MB-OFDM Interference Impact to In-band QPSK transmissions

1
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 Voice1 858
538-3434, FAX [phone],
E-Mailtlarsson_at_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 A
bstract 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
Impact of MB-OFDM and DS-UWB Inteference on
C-Band Receivers

• Torbjorn Larsson
• Paradiddle Communications, Inc.

3
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

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
DVB-S
6
Typical C-Band Downlink Channelization
(Telesat satellite Anik F2. Footprint North
America)
Horizontal Polarization Horizontal Polarization Vertical 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

• 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
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

8
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 (gt worst-case in this
  context!)
• Channel number 9 (Band group 1, TFC 1)
• Data rate 110 Mbps (106.7 Mbps)

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 (gt 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
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)

11
Interference Spectra Close Up
DTV center frequencies

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

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

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!

14
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)

15
BER versus Center Frequency (Code Rate 1/2)
Interference attenuation 67 dB
Center frequencies separated by 10MHz
16
Average BER (Code Rate 1/2)
17
Worst-Case BER (Code Rate 1/2)
18
BER versus Center Frequency (Code Rate 2/3)
Interference attenuation 67 dB
Center frequencies separated by 20MHz
19
Average BER (Code Rate 2/3)
20
Worst-Case BER (Code Rate 2/3)
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

22
Onward

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

tlarsson_at_san.rr.com