project: ieee p802.15 working group for wireless personal area networks (wpans)

Nov 2004

Torbjorn LarssonSlide 1

doc.: IEEE 802.15-04/0609r2

Submission

Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs)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:[[email protected]]

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.

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Impact of MB-OFDM and DS-UWB Inteference on C-Band Receivers

Torbjorn Larsson

Paradiddle Communications, Inc.

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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 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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

DVB-S

RSEncoder

Inter-leaver

204 bytes

188 bytes

Rate-1/2Conv

EncoderPuncturing

QPSKModulation

RRCPulse

Shaping

RadioTX

Code rates: 1/2, 2/3, 3/4, 5/6, 7/8

RSDecoder

De-Interleaver

De-Puncturing

Soft-InputViterbi

Matched Filter

RadioRX

188 bytes

204 bytes

Required BER = 2·10-4

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Typical C-Band Downlink Channelization

Horizontal Polarization Vertical Polarization

Channel Center Frequency (GHZ) Channel 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

(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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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)

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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)

2 2.5 3 3.5 4 4.5 5 5.5 6-90

-80

-70

-60

-50

-40

-30

GHz

dB

m

Power over 1 MHz Bandwidth

MB-OFDMDS-UWBFCC Mask

Resolution: 10 kHz

PSD averaged over 10 packets (roughly 0.9 ms)

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Interference Spectra – Close Up

• Both spectra exhibit substantial variations• Solution: run simulation for multiple DTV center frequencies

3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8 5

-45

-44

-43

-42

-41

-40

GHz

dB

m

Power over 1 MHz Bandwidth

MB-OFDMDS-UWBFCC Mask

DTV center frequencies

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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!

DTVTransmitter

3 dB above sensitivity

Atten 2

PDTV

MB-OFDMTransmitter

DS-UWBTransmitter

Atten 1 + DTVReceiver

BERCounter

RandomData

CenterFrequency

CenterFrequency

PMB-OFDM

PDS-UWB

Re-Sample

fs = 21.6 GHz

204 byte packets

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

-95 -94 -93 -92 -91 -90 -89 -8810

-6

10-5

10-4

10-3

10-2

10-1

100

BER After Viterbi Decoder (NF = 4 dB)

PDTV

[dBm]

BE

R

Code Rate 1/2Code Rate 2/3

BER Performance without Interference

• 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)

Noise Figure = 4 dB

Defines sensitivity

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

BER versus Center Frequency (Code Rate 1/2)

Interference attenuation = 67 dB

3.8 3.85 3.9 3.95 4 4.05 4.1 4.15 4.2 4.25 4.3

10-4

10-3

10-2

10-1

BER After Viterbi Decoder versus Center Frequency

Center Frequency [GHz]

BE

RMB-OFDMDS-UWB

Center frequencies separated by 10MHz

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Average BER (Code Rate 1/2)

65 65.5 66 66.5 67 67.5 68 68.5 69 69.5 70

10-4

10-3

10-2

10-1

Average BER After Viterbi Decoder versus Interference Attenuation

Interference Attenuation [dB]

BE

R

MB-OFDMDS-UWB

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Worst-Case BER (Code Rate 1/2)

65 65.5 66 66.5 67 67.5 68 68.5 69 69.5 70

10-4

10-3

10-2

10-1

Maximum BER After Viterbi Decoder versus Interference Attenuation


BE

R

MB-OFDMDS-UWB

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

BER versus Center Frequency (Code Rate 2/3)

3.7 3.75 3.8 3.85 3.9 3.95 4 4.05 4.1 4.15 4.210

-4

10-3

10-2

10-1

BER After Viterbi Decoder versus Center Frequency

Center Frequency [GHz]

BE

RMB-OFDMDS-UWBInterference

attenuation = 67 dB

Center frequencies separated by 20MHz

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Average BER (Code Rate 2/3)

65 65.5 66 66.5 67 67.5 68 68.5 69 69.5 70

10-4

10-3

10-2

10-1

Average BER After Viterbi Decoder versus Interference Attenuation


BE

RMB-OFDMDS-UWB

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Worst-Case BER (Code Rate 2/3)

65 65.5 66 66.5 67 67.5 68 68.5 69 69.5 70

10-4

10-3

10-2

10-1

Maximum BER After Viterbi Decoder versus Interference Attenuation


BE

R

MB-OFDMDS-UWB

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

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

Nov 2004


doc.: IEEE 802.15-04/0609r2

Submission

Onward…

• Run simulations for code rates 3/4, 5/6, 7/8• Run simulations for TFC 3 or 4• Include multipath• Suggestions?

[email protected]

project: ieee p802.15 working group for wireless personal area networks (wpans)

Documents