Transcript Project: IEEE P802.15 Working Group for Wireless Personal Area Networks...

Nov 2004
doc.: IEEE 802.15-04/0609r0
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.
Submission
Slide 1
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Impact of MB-OFDM and DS-UWB
Inteference on C-Band Receivers
Torbjorn Larsson
Paradiddle Communications, Inc.
Submission
Slide 2
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Motivation and Objective
•
Motivated by two contributions:
1)
2)
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
Submission
Slide 3
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
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
Submission
Slide 4
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
DVB-S
Code rates: 1/2, 2/3,
3/4, 5/6, 7/8
204
bytes
188
bytes
Interleaver
RS
Encoder
Rate-1/2
Conv
Encoder
Puncturing
QPSK
Modulation
RRC
Pulse
Shaping
Radio
TX
DeInterleaver
Soft-Input
Viterbi
DePuncturing
Matched
Filter
Radio
RX
204
bytes
188
bytes
RS
Decoder
Required BER =
2·10-4
Submission
Slide 5
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Typical C-Band Downlink Channelization
(Telesat satellite Anik F2. Footprint: North America)
Horizontal Polarization
• Total of 24 channels
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
• The center frequencies for the
two polarizations are offset by 20
MHz
5A
3.880
5B
3.900
6A
3.920
6B
3.940
• The result is 24 center
frequencies separated by 20 MHz
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
• Each polarization has 12
channels
• Transponder bandwidth is 36
MHz with a 4 MHz guard band
• The center frequencies are
separated by 40 MHz
Submission
Slide 6
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
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
Submission
Slide 7
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
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)
Submission
Slide 8
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
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
Submission
Slide 9
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Interference Spectra
Power over 1 MHz Bandwidth
-30
-40
PSD averaged over 10
packets (roughly 0.9 ms)
-50
dBm
Resolution: 10 kHz
MB-OFDM
DS-UWB
FCC Mask
-60
-70
-80
-90
2
2.5
3
3.5
4
GHz
4.5
5
5.5
6
• 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)
Submission
Slide 10
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Interference Spectra – Close Up
Power over 1 MHz Bandwidth
-40
MB-OFDM
DS-UWB
FCC Mask
DTV center
frequencies
dBm
-41
-42
-43
-44
-45
3
3.2
3.4
3.6
3.8
4
GHz
4.2
4.4
4.6
4.8
5
• Both spectra exhibit substantial variations
• Solution: run simulation for multiple DTV center frequencies
Submission
Slide 11
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
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
Submission
Slide 12
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Simulation Block Diagram
204 byte
packets
Random
Data
Center
Frequency
Center
Frequency
fs = 21.6 GHz
PDTV
DTV
Transmitter
MB-OFDM
Transmitter
3 dB above
sensitivity
Atten 1
DTV
Receiver
+
•
PMB-OFDM
Atten 2
ReSample
DS-UWB
Transmitter
•
•
PDS-UWB
•
Submission
Slide 13
BER
Counter
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 Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
BER Performance without Interference
BER After Viterbi Decoder (NF = 4 dB)
0
10
Code Rate 1/2
Code Rate 2/3
Noise Figure = 4 dB
-1
10
-2
BER
10
Defines sensitivity
-3
10
-4
10
-5
10
-6
10
-95
-94
-93
-92
-91
PDTV [dBm]
-90
-89
•
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)
Submission
Slide 14
-88
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
BER versus Center Frequency (Code Rate 1/2)
BER After Viterbi Decoder versus Center Frequency
-1
10
MB-OFDM
DS-UWB
Interference
attenuation = 67 dB
-2
10
BER
Center frequencies
separated by 10MHz
-3
10
-4
10
3.8
Submission
3.85
3.9
3.95
4
4.05
4.1
4.15
Center Frequency [GHz]
Slide 15
4.2
4.25
4.3
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Average BER (Code Rate 1/2)
-1
Average BER After Viterbi Decoder versus Interference Attenuation
10
MB-OFDM
DS-UWB
-2
BER
10
-3
10
-4
10
65
Submission
65.5
66
66.5
67
67.5
68
68.5
Interference Attenuation [dB]
Slide 16
69
69.5
70
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Worst-Case BER (Code Rate 1/2)
-1
Maximum BER After Viterbi Decoder versus Interference Attenuation
10
MB-OFDM
DS-UWB
-2
BER
10
-3
10
-4
10
65
Submission
65.5
66
68.5
68
67.5
67
66.5
Interference Attenuation [dB]
Slide 17
69
69.5
70
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
BER versus Center Frequency (Code Rate 2/3)
BER After Viterbi Decoder versus Center Frequency
-1
10
MB-OFDM
DS-UWB
Interference
attenuation = 67 dB
Center frequencies
separated by 20MHz
-2
BER
10
-3
10
-4
10
Submission
3.7
3.75
3.8
3.85
3.9
3.95
4
4.05
Center Frequency [GHz]
Slide 18
4.1
4.15
4.2
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Average BER (Code Rate 2/3)
-1
Average BER After Viterbi Decoder versus Interference Attenuation
10
MB-OFDM
DS-UWB
-2
BER
10
-3
10
-4
10
65
Submission
65.5
66
66.5
67
67.5
68
68.5
Interference Attenuation [dB]
Slide 19
69
69.5
70
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Worst-Case BER (Code Rate 2/3)
-1
Maximum BER After Viterbi Decoder versus Interference Attenuation
10
MB-OFDM
DS-UWB
-2
BER
10
-3
10
-4
10
65
Submission
65.5
66
66.5
67
67.5
68
68.5
Interference Attenuation [dB]
Slide 20
69
69.5
70
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
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
Submission
Slide 21
Torbjorn Larssson
Nov 2004
doc.: IEEE 802.15-04/0609r0
Onward…
•
•
•
•
Run simulations for code rates 3/4, 5/6, 7/8
Run simulations for TFC 3 or4
Include multipath
Suggestions?
tlarsson@san.rr.com
Submission
Slide 22
Torbjorn Larssson