Toward PicoCubes

1
TowardPicoCubes
1990
2010

• Dan Steingart
• Brian Otis Beth Reilly
• Eric Carleton
• Paul Wright Jan Rabaey

2
Vision/Challenge

• Small monolithic devices are here
• Small power supply development is underway
• Packaging comes next
• Not as an afterthought
• We will only be successful using proactive
  designs that integrate all sub-systems

3
Is Post Facto Assembly Feasible

• Size gt 50 cm3
• Easy
• Size 10 cm3
• Doable but limiting
• Size 1 cm3
• Battery/Capacitor sealing becomes bulky
• On-board power sources desirable

4
Common Traits

• What were giving up
• Not self propelled (no robotic flies)
• Zero / Very little human interface (usually)
• What were getting
• Zero physical maintenance
• Set it and forget it
• Lots of data

5
Very Abstract Differentiation

• Smart Networks
• Low duty cycle but for reconfiguration sequences
• Rx/Tx
• Multi Hop
• Complex Sensor/ Simple Actuator

• Dumb Networks
• Low Duty Cycle
• One Way
• Single Hop
• Simple Sensing
• Generally lower power

6
The Big Problem(s)

• Each component has specific fabrication and
  packaging needs that sometimes run in mutual
  opposition to those required by another component
• Example The protective barriers usually found on
  power sources and energy storage become a size
  liability

7
Devices

• Need dramatic expansion of 3rd dimension
• Before final packing connections are fragile

BWRC Low Power Radio - 1 mA - Brian Otis
8
Energy Production

• Energy Scavenging
• Solar
• Free exposed surface
• Piezoelectric
• Space to shake
• Thermoelectric
• Maintaining a thermal gradient

Outer Cage
9
Energy Storage

• Electrochemical
• Atmosphere control
• By default the last process step
• Internal Thermal Isolation
• Supercapacitive
• Fragile
• Heat Sensitive

10
Thought Experiment
Thermistor

• Ultimate Device 1 cm3
• 3-D Wiring
• x sides with solar panels
• y sides with Sensors
• On-Board Battery
• On-Board Piezo Electric Generation

Solar Panel
Photoresistor
11
Reality Check

• Device Limitations
• No through-chip vias
• Wire wrapping still new
• Battery/Capacitor
• Cant be violated
• Requires significant sealing
• Energy Scavenging
• New Tech Flexible
• Exploit open space between bender fingers
• RF/Bender interactions must be characterized

12
Justifying Approximation

• Assume modification of a Crossbow Mica 2
• With a simulated microprinted 100 µm by 1 cm2
  Li-Ion Cell
• 1 Sense-Transmit cycle every 3 minutes
• 5 hour life assuming no power injection

• Further Modication
• Now change Chipcon radio for Brian Otis design
• Same Duty cycle
• 2 day life assuming no power injection

13
Justifying Assumption
powerTOSSIM/DUALFOIL results for hypothetical
Picocubes
14
Step One

• 3 Separate Components
• 1 Bus/Backplane
• Overall
• Modular Design
• Simplifies Connection
• Steals a surface
• Component packing eats significant space

Power Bus
15
Step Two

• One Package
• One Bus
• Test bed for integration
• Wiring problem still avoided

Power Bus
Microbattery
16
Where to Start

• Consider design of individual components in steps
  but not as black boxes
• But rather anticipating full system integration
• Design a basic test platform
• Solar/Vibration Hybrid powered monolithic
  temperature sensor with 8 hour charge life (ala
  PicoNode)
• A proof of concept for future ideas