CHAP.4 Circuit Characteristics and Performance Estimation

Similar documents
Chapter 4. Circuit Characterization and Performance Estimation

Chapter 6 MOSFET in the On-state

Introduction to Digital Circuits

Outline. Chapter 2: DC & Transient Response. Introduction to CMOS VLSI. DC Response. Transient Response Delay Estimation

EE 560 MOS INVERTERS: DYNAMIC CHARACTERISTICS. Kenneth R. Laker, University of Pennsylvania

EECS 141: FALL 00 MIDTERM 2

Semiconductor Devices. C. Hu: Modern Semiconductor Devices for Integrated Circuits Chapter 6

CHAP.4 Circuit Characteristics and Performance Estimation

EEC 118 Lecture #15: Interconnect. Rajeevan Amirtharajah University of California, Davis

EECE251. Circuit Analysis I. Set 4: Capacitors, Inductors, and First-Order Linear Circuits

Physical Limitations of Logic Gates Week 10a

University of Cyprus Biomedical Imaging and Applied Optics. Appendix. DC Circuits Capacitors and Inductors AC Circuits Operational Amplifiers

L1, L2, N1 N2. + Vout. C out. Figure 2.1.1: Flyback converter

CHAPTER 12 DIRECT CURRENT CIRCUITS

Chapter 7 Response of First-order RL and RC Circuits

Chapter 4 AC Network Analysis

dv 7. Voltage-current relationship can be obtained by integrating both sides of i = C :

UNIVERSITY OF CALIFORNIA AT BERKELEY

EE100 Lab 3 Experiment Guide: RC Circuits

NDP4050L / NDB4050L N-Channel Logic Level Enhancement Mode Field Effect Transistor

Designing Information Devices and Systems I Spring 2019 Lecture Notes Note 17

2.4 Cuk converter example

Topic Astable Circuits. Recall that an astable circuit has two unstable states;

EE 330 Lecture 41. Digital Circuits. Propagation Delay With Multiple Levels of Logic Overdrive

Semiconductor Devices and Models

R.#W.#Erickson# Department#of#Electrical,#Computer,#and#Energy#Engineering# University#of#Colorado,#Boulder#

EE 330 Lecture 40. Digital Circuits. Propagation Delay With Multiple Levels of Logic Overdrive

( ) = Q 0. ( ) R = R dq. ( t) = I t

Capacitors & Inductors

Reading. Lecture 28: Single Stage Frequency response. Lecture Outline. Context

U(t) (t) -U T 1. (t) (t)

Basic Principles of Sinusoidal Oscillators

EE650R: Reliability Physics of Nanoelectronic Devices Lecture 9:

Introduction to AC Power, RMS RMS. ECE 2210 AC Power p1. Use RMS in power calculations. AC Power P =? DC Power P =. V I = R =. I 2 R. V p.

The problem with linear regulators

Non Linear Op Amp Circuits.

4. Electric field lines with respect to equipotential surfaces are

Silicon Controlled Rectifiers UNIT-1

Lecture 28: Single Stage Frequency response. Context

Direct Current Circuits. February 19, 2014 Physics for Scientists & Engineers 2, Chapter 26 1

NDH834P P-Channel Enhancement Mode Field Effect Transistor

NDS332P P-Channel Logic Level Enhancement Mode Field Effect Transistor

Reading from Young & Freedman: For this topic, read sections 25.4 & 25.5, the introduction to chapter 26 and sections 26.1 to 26.2 & 26.4.

6.01: Introduction to EECS I Lecture 8 March 29, 2011

Chapter 28 - Circuits

Topics to be Covered. capacitance inductance transmission lines

Technology Scaling. 9nm. Advanced Digital IC-Design. Content. What happens when technology is scaled? Progress: Described by Gordon Moore

CLOSED FORM SOLUTION FOR DELAY AND POWER FOR A CMOS INVERTER DRIVING RLC INTERCONNECT UNDER STEP INPUT

Chapter 10 INDUCTANCE Recommended Problems:

NDS356P P-Channel Logic Level Enhancement Mode Field Effect Transistor

Basic Circuit Elements Professor J R Lucas November 2001

Modeling the Overshooting Effect for CMOS Inverter in Nanometer Technologies

Phys1112: DC and RC circuits

AO V Complementary Enhancement Mode Field Effect Transistor

Lecture -14: Chopper fed DC Drives

Lecture 13 RC/RL Circuits, Time Dependent Op Amp Circuits

EE141. EE141-Spring 2006 Digital Integrated Circuits. Administrative Stuff. Challenges in Digital Design. Last Lecture. This Class

Capacitors. C d. An electrical component which stores charge. parallel plate capacitor. Scale in cm

XPT IGBT Module MIXA450PF1200TSF. Phase leg + free wheeling Diodes + NTC MIXA450PF1200TSF. Part number

LabQuest 24. Capacitors

Chapter 2: Principles of steady-state converter analysis

Electrical Circuits. 1. Circuit Laws. Tools Used in Lab 13 Series Circuits Damped Vibrations: Energy Van der Pol Circuit

Chapter 4 DC converter and DC switch

dv i= C. dt 1. Assuming the passive sign convention, (a) i = 0 (dc) (b) (220)( 9)(16.2) t t Engineering Circuit Analysis 8 th Edition

RC, RL and RLC circuits

PI74STX1G126. SOTiny Gate STX Buffer with 3-State Output. Features. Descriptio n. Block Diagram. Pin Configuration

Homework-8(1) P8.3-1, 3, 8, 10, 17, 21, 24, 28,29 P8.4-1, 2, 5

V AK (t) I T (t) I TRM. V AK( full area) (t) t t 1 Axial turn-on. Switching losses for Phase Control and Bi- Directionally Controlled Thyristors

Lab 10: RC, RL, and RLC Circuits

SOTiny Gate STX. Input. Descriptio n. Features. Block Diagram. Pin Configuration. Recommended Operating Conditions (1) Pin Description.

ES 250 Practice Final Exam

Problem Set #1. i z. the complex propagation constant. For the characteristic impedance:

HV513 8-Channel Serial to Parallel Converter with High Voltage Push-Pull Outputs, POL, Hi-Z, and Short Circuit Detect

INDEX. Transient analysis 1 Initial Conditions 1

Standard Rectifier Module

NDS355AN N-Channel Logic Level Enhancement Mode Field Effect Transistor

i L = VT L (16.34) 918a i D v OUT i L v C V - S 1 FIGURE A switched power supply circuit with diode and a switch.

3. Alternating Current

CHAPTER 6: FIRST-ORDER CIRCUITS

PI5A3157. SOTINY TM Low Voltage SPDT Analog Switch 2:1 Mux/Demux Bus Switch. Features. Descriptio n. Applications. Connection Diagram Pin Description

EE 330 Lecture 41. Digital Circuits. Propagation Delay With Multiple Levels of Logic Optimally driving large capacitive loads

Physics for Scientists & Engineers 2

Cosmic Feb 06, 2007 by Raja Reddy P

Sequential Logic. Digital Integrated Circuits A Design Perspective. Latch versus Register. Naming Conventions. Designing Sequential Logic Circuits

Lecture Outline. Introduction Transmission Line Equations Transmission Line Wave Equations 8/10/2018. EE 4347 Applied Electromagnetics.

Features / Advantages: Applications: Package: Y4

V L. DT s D T s t. Figure 1: Buck-boost converter: inductor current i(t) in the continuous conduction mode.

Top View. Top View. V DS Gate-Source Voltage ±8 ±8 Continuous Drain Current Pulsed Drain Current C V GS I D -2.5 I DM P D 0.

8.022 (E&M) Lecture 9

ELEN 624 Signal Integrity

Top View. Top View S2 G2 S1 G1

Chapter 8 The Complete Response of RL and RC Circuits

AC Circuits AC Circuit with only R AC circuit with only L AC circuit with only C AC circuit with LRC phasors Resonance Transformers

Features / Advantages: Applications: Package: Y4

Features / Advantages: Applications: Package: Y4

A Note of Widening on the Redshift Mechanism. June 23, 2010.

IE1206 Embedded Electronics

Application Note AN Software release of SemiSel version 3.1. New semiconductor available. Temperature ripple at low inverter output frequencies

2.9 Modeling: Electric Circuits

First Order RC and RL Transient Circuits

Transcription:

HAP.4 ircui haracerisics and Performance Esimaion 4. Resisance esimaion R ρ l w (ohms) where ρ Resisiviy Thickness l onducor lengh w onducor widh l R Rs w where Rs Shee resisance (Ω/square) in 0.5µm o 1.0µm MOS processes 00/4/18 1

* hannel resisance R c k W 1 gs 1 ( ) W µ ox gs 1 k, k 1,000~30,000Ω/square µ ( ) ox 1 gs Typical shee resisance for conducors (Table 4.1) hannel resisance + 0.5% per Meal and poly + 0.30% per Well diffusion + 1.00% per Resisance for nonrecangular regions (Fig 4.) 4.3 apaciance esimaion Running/Swiching speed of MOS Parasiic MOS capaciance + Runners (wired poly, meal, diffusion) x (device and conducor resisance) Toal loading capaciance 1. Gae capaciance (of oher inpu). Diffusion capaciance (of he drain regions) 3. Rouing capaciance (of connecions beween O/P & I/P) To esimae he speed of he device ( R and informaion) 4.3.1 MOS-apacior haracerisics (**wihou Source and Drain) Accumulaion g < 0 Gae volage Depleion g 0 Inversion g > 0 00/4/18

(a) Accumulaion layer is direcly conneced o subsrae. Gae capaciance can be approximaed by o ε ε SiO 0 ox A (4.4) where A Area of gae Relaive permiiviy Of SiO 3.9 ЄSiO Dielecric consan Є0 Permiiviy of free space (b) Depleion mode : Funcion of : (1) Doping concenraion (N) () Elecronic charge (q) (3) Depleion deph (d) (c) Inversion 00/4/18 gb o in series wih dep, where gb(gae-o-bulk(subs)) 0 dep gb + (varied as a funcion of gs) dep 0 0 0 ε 0 ( 1) dep 0 + εsi d dep dep A min ε Si, low frequency, high frequency ( < 100 Hz ) 3

MOS Device capaciance ircui symbol gs, gd Gae-o-channel capaciance (Inpu par) sb, db Source/drain diff-o-bulk (subs) capaciance (Oupu par) gb Gae-o-bulk capaciance (Inpu) Toal gae capaciance g of an MOS device (or so-called inpu capaciance) g gb + gs + gd (a) Off-region : ( gs < ) No channel gs gd 0 g gb o + dep (b) Non-sauraion region : ( gs - > ds ) gd gs ε0ε ox 0 gb 1 SiO A (c) Sauraed region : ( gs - < ds ) hannel is heavily invered pinch off gd 0 gs ε 0 ε SiO 3 ox A Approximaion of MOS gae capaciance 00/4/18 4

onservaive approximaion: g o ox A where ox : hin-oxide capaciance per uni area. ε0ε SiO 35 µ m ox ο ο 4 ( 17) 10 pf 100Α ~ 00Α ox () (W) (ox) 4 ( 1µ m) ( µ m) ( 5.5 10 pf µ m ) g 0. 005pF ( λ 0.5µm ) Uni ransisor: A ransisor ha can be convenienly conneced o meal a boh source and drain. 4.3.3 Diffusion (source and drain) capaciance d : Proporional o oal diffusion-o-subsrae juncion area 1. Base area +. he area of he ( ab ) + ( a + b ) ja jp " sidewall " periphery where 1. ja: Juncion capaciance 4 4 pf / µ m (3 10 NMOS,5 10 PMOS). jp: 4 4 Periphery capaciance pf / µ m (4 10 NMOS,4 10 PMOS) 3. a : Widh of diffusion area (µm) 4. b : engh of diffusion area (µm) 00/4/18 5

Thickness of depleion layer depends on he volage across he juncion, boh ja and jp are funcions of juncion volage (j) j ( a p) ( a, p) m j, jo 1 (m0.3~0.5) b j Juncion volage (negaive for reverse bias) jo(a,p) Zero-bias capaciance (j 0) b Build-in juncion poenial ~0.6 Diffusion capaciance form sb and db in Fig.4.5 4.3.5 Rouing capaciance Rouing capaciance beween meal and poly layers and he subsrae can be modeled as a parallel-plae capaciance. ε A plus fringing field ha occurs a he edges of he conducor due o is finie hickness. Toal capaciance (a) Parallel-plae capaciance of widh (w-/) + (b) A cylindrical capaciance of radiaes (/) w π ε + h h h h ln 1 + + + Empirical formula 0.5 0.5 w w ε + 0.77 + 1.06 + 1.06 h h h 00/4/18 6

In general, meal and poly lines have higher capaciance value han he prediced values (due o he fringing facor) 4.3.5. Muliple conducor capaciance (Fig.4.11 and 4.1) 4.3.6 Disribued R effecs ong wires arge R Transmission line effecs R-secion model dj d ( I I ) j 1 j ( j 1 j ) ( j j + 1) n( n + 1) R R n R (n: secion number) As number of secions increase Differenial equaions x disance from inpu d d rc r resisance per uni lengh d dx c capaciance per uni lengh x ( propagaion ime) kx ( wire of lengh x) rc 1 ( l lengh of he wire ) l Example : r 0 Ω µm, 4 c 4 10 pf µ m (a) Poly bus of -mm, 15 p 4 10 l 16ns (b) Poly bus (-mm) poly bus (1-mm) + buffer p 4 10 15 4 ns + τ buf 15 ( 1000 ) + τ + 4 10 ( 1000 ) + 4 ns buf 8ns + τ buf 00/4/18 7

onusions: (a) Use poly for only local inerconnecion or very slow global inerconnecion in wo-meal process. (b) As speed increases, meal layer will also have R effec Add buffer Widen he line ( R, increases a lile ) Shoren lengh ( l ) R delay model (imporan in calculaing clock disribuion in highspeed, high-densiy chips), R : oal lumped R, of he line. τ rcl apaciance design guide (Sec.4.3.7) Wire-lengh design guide (Table 4.7) Inducance (Sec.4.4) 4.5 Swiching haracerisics Swiching speed : 1. Time o charge he load capacior oward. Time o discharge oward SS (a) Rise ime (r): Time for a waveform o rise from 10%~90% of is seady-sae value. (b) Fall ime (f): fall from 90%~10% (c) Delay ime(d): ime difference beween inpu ransiion (50%) and he 50% oupu level. 00/4/18 8

Fall ime ( f ) : X1: NMOS u-off X: NMOS sauraed region X3: NMOS nonsauraion (inear mode) Analysis of f IN, (gs ), (iniial value) (a) f1 ou drops from 0.9 o (-n) (X dsgs-) (sauraion) (b) f ou drops from (-n) o 0.1 (dsgs- X3) (nonsauraion, linear mode) f1 (NMOS is in sauraion mode) d d ou ( ) 0 n + n (discharge curren is consan) f 1 f 1 0.9 n ( ) n ( n 0.1 ) ( ) n n n d ou 00/4/18 9

f (NMOS is in linear mode) IN DS ou I DS n ( ) ou n ou d d ou n ( ) ou n ou ln 19 0 [ n] f n ( 1 n), n n/ f f f 1 + f k ( Eq.4.37 ) n ( k 3 ~ 4) Increases speed (How o opimize MOS circui speed?) - Reduce load capaciance - Increase n - ow supply volage low speed 4.5.1. Rise ime ( p 0.1) ( 1 p) ( 19 p) r + ln 0 p p 1 p f k p ( k 3 ~ 4) Equal size n,p devices, np f r!! f r n p W p ( ~ 3) Wn 00/4/18 10

4.5.1.3 Delay ime * Approximaion: r f dr df An alernaive formula (a) 1 n ( 1 n) + ln ( ) df AN, AN n 1 n 1 n 0 (AN 0.83) wih nn/, 0ou/ 0 (b) dr A p A p p 0.83 (c) Average gae delay for rising and falling ransiion av dr + df SPIE simulaion (Figure 4.0) Empirical delay model: Back subsiue ino Eqs. (4.46) & (4.47) o obain AN and AP for Wp Wn, AN AP 0.36 dr 0.36 /n 4.5.3 Gae delays For pull-down case W n W p, µ n1 n n3 ox W 1 n neff n1 n 1 + 1 + 1 3 ( ) n3 00/4/18 11

For pull-up case: p,eff p (only one urns on), p,eff 3p (hree urn on) For p 0.3 n r k ( 0.3 ) Mobiliy difference Graphical undersanding n, f k n 3 串聯 f r 1 Graphical Rule n series 3 W µ ox τ series k n 3 Fall ime: m n-devices in series f`m f Rise ime: m p-devices in series r`m r Fall ime: m n-devices in parallel (all urn on) f`f/m Rise ime: m p-devices in parallel (all urn on) r`r/m 4.5.4.3 Swich-level R models 00/4/18 1

(A) Simple R model : (R, are lumped ogeher) df Σ Rpulldown Σ pulldown-pah dr (RN1+RN+ RN3+RN4) (ou+ab+ bc+cd) Rp ou Effecive Resisance R Reff (/W) (B) Penfield-Rubensein model : calculae delays in generalized R ree (ladder) Ri Summed Resisance from d Σ Ri i poin i o power or ground i apaciance a poin i df (RN1 cd)+[(rn1+rn) bc]+[(rn1+rn+rn3) ab]+[(rn1+rn+rn3+rn4) ou] 4.5.4.4 Macro modeling (Daa book) Tswin: Inpu waveform in: Inpu capaciance Tbeou: Delay hrough he gae Tswou: Oupu waveform : Oupu capaciance In ASI designs, logic gaes are reaed as simple delay elemens. d inernal + k oupu Oupu delay, which is proporional o oupu loading capaciance. Fixed inernal delay. 00/4/18 13

SPIE Simulaion Experimens: * Parameer Example: r 0.55 + k.1ns f 0.4 + k 3.8ns ( k is in pf ) 00/4/18 14

4.5.4.5 Body effec (1) Body Effec: γ sb (hange in hreshold volage is a funcion of Source-o-subsrae bias volage). Beer Bad () Poin D rises o abou 1.7vols before being discharged o ground ( 上圖所示 ) (3) Nodes cd, bc, and ab are a an n-hreshold below (~3.1vols). When N1B urns on. Nodes ab,bc,and cd are pulled o ground in ha order ( 下圖所示 ) (4) When >> inernal cap, his effec can be minimized. Sraegy o handle Body effec (1) Place ransisors wih laes arriving signals neares he oupu of he gae. The early signals discharge inernal nodes, and he body effec is minimized. () Minimize inernal capaciances: If diffusion wire is used o minimize he gae geomery, ry o use i a he oupu raher han on some inernal nodes. (3) Body effec is essenially a dynamic problem involving he charging of parasiic capaciances. * onclusion of Secion 4.5: Model of ransisors and parasiic capaciances mus be accurae so ha AD ools can work well. 00/4/18 15

4.6 MOS-Gae Transisor Sizing * For rf, Wp(-3)Wn Increase layou area and dynamic power consumpion. *Approximaed delay for an inverer pair (a) inv-pair f + r (WpWn) R(3eq) + ( R )(3eq) 3Req +3 Req 6Req R ON Res of a uni-size ransisor eqg+d (gae and diffusion cap.) (b) inv-pair (WpWn) R(eq)+()(R)(eq) fall rise 6Req * n/p raio rise / fall inv (Inverer hreshold volage) inv dd+p+n 1+ n p n p, (Eg.4) whennp, n p invdd/ In self-loaded circuis minimum-sized devices may be used o reduce power dissipaion and increase circui packaging densiy. When he circuis have o drive any significan rouing load, his opimizaion does no apply and he n- and p- devices should be sized o yield equal rise and fall ime. 00/4/18 16

4.6. Sage raio (Transisor Sizing skills) * Used in applicaions of (1) ong bus () I/O Buffers (3) Pads (4) Off-chip capaciive load To minimize he delay beween inpu and oupu while minimizing he area and power dissipaion. d: he average delay of a minimum-sized inverer driving anoher minimum-sized inverer. Delay of each sage (d ): R R/a, a d a d n-sage n(a d ) Toal delay /gra n ( produc of n sages) n lnr lna Toal delay lnr lna a d (Noe ha lnr, d are consans) Minimum delay ae (.714) 00/4/18 17

4.7 Power dissipaion MOS power sources: (1) Ps: Saic dissipaions (leakage curren) () Pd: Dynamic dissipaions - Swiching ransien curren from dd o GND. - The Energy/power o charge and discharge 4.7.1 Saic Power Dissipaion >No curren > Ps0, Reverse-biased parasiic diodes (leakage curren) eakage curren : i o is (e qv /KT -1) 0.1nA~0.5nA per device a room emperaure. where is: Reverse sauraion curren v: Diode volage Ps n leakage curren x supply volage, 1 where n oal device number 00/4/18 18

4.7. Dynamic Power Dissipaion 0 1,1 0 n,p devices are boh on Shor circui curren pulse from dd o ss As he increases, he discharge or charge currens sar o dominae he curren drawn from he power suppliers. 00/4/18 19

** Dynamic Power Dissipaion: harge/discharge *Pd 1 p/ p in() ou d + 1 p 0 p ip() (dd-ou) d p/ in() n-device ransien curren dou d ip() p-device ransien curren d(dd-ou) d Pd Average dynamic power o charge/discharge Pd dd p ou dou + 0 p (dd-ou) d(dd-ou) 0 dd dd / p dd fp 00/4/18 0

00/4/18 1

4.7.3 Shor-circui dissipaion PscImean * dd I mean [ 1 T 1 3 I() d + 1 T I() d ] Assume ha n-p, ßnßpß Imean * T (in()-) d 1 wih 1. in() dd r * (linear approximaion). 1 dd * r, r,(rfrf) (NMOS urn on) Psc rf 1 (dd-)3 p Funcions of ß, r, f of inpu waveform. * P oal Ps+Pd+Psc Percenage of aciviy * oal * dd p (Simulaor) (AD ool) I * R oal-swich * dd Toal-no-of-cycles * p * Power economy (achieve low-power and low-energy MOS designs) -Reduce leakage (use complemenary logic + minimize diffusion area) -dd (dd ) : mos effecive. - (design) and f clk (no desirable) -Noe ha dd --> Speed (r,f ) 00/4/18

4.9 harge Sharing -Bus is modeled as a capacior b Qbb* b, Qss * s Toal charge: Q T Qb+Qs b*b+s*s Toal capaciance: T b+s (when swich is closed) 4.11 Yield R Q T b*b+s*s T b+s If bdd, b>>s b R dd [ b+s ] (Why?) To make reliable ransfer from b o s,s<<b (In general, b> 10 s) Problem involving charge sharing occurs in dynamic logic gae designs (haper 5.) Funcion of (1) Technology () hip area (3) ayou Yield(Y) No. of oal chip on wafer Toal no. of chips x 100% (a) Y e AD Achip area (Seed s model) Ddefec densiy (lehal defecs per cm ) (for large chip and Y<30%) -AD 1- e (b) Y[ AD ], for small chips & Y>30% (Murphy s model) (c) Yield decreases significanly as chip area in increased!! 00/4/18 3