATOR

Material FluidReservoir

~30V,DC-10kHz

Deposition via Ink-Jet

~100um

Continuous pathsVolume Fill

High-resolution patterning, mixingThin films (60nm)

Some of our printed electromechanical / biological components: (a) elastic joint (b) zinc-air battery (c) metal-alloy wires, (d) IPMC actuator, (e) polymer field-effect transistor, (f) thermoplastic and elastomer parts, (g) cartilage cell-seeded implant in shape of sheep meniscus from CT scan.

Printed Active Materials

With Evan Malone

Zinc-Air Batteries

With Megan Berry

IPMC Actuators

Printed Agarose MeniscusCell Impregnated Alginate Hydrogel

CAT Scan

Direct 3D Print after 20 min.Sterile Cartridge

Multi-material 3D Printer

With Larry Bonassar, Daniel Cohen

The Universal Fabricator: Parallel to the Universal Computer

• In the 60’s, a computer– Cost > $100,000– Size: Refrigerator– Speed: Hours/job– Operation: Trained staff – Usability: Maintenance intensive

• Today: – Faster, cheaper, better, easier

Digital PDP-11, 1969

Stratasys FDM Vantage, 2005

Exponential Growth

Source: Wohlers Associates, 2004 report

RP Machine Sales

Critical Mass

• The computer took off when it infiltrated the home market

• Solved the chicken and egg problem:– People were motivated to write software for

their own needs because there was available hardware

– People were motivated to buy hardware because there was software to run on it

The First Home Computer

• ALTAIR 8800 microcomputer kit (1975)– $397 (2MHz, 256 bytes RAM)

Generally credited with launching the PC revolution

Fab@Home

Low cost, hackable, fablabable, open source

Bottom-up Fabrication

Self-assembling machines

• Fukuda et al: CEBOT, 1988

• Yim et al: PolyBot, 2000

• Chiang and Chirikjian, 1993

• Rus et al, 1998, 2001

Murata et al: Fracta, 1994

Murata et al, 2000

Jørgensen et al: ATRON, 2004

Zykov & Lipson, 2005

Modular Robotics: high complexity, do not scale in size

Stochastic Systems: scale in size, limited complexity

Whitesides et al, 1998

Winfree et al, 1998

Dynamically Programmable Self Assembly

Construction Sequence

High Pressure

Low Pressure

Construction Sequence

Construction Sequence

Construction Sequence

Construction Sequence

Construction Sequence

Reconfiguration Sequence

Reconfiguration Sequence

Implementation 2Inside of the

cube:• Servo-

actuated valves

• Basic Stamp II controller

• Central fluid manifold

• Communication, power transmission lines

Embossed fluid manifold

Hermaphroditic interface

Orifices for fluid flow

With Paul White, Victor Zykov

Implementation 2: Fluidic Bonding

Movie accelerated x16With Paul White, Victor Zykov

a) t = 18.8 s b) t = 19.3 s c) t = 19.5 s d) t = 19.7 s

e) t = 4.9 s f) t = 8.6 s g) t = 14.3s h) t = 15.6s Figure 5. Assembly and Disassembly of 500 μm Silicon Tiles on PDMS Substrate

With David Erickson, Mike Tolley

300 µm

Conclusions

• Universal Designer• Universal fabricator

– Makes shapes, circuits, sensors, actuators, energy & information processing

• Top-down approach– Printable machines

• Bottom-Up approach– Dynamical self–assembly

Computational Synthesis Labhttp://ccsl.mae.cornell.edu

Cornell UniversityCollege of Engineering

Credits

Viktor ZykovEvan Malone

Mike TolleyDaniel Cohen

Also: Paul White, David Erickson