Nov 2004 Project: IEEE P802.15 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:[+1 858 538-3434], FAX: [+1 858 538-2284], E-Mail:[tlarsson@san.rr.com] Re: [Analysis of the impact of MB-OFDM and DS-UWB interference on a DTV receiver made in earlier contributions, in particular 802.15-04/547r0 and 802.15-04/0412r0] 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 P802.15. 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 P802.15. Torbjorn Larsson

Impact of MB-OFDM and DS-UWB Inteference on C-Band Receivers Nov 2004 Impact of MB-OFDM and DS-UWB Inteference on C-Band Receivers Torbjorn Larsson Paradiddle Communications, Inc. Torbjorn Larsson

Motivation and Objective Nov 2004 Motivation and Objective Motivated by two contributions: 04/0412r0, In-band Interference Properties of MB-OFDM, by C. Razell, Philips 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 802.15.3a task group The author is an independent consultant, not affiliated with any UWB company. This work was not carried out under any consulting contract Torbjorn Larsson

C-Band DTV Systems The C-band downlink spans 3.7 – 4.2 GHz Nov 2004 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 300 421) 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 Torbjorn Larsson

Nov 2004 DVB-S Torbjorn Larsson

Typical C-Band Downlink Channelization Nov 2004 Typical C-Band Downlink Channelization (Telesat satellite Anik F2. Footprint: North America) Horizontal Polarization Vertical Polarization Channel Center Frequency (GHZ) Center Frequency (GHz) 1A 3.720 1B 3.740 2A 3.760 2B 3.780 3A 3.800 3B 3.820 4A 3.840 4B 3.860 5A 3.880 5B 3.900 6A 3.920 6B 3.940 7A 3.960 7B 3.980 8A 4.000 8B 4.020 9A 4.040 9B 4.060 10A 4.080 10B 4.100 11A 4.120 11B 4.140 12A 4.160 12B 4.180 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 Torbjorn Larsson

DTV Simulation Model Excludes Reed-Solomon coding and interleaving Nov 2004 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 Receiver noise figure: 4 dB Intend to run simulations for all code rates – Results presented only include rate 1/2 and 2/3 Torbjorn Larsson

MB-OFDM Transmitter Model Nov 2004 MB-OFDM Transmitter Model Based on the Sep. 2004 release of the MB-OFDM PHY Specifications (P802.15-04/0493r1) 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) Torbjorn Larsson

DS-UWB Transmitter Model Nov 2004 DS-UWB Transmitter Model Based on the July 2004 release of the DS-UWB PHY specifications (P802.15-04/0137r3) 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 (109.417 Mbps) BPSK modulation Spreading code for preamble and header (PAC): -1 0 +1 -1 -1 -1 +1 +1 0 +1 +1 +1 +1 -1 +1 -1 +1 +1 +1 +1 -1 -1 +1 Spreading code for frame body: +1 0 0 0 0 0 Torbjorn Larsson

Nov 2004 Interference Spectra Resolution: 10 kHz PSD averaged over 10 packets (roughly 0.9 ms) 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 -10.3 dBm DS-UWB transmit power is -10.8 dBm (data rate dependent) Torbjorn Larsson

Interference Spectra – Close Up Nov 2004 Interference Spectra – Close Up DTV center frequencies Both spectra exhibit substantial variations Solution: run simulation for multiple DTV center frequencies Torbjorn Larsson

Simulated DTV Center Frequencies Nov 2004 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 Torbjorn Larsson

Simulation Block Diagram Nov 2004 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! Torbjorn Larsson

BER Performance without Interference Nov 2004 BER Performance without Interference Noise Figure = 4 dB Defines sensitivity Sensitivity for rate 1/2 is -92.5 dBm (Eb/No = 3.2 dB) Sensitivity for rate 2/3 is -90.7 dBm (Eb/No = 3.7 dB) Torbjorn Larsson

BER versus Center Frequency (Code Rate 1/2) Nov 2004 BER versus Center Frequency (Code Rate 1/2) Interference attenuation = 67 dB Center frequencies separated by 10MHz Torbjorn Larsson

Average BER (Code Rate 1/2) Nov 2004 Average BER (Code Rate 1/2) Torbjorn Larsson

Worst-Case BER (Code Rate 1/2) Nov 2004 Worst-Case BER (Code Rate 1/2) Torbjorn Larsson

BER versus Center Frequency (Code Rate 2/3) Nov 2004 BER versus Center Frequency (Code Rate 2/3) Interference attenuation = 67 dB Center frequencies separated by 20MHz Torbjorn Larsson

Average BER (Code Rate 2/3) Nov 2004 Average BER (Code Rate 2/3) Torbjorn Larsson

Worst-Case BER (Code Rate 2/3) Nov 2004 Worst-Case BER (Code Rate 2/3) Torbjorn Larsson

Nov 2004 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 Torbjorn Larsson

Onward… Run simulations for code rates 3/4, 5/6, 7/8 Nov 2004 Onward… Run simulations for code rates 3/4, 5/6, 7/8 Run simulations for TFC 3 or 4 Include multipath Suggestions? tlarsson@san.rr.com Torbjorn Larsson