DFE Error Performance Under 1000BASE-T1 Noise Environments

Similar documents
Feasibility of 1 UTP for RTPGE : Impacts of Gauge, Temperature, and Modulation

Electro-Magnetic Emissions and Susceptibility for RTPGE with PAM-M & DFE

Precoding proposal for PAM4

10GBASE-T Transmitter SNDR Definition (System ID Approach) IEEE P802.3an Task Force Santa Clara, Feb 2005 Albert Vareljian, Hiroshi Takatori KeyEye

56+ Gb/s Serial Transmission using Duobinary Signaling

Transmission Proposal for 10GBASE-T G. Zimmerman, SolarFlare

Noise considerations for RTPGE objectives. Gavin Parnaby IEEE RTPGE Study Group Geneva September 2012

QAM-Based Transceiver Solutions for Full-Duplex Gigabit Ethernet Over 4 Pairs of UTP-5 Cable. Motivation for Using QAM

10GBASE-T Transmitter Key Specifications

Preliminary COM results for two reference receiver models

10GBASE-T T Tutorial. SolarFlare Communications IEEE Kauai, Hawaii. November 11, 2002

802.3bp 1000BASE-T1 EMC & Noise Ad Hoc Report

SERDES for 100Gbps. May 24, 2017 Bart Zeydel, Francesco Caggioni, Tom Palkert

OIF CEI 6G LR OVERVIEW

Lecture 17 Components Principles of Error Control Borivoje Nikolic March 16, 2004.

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

Auto-Negotiation Update

QAM-Based 1000BASE-T Transceiver

Towards 100G over Copper

Communication Efficiency of Error Correction Mechanism Based on Retransmissions

ECEN720: High-Speed Links Circuits and Systems Spring 2017

Ultra320 SCSI with Receiver Equalization, 25 meters into a Backplane with 6 loads. Russ Brown Quantum Corporation

For IEEE 802.3ck March, Intel

Optimal Transmit Spectra for Communication on Digital Subscriber Lines

Modulation. Digital Data Transmission. COMP476 Networked Computer Systems. Sine Waves vs. Square Waves. Fourier Series. Modulation

Project: IEEE P Working Group for Wireless Personal Area Networks N

Channel operating margin for PAM4 CDAUI-8 chip-to-chip interfaces

Channel coding for polarization-mode dispersion limited optical fiber transmission

Enabling Improved DSP Based Receivers for 100G Backplane

Line Coding Considerations for 10BASE-T1S

Optical transmitter characteristics for GEPOF technical feasibility

Part A: Spread Spectrum Systems

10 Mb/s Single Twisted Pair Ethernet Conducted Immunity Steffen Graber Pepperl+Fuchs

Arash Farhood Cortina systems. IEEE Next Gen 100G Optical Ethernet Task Force

OFDMA PHY for EPoC: a Baseline Proposal. Andrea Garavaglia and Christian Pietsch Qualcomm PAGE 1

On the Predictability of Underwater Acoustic Communications Performance: the KAM11 Data Set as a Case Study

UWB Impact on IEEE802.11b Wireless Local Area Network

The Measurement and Characterisation of Ultra Wide-Band (UWB) Intentionally Radiated Signals

06-011r0 Towards a SAS-2 Physical Layer Specification. Kevin Witt 11/30/2005

RF Interference Cancellation - a Key Technology to support an Integrated Communications Environment

1000BASE-RH PHY system simulations

Performance of Wideband Mobile Channel with Perfect Synchronism BPSK vs QPSK DS-CDMA

EQUALIZERS. HOW DO? BY: ANKIT JAIN

Resource Blocks for EPoC Considerations. Avi Kliger, BZ Shen, Leo Montreuil Broadcom

Ultra640 SCSI with Receiver Equalization, 25 meters into a Backplane with 6 loads. Russ Brown Quantum Corporation

Rate and Power Adaptation in OFDM with Quantized Feedback

To learn fundamentals of high speed I/O link equalization techniques.

Performance Evaluation Of Digital Modulation Techniques In Awgn Communication Channel

Comparison of Time Domain and Statistical IBIS-AMI Analyses Mike LaBonte SiSoft

Comparison of Time Domain and Statistical IBIS-AMI Analyses

Adaptive Modulation and Coding

Technology comparison matrix for duplex SMF PMDs. Yoshiaki Sone NTT IEEE802.3bs 400 Gb/s Ethernet Task Force, Ottawa, September 2014.

Part A: Spread Spectrum Systems

A Study on the SPIHT Image Coding Technique for Underwater Acoustic Communications

TDECQ update noise treatment and equalizer optimization (revision of king_3bs_02_0217_smf)

5Gbps Serial Link Transmitter with Pre-emphasis

ECEN720: High-Speed Links Circuits and Systems Spring 2017

NRZ Bandwidth (-3db HF Cutoff vs SNR) How Much Bandwidth is Enough?

Data and Computer Communications Chapter 3 Data Transmission

Does PAM-4 or NRZ Require an Intra-Baud Clipping Penalty?

INTRODUCTION TO COMMUNICATION SYSTEMS LABORATORY IV. Binary Pulse Amplitude Modulation and Pulse Code Modulation

Project: IEEE P Working Group for Wireless Personal Area Networks N

Data Communication. Chapter 3 Data Transmission

10 Mb/s Single Twisted Pair Ethernet Evaluation Board Noise Measurements Marcel Medina Steffen Graber Pepperl+Fuchs

PHYSICS 330 LAB Operational Amplifier Frequency Response

T10/05-428r0. From: Yuriy M. Greshishchev, PMC-Sierra Inc. Date: 06 November 2005

Department of Electronics and Communication Engineering 1

Bit Error Rate Performance Evaluation of Various Modulation Techniques with Forward Error Correction Coding of WiMAX

Electronic Dispersion Compensation of 40-Gb/s Multimode Fiber Links Using IIR Equalization

Performance Analysis of n Wireless LAN Physical Layer

10 Mb/s Single Twisted Pair Ethernet Noise Environment for PHY Proposal Evaluation Steffen Graber Pepperl+Fuchs

Carrier Frequency Offset Estimation Algorithm in the Presence of I/Q Imbalance in OFDM Systems

Signal Encoding Techniques

High Speed I/O 2-PAM Receiver Design. EE215E Project. Signaling and Synchronization. Submitted By

EC 554 Data Communications

Implementation of MIMO-OFDM System Based on MATLAB

To learn statistical bit-error-rate (BER) simulation, BER link noise budgeting and using ADS to model high speed I/O link circuits

TDECQ results is function of the 4th-order B-T filter roll-off stop frequency. We are proposing to mandate the minimum roll-off stop frequency.

6.976 High Speed Communication Circuits and Systems Lecture 20 Performance Measures of Wireless Communication

10 Mb/s Single Twisted Pair Ethernet Noise Environment for PHY Proposal Evaluation Steffen Graber Pepperl+Fuchs

A possible receiver architecture and preliminary COM Analysis with GEL Channels

Implementation of Reed Solomon Encoding Algorithm

CAUI-4 Chip Chip Spec Discussion

IEEE p802.3bn EPoC. Channel Model Ad Hoc committee Baseline Channel Model

University of Bristol - Explore Bristol Research. Peer reviewed version. Link to publication record in Explore Bristol Research PDF-document

PHY PMA electrical specs baseline proposal for 803.an

ULTRA WIDE BAND(UWB) Embedded Systems Programming

802.3CG EMISSION LIMITS AND PSD MASK. Rosemont, March 5-6th 2018 Conrad Zerna

Chapter 2 Overview - 1 -

IEEE 802.3ap. Transmitter Tap Range Selection. Brian Brunn, Xilinx Robert Brink, Agere Systems. 21 June 2005

Backchannel Modeling and Simulation Using Recent Enhancements to the IBIS Standard

Chapter 3 Data Transmission COSC 3213 Summer 2003

GIGABIT ETHERNET CONSORTIUM

i) Physical, ii) Data-link, iii) Network, iv) Transport, v) Session, vi) Presentation, and vii) Application

A New Approach to Measure Tx Signal Strength and Penalty

Digitally Enhanced Inter-modulation Distortion Compensation in Wideband Spectrum Sensing. Han Yan and Prof. Danijela Cabric Nov.

OFDMA Networks. By Mohamad Awad

10 Mb/s Single Twisted Pair Ethernet 10BASE-T1L PSD Mask Steffen Graber Pepperl+Fuchs

Chapter 9. Digital Communication Through Band-Limited Channels. Muris Sarajlic

Performance Optimization of Hybrid Combination of LDPC and RS Codes Using Image Transmission System Over Fading Channels

Transcription:

DFE Error Performance Under 1000BASE-T1 Noise Environments Xiaofeng Wang, Qualcomm Inc wangxiao@qti.qualcomm.com IEEE 802.3bp RTPGE --- May 2014 1

Supporters Sujan Pandey, NXP Benson Huang, Realtek Shaoan Dai, Marvell IEEE 802.3bp RTPGE --- May 2014 2

Agenda Motivations DFE Over 2m Channel DFE Over 15m Channel Discussions DFE Simulation Model Error Performance Under RFI Error Performance Under Transient Noise Conclusions IEEE 802.3bp RTPGE --- May 2014 3

Motivations Previous SNR results provide estimates of BER only when DFE error propagation is negligible. Actual error performance are required under 1000BASE-T1 noise environments and with realistic DFE taps for evaluation of symbol mapping schemes. finalization of Tx PSD mask. design of FEC. IEEE 802.3bp RTPGE --- May 2014 4

DFE Results Over 2m Channel First-order transmit filter to fit the agreed Tx-PSD mask with reduced DC level at 80.5dBm/Hz. Agreed Xtalk models and thermal noise @ -140dBm/Hz and receive HPF with fc=10 MHz. 10% overhead in addition to mapping, two mapping schemes: Map 1, 2D-PAM3 with out 00 ; Map 2, 10 bits to 7 symbols mapping [xiaofeng_3bp_02_0314.pdf.] Table - Results for 2m channel @ 125 0 C, Infinite-length DFE. Eye Height (mv) SNR (db) DFE Tap 1 DFE Tap 2 Map 1, SQNR=32.8 db 270 30.16 0.67 0.13 Map 1, SQNR=29.8 db 270 27.49 0.66 0.10 Map 2, SQNR=32.8 db 275 30.35 0.61 0.07 Map 2, SQNR=29.8 db 275 27.8 0.59 0.04 IEEE 802.3bp RTPGE --- May 2014 5

DFE Results Over 15m Channel Table - Results for 15m channel @ 20 0 C, Infinite-length DFE. Eye Height (mv) SNR (db) DFE Tap 1 DFE Tap 2 Map 1, SQNR=32.8 db 174 29.06 0.88 0.32 Map 1, SQNR=29.8 db 174 26.4 0.87 0.30 Map 2, SQNR=32.8 db 177 29.4 0.82 0.25 Map 2, SQNR=29.8 db. 177 26.8 0.81 0.22 IEEE 802.3bp RTPGE --- May 2014 6

Discussions The first DFE feedback tap is greater than 0.5 in all cases. Finite-length implementation may lead to smaller DFE tap at the cost of reduced SNR and non-flat FFE gain. In the absence of narrowband noises, large SNR renders DFE error propagation a nonissue. Two worst-case scenarios need to be investigated 2m channel with 200mVpp RFI. ISO-7637-3 transient noise. To measure error propagation, number of error code symbols per codeword will be used. It is assumed that a codeword has 255 code symbols and a code symbol has 6 PAM-3 symbols (or 9 bits) for Map1 and 7 PAM-3 symbols (0r 10 bits) for Map 2. IEEE 802.3bp RTPGE --- May 2014 7

DFE Simulation Model B(z) with two taps. When a decision error happens, the signal before decision is: x(n)+r(n)+z(n)-b(1)e(n-1)-b(2)e(n-2) Observations: Error propagation can happen even when b(1) <0.5 and b(2)=0 because the effect of r(n). When r(n) and r(n-1) have the same sign, error propagation unlikely occurs. This indicates that error propagation will happen more likely when RFI frequency is over fb/4 and less likely in transient noise and RFI with frequency less than fb/4. The worst case happens when RFI with frequency fb/2 and e(n-1) and r(n) always push the error toward the same direction. IEEE 802.3bp RTPGE --- May 2014 8

Error Performance Under RFI B(z)=0.67z -1 +0.13z -2 RFI peak to peak amplitude =0.70 vs. eye height =1. RFI frequency at 0.47fb with fb as symbol rate. If RS code, Map 1 needs t=11 and Map 2 needs t=4. Probability 10-1 10-2 10-3 10-4 10-5 Map 1: 6 PAM-3 symbols or 9 bits per code symbol Map 2: 7 PAM-3 symbols or 10 bits per code symbol Mapping Scheme Map 1 Map 2 SER 3.7e-4 5.6e-5 10-6 10-7 0 2 4 6 8 10 12 Number of Error Code Symbols Out of 255 Code Symbols IEEE 802.3bp RTPGE --- May 2014 9

Error Performance Under Transient Noise I 2m channel B(z)=0.67z -1 +0.13z -2 Transient noise modeled as a cycle of sine wave with duration 80ns and peak to peak amplitude =0.70 vs. eye height =1. SNR = 26 db. No error propagation observed and hence RS code with t=1 is sufficient. Mapping Schemes SER Map 1 Map 2 3.4e-8 1.36e-8 IEEE 802.3bp RTPGE --- May 2014 10

Performance Under Transient Noise II 15m channel B(z)=0.88z -1 +0.32z -2 Transient noise a cycle of sine wave with duration 80ns and peak to peak amplitude =1.14 vs. eye height =1. SNR = 26 db. RS code with t=8 for Map 1 and RS code with t=7 for Map 2. 10-1 Map 1: 6 PAM-3 symbols or 9 bits per code symbol Map 2: 7 PAM-3 symbols or 10 bits per code symbol Mapping Scheme Map 1 Map 2 SER 1.02e-4 9.2e-5 Probability 10-2 10-3 10-4 10-5 10-6 1 2 3 4 5 6 7 8 Number of Error Code Symbols Out of 255 Code Symbols IEEE 802.3bp RTPGE --- May 2014 11

Conclusions HPF significantly reduce DFE tap gains. Even with HPF in 2m channel, the first DFE feedback tap tends to be large. Severe error propagation occurs in the presence of high-frequency RFI. RFI with frequency at half of symbol rate causes error propagation that is undecodable. Because ISO-7637 transient noise primarily consists of components of frequency far lower than half of the symbol rate, error propagation is not an issue. Appropriate line coding can significantly reduces the required coding overhead. With the proposed 10 bits to 7 symbol mapping scheme, RS code with t=4 or 5 is sufficient. IEEE 802.3bp RTPGE --- May 2014 12