doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 1 Project: IEEE P Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Impact of MB-OFDM and DS-UWB Interference on C Band Receivers] Date Submitted: [] Source: [Torbjorn Larsson] Company [Paradiddle Communications] Address [13141 Via Canyon Drive, San Diego, CA 92129, USA] Voice:[ ], FAX: [ ], Re: [Analysis of the impact of MB-OFDM and DS-UWB interference on a DTV receiver made in earlier contributions, in particular /547r0 and /0412r0] Abstract:[The impact of MB-OFDM and DS-UWB interference on a C-band DTV receiver is investigated by simulation] Purpose:[To present an unbiased comparison of the impact of MB-OFDM and DS-UWB interference based on a minimal set of universally accepted assumptions] Notice:This document has been prepared to assist the IEEE P It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release:The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 2 Impact of MB-OFDM and DS-UWB Inteference on C-Band Receivers Torbjorn Larsson Paradiddle Communications, Inc.

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 3 Motivation and Objective Motivated by two contributions: 1)04/0412r0, In-band Interference Properties of MB-OFDM, by C. Razell, Philips 2)04/547r0, Responses to “In-Band Interference Properties of MB-OFDM”, by C. Corral, G. Rasor, S. Emami, Freescale Semiconductor The emphasis in the above contributions is on qualitative analysis In contrast, the approach here is “brute force” simulation Our hope is that the assumptions made are universal enough to be accaptable to the entire a task group The author is an independent consultant, not affiliated with any UWB company. This work was not carried out under any consulting contract

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 4 C-Band DTV Systems The C-band downlink spans 3.7 – 4.2 GHz C-band antennas are typically 6 – 12 feet in diameter Based on the DVB-S (Digital Video Broadcasting – Satellite) standard (EN ) DVB-S was designed for MPEG-2 broadcasting in the Ku-band, but is also used in the C-band DVB-S does not specify a unique set of data rates or symbol rates; However… Typical transponder bandwidth is 36 MHz (33 MHz also used) Typical symbol rate 27 – 29 Msps DVB-S2 is the next generation with improved bandwidth efficiency and FEC

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 5 DVB-S

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 6 Typical C-Band Downlink Channelization Horizontal PolarizationVertical Polarization ChannelCenter Frequency (GHZ)ChannelCenter Frequency (GHz) 1A3.7201B A3.7602B A3.8003B A3.8404B A3.8805B A3.9206B A3.9607B A4.0008B A4.0409B A B A B A B4.180 (Telesat satellite Anik F2. Footprint: North America) Total of 24 channels Each polarization has 12 channels Transponder bandwidth is 36 MHz with a 4 MHz guard band The center frequencies are separated by 40 MHz The center frequencies for the two polarizations are offset by 20 MHz The result is 24 center frequencies separated by 20 MHz

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 7 DTV Simulation Model Excludes Reed-Solomon coding and interleaving –Impossible to simulate error rates with RS coding –Will probably favor DS-UWB Symbol rate: 27 Msps No quantization (including input to Viterbi decoder) Ideal pulse shaping/matched filters (0.35 roll-off) No nonlinarity No frequency offset No phase noise Pre-computed phase error and time offset Intend to run simulations for all code rates – Results presented only include rate 1/2 and 2/3

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 8 MB-OFDM Transmitter Model Based on the Sep release of the MB-OFDM PHY Specifications (P /0493r1) Complete Matlab implementation of the specifications System operating in band-hopping mode Includes (5-bit) DAC and realistic filter characteristics Spectral pre-shaping to compensate for non-ideal filter characteristics (=> worst-case in this context!) Channel number 9 (Band group 1, TFC 1) Data rate “110” Mbps (106.7 Mbps)

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 9 DS-UWB Transmitter Model Based on the July 2004 release of the DS-UWB PHY specifications (P /0137r3) Complete Matlab implementation of the specifications No DAC Ideal RRC pulse shaping filter truncated to 12 chip periods (=> worst-case!) Channel number 1 (chip rate: 1313 Mcps) Data rate: “110” Mbps ( Mbps) BPSK modulation Spreading code for preamble and header (PAC): Spreading code for frame body:

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 10 Interference Spectra Transmit power is set so as to push each spectrum as close as possible to the FCC limit (worst-case condition) MB-OFDM transmit power is dBm DS-UWB transmit power is dBm (data rate dependent) Resolution: 10 kHz PSD averaged over 10 packets (roughly 0.9 ms)

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 11 Interference Spectra – Close Up Both spectra exhibit substantial variations Solution: run simulation for multiple DTV center frequencies DTV center frequencies

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 12 Simulated DTV Center Frequencies Rate 1/2 simulations: 3.8 – 4.3 GHz in steps of 10 MHz –Arbitrary choice across 500 MHz bandwidth Rate 2/3 simulations: 3.72 – 4.18 GHz in steps of 20 MHz –According to channelization plan on slide 6

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 13 Simulation Block Diagram Attenuation 1 is set so that the received DTV power is 3 dB above sensitivity Each simulation is performed with multiple DTC center frequencies Simulation results are plotted as a function of center frequency and attenuation 2 No multipath!

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 14 BER Performance without Interference Sensitivity for rate 1/2 is dBm (Eb/No = 3.2 dB) Sensitivity for rate 2/3 is dBm (Eb/No = 3.7 dB) Noise Figure = 4 dB Defines sensitivity

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 15 BER versus Center Frequency (Code Rate 1/2) Interference attenuation = 67 dB Center frequencies separated by 10MHz

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 16 Average BER (Code Rate 1/2)

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 17 Worst-Case BER (Code Rate 1/2)

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 18 BER versus Center Frequency (Code Rate 2/3) Interference attenuation = 67 dB Center frequencies separated by 20MHz

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 19 Average BER (Code Rate 2/3)

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 20 Worst-Case BER (Code Rate 2/3)

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 21 Conclusions For the two simulated cases (rate 1/2 and 2/3), the difference in average BER across the C-band is 1 dB or less The difference in worst-case BER is less than 0.5 dB More general conclusions should be postponed until all code rates have been simulated

doc.: IEEE /0609r0 Submission Nov 2004 Torbjorn LarsssonSlide 22 Onward… Run simulations for code rates 3/4, 5/6, 7/8 Run simulations for TFC 3 or4 Include multipath Suggestions?