- Buyers Guide
802.11ac Design Considerations for Mobile Devices
The rapid growth of smart mobile devices is driving mobile data usage and Wi-Fi proliferation, creating an ever-increasing demand for faster wireless networks to support bandwidth-intensive applications, such as web browsing and video streaming. The new IEEE 802.11ac standard is designed to meet this demand, by providing major performance improvements over previous Wi-Fi generations, as shown in Table 1, which shows theoretical maximum data rates.
Mobile devices supporting 802.11ac are already becoming available. IMS Research predicts rapid adoption of 802.11ac and expects that over 400 million devices supporting the new standard will be shipped in 2016. For mobile devices, 802.11ac provides significant performance improvements over its predecessor, 802.11n. Single-antenna devices, such as smartphones, can achieve three times the data-transmission rate, compared with 802.11n, by using the wider 80 MHz channels with an advanced modulation (MCS 9), defined in 802.11ac. The additional performance means that 802.11ac will allow new types of WLAN usages as well as accelerating existing applications.
These new usages include simultaneous streaming of HD video to multiple clients throughout the home and instantaneous data sync and backup. 802.11ac achieves its performance improvements primarily by extending WLAN concepts initially introduced in previous 802.11 generations, such as 802.11n. In addition, 802.11ac uses the 5 GHz band, which is less crowded than the 2.4 GHz band used by most previous WLAN generations. This facilitates higher WLAN throughput with less interference. The biggest factors contributing to 802.11ac performance advantages are:
802.11ac introduces mandatory 80 MHz and optional 160 MHz channel bandwidths in the 5 GHz band, in addition to the 20 MHz and 40 MHz channels specified in 802.11n.
Increased Modulation Complexity
With 802.11ac, the maximum modulation complexity increases from 64 QAM to 256 QAM, with a 5/6 coding rate. This change alone provides a 33 percent boost in the data rate, compared with 802.11n. To take advantage of this benefit over existing Wi-Fi networks, some equipment makers are expected to provide support for 256 QAM 802.11ac modulation over the 2.4 GHz band, as well as the 5 GHz band.
Expanded Support for MIMO and Beamforming
The number of simultaneous multiple-input multiple-output (MIMO) spatial streams increases from 4 4 to 8 8, allowing for a total maximum aggregate WLAN throughput of 6.93 Gbps. The 802.11ac standard also adds support for multi-user MIMO, allowing more efficient allocation of network capacity among multiple devices: each antenna of a multi-antenna device, such as a wireless router, can simultaneously communicate with a different single-antenna device such as a smartphone. 802.11ac supports standardized beamforming, which takes advantage of multiple antennas to increase signal strength and signal transmission range between communicating devices.
While the new standard delivers significant performance improvements, it also creates a variety of new challenges for designers of mobile devices. These challenges make achieving the desired performance more difficult than ever. Key issues include:
- Maximizing linearity and achieving low dynamic Error Vector Magnitude (EVM)
- Generating high power output while minimizing power consumption
- Coexisting with cellular communications in the 2.4 GHz band
Each of these challenges is explained in more detail.
Linearity and Dynamic EVM
The increase in modulation complexity to 256 QAM contributes significantly to the increased performance of 802.11ac. With 256 QAM, the data rate is achieved by representing eight coded bits per Orthogonal Frequency Division Multiplexing (OFDM) symbol, compared with six bits with 64 QAM (the maximum supported by 802.11n). However, this higher-order modulation also creates significant design challenges.
As the data rate increases, signal quality becomes even more important. To successfully receive an 802.11ac 256 QAM signal, a receiver will require a significantly less distorted signal than with 802.11n 64 QAM. To deliver this signal, the power amplifier (PA) of the transmitting device will need to provide greater linearity.
This linearity requirement is reflected by a more stringent EVM specification. EVM is an important figure of merit used to measure the linearity of a transmitted OFDM signal. EVM is the magnitude of the difference between the measured vector and the ideal (reference) vector. In 802.11n, the maximum allowed system EVM for a transmitting device is –28 dB (4 percent), when using the highest allowed modulation (64 QAM) and a 5/6 coding rate. With the higher-order 256 QAM modulation supported by 802.11ac and a modulation bandwidth of 160 MHz, the maximum system EVM required is further tightened to –32 dB or 2.5 percent.
To meet this requirement, careful system budgeting is needed for each of the principal components of the device, such as the transceiver and the front-end module. The transceiver must provide sufficient linear drive for the front-end module; any shortcoming will need to be compensated by a higher gain from the PA. At the same time, the transceiver is expected to deliver a low EVM residual floor that is several dB below the overall system EVM requirement. Based on these assumptions, the PA is expected to achieve less than 2.5 percent EVM at the rated power.
When assessing PA linearity, it is essential to measure EVM in a way that is representative of actual use. Wi-Fi is a Time Division Duplexing (TDD) system, in which the PA continuously pulses on and off during use; it pulses on each time the device transmits data, then pulses off to save power. It is therefore important to measure EVM in this pulsed condition (dynamic EVM) rather than when the PA is always on (static EVM). In an ideal case, dynamic EVM should be the same as static EVM. But in reality, dynamic EVM is usually worse than static EVM due to the PA’s transient response.
802.11ac defines a very strict transient response requirement. At the start of each packet, the transmitting device sends an OFDM training sequence used to calibrate the phase, amplitude and timing of the signal. The sequence consists of 10 short symbols and two long symbols. The short symbols include an automatic gain control (AGC) symbol used by the receiver to set its AGC amplifier so that it applies the correct gain to the received signal. The receiving device then assumes that the transmit power level remains constant until the end of the packet. The total transmission time for the training sequence is approximately 20 µs.
If the PA of the transmitting device exhibits slow power turn-on or power overshoot during this critical period, the constellation after OFDM demodulation will expand or shrink respectively, resulting in poor dynamic EVM. Figure 1 illustrates the transient PA transmit power responses, resulting from slow turn-on and power overshoot respectively; Figure 2 shows the constellations that result from these transient responses. Theoretically, if all points were centered perfectly within the circles, EVM would be 0 percent. These figures are based on lab test results with 64 QAM; the problem is even more pronounced when using 256 QAM.
Because of this effect, a PA used in an 802.11ac device must exhibit an extremely stable transient power response, in order to minimize dynamic EVM. By designing for stable response, very low dynamic EVM has been achieved in PAs and front-end modules. For example, the TriQuint TriConnect™ WLAN/Bluetooth front-end module achieves 2.2 percent EVM at 18 dBm with 802.11ac 256 QAM (MCS 9, 40 MHz channel bandwidth).
As WLAN performance requirements will inevitably continue to increase in the future, more sophisticated techniques will be required to further improve signal range and quality. An example of such a technique is digital pre-distortion, in which the transceiver chipset uses feedback from the PA to adjust the signal so that the system EVM is minimized. FEMs will need to include the capabilities necessary to provide this feedback.
Generating High Power Output While Minimizing Current Consumption
Designers of mobile devices must balance a range of conflicting requirements to optimize WLAN performance, while managing power consumption. The challenge with 802.11ac is to achieve a design that successfully meets requirements for very low EVM, along with high transmit power output, while minimizing current consumption to lengthen battery life.
The challenges are accentuated as mobile devices become more complex and need to support a wider range of cellular and WLAN standards. Greater levels of integration are needed to make the best use of limited space, simplify design and enable faster development of next-generation wireless products. An example is how TriQuint’s integrated single-band WLAN modules can replace as many as four discrete components. This highly integrated Wi-Fi front-end module, shown in Figure 3, combines the discrete components required for WLAN transmit and receive chains into a single compact module. It houses a PA and a low noise amplifier (LNA) with a switched bypass mode, RF detector (Vdet) and switch.
In these highly complex, integrated mobile devices, the WLAN signal must share the RF chain with crowded cellular bands. This in turn increases the need for filters and switches, resulting in greater signal losses in the transmission path of the PA. The resulting reduction in the signal-to-noise ratio makes it more difficult to achieve the required EVM. To compensate for these losses, the power output may be increased. However, as the PA approaches the power saturation level, linearity degrades. The relationship between power output and linearity typically resembles a “hockey-stick” curve; once a specific power level is exceeded, the EVM increases exponentially. The PA must therefore operate at a sufficiently backed-off power level to maintain the very low EVM required for complex 802.11ac modulation schemes.
The challenge for FEMs and PAs is to meet the EVM requirement, while producing adequate power output and minimizing current consumption. Focusing solely on WLAN performance, at the expense of power management, may compromise battery life. Conversely, focusing solely on power management may reduce WLAN performance. The trade-off between performance and power consumption varies depending on the application; for mobile applications, the preferred tradeoff between current consumption and transmit power has been in the range of 16 to 20 dBm.
The selection of the manufacturing process technology is critical to meeting the needs of different applications. Many technologies are available for WLAN PA designs. GaAs design technologies have been developed to offer faster data-exchange rates, extended battery life and better amplification. For PAs, InGaP/GaAs HBT offers the greatest power handling capability and ruggedness, while GaAs PHEMT allows a high level of integration with switches and LNAs.
Coexistence with 2.4 GHz Cellular Networks
Though 802.11ac operates in the 5 GHz band, mobile devices will typically provide dual-band 5/2.4 GHz WLAN capability, for compatibility with existing Wi-Fi networks based on earlier standards such as 802.11n and 802.11g. To increase data rates over existing 2.4 GHz networks, some device manufacturers are expected to implement 256 QAM modulation – a key element of the 802.11ac standard – over 2.4 GHz WLAN channels. This will significantly increase 2.4 GHz WLAN performance, compared with the less-complex modulation schemes previously available with the earlier Wi-Fi standards, though the data rates will not match those available in the 5 GHz band due to other constraints of the 2.4 GHz band, such as narrower channels.
While dual-band operation offers obvious advantages, it also introduces coexistence challenges. Operating in the 2.4 GHz band increases the potential for interference with cellular communications – notably LTE networks – that utilize closely adjacent frequency bands, as shown in Figure 4. Potential interference issues are multiplying quickly.
Designers, therefore, must address the potential for transmitted WLAN signals to desensitize LTE reception on the device and for LTE signals to interfere with WLAN communications. The challenges will increase as mobile devices grow in complexity; many next-generation smartphones will support up to 14 radio frequency bands.
Solving this coexistence challenge requires RF filters that are capable of rejecting closely adjacent frequencies. At the same time, the filters must minimize insertion loss in the WLAN transmission pathway, to help maintain the high signal-to-noise ratio and correspondingly low EVM required for 256 QAM modulation. Bulk Acoustic Wave (BAW) filters are particularly effective at meeting these requirements, offering significant advantages over the Surface Acoustic Wave (SAW) and ceramic filters, traditionally used for cellular applications. In the GHz range, BAWs can achieve quality factors (Q-values) that are superior to other traditional acoustic technologies. As a result of the high Q-values, the filter skirts will be very steep, while insertion losses remain low, even at the edges of the passband. While the performance of conventional SAW filters degrades at higher frequencies and reaches its practical limits at 2.5 GHz, BAW filters can address frequencies up to 6 GHz. BAW devices also have significantly less temperature dependency and much greater electrostatic discharge (ESD) robustness.
Double-digit growth is forecasted for devices enabled with 802.11ac, with annual shipments estimated to exceed 400 million by 2016. The 802.11ac standard provides substantial performance improvements over its predecessor, 802.11n. But achieving this performance will be more challenging than ever. Key design concerns include achieving low dynamic EVM, generating high power output while minimizing current consumption and coexisting with LTE cellular communications in the 2.4 GHz band. The good news is that highly integrated 802.11ac front-end modules and BAW filters are becoming available to help solve these challenges.