OFDM System Performance

1
OFDM System Performance

• Karen Halford, Steve Halford and Mark Webster

2
Outline of Proposal Presentations

• TGg Regulatory Approval Plan Speaker Jim
  Zyren
• Overview of OFDM for High Rate Speaker Steve
  Halford
• Reuse of 802.11b Preambles with OFDM Speaker
  Mark Webster
• Ultra-short Preamble with HRb OFDM Speaker Mark
  Webster
• OFDM System Performance Speaker Steve Halford
• Power Am Effects for HRb OFDM Speaker Mark
  Webster
• Channelization for HRb OFDM Speaker Mark
  Webster
• Phase Noise Sensitivity for HRb OFDM Speaker Jim
  Zyren
• Implementation and Complexity Issues for OFDM
  Speaker Steve Halford
• Why OFDM for the High Rate 802.11b Extension?
  Speaker Jim Zyren

3
Outline of Presentation

• 5.1 AWGN Performance
• 5.2 Rayleigh Fading Performance
• 5.3 Multipath Performance
• 5.3.1 Exponential Channel with Flat Fading
• 5.3.2 Exponential Channel without Flat Fading
  (Normalized)
• 5.3.3 PER sweeps from 1 to 10
• 5.4 Throughput Performance
• 5.5 Performance Against CW Jammer (FCC15.247 Test)

4
5.1 AWGN Performance 100 Byte Packets
5
5.1 AWGN Performance 1000 Byte Packets
6
5.1 AWGN Performance 2346 Byte Packets
7
5.1 AWGN Performance 1 and 10 PER for 1000
Byte Packets
Eb/No required for 1 PER
Eb/No required for 10 PER
8
5.2 Rayleigh Fading Performance Block Diagram
Generate Noise
Calculate Noise Power (N0)
Measure energy per bit
Measure Packet Error Rate
Packet Error Rate
Transmitter Model
Receiver Model
x

Rayleigh Coefficient
Packet Length Data Rate
9
5.2 Rayleigh Fading Performance 1000 Byte
Packets
10
5.2 Rayleigh Fading Performance 1 and 10 PER
for 1000 Byte Packets
Eb/No required for 1 PER
Eb/No required for 10 PER
11
5.3.1 Multipath Performance with Flat
FadingBlock Diagram
Generate Noise
Calculate Noise Power (N0)
Measure energy per bit
Measure Packet Error Rate
Packet Error Rate
Exponential Channel Model
Receiver Model
Transmitter Model
Packet Length Data Rate
Sample Rate Delay Spread
12
5.3.1 Multipath Performance with Flat Fading
Matlab Code
13
5.3.1 Multipath Performance with Flat Fading
Eb/No
14
5.3.1 Multipath Performance with Flat Fading SNR
15
5.3.2 Multipath Performance without Flat
FadingBlock Diagram
Generate Noise
Calculate Noise Power (N0)
Measure energy per bit
Measure Packet Error Rate
Exponential Channel Model
Transmitter Model
Receiver Model
Packet Error Rate
Packet Length Data Rate
Sample Rate Delay Spread
16
5.3.2 Multipath Performance without Flat Fading
Eb/No
17
5.3.2 Multipath Performance without Flat Fading
SNR
18
5.3.3 Multipath Sweeps 1 to 10
Comparison Item 24

• For each modulation mode detemine and state the
  SNR (Es/No) at which in AWGN only, the waveform
  can achieve a PER of 0.01 for packets lengths of
  1000B. Using the multipath model used in 23b
  above, fix the amount of AWGN at the 0.01 PER
  level for AWGN only. Increase the RMS delay
  spread until the PER for 1000B packets reach 0.1.
  State the RMS delay spread at this point.

Answer 0.0 nSeconds for all rates
Why ?
19
5.3.3 Multipath Sweeps 1 to 10
PER Curves are very steep -- about 2 dB
separates the 1 from the 10 point
20
5.3.3 Multipath Sweeps 1 to 10
Rayleigh fading causes frequent swings to low SNR
level
21
5.3.3 Multipath Sweeps 1 to 10
What we ran in place of Comparison Item 24

• For each modulation mode detemine and state the
  SNR (Es/No) at which 25 nSeconds RMS delay, the
  waveform can achieve a PER of 0.01 for packets
  lengths of 1000B. Using the multipath model used
  in 23c above, fix the amount of AWGN at the 0.01
  PER level for 25 nSeconds RMS delay. Increase
  the RMS delay spread until the PER for 1000B
  packets reach 0.1. State the RMS delay spread at
  this point.

22
5.3.3 Multipath Performance PER sweeps from 1
to 10
23
5.4 Throughput Performance

• 5.4.1 Preamble Structures
• 5.4.2 ACK Assumptions
• 5.4.3 Throughput Analysis
• 5.4.3.1 Tables of 100, 1000, 2346 Byte Packets
• 5.4.3.2 Plots for full range of packet sizes
• 5.4.4 Throughput analysis for varying durations
  of overhead

24
5.4.1 Preamble StructuresLong and Short
Preambles
802.11 HRb LONG PREAMBLE
Signal Extension
PSDU SELECTABLE OFDM Symbols _at_ 6.6, 9.6, 13.2,
19.8, 26.4, 39.3, 52.8 or 59.4 Mbps
OFDM SYNC
PREAMBLE/HEADER
192 usecs
10.9 usecs
6 usecs
Data Payload
802.11 HRb SHORT PREAMBLE
Signal Extension
PREAM/HDR 72 BITS _at_ 1 Mbps
PSDU SELECTABLE OFDM Symbols _at_ 6.6, 9.6, 13.2,
19.8, 26.4, 39.3, 52.8 or 59.4 Mbps
OFDM SYNC
6 usecs
96 usecs
10.9 usecs
25
5.4.1 Preamble StructuresUltra-Short Preamble

• Proposed Ultra-Short Preamble

Signal Extension
Data Payload

• Data Rate
• bytes of data

PSDU SELECTABLE _at_ 6.6, 9.9, 13.2, 19.8, 26.4,
39.6, 52.8 or 59.4 Mbps
Long SYNC
SIGNAL SYMBOL
12 Short Syncs Reps
6 usecs
16 usecs
3.6 usecs
26
5.4.2 ACK Assumptions
1) No RTS/CTS OR MPDU lt RTS_Threshold
Many different scenarios, but the constant is
MPDU, SIFS, ACK
source
DIFS
Data
SIFS
ACK
destination
2) RTS/CTS and/or MPDU gt RTS_Threshold
source
RTS
Data
SIFS
DIFS
destination
SIFS
ACK
SIFS
CTS
3) Middle of Fragmented Transmission
source
SIFS
Fragment 1
SIFS
ACK 1
destination
27
5.4.2 ACK Assumptions (continued)
112 Bits _at_ 6.6 Mbps 20 usec
SIFS
PSDU SELECTABLE OFDM Symbols _at_ 6.6, 9.6, 13.2,
19.8, 26.4, 39.3, 52.8 or 59.4 Mbps
Packet Header
Packet Header
ACK
OFDM PAD 6 usecs
28
5.4.3.1 Throughput for 100 Byte Packets
29
5.4.3.1 Throughput for 1000 Byte Packets
30
5.4.3.1 Throughput for 2346 Byte Packets
31
5.4.3.2 Throughput with ACK
32
5.4.3.2 Throughput without ACK
33
5.4.4 Comparison of Throughput for Variable
Overhead for 100 Byte MPDU
34
5.4.4 Comparison of Throughput for Variable
Overhead for 1000 Byte MPDU
35
5.4.4 Comparison of Throughput for Variable
Overhead for 2346 Byte MPDU
36
5.4.5 Aggregate Throughputs for 2.4 GHz

• Our proposal allows for 3 channels in US 2.4
  GHz band
• Each channel can coexist in the same area
• Aggregate throughput is 3 times single channel
  throughput

37
5.5 CW Jammer Test Description
CW jammer test steps a CW tone across the signal
band in 50 kHz steps. At each step, the jamming
level required to to produce the recommended BER
is determined. The worst 20 of the J/S levels
are discarded and the smallest of the remaining
J/S is used as the jamming margin. Processing
gain is then calculated according to the
following
38
5.5 Performance Against CW Jammer
Gp (S/N)0 Mj Lsys