InvenSense: MEMS Start-Up to Market Leader InvenSense ...
Embed Size (px)
Transcript of InvenSense: MEMS Start-Up to Market Leader InvenSense ...

InvenSense: MEMS Start-Up to Market LeaderJoe Seeger
September 15, 2010
9/20/2010
InvenSense: MEMS Start Up to Market Leader
1
Agenda
Introduction & History
i iMotion Processing
MEMS Startup Challenges
InvenSense Solutionse Se se So u o s
Summary & Conclusions
2
BSAC Thrust Session Fall 2010 September 15, 2010

InvenSense Overview
• Funded in 2004, nearly 200 employees todayStatus
• Consumer ElectronicsMarket
• Major consumer electronic OEM and ODM companiesCustomer
• New generation of motion based interfacesSolution
• Nasiri‐Fabrication and MotionProcessing PlatformsTechnology
• Fabless CMOS & MEMS, nearly 100MU shippedManufacturing
• 56 filed and issued with many core IPPatents
3
InvenSense Precursors, 2003
S. Bhave, J. Seeger, X. Jiang, B. Boser, R. Howe, J. Yasaitis, Transducers 2003 S. Nasiri,
Transparent Networks Inc.
4
BSAC Thrust Session Fall 2010 September 15, 2010

MEMS Gyroscope Market, 2003
• First demonstrated by Draper Labs in early 1990
• Bosch commercialized it for automotive VSC in 1997
• First demonstrated by Draper Labs in early 1990
• Bosch commercialized it for automotive VSC in 1997
Historical Perspective
• >10 MEMS gyro companies at various stages, all for automotive
• All single axis, mostly Z‐axis at around $15/axis
• Bulky, high current, and costly for consumer products
• >10 MEMS gyro companies at various stages, all for automotive
• All single axis, mostly Z‐axis at around $15/axis
• Bulky, high current, and costly for consumer products
Key Market Drivers in 2003 for MEMS Gyro
• Driven by DSC/DVC image quality, all Japanese companies
Si l i Pi $3 5/ i M NEC d F ji
• Driven by DSC/DVC image quality, all Japanese companies
Si l i Pi $3 5/ i M NEC d F ji
Opportunity: Low Cost Consumer Grade MEMS Gyros
• Single axis Piezo gyros ~$3.5/axis; Murata, NEC and Fujitsu
• Key issues: Performance, Bulky, Fragile and need for external electronics
• Single axis Piezo gyros ~$3.5/axis; Murata, NEC and Fujitsu
• Key issues: Performance, Bulky, Fragile and need for external electronics
• MEMS gyros were thought as unfeasible to get below $3/axis• MEMS gyros were thought as unfeasible to get below $3/axis
Consumer Grade MEMS Gyro
5
InvenSense Prototype Gyroscope
MEMS Wafer
Integrated IC
6
BSAC Thrust Session Fall 2010 September 15, 2010

State‐of‐the‐Art Solutions2004
(inception)2006 2008 2009 2010 2011
1st generationX/Y gyro6x6x1.4mm
2nd generationX/Y Gyro4x5x1.2mm
X/Z Gyro 4x5x1.2mm
MPU-3000with X/Y/Z Gyro4x4x0.9mm
MotionProcessing
7
Image Stabilization Gaming TV Remote
What is MotionProcessing?
8
BSAC Thrust Session Fall 2010 September 15, 2010

MotionProcessing: “Interface to Digital World”
5 6,71-4
18
8,9
10
17
11
121314
15,16
9
Consumer Market Dynamics
• Need for product differentiation and new features
• Adoption by thought leaders and acceptance by customers
M l l d i i h
• Need for product differentiation and new features
• Adoption by thought leaders and acceptance by customers
M l l d i i h
Highly Competitive and Feature Driven
• Must solve a real need to get traction with customers• Must solve a real need to get traction with customers
• Constant and relentless drive to lower prices
• Unpredictable demands and shipment schedules
• Actively encourages multiple suppliers and fierce competition
• Need for complete solutions and thorough customer education
• Constant and relentless drive to lower prices
• Unpredictable demands and shipment schedules
• Actively encourages multiple suppliers and fierce competition
• Need for complete solutions and thorough customer education
Very Demanding for Size, Cost and Volume
Late Adopter of New Technologies• What are the real killer applications?
• Skeptical of consumer behavior and willingness to change
• Market is driven by only a few thought leaders
• Need for delivering total solution and lots of support
• What are the real killer applications?
• Skeptical of consumer behavior and willingness to change
• Market is driven by only a few thought leaders
• Need for delivering total solution and lots of support
Late Adopter of New Technologies
10
BSAC Thrust Session Fall 2010 September 15, 2010

11
MEMS Startup Challenges
• Understand market need and application
Market Knowledge
MEMS Production
• Address Fabrication, Packaging, Test, & Capacity challenges
• High‐volume MEMS production experience?
• Address MEMS simulation, IC design specific to MEMS
• MEMS is one component of total system
• Address cost, yield, performance and develop new products
MEMS d i k l d & i l d t i ?
Engineering & MEMS Development
• MEMS design knowledge & commercial product experience?
• Need some “unfair” advantage to compete against established companies
• Is it a Feature, Product, or Platform?
Disruptive Technology
12
BSAC Thrust Session Fall 2010 September 15, 2010

Challenges With Low Cost MEMS
Presented by Leading MEMS Company in 2005
13
Nasiri‐Fabrication Platform
Si Cavity etch (M2)
Si Wafer 1
MEMS Substrate
Backside Marks (M1)Engineered SOI WaferSi Cavity etch (M2)
Fusion Bond Si Wafer 2
Grind and Polish
Hermetic seal ring and Interconnect (M3)
Structure layer etch (M5)
ProprietaryWafer Bond
Ge based coating (M4)
Tab Removal CMOS
CMOS Wafer
Cavity Etch (M6)
14
BSAC Thrust Session Fall 2010 September 15, 2010

Wafer‐Scale Integration
15
Inherent Benefits of Nasiri Fabrication
• Most suitable for inertial sensors
• Using off‐the‐shelf equipment
• CMOS compatible process allows easy porting
• Most suitable for inertial sensors
• Using off‐the‐shelf equipment
• CMOS compatible process allows easy porting
Bulk Silicon Fabrication
• Enables wafer‐level integrations with CMOS
• Provides wafer level reliable hermetic vacuum seals
• Metallic eutectic bonding , no need for getter
• ~50% die size w/ small seal rings and interconnects,
• Enables wafer‐level integrations with CMOS
• Provides wafer level reliable hermetic vacuum seals
• Metallic eutectic bonding , no need for getter
• ~50% die size w/ small seal rings and interconnects,
Wafer Level Eutectic Bonding
Wafer Level Testing
• Lowest cost semiconductor level testing
• Timely feedback for process and quality monitors
• Lowest cost semiconductor level testing
• Timely feedback for process and quality monitors
• Easily ported to multiple MEMS or CMOS foundries
• Suitable for Gyro, Accel, Pressure, Resonators, more
• Easily ported to multiple MEMS or CMOS foundries
• Suitable for Gyro, Accel, Pressure, Resonators, more
Portable and Versatile
1616
BSAC Thrust Session Fall 2010 September 15, 2010

Opportunity in Packaging
Costly pick and Costly ceramic Costly capping assy Costly integrationCostly pick and place
Costly ceramic package
Costly test and cal
Costly ceramic package
Costly vacuum seal
Costly test
Costly capping assy
Costly getter
Inefficient use of Si
Costly integration
Waste of Si for interconnect
Costly LGA packaing
Inherent Cost Benefit
CMOS‐MEMSWafer Sort
Testing
Low Cost QFN
Assy17
Semiconductor Like Process Flow
CMOS‐MEMS Wafer Sort Packaging FT & Ship
1818
BSAC Thrust Session Fall 2010 September 15, 2010

Mixed‐signal
Nasiri‐Fabrication
MotionProcessing™ Technology
Motion Processor
Integrated
signal CMOS
Fabrication
Integrated MEMS Sensors
Firmware
19
MEMS is one component in the system…
…but tightly coupled to CMOS design20
BSAC Thrust Session Fall 2010 September 15, 2010

MEMS & IC layouts are coupled too
21
MEMS Requires IC‐like Design FlowArchitecture
Mech Design
MEMS + IC Simulation
Si Validation
Mech Verification
22
BSAC Thrust Session Fall 2010 September 15, 2010

Development Achievements
2006X/Y-Axis gyro
Die Size 12.2mm2
CMOS 0.5um
Output Analog
Power 27mW
2008
2009
X/Y-Axis Gyro
Die Size 7.4mm2
CMOS 0.35um
Output Analog
Power 20mW
Die Size 7.4mm2
CMOS 0.35um
Output Analog
2010
X/Z-Axis Gyro
ITG/MPU-3000X/Y/Z-Axis Gyro
Die Size 6.7mm2
CMOS 0.18um
Output Digital
Power 14mW
p g
Power 20mW
23
Lessons Learned…
• Must have some real competitive edge to succeed
Ai t h l t th i ti l ti (2X t 3X)
• Must have some real competitive edge to succeed
Ai t h l t th i ti l ti (2X t 3X)
Go for Breakthrough Solutions
• Aim at much lower cost than existing solutions (2X to 3X)• Aim at much lower cost than existing solutions (2X to 3X)
• Develop proprietary process fabrication IP that can scale
• Consider 2nd source foundries early on
• Develop proprietary process fabrication IP that can scale
• Consider 2nd source foundries early on
Develop Multiple Foundries
D i f L C t S l ti
• Stay with standard packages and test equipment
• Invest in your in‐house testing and calibration facilities
• Stay with standard packages and test equipment
• Invest in your in‐house testing and calibration facilities
Design for Low Cost Solution
24
BSAC Thrust Session Fall 2010 September 15, 2010

In Summary…
• Very large and rapidly growing market
• Requires complete solution to help customers with time to market
• Very large and rapidly growing market
• Requires complete solution to help customers with time to market
MEMS Motion Processing Market: High Growth Opportunity
• High performance at low cost with rapid production ramp
• Adoption is dependent on few thought leaders
• High performance at low cost with rapid production ramp
• Adoption is dependent on few thought leaders
Challenges with the Consumer Market
• Has taken nearly 20 years for MEMS gyro to meet consumer market needs
• Had to provide over 20X reduction in price, and size
• Has taken nearly 20 years for MEMS gyro to meet consumer market needs
• Had to provide over 20X reduction in price, and size
High Volume MEMS Commercialization
• Fabless MEMS is the only way for new start up to offer new MEMS products
• Must follow fabless semiconductor model as closely as possible
• CMOS‐MEMS and wafer level packaging and testing is the key
• Fabless MEMS is the only way for new start up to offer new MEMS products
• Must follow fabless semiconductor model as closely as possible
• CMOS‐MEMS and wafer level packaging and testing is the key
Development of Low Cost and High Volume MEMS Products
25
Thank you
InvenSense Headquarters1197 Borregas Ave.Sunnyvale, California 94089408.988.7339 Main408.988.8104 [email protected]@invensense.com
InvenSense Taiwan1F, 9 Prosperity 1st RoadHsinchu Science Park30076, Taiwan+886.3.6686999 Main+866.3.6686777 Fax
InvenSense Japan G.K.Nisso Building No. 16, 2F3-8-8 Shin YokohamaKohoku-ku, Yokohama222-0033, Japan+81.045.308.9721 Main+81.045.534.3005 Fax
InvenSense Korea Ltd.#601, Jungle Bldg. 206-17Nonhyun-Dong, Gangnam-Gu,Seoul 135-833, Korea+82.2.541.2900 Main+82.2.541.2955 Fax
26
BSAC Thrust Session Fall 2010 September 15, 2010