1. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 1 Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Impact of MB-OFDM and DS-UWB Interference — Part 2] 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 802.15-04/0412r0, 802.15-04/547r0 and 04/451/r2] 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.

2. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 2 Impact of MB-OFDM and DS-UWB Inteference — Part 2 Torbjorn Larsson Paradiddle Communications, Inc.

3. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 3 Motivation and Objective Motivated by the following three 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 and S. Emami, Freescale Semiconductor 3)04/451/r2, “Multiband OFDM Interference on In-band QPSK Receivers Revisited”, by C. Corral, S. Emami and G. Rasor,, Freescale Semiconductor The above contributions focused on the impact of MB-OFDM interference on a DTV victim receiver In contrast, the objective here is to quantify the difference in the impact of interference from MB- OFDM and DS-UWB

4. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 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 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

5. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 5 DVB-S

6. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 6 Typical C-Band Downlink Channelization Horizontal PolarizationVertical Polarization ChannelCenter Frequency (GHZ)ChannelCenter Frequency (GHz) 1A3.7201B3.740 2A3.7602B3.780 3A3.8003B3.820 4A3.8404B3.860 5A3.8805B3.900 6A3.9206B3.940 7A3.9607B3.980 8A4.0008B4.020 9A4.0409B4.060 10A4.08010B4.100 11A4.12011B4.140 12A4.16012B4.180 (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

7. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 7 DTV Simulation Model Excludes Reed-Solomon coding and interleaving –Impossible to simulate error rates with RS coding Symbol rate: 28 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 Code rates 2/3 and 7/8

8. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 8 MB-OFDM Transmitter Model Based on the Jan. 2005 release of the MB-OFDM PHY spec 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)

9. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 9 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

10. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 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 -10.3 dBm DS-UWB transmit power is -10.8 dBm (data rate dependent) Resolution: 10 kHz PSD averaged over 10 packets (roughly 0.9 ms)

11. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 11 Interference Spectra – Close Up Both spectra exhibit substantial variations Solution: run simulation for multiple DTV center frequencies DTV center frequencies

12. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 12 Amplitude Histogram: Wideband (without Multipath) PAR = 11.9 dB PAR = 12.0 dB Data rate: 110 kbps

13. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 13 Amplitude Histogram: Wideband (with Multipath) PAR = 14.1 dB PAR = 14.2 dB Data rate: 110 kbps 100 multipath channel realizations

14. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 14 Amplitude Histogram: Output of DTV Matched Filter (with Multipath) Center frequency = 4 GHz 100 multipath channel realizations

15. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 15 Output of Matched Filter (Close-Up)

16. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 16 Changes Since November 2004 All simulations carried out with center frequencies according to channelization plan on slide 6  3.72 GHz to 4.18 GHz in steps of 20 MHz Added multipath (CM3, no shadowing) Increased symbol rate from 27 to 28 Msps

17. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 17 Simulation Block Diagram Attenuation 1 is set so that the received DTV power is 3 dB above sensitivity Each simulation is performed with all 24 DTV center frequencies Simulation results are plotted as a function of center frequency and attenuation 2

18. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 18 DTV Sensitivity (NF = 4 dB) Defines sensitivity

19. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 19 DTV Sensitivity (NF = 4 dB) Code RateSensitivity [dBm] 1/2-92.3 2/3-90.5 3/4-89.5 5/6-88.5 7/8-87.9 (Symbol rate: 28 Msps)

20. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 20 BER versus Center Frequency Interference attenuation = 67 dB Code Rate 1/2, without multipath

21. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 21 BER versus Center Frequency Interference attenuation = 71 dB Code Rate 1/2, with multipath

22. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 22 Two Performance Metrics 1.Compute average error rate over the 24 center freqeuencies 2.Select maximum error rate from the 24 center frequencies

23. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 23 Average BER Code Rate 1/2, with multipath

24. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 24 Maximum BER Code Rate 1/2, with multipath

25. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 25 BER versus Center Frequency Interference attenuation = 71 dB Code Rate 7/8, with multipath

26. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 26 Average BER Code Rate 7/8, with multipath

27. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 27 Worst-Case BER Code Rate 7/8, with multipath

28. doc.: IEEE 802.15-05/0039r0 Submission Jan 2005 Torbjorn LarssonSlide 28 Conclusions For DTV code rate 1/2, there is practically no difference between the two interfering waveforms The difference increases with code rate –MB-OFDM interference is more bursty in nature and thus has more impact for higher code rates For DTV code rate 7/8, MB-OFDM is ≈ 1.3 dB worse than DS-UWB @ BER 2·10 -4