Minh Quach. Signal Integrity Consideration and Analysis 4/30/2004. Frequency & Time Domain Measurements/Analysis

Similar documents
Signal Integrity Tips and Techniques Using TDR, VNA and Modeling. Russ Kramer O.J. Danzy

Aries QFP microstrip socket

Taking the Mystery out of Signal Integrity

A Technical Discussion of TDR Techniques, S-parameters, RF Sockets, and Probing Techniques for High Speed Serial Data Designs

Aries Kapton CSP socket

High Speed Digital Systems Require Advanced Probing Techniques for Logic Analyzer Debug

The Challenges of Differential Bus Design

VLSI is scaling faster than number of interface pins

Calibration and De-Embedding Techniques in the Frequency Domain

Custom Interconnects Fuzz Button with Hardhat Test Socket/Interposer 1.00 mm pitch

Advanced Signal Integrity Measurements of High- Speed Differential Channels

Limitations And Accuracies Of Time And Frequency Domain Analysis Of Physical Layer Devices

Improving TDR/TDT Measurements Using Normalization Application Note

The Practical Limitations of S Parameter Measurements and the Impact on Time- Domain Simulations of High Speed Interconnects

Agilent Time Domain Analysis Using a Network Analyzer

Probing Techniques for Signal Performance Measurements in High Data Rate Testing

Keysight Technologies Signal Integrity Tips and Techniques Using TDR, VNA and Modeling

The Facts about the Input Impedance of Power and Ground Planes

DesignCon Design of Gb/s Interconnect for High-bandwidth FPGAs. Sherri Azgomi, Altera Corporation

Application Note 5525

Demystifying Vias in High-Speed PCB Design

Bill Ham Martin Ogbuokiri. This clause specifies the electrical performance requirements for shielded and unshielded cables.

Logic Analyzer Probing Techniques for High-Speed Digital Systems

T est POST OFFICE BOX 1927 CUPERTINO, CA TEL E P H ONE (408) FAX (408) ARIES ELECTRONICS

High Speed Characterization Report

The data rates of today s highspeed

High Speed Characterization Report

High Speed Characterization Report

FIBRE CHANNEL CONSORTIUM

Measuring PCB, Cable and Interconnect Impedance, Dielectric Constants, Velocity Factor, and Lengths

Time Domain Reflectometry (TDR) and Time Domain Transmission (TDT) Measurement Fundamentals

Optimizing On Die Decap in a System at Early Stage of Design Cycle

IEEE CX4 Quantitative Analysis of Return-Loss

Microcircuit Electrical Issues

Design and experimental realization of the chirped microstrip line

High Speed Characterization Report

Where Did My Signal Go?

Characterization and Measurement Based Modeling

Signal Integrity Testing with a Vector Network Analyzer. Neil Jarvis Applications Engineer

EE273 Lecture 3 More about Wires Lossy Wires, Multi-Drop Buses, and Balanced Lines. Today s Assignment

SIGNAL INTEGRITY ANALYSIS AND MODELING

B2501 B Series 0.5mm (.0197) Pitch

High Data Rate Characterization Report

High Data Rate Characterization Report

Improving CDM Measurements With Frequency Domain Specifications

Application Note. Signal Integrity Modeling. SCSI Connector and Cable Modeling from TDR Measurements

DL-150 The Ten Habits of Highly Successful Designers. or Design for Speed: A Designer s Survival Guide to Signal Integrity

High Speed Characterization Report

if the conductance is set to zero, the equation can be written as following t 2 (4)

EE290C - Spring 2004 Advanced Topics in Circuit Design

Aries Center probe CSP socket Cycling test

Relationship Between Signal Integrity and EMC

ECE 497 JS Lecture - 22 Timing & Signaling

Aries Kapton CSP socket Cycling test

High Speed Characterization Report. Contact Plating Effects on Signal Integrity Gold on Post / Gold on Tail vs. Gold on Post / Matte Tin on Tail

Impedance and Electrical Models

Keysight MOI for USB Type-C Connectors & Cable Assemblies Compliance Tests (Type-C to Legacy Cable Assemblies)

TDR Primer. Introduction. Single-ended TDR measurements. Application Note

Aries CSP microstrip socket Cycling test

High Data Rate Characterization Report

High Speed Characterization Report

Ensuring Signal and Power Integrity for High-Speed Digital Systems

High Speed Competitive Comparison Report. Samtec MMCX-J-P-H-ST-TH1 Mated With MMCX-P-P-H-ST-TH1 Competitor A (Mated Set) Competitor B (Mated Set)

Burn-in & Test Socket Workshop

Characterization Methodology for High Density Microwave Fixtures. Dr. Brock J. LaMeres, Montana State University

Design Guide for High-Speed Controlled Impedance Circuit Boards

High Data Rate Characterization Report

High Speed Characterization Report

TileCal Analogue Cable Measurement Report

Keysight MOI for USB Type-C Connectors & Cable Assemblies Compliance Tests (Type-C to Legacy Cable Assemblies)

Migrating 4195A to E5061B LF-RF Network Analyzer. April 2010 Agilent Technologies

High Speed Characterization Report

Differential Signal and Common Mode Signal in Time Domain

Advanced Transmission Lines. Transmission Line 1

Analysis on the Effectiveness of Clock Trace Termination Methods and Trace Lengths on a Printed Circuit Board

High Speed Characterization Report

Low Jitter, Low Emission Timing Solutions For High Speed Digital Systems. A Design Methodology

AUTOMOTIVE ETHERNET CONSORTIUM

High Speed Characterization Report

DesignCon 2003 High-Performance System Design Conference (HP3-5)

High Speed Digital Design & Verification Seminar. Measurement fundamentals

Challenges and Solutions for Removing Fixture Effects in Multi-port Measurements

DL-150 The Ten Habits of Highly Successful Designers. or Design for Speed: A Designer s Survival Guide to Signal Integrity

High Speed Characterization Report

High Speed Characterization Report

Serial ATA International Organization

Validation & Analysis of Complex Serial Bus Link Models

A VIEW OF ELECTROMAGNETIC LIFE ABOVE 100 MHz

Data Mining 12-Port S- Parameters

High Speed Characterization Report

Student Research & Creative Works

Agilent E2695A SMA Probe Head for InfiniiMax 1130 Series Active Oscilloscope Probes. User s Guide

High Speed Characterization Report

How to Read S-Parameters Like a Book or Tapping Into Some Of The Information Buried Inside S- Parameter Black Box Models

PDN Probes. P2100A/P2101A Data Sheet. 1-Port and 2-Port 50 ohm Passive Probes

06-496r3 SAS-2 Electrical Specification Proposal. Kevin Witt SAS-2 Phy Working Group 1/16/07

RF Characterization Report

Keysight Technologies Using the Time-Domain Reflectometer. Application Note S-Parameter Series

A Novel Embedded Common-mode Filter for above GHz differential signals based on Metamaterial concept. Tzong-Lin Wu

Craig Rickey 8 June Probe Card Troubleshooting Techniques

Transcription:

Minh Quach. Signal Integrity Consideration and Analysis 4/30/2004 Frequency & Time Domain Measurements/Analysis

Outline Three Measurement Methodologies Direct TDR (Time Domain Reflectometry) VNA (Vector Network Analyzer) Fundamental Transmission line & TDR Theory Lossless and lossy line Filter Rise Time calculation Bandwidth Interpretation Page 2

Outline (Cont) Signal Integrity Reflection Dispersion AC loss (skin effect and dielectric loss) Cross Talk Ground Bounce Inadequate power bus decoupling Differential Pair Common & Differential Mode Even & Odd Mode Page 3

Test Environment Hardware Page 4

Interconnect Vs Vin_open VL_open Vout_open VtStep SRC4 Vlow=0 V Vhigh=1 V Delay=0 nsec Rise=1 nsec t R R9 R=50 Ohm L L13 L=.01 nh R= C C10 C=.3 pf SLIN TL9 Subst="SSub2" W=10.0 mil L=6 in L L12 L=.01 nh R= C C9 C=.3 pf SLIN TL10 Subst="SSub2" W=10.000 mil L=3 in L L11 L=4.34 nh R= R R8 R=50 kohm Page 5

Measurement Methods Direct Measurement TDR (Time Domain Reflectometry, Time Domain) VNA (Vector Network Analyzer, Frequency Domain) Page 6

Measurement Method TDT (Transmitted Signal) V inc Incident Wave DUT TDR (Reflected signal) ZLoad Page 7

Direct Measurements Page 8

Freq = 100MHZ Page 9

Freq = 200MHZ Page 10

Freq =400 MHZ Page 11

Fundamental Differences Between Analog and Digital in Interconnect Applications Analog Digital -- Impedance Matching -- Impedance Matching for max power transfer to minimize signal distortion -- Frequency Domain -- Time Domain -- VNA -- TDR -- Amplitude loss -- Edge timing degradation -- Frequency -- Edge rate and frequency Page 12

Using the Reflected Waveform to Construct the Spice Model Direct Measurement TDR (Time Domain Reflectometry, Time Domain) Using the Reflected Waveform to construct the spice model Using ADS (Advanced Design Software) Transmission Line Theory VNA (Vector Network Analyzer, Frequency Domain) Page 13

V inc Incident Wave Lump or Distributed DUT ZLoad TDT (Transmitted Signal) TDR (Reflected signal) Page 14

Lump or Distributed T prop delay T rise Lump Parameter if: T rise > T prop delay * 6 Page 15

High Frequency Socket Model in Lump Parameters Valid up to 7GHz R R1 R=300 Ohm R R4 R=0 Ohm L L2 L=.1 nh R= C C1 C=.170 pf L L1 L=.80 nh R= C C2 C=.170 pf L L3 L=.10 nh R= Page 16

What is a transmission line A transmission line is any pair of conductor that has dimensions are constant for the length of the line that are used to move electromagnetic energy from one place to another. In printed circuit boards, this is typically a trace and one or more power planes. Power lines are transmission lines. Coaxial cable is a transmission lines. Twisted pairs are transmission lines. Electromagnetic waves are moving in a transmission line not electron Page 17

Characteristic Impedance Zo Zo Transmission line V = (R+jWL)I Short circuit Zo I = (G+jWC)V Open circuit Zo = (R+jWL) (G+jWC) For Low Loss L C Page 18

Transmission Line Fundamental Software Page 19

Time Domain Reflectometry (TDR) Page 20

Transmission Line Basic * p Page 21

Discontinuity Examples Page 22

Discontinuity Examples Page 23

Reflection Response is a function of System Rise Time Page 24

Impedance Profile Z 1 Z 2 Z 3 Propagation Page 25

TDR responses different due to lossy line SMA connector SMA connector 0.220 PCB 0.210 CHANNEL3 0.200 0.190 0.180 0.170 50 51 52 53 time, nsec Page 26

Filter Rise Time Calculation T risetime = 2.2 Tc = 2.2 RC T risetime = 2.2 Tc = 2.2 L/R T risetime = 3.4(LC)^1/2 Rise time of RC filter Rise time of RL filter Rise time of LC filter Page 27

Filter Rise Time Calculation (cont) C = 0.5pF R=50 Tr in = 400ps Tr filter =2.2 X 50 X.5 pf = 55pS Tr filter =2.2 X 25 X.5 pf = 27.5pS T rise composite = ( Tr in^2 + Tr filter^2 ) ^1/2 T rise composite = 403.8 ps T rise composite = 400.9pS Page 28

T risetime = 2.2 Tc = 2.2 RC R=Zo/2 T risetime = 2.2 Tc = 2.2 L/R R=2*Zo Page 29

Filter Rise Time Calculation (cont) Note:Tr=k/BW3db Where: K from 0.338 to 0.35 for gaussian pulse and single pole exponential decay respectively C = 0.5pF R=50 3 GHZ SCOPE Tr in = 400ps Tr filter =2.2 X 50 X.5 pf = 55pS Tr scope = 0.361/Frms = 120ps T rise composite = ( Tr in^2 + Tr filter^2 ) ^1/2 T rise composite = 403.8 ps Output of filter Note:Tr=k/BWrms Where: K from 0.36 to 0.549 for gaussian pulse and single pole exponential decay respectively T rise composite = ( Tr in^2 + Tr filter^2 + Tr scope^2 ) ^1/2 T rise composite = 421 ps Scope Display Page 30

Transmitted Rise Time of 2 different type of SMA cables Transmitted Rise Time of 2 different type SMA cables 1000 900 800 700 y = 0.032x 2 Rise Time in ps 600 500 400 T risetime = K * L 2 300 200 y = 0.004x 2 100 0 0 50 100 150 200 250 300 350 Cable Length in inches 20GHZ BW SMA cable low BW SMA cable Page 31

Cable Length VS Rise Time of high BW sma cable. Using Reflected Signal to Predict Transmitted Signal Rise Time VS SMA (18GHZ Bandwidth) Cable Length 1600 1400 1200 Tr = KL 2 ; Reflect signal travels twice as the distance on the same cable (2L); Reflect rise time / Transmit rise time = 4 Reflected Signal Rise Time Rise Time in ps 1000 800 600 Tr = K*L^2 TR=0.016L^2 Transmitted Signal Rise Time 400 200 TR=0.004L^2 0 0 50 100 150 200 250 300 350 Cable Length in inches Transmit Reflect Best Fit Transmit Best Fit Reflect T incident TDT TDR Page 32

TDR indicates a large Capacitor presents at the load causing rise time to degrade to 5ns Page 33

Q/A Page 34

When Speed or Wire/Net length are important The edge rate (rise or fall time) of a signal is fast enough that the signal can change from one logic state to the other less time than it takes the signal to travel the length of the wire/net. Example: Rise time = 1ns, pcb trace length = 6 inch. v = 6 inch/ns. Overshoot and undershoot begins to show at ¼ of this length (1.5 inch). Page 35

When should impedance be controlled When the length of the transmission line exceeds ¼ of the transition electrical length ( TEL = Tr X Velocity) Tr= 1ns, Velocity = 6 inch/ns; TEL = 6 And When reflections may cause malfunction from: Overshoot or undershoot and The logic technology support termination Page 36

Reflection causes by un-proper termination with fast rise time R R2 R=10 Ohm Vin SLIN TL2 Subst="SSub1" W=6 mil L=6 in vout Term Term2 Num=2 Z=400 Ohm Rise time=0.5ns Period=10ns tprop =1ns (6 FR4 PCB) Rise time=25ns R R3 R=10 Ohm Vin2 SLIN TL3 Subst="SSub1" W=6 mil L=6 in vout2 Term Term3 Num=3 Z=400 Ohm Period=100ns Page 37

Reflection at Vo cause by Electrical length is longer than rise time without proper termination 1.6 1.4 1.2 1.0 0.8 L < ¼ TEL L > ¼ TEL vout_ref, V Vin2, V vout2, V vout, V 0.6 0.4 0.2 0.0-0.2-0.4-0.6 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 time, nsec Page 38

Frequency/VNA VS. Time Domain/TDR/Scope Frequency Domain Time Domain Steady state Narrow Bandwidth Better dynamic range More accurate Calibrate out fixture Can convert to time domain Data analysis more insightful Require calibration Longer to setup For lump parameter Transition Analysis Large Bandwidth Faster setup Less expensive/more common Direct representation For long transmission line Page 39

Square wave composition Square Wave Composition 0.8 0.6 0.4 0.2 Amplitude -0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4-0.4-0.6 f(t) = 2/3.14* SUM [ 1/n *SIN(2*3.14*F*t*n)]; Where n is 1,3 etc (odd number) -0.8 Time 1st har 3nd har 5rd har 7rd har 9rd har 11th har 13th har 15th har 17th har sum-to-5th sum-to-11th sum-to-17th Page 40

Frequency Spectrum of square wave Frequency Spectrum of the Sqaure Wave 1.2 1 Relative Amplitue 0.8 0.6 0.4 0.2 0 1 3 5 7 9 11 13 15 17 19 21 Harmonic # Page 41

S parameter Definition S 11 = V ref,1 / V inc,1 ; S 11db = 20 log (V ref,1 / V inc,1 ) S 21 = V port2 / V port1 ; S 21db = 20 log (V port2 / V port1 ) For passive DUT (linear system) S 22 = S 11 ; S 12 = S 21 Port 1 S 21 DUT Port 2 S 11 S 22 S 12 Page 42

Edge Rate VS Frequency Tr = 0.35/F 3db 1.5 Tr @90% 1 0.5 0 0 2 4 6 8 10 12 14 16-0.5-1 Tr @10% -1.5 Page 43

Bandwidth Selection Rise Time Frequency BW rise time = 0.35/Tr BW1 system = 3*BW rise time BW2 system = 6 *Frequency BW system = MAX ( BW1, BW2 ) Page 44

Bandwidth Selection Clock Frequency @ Rise Time Tr. BW rise time = 0.35/Tr Interconnect BW with minimum signal distortion and loss BW interconnect = MAX ( 3 X BW rise time, 6 X Clock Frequency ) 400MHz frequency with: Rise time 3 X BW rise time 6 X BW signal BW interconn 200ps 5.25 GHZ 2.4 GHZ 5.25 GHZ 400ps 2.62 GHZ 2.4 GHZ 2.62 GHZ 800ps 1.31 GHZ 2.4 GHZ 2.4 GHZ Page 45

Bandwidth Selection Clock Frequency @ Rise Time Tr. BW rise time = 0.35/Tr Interconnect BW with minimum signal distortion and loss BW interconnect = MAX ( 3 X BW rise time, 6 X Clock Frequency ) Results from simulation: 3 X BW rise time causes approximately 3 % rise time degradation 6 X Frequency causes approximately 2 % amplitude loss Page 46

Signal Integrity Issues Impedance mismatch Reflection Crosstalk Ground bounce Inadequate power bus decoupling Propagation delay Dispersion Loss Page 47

Signal Integrity Components Signal Integrity Discontinuity (Connectors, via) Dispersion AC loss Skin Effect Dielectric Loss (loss tangent) Ground Bounce Inadequate power bus decoupling Differential Pair Less sensitive to components above Page 48

Dispersive Loss Reflection at the Interface of 2 Transmission lines: One cable is different loss than the other (Even the characteristics impedance is the same) Reflection coefficient is not zero and also frequency dependence. Page 49

AC LOSS Skin Effect Loss Dielectric Loss Page 50

Rise Time Degradation After Signal Propagates Through 5mil Width, 32 Long for the FR4 PCB 220 200 Vout_32in_pcb, mv vin, mv 180 160 140 120 100 80 60 Input Rise Time = 30pS Output Rise Time of 32in pcb = 1nS 40 20 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 time, nsec Page 51

Skin Effect Highest inductance Lowest inductance Med inductance Page 52

Zo = (R+jWL) (G+jWC) Excerpt from PCB Design West Conf 3/21/ 2001 (GigaTest Labs) Page 53

Excerpt from PCB Design West Conf 3/21/ 2001 (GigaTest Labs) Page 54

Excerpt from PCB Design West Conf 3/21/ 2001 (GigaTest Labs) Page 55

AC Loss (Cont) Skin Effect Loss Dielectric Loss Page 56

Excerpt from PCB Design West Conf 3/21/ 2001 (GigaTest Labs) Loss Tangent Page 57

Zo = (R+jWL) (G+jWC) Excerpt from PCB Design West Conf 3/21/ 2001 (GigaTest Labs) Page 58

Dielectric Loss 0.025 Loss Tangent of Various Materials 0.02 0.02 Loss Tangent 0.015 0.01 0.008 0.012 0.005 0 0.0002 0.0012 Teflon Roger6002 Getek Nelco/N70 00-1 FR4 Page 59

Skin Effect & Dielectric Loss of an FR4, 4 mils, 10mils trace width and 24 in length (ADS simulation) 0-2 Skin Loss -4-6 db(s(13,14)) db(s(3,4)) db(s(1,2)) -8-10 -12 Skin & Dielectric Loss -14-16 -18 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 4.4 4.8 5.0 freq, GHz Dielectric Loss Page 60

Skin & Dielectric Loss of an FR4, Roger of 4mils and 10mils Width, and 24 long (ADS simulation) 0-2 -4-6 Roger Dielectric Loss Roger, 10mils trace W db(s(17,18)) db(s(11,12)) db(s(15,16)) db(s(13,14)) db(s(9,10)) -8-10 -12-14 FR4, 4mils trace W FR4, 10mils trace W Roger, 4mils trace W -16-18 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 4.4 4.8 5.0 freq, GHz Page 61

2.5Gb/s data for FR4 PCB with 4mils trace width, 24 long. 5Gb/s Output Signal Input Signal 2.5Gb/s Output Signal Page 62

Compensation Circuits R R11 R=50 Ohm C C1 C=800 pf C C2 C=40 pf C C3 C=2 pf t Vs_equal SLIN TL3 Subst="SSub1" W=4.0 mil L=24 in Vin_before_equal R R1 R=1.77 Ohm R R2 R=1.77 Ohm R R3 R=4.42 Ohm R R4 R=4.42 Ohm R R5 R=10.9 Ohm Vo_equal R R6 R=10.9 Ohm R R10 R=50 Ohm VtPulse SRC1 Vlow=0 V Vhigh=2 V Delay=0 nsec Edge=linear Rise=100 psec Fall=100 psec Width=3 nsec Period=9 nsec R R7 R=706 Ohm L L1 L=2 uh R= R R8 R=281 Ohm L L2 L=100 nh R= R R9 R=109 Ohm L L3 L=5 nh R= Compensation circuit from ITC 2002 by Wolfram Humann, Agilent Technologies Page 63

S-Par of before and after compensation Frequency Spectrum of the Sqaure Wave db(s(3,4)) db(s(1,2)) 0-5 -10-15 -20-25 -30-35 -40-45 -50-55 -60 Before Compensation After Compensation Relative Amplitue 1.2 1 0.8 0.6 0.4 0.2 0 1 3 5 7 9 11 13 15 17 19 21 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 freq, GHz Harmonic # Page 64

Time Domain of Before and After Compensation Vi compensation = 2 Vin Vin_s_equal, V Vin_equal, V Vin, V Vo, V Vo_equal, V 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0-0.1 Vi After compensation Vi Before compensation Vo After compensation Vo Before compensation 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 time, nsec Vi at the compensation circuit Page 65

5Gb/s data for FR4 PCB with 4mils trace width, 24 long. Before Compensation After Compensation Page 66

Plot excerpt from the High-Speed Signal Propagation book, By Howard Johnson. Martin Graham Page 67

Critical Interconnect Components PCB vias Connectors and Cable Connectors Sockets Probe tips Packages and Multichip modules PCB traces Single Run Differential Run Page 68

Q/A Page 69

Signal Integrity Issues Impedance mismatch Reflection Crosstalk Ground bounce Inadequate power bus decoupling Propagation delay Dispersion Loss Page 70

Ground Bounce / Power Droop Vchip Vps di/dt = 50pf*1.52 (2V) / Trise^2 = 1.52E8 A/s PS_noise = L*di/dt Where: Trise = 1ns = 2nH * 1.52E8 A/s = 0.3V Vref Vground Page 71

Impedance VS Frequency of Real capacitor model 1.0K 100 Z in OHMS 10 1.0 100m 1.0KHz 10KHz 100KHz 1.0MHz 10MHz 100MHz 1.0GHz 10GHz V(R9:1)/ I(R9) Frequency Page 72

Decoupling and Board Capacitors C1,C2,C3 and C4 are decoupling caps Z Page 73

Impedance Response of Decoupling and Board Capacitance 3.0 R_ESR = 0.5 OHMS 1.0 Parallel Resonance Frequency Z OHMS 300m 100m 30m R_ESR = 0.05 OHMS C4 Series Resonance Frequency C3 C2 C1 10m 1.0KHz 10KHz 100KHz 1.0MHz 10MHz 100MHz 1.0GHz V( R9:1)/ I(R9) Frequency Page 74

Impedance Response of Decoupling and Board Capacitance 3.0 C_board = 100nF 1.0 C_board = 10nF Z OHMS 300m 100m 30m 10m 1.0KHz 10KHz 100KHz 1.0MHz 10MHz 100MHz 1.0GHz 10GHz V(R9:1)/ I(R9) Frequency Page 75

Differential pairs on PCB Not very good in common mode rejection with local cross talk. Coupling about 20% to 50% Match to an external balanced differential transmission medium. Defeat Ground Bounce Improve routability Reduce EMI Can be pushed very closed together to save board space. Compensate for differential impedance Once signals are paired, they can not be separated without messing up the impedance. Page 76

TDR Differential Measurement VtStep SRC1 Vlow=0 V Vhigh=1 V Delay=0 nsec Rise=40 psec C C2 C=1.0 pf t R R1 R=50 Ohm TLIN TL1 Z=50.0 Ohm E=90 F=1 GHz Vin_plus DUT Vo_plus R R3 R=50 Ohm t Vin_minus R R4 R=50 Ohm VtStep SRC2 Vlow=0 V Vhigh=1 V Delay=0 nsec Rise=40 psec R R2 R=50 Ohm TLIN TL2 Z=50.0 Ohm E=90 F=1 GHz SCLIN CLin1 Subst="SSub1" W=5.0 mil S=8.000 mil L=6 in Vo_minus C C1 C=1.0 pf Page 77

TDR Differential Waveforms Vo_plus, mv Vo_minus, mv Vin_plus, mv Vin_minus, mv 600 500 400 300 200 100 0-100 -200-300 -400-500 -600 0.0 0.2 0.4 Vin 0.6 0.8 1.0 1.2 1.4 1.6 Vo 1.8 2.0 2.2 time, nsec 2.4 2.6 2.8 3.0 3.2 3.4 3.6 4.0 3.8 Page 78

4 Port single ended to differential S parameters S13 S31 S14 S41 Port1 Port3 S11, S12, S21,S22 S33, S34, S43,S44 Port2 Port4 S24 S42 S23 S32 SDD11, SDD12,SDD21,SDD22 Port1 Port2 SCC11, SCC12,SCC21,SCC22 Page 79

4 Ports single ended S parameters Page 80

2 Balanced Mode S parameters Page 81

Differential S parameters 10mils 2in Page 82

Poor Common Mode Rejection for PCB differential Pair Differential pair Induced Trace H GROUND K Crosstalk = ---------------- ; Where K is always less than 1 1+ ( D/H)^2 D Page 83

Proximity Effect Ground Plane Skin Depth and Proximity Effect Ground Plane Page 84

Proximity Effect (Cont) Ground Plane Ground Plane Page 85

Odd and Even mode characteristic impedance C increases C decreases L decreases L increases Page 86

Impedance of Differential Pair Z odd = L self - L m C self + C m Z even = L self + L m C self - C m t (L self - L m )( C self + C m ) odd = t (L self + L m )(C self C m ) even = Z odd =<Zo Z diff = 2 * Z odd Z even >=Zo Z common = 1/2 Z even Page 87

Cross Talk Between PCB Traces Causing Shift in Propagation Delay vo_ustrip_neg_coupling, mv vo_ustrip_pos_coupling, mv vo_ustrip, mv 550 500 450 400 350 300 250 200 150 100 50 0 Victim - + + + Timing shift due to aggressor with complement bit -50 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 time, nsec No Cross Talk Original timing Timing shift due to aggressor with same bit pattern as victim Page 88

Modeling Circuit VtPulse SRC3 Vlow=0 V Vhigh=.4 V Delay=0 nsec Edge=linear Rise=45 psec Fall=45 psec Width=70 nsec Period=100 nsec t Var Eqn R R6 R=50 Ohm VAR VAR3 CS=.4pf CR=0.1pf CA=.2pf LS=.01nh LR=.1nh LA=1.4 nh vin_smaa TLIN TL8 Z=50 Ohm E=360 F=39.5 MHz C C12 C=.43pf L L5 L=LA R= C C11 C=CA SLIN TL7 Subst="SSub1" W=10 mil L=8 in C C10 C=CA L L6 L=LA R= C C9 C=.43pf Vout_smaA R R7 R=50 Ohm Page 89

Modeling: TDR simulation VS. Measurement 0.225 0.220 0.215 0.210 0.205 0.200 0.195 0.190 0.185 0.180 0.175 50.0 50.5 51.0 51.5 52.0 52.5 53.0 53.5 54.0 54.5 55.0 55.5 56.0 56.5 57.0 57.5 58.0 Page 90

Simulation and measurements of S11 and S12 of w=10mils and L=2in; standard sma 0-2 -4-6 -8-10 m2 ind Delta=-6.500E8 dep Delta=-3.380 delta mode ON m1 freq=4.860ghz db(sim_single_10mils_2in_s_par..s(1,2))=-1. m2 m1 db(sin_10mils_2in_smas_2..s(1,2)) db(sin_10mils_2in_smas_2..s(1,1)) db(sim_single_10mils_2in_s_par..s(1,2)) db(sim_single_10mils_2in_s_par..s(1,1)) -12-14 -16-18 -20-22 -24-26 -28-30 -32-34 -36-38 Wave L = v / F = C / [Sqr(Er) * F] -40 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 freq, GHz Page 91

Critical Interconnect Components Which one is important? PCB via Connectors and Cable Connectors Socket Probe tips Packages and Multichip modules PCB trace Single Run Differential Run Page 92

S12 of 10mils width and 2in length of 3 types of sma connectors 0-2 -4 db(sin_10mils_2in_smas_2..s(1,2)) db(sin_10mils_2in_smar_2..s(1,2)) db(sin_10mils_2in_smaa..s(1,2)) -6-8 -10-12 -14-16 -18-20 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 freq, GHz Page 93

S12 of 4mils width and 32in length of 3 types of sma connectors 0-2 -4-6 -8 db(sin_4mils_32in_smar..s(1,2)) db(sin_4mils_32in_smas_2..s(1,2)) db(sin_4mils_32in_smaa_2..s(1,2)) -10-12 -14-16 -18-20 -22-24 -26-28 -30-32 -34 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 freq, GHz Page 94

THE END Page 95