A presentation on Wi-Fi6 or 802.11ax technology and RF design challenges. A 'black box' method to measure Error Vector Magnitude is also presented. OFDMA, MU-MIMO, OFDM.

1. RF DESIGN CHALLENGES AND TESTING
Center of Wireless Technologies Colloquium
Technical University of Eindhoven
7th December, 2020
WI-FI6 / 802.11AX

2. • Masters in Electrical Engineering - Mixed Signal Microelectronics(MSM), TU/E, 2011
• Wireless engineer at TP Vision(formerly Philips TV’s), 2013-2015
– Evaluated the first 802.11ac radio systems for Smart TV product line
– Wireless peripherals development (BT audio, RF4CE/BLE remote control)
• Systems engineering at Qorvo, 2015-2020
– Wi-Fi- IOT coexistence solutions (patent)
– IOT Design win in Amazon products
– Wi-Fi filters and FEM development
• Wi-Fi Architect at Liberty Global-Current
– Looking into Wi-Fi6 based CPE and mesh systems for home segment
Love affair with wireless technologies still strong!
ABOUT ME
2
MAYUR SARODE

3. This presentation will highlight Wi-Fi6 RF design challenges and discuss some test results
measured with commercial Wi-Fi6 routers.
To support new features, Wi-Fi chipset / RFFE* vendors had to make significant improvements on
Wi-Fi6 hardware design. Key features promised by Wi-Fi6 standard
• OFDM improvement
• 1024 QAM modulation (MCS11/10)
• OFDMA scheduling
• UL/DL MU-MIMO
• TWT and BSS coloring
PRESENTATION ABSTRACT
RFFE: RF Front End 3

4. THE WI-FI6 PROMISE
4
Graphic source: Wi-Fi alliance

5. Wi-Fi6 R1
Wi-Fi6E
Wi-Fi6 R2
WI-FI TIMELINE
5
2019
2020
2021
UL MU-MIMO
Preamble Puncturing
6 GHz spectrum
DL/UL OFDMA
DL MU-MIMO
2015
DL MU-MIMO
Wi-Fi5
802.11n 802.11ac 802.11ax
2009
MIMO
Wi-Fi4

6. FREQUENCY BANDS
WI-FI6/E
6
2.4- 2.48
GHZ
5.17-5.33
GHz
5.490-5.835
GHz
5.925 – 6.425
GHz
• With the opening up of 6 GHz band, routers/APs moving towards Tri-band architecture
– Bandpass filters necessary to separate out the bands
• Wi-Fi transceivers need to support 6GHz bands
– Separate silicon Vs 5 GHz transceiver update
Legacy
&
IOT
Low latency/ high throughput applications
Legacy and High throughput applications
Dual band
Tri band

7. Wi-Fi6
DESIGN
Amplitude
Time
Spatial
Frequency
RF PERFORMANCE METRICS
7
Output power
Spectral mask/flatness
Inter Carrier interference
Adjacent channel interfere
Antenna isolation
Timing
drift
Carrier
Frequency
accuracy
UL MU-MIMO
UL OFDMA
8 spatial streams
1024 QAM
OFDM
DL-OFDMA

8. EVM | CFO | SNR
SOME TERMINOLOGY
8
EVM -35 dB (MCS11/MCS10)
CFO 2.4 GHz: +/- 25 ppm
5 GHz: +/- 20 ppm
SNR 35 dB
Error Vector Magnitude (EVM)
Carrier Frequency Offset (CFO)
Signal to Noise Ratio (SNR)
SNR

9. WI-FI IMPAIRMENTS
Graphic source: wlanpedia 9
• WI-Fi6 MCS11/10 index mandates a tough -35 dB EVM limit
• System design goal
– Meet EIRP limits by maximizing transmit power and designing Omni-directional antennas
Wi-Fi chipset
Front End Module Filtering & Matching Antenna
2.4 GHz X 4
5 GHz X 8
iPA
LO leakage
TX PATH

10. • ¼ of 802.11ac Subcarrier spacing
– Local Oscillator with low phase noise to minimize ICI
– Sensitive to Carrier Frequency Offset
• 77% increase in OFDM data channels
– Impact on frequency synthesizer/mixer design
• 4 times larger FFT size
– 160 MHz bandwidth mandatory
– 30% more efficient than 802.11ac
– Longer symbol duration
– Higher power consumption and larger area
802.11 AX OFDM
ICI: Inter Carrier Interference
FFT: Fast Fourier Transform
10
Sub carrier
spacing
312.5 KHz 78.125 KHz
FFT
size(max)
512 2048
Orthogonal Frequency Division Multiplexing
Graphic source: wlanpedia

11. iPA + RFFE
• Wi-Fi6 demands tighter EVM (-43 dB) on Front end PA
– Designed to work with 10 dB return loss (VSWR=2:1) antennas
• 80/160 MHz bandwidth operation
– 0.1 dB amplitude droop causes 6 dB EVM degradation
• Higher current to meet linearity requirements→ DPD saves the day!
POWER AMPLIFIER
11
Transmit Power→
EVM
→
PA power profile
PA frequency response
Source: Litepoint
EVM
→
Amplitude
Distortion
→

12. • Allows for significant power savings when used with an external non-linear FEM
– Reduces system power consumption for MIMO systems
• 7% of the total CPE power budget (0.5W per chain)
– Better thermal management for CPE devices
• Contribute to smaller and greener CPE devices in the future!
DIGITAL PRE DISTORTION
12
Wi-Fi chipset
DPD
+
Non-linear PA
External FEM
=
Wi-Fi chipset implements a PA gain/phase correction algorithm to improve cascaded PA’s linearity
FEM: Front End Module
Wi-Fi chipset
Highly Linear PA

13. SPECTRAL SHAPING/TX POWER
DPD IMPROVEMENTS
Source: Qorvo 13
Spectral Shaping
• Increase MCS11/10 coverage ( larger TX power)
• Improve spectral mask
– meet FCC band-edge requirements at higher TX power for channel 1 and 11
TX power improvement

14. ORTHOGONAL FREQUENCY DIVISON MULTIPLE ACCESS
802.11AX OFDMA
14
User 1
• Resource Units (RU) as small as 2 MHz
– 37 simultaneous users in 80 MHz band!
• Ideal for applications requiring low latency/jitter
• Basis for Preamble puncturing
– Potential to improve 80/160 MHz channel utilization
RU
OFDMA
MU-
MIMO
SU-
MIMO
Wi-Fi vendor’s
special sauce
Graphic source: wlanpedia

15. • RU size is dynamically reconfigured over time
• Different MCS rates/output power per user!
802.11AX OFDMA
Source: Litepoint 15

16. • AP adjusts power level for each Resource Unit(RU)
• PA’s may need upto 12 dB better linear range to avoid co-channel interference
• Degradation in EVM of lowered power RU’s expected
OFDMA DOWNLINK
Graphic source: Litepoint 16
TX power@AP
Resource
Units
12 dB

17. DYNAMIC POWER CONTROL
• Similar to 4G LTE uplink communication
– GPS guided clocks to sync all devices
• 802.11ax AP’s/routers dependent on their own built-in oscillators as the reference
– Clients adjust their internal clock and frequency references by extracting offset information via TRIGGER they
receive
OFDMA UPLINK
17
RX power@AP
Users
TRIGGER
Graphic source: Litepoint
RXpower@AP
Users
Noise
floor
BEFORE POWER CONTROL AFTER POWER CONTROL

18. TRIGGER FRAME CONTROL
OFDMA UPLINK
Graphic source: Litepoint 18
• Inter Carrier Interference causes
– Receiver compression, Signal leakage, CFO
RSSI accuracy
• RSSI measurement accuracy:+/- 2dB
• Transmit power accuracy: +/- 3dB
Timing/Frequency Error
• Transmit within < 0.4 usec relative to TRIGGER frame
• Relative Frequency Error< +/-350 Hz (0.07 ppm@ 5.2 GHz)
TRIGGER
TRIGGER FRAMES ALSO USED in UL MU-MIMO!

19. TRIGGER
DL & UL
MU-MIMO
19
• 4 users (2 spatial stream) support on the 5 GHz band
– Upto 12 spatial streams ( 4 in 2.4 GHz, 8 in 5 GHz band)
• Challenge to integrate many antennas in a small CPE devices
– minimum 20 dB antenna-antenna isolation necessary
• MU-MIMO uses TRIGGER frames to synchronize uplink from stations
RX power@AP
Time
WI-FI6 TX beamforming
Upto 3 dB higher gain
More accurate beam steering
USER 1 USER 2
USER 3
USER 4
Graphic source: Litepoint
freq

20. OFDM TEST SETUP
WI-FI6 PHY TESTING
20
• Fully CONDUCTED black-box test-setup to measure critical PHY level properties of Wi-Fi5/6 design (without
antenna)
– Data Modulation Code Scheme(MCS) count
– Data Error Vector magnitude (EVM)
– Carrier Frequency Offset (CFO)
– Signal to Noise Ratio (SNR)
• Measurements done with MATLAB WLAN tool box
– Captures TCP I/Q samples from Spectrum Analyser(SA)
• Variable attenuator introduced to induce MCS drop
REFERENCE
2.4 GHz Channel 1
HT 20
5 GHz Channel 36
HT40
ASUS
RTAX88U
Wi-Fi station
BLACK BOX
testing

21. TIME DOMAIN
• Analysis done on Data packets collected during 25 msec TCP session
WI-FI PACKETS
Graphic source: Litepoint 21

22. Sl no PPDU Format
Frame Type
L-SIG
EVM
(rms)
L-SIG
EVM
(max) MCS SNR
Data
EVM
Stream 1:
(RMS)
Data
EVM
Stream 2:
(max)
Data
EVM
Stream 2:
(RMS)
Data
EVM
Stream 2:
(max) CFO (Hz) PPM
Spatial
Streams
4 HE-SU ampdu -41.2 -34.4 11 36.12 -0.6 0 -37.3 -28.8 -18694.5 -8 2
12 HE-SU ampdu -40.3 -33.7 11 36.1 -0.6 0 -38.9 -31.4 -18708.2 -8 2
24 HE-SU ampdu -41 -35.1 11 36.48 -0.7 0 -38.4 -29.1 -18654.3 -8 2
37 HE-SU ampdu -39.1 -34.4 11 34.53 -0.6 0 -37.1 -26.7 -18600.8 -8 2
45 HE-SU ampdu -41.5 -34.4 11 33.91 -0.7 0 -37.4 -23.8 -18677.9 -8 2
DATA PACKET ANALYSIS
MATLAB SIGNAL PROCESSING
22
EVM/Constellation diagram
• Key Wi-Fi6 PHY parameters are measured and averaged over multiple data packets ( >100)
MCS 11 constellation diagram MCS 11 Spectral Mask

23. 0
5
10
15
20
25
30
35
40
45
50
0 6 9 12 15 18 24 27 30 33 36 42
Packets
[%]
Attenuation [dB]
MCS rates
ASUS RTAX88U (MU-MIMO/TX beamforming OFF)
MCS 11 MCS 10 MCS 9 MCS 8 MCS 7 MCS 6 MCS 5 MCS 4
MCS SELECTION
ASUS RTAX88U
23
5 GHz
• 10% of the Data packets (HE-SU) are sent on MCS11 index
-25 dBm
MU-MIMO DISABLED
BEAMFORMING DISABLED

24. EVM
ASUS RTAX88U
24
MCS11 MCS 10 MCS 9 MCS 8 MCS 7 MCS 6 MCS 5 MCS 4 MCS 3 MCS 2 MCS 1 MCS 0
-35 -35 -32 -30 -27 -25 -22 -19 -16 -13 -10 -5
802. 11 ax EVM regulatory limit
5 GHz
-25 dBm
MU-MIMO DISABLED
BEAMFORMING DISABLED
PASS at all
attenuation points
NOT PASS at all
attenuation points

25. CFO
• CFO well within the spec!
ASUS RTAX88U
25
5 GHz
MU-MIMO DISABLED
BEAMFORMING DISABLED
-25 dBm

26. • Investigate Wi-FI6 receiver performance for adjacent channel rejection. Impact on
– BSS color implementation
– Dynamic Bandwidth selection
• Benchmark RF specs. on Downlink use cases
– Introduce butler matrix to test MIMO
– Evaluate selection of SU-MIMO, OFMDA, MU-MIMO mode of transmission
• Create a “Black box” testing methodology for Uplink uses cases
CONCLUSION
26

27. Same spec but smaller RBW
• 802.11 ac: 312.5 RBW
• 802.11ax: 78.125 RBW
LO LEAKAGE SPEC
Source: Litepoint and Rhode & Schwarz 27
Ptot= transmit power per antenna (dBm)

28. PASS FAIL criteria
• CFO
– 2.4 GHz:+/- 25 ppm*
– 5 GHz: =/- 20 ppm
IEEE 802.11AX
28
20 MHz BW spectral mask
EVM Vs MCS rate
*Parts Per Million

29. • https://www.litepoint.com/wp-content/uploads/2018/12/PA-Testing-Application-Notes-092517.pdf
• https://www.litepoint.com/wp-content/uploads/2019/09/Wi-Fi-6-OFDMA-App-Notes-091319-
web.pdf
• https://www.qorvo.com/design-hub/technical-articles/the-new-wi-fi-6-standard-combining-
software-and-hardware-for-best-in-class-solutions
• http://download.ni.com/evaluation/rf/Introduction_to_WLAN_Testing.pdf
APPENDIX
29