ASIC & FPGA Chip Design:

Introduction

Mahdi Shabany
Department of Electrical Engineering
Sharif University of technology
Course Code: 25776

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 2

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 3

Course Description
 How to go from Idea/Algorithm to the actual hardware

module RippleCarryAdderII (Cin, X, Y, S, Cout);

parameter n = 4; input Cin; input [n-1:0] X, Y; output [n-1:0] S;
wire [n-1:0] C;
Full_Adder stage0 (Cin, X[0], Y[0], S[0], C[1]);
Full_Adder stage1 (C[1], X[1], Y[1], S[1], C[2]);
Full_Adder stage2 (C[2], X[2], Y[2], S[2], C[3]);
Full_Adder stage3 (.Cout(Cout), .Cin(C[3]), .x(X[3]), .y(Y[3]), .S(S[3]));
endmodule

ASIC FPGA
© M. Shabany, ASIC & FPGA Chip Design
Course Description
 Hardware Description Language (HDL) :Verilog
 Professional Verilog Coding for Synthesis
 Verification Techniques
 FPGA Architectures
 Digital System Design with Xilinx FPGAs
 ASIC Digital Design Flow (from Verilog to the actual Chip!)
 Synthesis Algorithms
 Power Dissipation
 Power Grid and Clock Design
 Fixed-point Simulation Methodology
 Detailed Design Optimization Workshop with ISE (for the fist time!)

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:

Standard
Specifications
Cells

Pre-Layout Post-Layout
Simulation Yes Timing Yes Back Yes
RTL Coding Synthesis APR Timing Logic
Pass? Alanysis Annotation Alanysis verification
Pass? Pass?
NO NO
Test Bench Timing NO
Constraints
Tapeout

1. HDL Coding 2. Simulation 3. Synthesis 4. Placement & routing 5. Timing Analysis & Verification

Front-End Back-End
 In this course we learn all the above steps in detail for
 ASIC Platform
 FPGA Platform

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Hardware Description Language
Verilog Fundamentals
 Language Fundamentals
 Modeling Combinational/Sequential Logic Circuits
 Modeling Finite State Machines
Verilog for Verification
 Verification/Simulation techniques with test‐benches
Verilog for synthesis
 Verilog Styles for Synthesis
 Architectural techniques for high‐speed designs
o Parallel proc., pipelining, retiming, …
 Implementations of common operations
o Complex multiplication, division, complex norm, CORDIC
 Fundamentals of fixed‐point realization

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 PLDs & FPGA Architectures
FPGA Technologies
 SPLDs (PAL and PLA architectures)
 Commercial CPLD Architectures
 SRAM/LUT Based FPGAs
 Anti‐fuse/MUX Based FPGAs
 Flash Based FPGAs
FPGA Architectures
 Heterogeneous/Homogeneous FPGAs
 Fine‐grained, coarse‐grained and platform FPGAs
FPGA Elements & Design Trade‐offs
 Logic Cells Common Architectures
 Programmable Routing Channels Design
 I/O & Pad architectures
Commercial FPGAs
 Altera (FLEX 10K, Stratix III) , Xilinx (XC4000, Virtex II,4,5), Actel (Act3, Axcelerator)

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Advanced Digital System Design with Xilinx FPGAs
 Design Creation
 Synthesize
 Simulation
 Constraints Entry
 Implementation
 Implementation Results Analysis – Timing Analysis
 Implementation Results Analysis – Power Analysis
 Implementation Results Improvement
 Device Configuration and Programming
 Design Debugging

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Core Generator
 CORE Generator Tool
 Intellectual Property (IP) Cores
 CORE Generator Tool files
 Design Flows
 Defining Memory Contents for RAM and ROM
 Defining Coefficient Values in a COE File

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 ASIC Design Flow:
HDL Coding & Verification
Synthesis & Timing Optimization
 Complete Synopsys Design Complier Design Flow
Physical Design
 Cadence First Encounter
 Floorplan (Initial floorplan and power planning)
 Placement (Full‐scale floorplan and clock tree insertion)
 Routing (power routing & Nanoroute)
 Timing Closure (Analysis & Optimization of setup and hold time violations)
 Fill (Filler Cells, Metal Fill, and Verify Geometry)

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 CAD Tool Algorithms:
Synthesis Algorithms
Two-level Optimization
Multi-level Logic Optimization
Technology Mapping

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Power Dissipation
 Power Dissipation concept
 Dynamic Power
 Static Power
 Challenges

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Power Grid and Clock Design
 Power Distribution Design
 Introduction
 IR Drop
 Ldi/dt Drop
 Decoupling Capacitances
 Clock Considerations
 PLL/DLL Architecture

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Prerequisites:
Only Digital Logic!
All the skills you need will be taught in the course

 Softwares you will learn:

Altera Quartus
Xilinx ISE
Mentor Graphics Modelsim
TCL Scripting
Synopsys Design Compiler
Cadence SOC Encounter

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Implementation Platform:
Altera DE2 Board
Atlys Xilinx Board

 You will do several practical assignments including Verilog Coding,

FPGA implementation and testing based on the DE2 and Atlys Boards

© M. Shabany, ASIC & FPGA Chip Design

 Course Description:
 Assignments & Labs (12)
4 implementation assignments on DE2 Board (8 Marks)
6 Implementation assignments on Atlys boards (12 Marks)
2 Paper Assignments (4 marks)
 1 comprehensive project from A to Z (5 marks)

Soft deadlines: 20% daily penalty for the late submissions

© M. Shabany, ASIC & FPGA Chip Design

Course Evaluation:

Section Quantity Value each Value

Assignments 12 2 24
Project 1 6 6
Midterm 1 30 30
Final Exam 1 40 40

Total: 100

If both the midterm and final marks are less than 50%, the total mark is calculated only
based on the exams and assignments will not be taken into account.

© M. Shabany, ASIC & FPGA Chip Design

 Course Outline : References
 HDL:
Thomas & Moorby, The Verilog Hardware Description Language, 3rd ed, Kluwer Academic.
Introduction to Logic Synthesis using Verilog HDL, Robert Reese, Mitchell Thornton, 2006.
Advanced FPGA Design, Architecture, Implementation, and Optimization, Steve Kilts, 2007.

 IC Design Flow:
Course Lecture notes, 2014
Digital IC Design Flow, provided by the instructor, 2014

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 20

Integrated Circuit (IC) History

Discrete Circuits Integrated Circuits

Limited applications Many applications
Power hungry Low power
Large area Small area
Moderate speed High speed

© M. Shabany, ASIC & FPGA Chip Design

Integrated Circuits (IC) History
1940 1945 1947 1951 1955 1957 1958 1959 1961 1962 1968 1970 1971

Ohl built the PN Junction: 1940

Shockley Lab was established: 1945
Brattain and Bardeen invented the first transistor (US Patent 2524035): 1947

A picture of the filed patent

© M. Shabany, ASIC & FPGA Chip Design

Integrated Circuits (IC) History
1940 1945 1947 1951 1955 1957 1958 1959 1961 1962 1968 1970 1971

 1951: Shockley invented the first junction transistor for mass production
(US Patent 2623105)

A picture of the filed patent

© M. Shabany, ASIC & FPGA Chip Design

‫‪Integrated Circuits (IC) History‬‬
‫‪1940‬‬ ‫‪1945‬‬ ‫‪1947‬‬ ‫‪1951‬‬ ‫‪1955‬‬ ‫‪1957 1958 1959 1961 1962 1968 1970 1971‬‬

‫‪ Noyce :1955‬به آزمایشگاه ‪ Shockley‬پیوست‬

‫‪ Noyce :1957‬آزمایشگاه ‪ Shockley‬را به هدف تاسیس ‪ Fairchild‬با کمک ‪Hoerni‬‬
‫و ‪ Gorden Moore‬ترک کرد‬
‫‪ Hoerni :1958‬تکنیکی را برای افشاندن ناخالصی ها بر روی ‪ Si‬ابداع کرد و به این‬
‫ترتیب توانست یک ترانزیستور مسطح با کمک عایق ‪ SiO2‬بسازد‬
‫‪ Noyce :1959‬اولین مدار مجتمع را با استفاده از ترانزیستورهای مسطح ساخت‬

‫‪© M. Shabany, ASIC & FPGA Chip Design‬‬

‫‪Integrated Circuits (IC) History‬‬
‫‪1940‬‬ ‫‪1945‬‬ ‫‪1947‬‬ ‫‪1951‬‬ ‫‪1955‬‬ ‫‪1957 1958 1959 1961 1962 1968 1970 1971‬‬

‫‪ Jack Kilby :1959‬که در )‪ Texas Instrument (TI‬کار می کرد ایده مدار مجتمع یک‬
‫پارچه را مطرح کرد‪ .‬در این طرح قطعات به وسیله سیم های لحیم شده به یکدیگر متصل‬
‫می شدند و از دیودهای ‪ PN‬به عنوان مقاومت استفاده می شد ( ‪(U.S Patent‬‬
‫‪)3,138,473‬‬

‫‪A picture of the filed patent‬‬

‫‪© M. Shabany, ASIC & FPGA Chip Design‬‬
Integrated Circuits (IC) History
1940 1945 1947 1951 1955 1957 1958 1959 1961 1962 1968 1970 1971

)50$ ‫ منطقی را تولید انبوه کردند (هر کدام‬IC ‫ اولین‬Fairchild ‫ و‬TI :1961
‫ شانزده بیتی ساخته شد‬Ripple Carry Adder :1962
Fairchild bipolar RTL Flip-Flop RCA 16-transistor MOSFET IC

© M. Shabany, ASIC & FPGA Chip Design

Integrated Circuits (IC) History
1940 1945 1947 1951 1955 1957 1958 1959 1961 1962 1968 1970 1971

‫ را پایه گذاری کردند‬Intel ‫ را رها کرده و‬Fairchild ‫ موسسه‬Moore ‫ و‬Noyce :1968

‫ را تولید کرد‬Static RAMs 256-bit , Fairchild :1970
‫ را فروخت‬Dynamic RAMs 1-K bit , Intel :1970

Fairchild 4100 256-bit SRAM Intel 1103 1K-bit DRAM

© M. Shabany, ASIC & FPGA Chip Design

‫‪Integrated Circuits (IC) History‬‬
‫‪1940‬‬ ‫‪1945‬‬ ‫‪1947‬‬ ‫‪1951‬‬ ‫‪1955‬‬ ‫‪1957 1958 1959 1961 1962 1968 1970 1971‬‬

‫‪ 4004 , Intel :1971‬را تولید کرد‬

‫اولین کامپیوتر قابل برنامه ریزی چند منظوره برای شرکت ماشین حساب ژاپنی‬

‫‪© M. Shabany, ASIC & FPGA Chip Design‬‬

 IC History
1971 ……………………………………………………………………………………… 2010

K 1 Billion Transistors
1,000,000
Projected
100,000
Pentium® III
10,000 Pentium® II
Pentium® Pro
1,000 Pentium®
i486
100 i386
80286
10 8086

1
1975 1980 1985 1990 1995 2000 2005 2010
Source: Intel

Number of Transistors doubled every 18 months!

(Moore’s Law)
Scaling still continues…
© M. Shabany, ASIC & FPGA Chip Design
IC History
1971 ………………………………………………..……………………… 2015

 Amazingly visionary – million transistors/chip barrier was

crossed in the 1980’s.
 2300 transistors, 1 MHz clock (Intel 4004) - 1971
 16 Million transistors (Ultra Sparc III)
 42 Million, 2 GHz clock (Intel P4) - 2001 Integration Levels
 140 Million transistor (HP PA-8500) SSI: 10 gates
MSI: 1000 gates
LSI: 10,000 gates
VLSI: > 10k gates

© M. Shabany, ASIC & FPGA Chip Design

IC History
Early Designs Advanced Designs

(1971) (2017)
 1000 Transistors (1 MHz)  Intel Corei7 -7700K(~ 4.2 GHz)
 Process: 10um  Process: 14nm
 Fully Handcrafted  Fully Automated
 Manual Layout  Automated Layout
 Individually Optimized  Hierarchical Design
Digital Very Large Scale Integration (VLSI) (Not Analog!)

© M. Shabany, ASIC & FPGA Chip Design

Process Improvement

Power density too high to keep up frequency doubles every 2 years

junctions at low temp

© M. Shabany, ASIC & FPGA Chip Design

Technology Direction

Year 1999 2002 2005 2008 2011 2014 2017

Feature size (nm) 180 130 100 70 50 35 14
M trans/cm2 7 14-26 47 115 284 701 900
Chip size (mm2) 170 170-214 235 269 308 354 460
Signal pins/chip 768 1024 1024 1280 1408 1472 ~1600
Clock rate (MHz) 600 800 1100 1400 1800 2200 4200
Wiring levels 6-7 7-8 8-9 9 9-10 10 10
Power supply (V) 1.8 1.5 1.2 0.9 0.6 0.6 0.8

© M. Shabany, ASIC & FPGA Chip Design

Why Scaling?

 Technology shrinks by ~0.7 per generation

 With every generation can integrate 2x more functions on a chip; chip cost
does not increase significantly

 Cost of a function decreases by 2x

 But …
 How to design chips with more and more functions?
 Design engineering population does not double every two years…
 Hence, a need for more efficient design methods
 Exploit different levels of abstraction

© M. Shabany, ASIC & FPGA Chip Design

 IC Manufacturing Process

Silicon Wafer Patterned Silicon

Wafer

Silicon Cylinder

Circuit Core
Packaged Die Final Test
Unpackaged Die
© M. Shabany, ASIC & FPGA Chip Design
 IC Manufacturing Process

Circuit core
Cavity
Die Bond wire
Package Trace
Patterned Silicon
PAD Wafer

Packaged Die
Lead

Circuit Core

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 37

 Digital vs. Analog Design : Analog Circuits
 Transmitter: Analog/RF

Voice
Modulator

Power
Amplifier
High-frequency
carrier

 Receiver:

Down
Demodulator
Converter
Low-Noise Audio
Amplifier Amplifier

Analog/RF

© M. Shabany, ASIC & FPGA Chip Design

 Digital vs. Analog Design : Analog-Digital Circuit
 Transmitter:
Digital (Baseband) Analog/RF Front-end

Voice Voice Coding Pulse

ADC Modulator
Compression Interleaving shaping
Power
Amplifier

 Receiver: This Course

Down
ADC Demodulator Equalizer
Converter
Low-Noise
Amplifier

Voice De-interleaving
DAC
Decompression Decoding
Audio
Amplifier
Digital (Baseband)

© M. Shabany, ASIC & FPGA Chip Design

 Digital vs. Analog Design
 Digital Design Advantages:
More noise reliability
Allows signal decoding and amplification
Allows coding to achieve higher performance
Allows encryption for higher security
Provides a perfect vehicle for digital signal processing
Allows modular chip design
Enjoys the benefit of advanced sophisticated CAD tools
Can be verified on programmable devices before tape-out
Provides a platform to merge multiple networks such as telephone,
terrestrial TV and computer networks.

© M. Shabany, ASIC & FPGA Chip Design

Digital vs. Analog Design

© M. Shabany, ASIC & FPGA Chip Design

 Digital System Design Steps
 Specific application (e.g., Digital Image Processor)
As opposed to a general microprocessor
 Design requirements
Images to be processed per second
Required processing
RTL
Interfacing
Dimension, power, price,…
 General design and simulation
 RTL level design to verify and simulate the design

 Higher levels: Few complicated blocks

 Lower levels: Many simple blocks

To avoid any problem in lower levels, many simulations

in higher levels

© M. Shabany, ASIC & FPGA Chip Design

Design Flow
IC Fabrication
Idea
Wafer
Architecture Design (hundreds of dies)

Sawing & Packaging

Block
diagram Final chips

Circuit & Layout Design Testing

customers
Layout

Bad chips Good chips

© M. Shabany, ASIC & FPGA Chip Design

Digital Systems Components
 Printed Circuit Board (PCB)
 Embedded Software
 Microprocessor (general Purpose)
 Microcontroller
 Digital Signal Processor (DSP)
 Programmable Logic Devices
 Simple Programmable Logic Device (SPLD)
 Complex Programmable Logic Device (CPLD)
 Field Programmable Gate Arrays (FPGAs)
 Application Specific Integrated Circuits (ASICs)

© M. Shabany, ASIC & FPGA Chip Design

Digital Systems Components
DSPs

 Easy to program (usually standard C)

 Very efficient for complex sequential math-intensive tasks

 Fixed datapath-width. Ex: 24-bit adder, is not efficient for 5-bit addition

 Limited resources

© M. Shabany, ASIC & FPGA Chip Design

Digital Systems Components
Microprocessors/Microcontrollers

 Lessens the risk of system development by reducing design complexity

 Fixed hardware so more effort on developing a good code

 Fixed HW not suitable for high level of parallelism/computations

 Many sequential nature
 One order of magnitude less performance than to-date FPGA/ASICs

© M. Shabany, ASIC & FPGA Chip Design

Digital Systems Implementation Platforms

Digital IC

PLDs ASIC

SPLD CPLD FPGA Semi-Custom Full-Custom

PLA PAL Standard cell Gate Array

This course

© M. Shabany, ASIC & FPGA Chip Design

Technology Timeline

The white portions of the timeline bars indicate that although early incarnations of these technologies
may have been available, they weren’t enthusiastically received by the engineers working in the
trenches during this period. For example, although Xilinx introduced the world’s first FPGA as early as
1984, design engineers didn’t really start using it until the early 1990s.

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 49

 FPGA vs. ASIC
 Field Programmable Gate Array (FPGA) Advantages:
Fast programming and testing time by the end user (instant turn-around)
Excellent for prototyping
 Easy to migrate from prototype to the final design
Can be re-used for other designs
Cheaper (in small volumes) lower start-up costs
Re-programmable
Lower financial risk
Ease of design changes/modifications
Cheaper design tools

© M. Shabany, ASIC & FPGA Chip Design

 FPGA vs. ASIC

 FPGA Drawbacks:
Slower than ASIC (2-3 times slower)
Power hungry (up to 10 times more dynamic power)
Use more transistors per logic function
More area (20 to 35 times more area than a standard cell ASIC)

© M. Shabany, ASIC & FPGA Chip Design

 FPGA vs. ASIC
 Application Specific Integrated Circuit (ASIC) Advantages:
Faster
Lower power
Cheaper (if manufactured in large volumes)
Use less transistors per logic function

 ASIC Drawbacks:
Implements a particular design (not programmable)
Takes several months to fabricate (long turn-around)
More expensive design tools
Very expensive engineering/mask cost for the first successful design

© M. Shabany, ASIC & FPGA Chip Design

 ASIC Mask Generation Cost

 The ASIC is accompanied by increasing nonrecurrent engineering (NRE) costs which meant
that there was an increased emphasis on “right first time” design.

 These NRE costs is largely due to the cost of generating masks as it is becoming more
expensive to generate the masks for finer geometries needed by shrinking silicon
technology dimensions.

© M. Shabany, ASIC & FPGA Chip Design

Implementation Approaches (ASIC vs. FPGA)

ASIC FPGA
Application Specific Field Programmable
Integrated Circuit Gate Array
 Bought off the shelf & reconfigured
 Expensive & time consuming
by the end designers
fabrication in semiconductor
foundry
 No physical layout design
 Designed all the way from
behavioral description to  Design ends with a bitstream used
physical layout to configure a device

© M. Shabany, ASIC & FPGA Chip Design

Implementation Approaches (ASIC vs. FPGA)
ASICs FPGAs

Off-the-shelf
High performance

Low development cost

Low power
Short time to market

Low cost in
high volumes Re-configurability

© M. Shabany, ASIC & FPGA Chip Design

ASIC vs. FPGA

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 57

 Design Abstraction Levels
 Divide-and-Conquer
Design modules once
Instantiate them thereafter SYSTEM

Standard Cells
 Already laid out MODULE
Avoid re-design +
Same as programming GATE

 Designer cares about module’s: CIRCUIT

Functionality
Delay characteristics G
DEVICE

Area S
n+ n+
D

 NOT:
How the module was designed This Course VLSI Course
Detailed solid-state behavior

© M. Shabany, ASIC & FPGA Chip Design

Design Abstraction Levels
Y-Chart System Level Algorithmic Level
A:=A*B+F
I/O Processor IF(C=TRUE) THEN
A:=A+2*F
ELSE A:=A-1
RAM ROM ENDIF

Register Transfer Level Logic Level

MUX A C Q

Register
Circuit &
L |
Logic ROM D Qn
U
Register Transfer
Algorithmic Circuit Level
Architectural
(System)

© M. Shabany, ASIC & FPGA Chip Design

 Device Metrics
Performance Metrics of a Digital Chip:
 Cost
NRE (fixed) costs - design effort
RE (variable) costs - cost of parts, assembly, test
 Speed Single Die
Delay (ns) → Operating Frequency (MHz)
 Power Dissipation
 Energy to Perform a Function
Energy per bit (nJ/b)
 Reliability
Noise immunity
Noise margin
 Scalability Wafer
Larger Designs Diameter: 10-30 cm
 Time-to-Market Thickness: 1 mm

© M. Shabany, ASIC & FPGA Chip Design

 Device Metrics
Performance Metrics of a Digital Chip:
 Cost
NRE (fixed) costs - design effort
RE (variable) costs - cost of parts, assembly, test
 Speed Single Die
Delay (ns) → Operating Frequency (MHz)
 Power Dissipation
 Energy to Perform a Function
Energy per bit (nJ/b)
 Reliability
Noise immunity
Noise margin
 Scalability Wafer
Larger Designs
 Time-to-Market

© M. Shabany, ASIC & FPGA Chip Design

 Cost of Integrated Circuits
 NRE (non-recurring engineering) costs
 Fixed cost to produce the design
• Design effort
• Design verification effort
• Mask generation
 Influenced by the design complexity and designer productivity
 More pronounced for small volume products
 RE (Recurring costs) – proportional to the product volume (i.e., Variable)
 Silicon processing
• also proportional to chip area
 Assembly (packaging)
 Test
fixed cost
Cost per IC = variable cost per IC + -----------------
volume

© M. Shabany, ASIC & FPGA Chip Design

Recurring Costs
cost of die + cost of die test + cost of packaging
variable cost = ----------------------------------------------------------------
final test yield
cost of wafer
cost of die = -----------------------------------
dies per wafer × die yield
 × (wafer diameter/2)2  × wafer diameter
Dies per wafer = ----------------------------------  ---------------------------
die area  2 × die area

α depends on the complexity of the

manufacturing process (roughly
proportional to the number of masks)

die yield = (1 + (defects per unit area × die area)/)-

© M. Shabany, ASIC & FPGA Chip Design
Yield Example

 Example
 Wafer size of 12 inches, die size of 2.5 cm2, 1 defects/cm2,
  = 3 (measure of manufacturing process complexity)

 252 dies/wafer (remember, wafers round & dies square)

 Die yield of 16%
 252 x 16% = only 40 dies/wafer die yield !

 Die cost is a strong function of the die area

 Proportional to the third or fourth power of the die area

Cost of die = f(die area)4

© M. Shabany, ASIC & FPGA Chip Design

 Examples of Cost Metrics (1994)

Chip Metal Line Wafer Defects Area Dies Yield Die

layers width cost (mm2) cost
/cm2 /wafer
386DX 2 0.90 $900 1.0 43 360 71% $4
486DX2 3 0.80 $1200 1.0 81 181 54% $12
PowerPC 601 4 0.80 $1700 1.3 121 115 28% $53
HP PA 7100 3 0.80 $1300 1.0 196 66 27% $73
DEC Alpha 3 0.70 $1500 1.2 234 53 19% $149
Super SPARC 3 0.70 $1700 1.6 256 48 13% $272
Pentium 3 0.80 $1500 1.5 296 40 9% $417

© M. Shabany, ASIC & FPGA Chip Design 65

 Device Metrics
Performance Metrics of a Digital Chip:
 Cost
NRE (fixed) costs - design effort
RE (variable) costs - cost of parts, assembly, test
 Speed Single Die
Delay (ns) → Operating Frequency (MHz)
 Power Dissipation
 Energy to Perform a Function
Energy per bit (nJ/b)
 Reliability
Noise immunity
Noise margin
 Scalability Wafer
Larger Designs
 Time-to-Market

© M. Shabany, ASIC & FPGA Chip Design

Performance

 Frequency of operation: 1/T

 Dependent on the propagation delay of signal through the logic
 Time to get the data out/in of the registers
 Clock uncertainty
 tp is a typical measure (not an accurate one)
 the delay experienced by a signal when passing through a gate
 50% transition points of the input and output waveforms
• Two types LH and HL
• Good for comparison
 Rise time/fall time affects delay
 10% - 90% definitions

© M. Shabany, ASIC & FPGA Chip Design

Delay Definitions
Vin Vout

Vin
Propagation delay
input 50% tp = (tpHL + tpLH)/2
waveform

t
tpHL tpLH
Vout
90%
output
50% signal slopes
waveform
10%
t
tf tr
© M. Shabany, ASIC & FPGA Chip Design
 Device Metrics
Performance Metrics of a Digital Chip:
 Cost
NRE (fixed) costs - design effort
RE (variable) costs - cost of parts, assembly, test
 Speed Single Die
Delay (ns) → Operating Frequency (MHz)
 Power Dissipation
 Energy to Perform a Function
Energy per bit (nJ/b)
 Reliability
Noise immunity
Noise margin
 Scalability Wafer
Larger Designs
 Time-to-Market

© M. Shabany, ASIC & FPGA Chip Design

Power Consumption

 Peak transient power

 Power line sizing, decoupling, etc. Ppeak = ipeakVsupply = max[p(t)]

 Average power
 Battery current delivery, cooling system
 Static power vs. Dynamic power
 Static current  no computation, etc
 Dynamic current  Switching on/off the gates
 The higher the number of switching events, the higher the dynamic power consumption

 Dynamic Energy  Amount of energy that is needed to be spent to do a job

 Time is no matter

© M. Shabany, ASIC & FPGA Chip Design

 Power and Energy Dissipation
 Propagation delay and the power consumption of a gate are related
 Propagation delay is (mostly) determined by the speed at which a
given amount of energy can be stored on the gate capacitors
 the faster the energy transfer (higher power dissipation) the faster the gate
 For a given technology and gate topology, the product of the power
consumption and the propagation delay is a constant
 Power-delay product (PDP) – energy consumed by the gate per switching event
 An ideal gate is the one that is fast and consumes little energy, so the
ultimate quality metric is
 Energy-delay product (EDP) = power-delay

© M. Shabany, ASIC & FPGA Chip Design

 Device Metrics
Performance Metrics of a Digital Chip:
 Cost
NRE (fixed) costs - design effort
RE (variable) costs - cost of parts, assembly, test
 Speed Single Die
Delay (ns) → Operating Frequency (MHz)
 Power Dissipation
 Energy to Perform a Function
Energy per bit (nJ/b)
 Reliability
Noise immunity
Noise margin
 Scalability Wafer
Larger Designs
 Time-to-Market

© M. Shabany, ASIC & FPGA Chip Design

 Reliability: Noise in Digital Integrated Circuits
 Noise : unwanted variations of voltages and currents at the logic nodes

 from two wires placed side by side

 Capacitive coupling v(t)

 Voltage change on one wire can

influence signal on the neighboring wire
 Cross talk
i(t)
 Inductive coupling
 Current change on one wire can
influence signal on the neighboring wire

 from noise on the power and ground supply rails VDD

 May influence signal levels in the gate

© M. Shabany, ASIC & FPGA Chip Design

Noise Margins
 For robust circuits, want the “0” and “1” intervals to be as large as possible
VDD VDD

VOH "1"
NMH = VOH - VIH
VIH
Noise Margin High Undefined
Region
Noise Margin Low VIL
NML = VIL - VOL
VOL
"0"
Gnd Gnd
Gate Output Gate Input

Noise margin represents the levels of noise that can be sustained

when gates are cascaded
© M. Shabany, ASIC & FPGA Chip Design
 Noise Immunity
 Noise margin expresses the ability of a circuit to overpower a noise source
 Noise sources: supply noise, cross talk, interference, offset
 Absolute noise margin values are deceptive
 Floating node is more easily disturbed than a node driven by a low
impedance (in terms of voltage)

 Noise immunity expresses the ability of the system to process and transmit
information correctly in the presence of noise (noise rejection)

 For good noise immunity, the signal swing (i.e., the difference between VOH
and VOL) and the noise margin have to be large enough to overpower the
impact of fixed sources of noise

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 78

 MOS Device Theory

 Bipolar Junction Transistor (BJT)

 Small current in Base drives a larger current b/w Emitter & Collector.
Mostly Quiescent power dissipation due to the Base current
Analog High power dissipation limits the number of transistors on a single chip
Not suitable for Very Large Scale Integration (VLSI)

 Metal Oxide Semiconductor Field Effect Transistors (MOSFET)

 Come with almost zero control current (Gate voltage controls the drain current)
Higher integration
Mostly Much lower power consumption than BJT
Digital Come in two flavors : n-MOS (n-type dopants), p-MOS (p-type dopants)

 Complementary Metal Oxide Semiconductor (CMOS)

Utilizing both n-MOS and p-MOS transistors

© M. Shabany, ASIC & FPGA Chip Design

 MOS Device Theory
 Transistors are built on a silicon substrate
 Silicon is a Group IV material → covalent bonds with 4 adjacent atoms
 Silicon is a poor conductor → can be raised by adding dopants

Si Si Si
n-type p-type
 Group V dopants Si Si Si  Group III dopants
 Five valence electrons  3 valence electrons
 Extra electron free to move Si Si Si  Missing electron(hole) free to move
 Negative carrier  Positive carrier
 Example:  Example:
Arsenic, Phosphorus Boron
Si Si Si Si Si Si
- +
Si As + Si Si B- Si

Si Si Si Si Si Si

© M. Shabany, ASIC & FPGA Chip Design

 MOS Device Theory
 Four terminals: Body Source Gate Drain
1. Gate Polysilicon
2. Source SiO 2
3. Drain
4. Body G
n+ n+

p bulk Si D S
B
 Consists of :
 Gate (Metal (old), Polysilicon (now)) nMOS
 Insulating layer (SiO2 (oxide- glass))
 Source (n+ in nMOS, p+ in pMOS)
 Drain (n+ in nMOS, p+ in pMOS)
 Body (conductor)
 n+: Heavily doped n-type
 P+: Heavily doped p-type

© M. Shabany, ASIC & FPGA Chip Design

 nMOS Transistors
 Gate–oxide–body stack looks like a capacitor

 Body is commonly tied to ground (0 V)

 Gate at low voltage:

 Body is at low voltage
 Source-body diode is OFF
 Drain-body diode is OFF
 No current flows Body Source Gate Drain
 Transistor is OFF

OFF
n+ n+

p bulk Si

© M. Shabany, ASIC & FPGA Chip Design

 nMOS Transistors

 Gate at high voltage:

 Positive charge on gate
 Negative charge attracted to body
 Inverts a channel under gate to n-type
 Current flow in this channel b/w source and drain when drain voltage is nonzero
 Transistor is ON

Body Source Gate Drain

ON
n+ n+

p bulk Si

© M. Shabany, ASIC & FPGA Chip Design

pMOS Transistors

 Similar to nMOS with reversed doping and voltages

 Body is commonly tied to high voltage (VDD)
 Gate low: transistor ON
 Gate high: transistor OFF
 Bubble indicates inverted behavior

Body Source Gate Drain

Polysilicon
SiO 2

G
p+ p+

p bulk Si D S
B
pMOS
© M. Shabany, ASIC & FPGA Chip Design
MOS Transistors
 L is the channel length
 L : Process parameter, technology
 Smaller L → Faster transistors → higher speed circuits
 Typical process values: 0.35μm, 0.18μm, 0.13μm, 90nm, 60nm, …
 VDD decreases by technology
 1.5 V for 0.18 μm
 1.2 V for 0.13 μm
 Lower VDD saves power consumption
 GND = 0 V

© M. Shabany, ASIC & FPGA Chip Design

 CMOS
 Silicon wafer is the base material
 Diameter: 10-30 cm
 Thickness: 1 mm

 CMOS : both nMOS and pMOS transistors fabricated on a single wafer

 Wells : special regions to separate bulks of nMOS and pMOS

SiO2

n+ diffusion

p+ diffusion
n+ n+ p+ p+
polysilicon
n well
p substrate
metal1

nMOS transistor pMOS transistor

© M. Shabany, ASIC & FPGA Chip Design

 CMOS
 Substrate must be tied to GND and n-well to VDD
 Poor connection of metal to lightly-doped semiconductor (Shottky Diode)
 Use heavily doped well and substrate contacts
GND V DD

p+ n+ n+ p+ p+ n+

n well
substrate tap p substrate
well tap
Cross Section along Dashed Line

GND VDD
Top View

© M. Shabany, ASIC & FPGA Chip Design

‫‪CMOS‬‬
‫•ارزان‬
‫•اتصال ‪sub‬به‬
‫زمین مشترک‬

‫‪NWELL‬‬
‫اتصال ‪sub‬به‬
‫بزرگترین منبع‬ ‫ارزان‬

‫‪PWELL‬‬

‫گران‬ ‫•دو درجه آزادی‬

‫•ایزوله از ‪sub‬‬
‫‪Twin-WELL‬‬

‫‪© M. Shabany, ASIC & FPGA Chip Design‬‬

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 89

 Layout
 Chips are specified with set of masks
 Minimum dimensions of masks determine transistor size (and hence
speed, cost, and power) SiO gate oxide 2

 “Feature size”
polysilicon

Lmin= distance between source and drain (channel) gate

W

(minimum width of Polysilicon) t ox

L
n+ n+
p-type body

 Feature size improves 30% every 3 years or so

Feature Size

 Can integrate 2× more functions per chip → ~2× less cost per function
 Normalize for feature size when describing design rules (  Lmin / 2 )
 E.g.,  = 90nm in 0.18 μm process

© M. Shabany, ASIC & FPGA Chip Design

 Layout Layers
 Each layout consists of various levels described by different colors.

Top View

Cross Section View

n well
P substrate
wafer

© M. Shabany, ASIC & FPGA Chip Design

Design Rules
 Circuit engineer designs a circuit Deign Rule Check (DRC)
 Process engineer fabricates the design is the interface between them
 DRC : constraints on patterns in terms of minimum width and separation
 DRC: guarantees that the circuit to be manufacturable

© M. Shabany, ASIC & FPGA Chip Design

 Digital VLSI Layout
 Designed and laid-out standard cells
 They are placed & abutted in a chip

Metal 2

Via 12

Metal 1

Several metal layers used for routing

© M. Shabany, ASIC & FPGA Chip Design

 Digital VLSI Layout
 Standard cells are automatically placed and routed using different metal
layers through the corresponding CAD tools

© M. Shabany, ASIC & FPGA Chip Design

 Chip Floorplan
 Start with the pin count (# of I/O and VDD, VSS pads)
 Pads are already designed (provided in the design kit)
Corner
Pads I/O Pads Corner VSS Core
Pads
VDD Core

I/O Pads
Core
I/O Pads

Design

Corner Corner
Pads I/O Pads Pads

© M. Shabany, ASIC & FPGA Chip Design

Chip Floorplan : Pads

Layout View Design View

© M. Shabany, ASIC & FPGA Chip Design

Course Outline
• Course Outline
• Introduction to ASIC/FPGA IC Design
 Integrated Circuits (IC) History
 Digital Design vs. Analog Design
 ASIC vs. FPGA
 Design Abstraction and Metrics
 CMOS as the building block of Digital ASICs
 Layout
 Packaging

© M. Shabany, ASIC & FPGA Chip Design 97

Packaging
 Package:
 Die interface to outside world
 Removes heat from chip
 Mechanical support
 Protects die against humidity
× Introduces delay/parasitics to the chip

 Advanced Package Requirements:

 Electrical: Low Parasitics
 Mechanical: Reliable and Robust
 Thermal: Efficient Heat Removal
 Economical: Cheap

© M. Shabany, ASIC & FPGA Chip Design

Die to Package Substrate
Pad Pitch

Substrate Die

Die Solder bumps

Interconnect
Layers
Pad

Lead Frame
Substrate

Wire Bonding Flip-Chip Bonding

 Small pin count  Large pin count
 Must be attached serially  Connections made simultaneously
 Large inductance  Superior electrical performance
 Cross talk b/w wires  Small cross talk
 Pads on four sides of the chip  Pads at any position on the chip

© M. Shabany, ASIC & FPGA Chip Design

 Packaging
 Package-Socket-PCB interconnection:
Package
 Performance bottlenecks:
 Chip itself
 Package
Socket
 Socket
 Board

© M. Shabany, ASIC & FPGA Chip Design

Package/Socket to Board Interconnection

Through-Hole Mounting Surface Mount

 Mechanically reliable  More wiring space
× Low package density  Higher package density
× Limits routing on the board  Chips at both sides of the board
 Dualin-Line (DIP) (up to 64 pins) × Weak chip-board connection
 PGA (up to 400 pins) × Non-accessible pins for testing

© M. Shabany, ASIC & FPGA Chip Design

 Package/Socket to Board Interconnection
 For very large pin-counts, even surface-mount packaging is not enough!
 Ball Grid Array (BGA)

Solder bumps are used to connect both the die to the

package substrate and the package to the board

© M. Shabany, ASIC & FPGA Chip Design

 Packaging Material/Type
 Package Material:
 Plastics (cheaper)
 Ceramic (Better heat removal)

 Package Types:

© M. Shabany, ASIC & FPGA Chip Design

Packaging Types

Pin Grid Array (PGA) Quad Flat Pack (QFP)

Die
Solder bumps
Interconnect
Layers

Substrate

Ball Grid Array (BGA) Plastic Leader Chip Carrier

(PLCC)

© M. Shabany, ASIC & FPGA Chip Design