Timing matters for 802.11ax users
Article By : John Lukez
The 802.11ax standard requires that all STA devices transmit within 400ns of each other, with the synchronisation provided by a trigger frame.
« Previously: Why 802.11ax is not just higher data rate
Users of the 802.11ax standard share the spectrum at the same time. Thus, interference between simultaneous users must be minimised to achieve the highest system capacity. Accurate timing across users is required to minimise interference and maximize system capacity.
The 802.11ax standard requires that all STA devices transmit within 400ns of each other. The synchronisation in the 802.11ax system is provided by the trigger frame sent by the AP. All the STA devices are synchronised to the AP, which acts as the “master” in an 802.11ax system. The trigger frame contains information sent by the AP about which stations can transmit during the specified time and which subsets of OFDMA sub-carriers resource units to use. A simple analogy for the trigger frame is to think of the coxswain on a rowing team controlling the direction of the boat and the pace of the crew.
In 802.11ax, this precise user-to-user timing is important because larger timing errors with residual CFO and sampling frequency offset (SFO) generate inter-carrier interference (ICI) and intersymbol interference (ISI), all of which reduce system capacity. Even in 802.11ac systems (OFDM), the frequency error was an important parameter to control. Because one user gets all the sub-carriers at a moment in time, there is only one user's frequency error that matters. When multiple STA devices transmit at once, the frequency error of each user matters. Figure 1 illustrates how CFO and timing errors in the time domain lead to interference in the frequency domain and degrade the receiver’s ability to correctly decode the information.
Figure 1: Timing offsets produce the frequency degradation shown in the spectrum plots on the right.
Testing this key behaviour requires that the tester can act like an access point and generate the trigger frames to the STA devices while measuring the time from the sending of the trigger packet to the reception of the DUT (STA) packet within tens of nanoseconds. Figure 2 illustrates just such a test sequence.
Figure 2: An 802.11ax timing test requires that the DUT properly respond to a trigger frame.
In this test, you need to send the trigger frame with encoded information. The DUT then receives the trigger frame, decodes the information and responds with a packet. You must then measure the time from when the trigger frame packet left the tester to when the DUT packet was received. This time needs to be the short interframe space (SIFS) time ±400ns (16,000ns ±400ns).
CFO is another parameter that you must validate in 802.11ax systems. The CFO is the remaining frequency error in the station device after it has synchronised to the trigger packet from the AP. An increase in CFO leads to more ICI in the 802.11ax system and data decoding errors increase. The 802.11ax standard requires that CFO be less than 350Hz.
Measurement of CFO uses the same process as the timing-error test. You must send a trigger-frame packet, which the DUT (STA) decodes. The DUT aligns its clock to the information within the trigger frame and responds with a packet to the tester. You'll then need to decode this information and calculate the residual CFO.
Conclusion
802.11ax utilising OFDMA employs several characteristics to dramatically increase system capacity over past Wi-Fi systems. The addition of timing synchronisation, central coordination of users with the AP and power control adds complexity never-before seen in Wi-Fi systems. Thus, new tests are required to ensure these systems will operate as expected.
Many of these parameters could fail in a way that would only be visible as part of a larger, multi-user system, as opposed to a hard failure of any one device. Depending on the final behaviour of a system, these offending devices may be assigned to a low-quality selection of subcarriers, refused connections or dropped from the network entirely due to their poor performance. You could imagine a cell phone being dropped from the network if it begins to harm network performance.
By employing test approaches described in this article, makers of 802.11ax systems can ensure they will have products that work seamlessly on 802.11ax networks and deliver the highest performance while maximising network capacity.
John Lukez is Vice President of Applications Engineering at LitePoint.
First published by EDN.
« Previously: Why 802.11ax is not just higher data rate
