Realizing the Full Potential of MEMS Design Automation. Steve Breit, Ph.D., V.P. Engineering

Size: px
Start display at page:

Download "Realizing the Full Potential of MEMS Design Automation. Steve Breit, Ph.D., V.P. Engineering"

Transcription

1 Realizing the Full Potential of Design Automation Steve Breit, Ph.D., V.P. Engineering MEPTEC Symposium 2012

2 Coventor Overview Founded in 1996 with a focus on software for design and simulation Management team from and EDA Accelerate commercial success by reducing Build and Test cycles Offering software tools and expertise in design and simulation methodology Consumer Automotive Solutions for product development Accelerometers, gyros, resonators, microphones, displays -specific multi-physics and +IC coupling Established proven track record with market leaders Top tier device makers and specialized manufacturers Industrial & Defense 11 of top 15 companies* use Coventor *Source: Yole Development (Feb. 2010): Top 15 companies = 80% of market SLIDE # 2

3 Outline Coventor s Vision for Market Acceleration: 3 waves in the Market 3 waves in market differentiators The Role of Design Automation in Accelerating Market Growth Complementary to A/MS flow Enabling design kits 1. Invention 2. Proprietary Manufacturing 3. Design & Integration Innovation We re here SLIDE 3

4 The First Wave: Invention One Product, One Process, One Package Jean-Christophe Eloy Yole Development In one short phrase, an understanding of why the industry has been slow to develop This law was true of the 1 st wave, dominated by IDM s $100 s Millions, if not Billions, invested developing processes, packaging and testing for seminal market successes Automotive (Bosch, Analog, Freescale) Inkjets (HP, Canon) Projection (Texas Instruments) Successes, but not readily scalable internally within IDMs or to a broader market SLIDE 4

5 The Second Wave: Manufacturing Differentiation Started in 2006 with the Nintendo Wii, the first consumer product success for motion sensors Larger volumes Shorter product design cycles More competition driven by lower cost The Second Wave weakened the One Product, One Process, law To remain competitive, IDM s had to leverage their significant investments in processes and manufacturing: How can we re-use our process and manufacturing investment into new or derivative products? Now, ADI, ST, Bosch, Freescale, etc. have processes and manufacturing capabilities that break the law Market Pull SLIDE 5

6 The end of Manufacturing Differentiation 2 nd wave market leaders differentiate on proprietary processes This will end because Mass market pull drove leading IDMs to outsource fabrication, spreading know-how to independent foundries 2 nd -wave entrants (Knowles, Invensense, ) leveraged growing eco system, spreading more know-how Large semi foundries want in on the action market growing faster than the semi market Lots of 200mm equipment to amortize still good for The end of manufacturing differentiation is inevitable Large semi foundries will come up the curve, sooner or later Standard(ized) process modules will be available Volume manufacturing will drive down costs SLIDE 6

7 The Third Wave: The Second Wave showed that the industry could grow at 15% year and could reach $15 billion market size in a few years Industry leaders say we are at the beginning of the Third Wave HP s vision for 1 trillion sensors by 2020: Central Nervous System for the Earth Bosch s vision for 7 trillion devices to serve 7 billion people in 2017 Projected to dwarf the second wave The Third-Wave, the trillion sensor vision, the $300B - $1Trillion vision, needs more than the first 2 waves What will drive the transition to hyper growth? Competencies Have to be Accessible and Scalable SLIDE 7

8 The Third Wave: Design and Integration Innovation Design Innovation = New applications Derivative designs Optimized designs New types of devices Integration Innovation = New applications Multi-device systems + A/MS + RF + logic Wafer or Package level technologies made accessible & scalable by the Eco System IDMs / Fabless Software Tools Equipment CMOS Foundries Testing Foundries Packaging SLIDE 8

9 Integration Challenges +CMOS SIP stacked die CMOS SIP: F2F IC CMOS SIP: TSV Monolithic CMOS +IC Source: imec.be +Package environmental effects, parasitics Single die WLP Coventor Inc Slide 9

10 Denser Integration is Inevitable Denser integration driven by Higher performance Improved SNR More dynamic range etc. Lower cost Smaller form factor Reduced power use Derived from G.Q. Zhang, NXP SLIDE 10

11 The Role of Design Automation In the 2 nd Wave:

12 The Role of Design Automation In the 3 rd Wave: 1. Simulate the complete system device behavior vs. process variations +IC coupled system Packaging effects Yield 2. Enable the ecosystem (foundries IDMs, fabless) Use simulation to compress development cycle Reference design flow, & EDA software Design Kits (MDKs) SLIDE 12

13 Goal: Simulate All Specs in a Virtual Product Data Sheet Gyro Data Sheet Yield / Cost Many specs involve packaging effects Most specs are electrical, requiring +IC simula8on Must know process varia8ons /IC coupling parasi8cs ma>er ADI ADXRS620

14 Analog / Mixed-Signal Reference Design Flow Levels of Abstraction Design Flow A/MS Designers High Algorithm Design (MATLAB / Simulink) More complexity Fewer assump?ons Schema8c Design (Cadence Virtuoso) Func8onal Verifica8on (Cadence Spectre, ) Layout (Cadence Virtuoso) Physical Verifica8on (DRC & LVS tools) PDK Process Development Low Parasi8c Extrac8on (Hercules, RCX) (TCAD tools) Tape out Coventor Inc Slide 14

15 Coventor s Vision for a +IC Reference Flow Designers Design Flow A/MS Designers MATLAB Interface Algorithm Design (MATLAB / Simulink) Symbol Schema8c Design (Cadence Virtuoso) Design Automa8on PlaTorm Netlist Pcell/PyCell Func8onal Verifica8on (Cadence Spectre, ) Layout (Cadence Virtuoso) PDK Abstract Cells Parasi8cs Physical Verifica8on (DRC & LVS tools) Parasi8c Extrac8on (Hercules, RCX) Process Development (TCAD tools) Tape out Coventor Inc Slide 15

16 Completing the Vision: Design Kits Designers Design Flow A/MS Designers MATLAB Interface Algorithm Design (MATLAB / Simulink) Parametric Device Models MDK Design Automa8on PlaTorm Symbol Netlist Pcell/PyCell Schema8c Design (Cadence Virtuoso) Func8onal Verifica8on (Cadence Spectre, ) Layout (Cadence Virtuoso) PDK Standard Process Modules Abstract Cells Parasi8cs Physical Verifica8on (DRC & LVS tools) Parasi8c Extrac8on (Hercules, RCX) Process Development (TCAD tools) Tape out Coventor Inc Slide 16

17 What s in a Design Kit? MDK Contents High-level process description Deposit thickness, conformality Etch depth, sidewall profiles Mask names Example: DALSA CMOS- PDK -specific design rules Release etch hole specs Material properties Process-specific Pre-stress, stress gradient Device IP Parametric device models SLIDE 17

18 Standardized process + reference flow + MDKs Design Automation Eco System Partnership Monolithic SiGe +CMOS Process CoventorWare SEMulator3D + Capping layer Mechanical layer Electrode layer Any CMOS (on 200 mm) Design Kits (MDKs) for Coventor tools Reference flow for integrated +IC design Coventor Inc Slide 18

19 Conclusions Wave 2 leaders rely on proprietary processes Wave 3 leaders will rely on Design & Integration Innovation Think Call to Action: Collaboration among eco-system leaders to make competencies accessible & scalable 1. Invention 2. Manufacturing 3. Design & Integration Innovation SLIDE 19