Optimizing Performance and Efficiency of Power Amplifiers in Wireless Base Stations

The rapid growth of both users and digital content is putting enormous pressure on wireless infrastructure as the need to carry increasing amounts of data traffic on limited wireless spectrum is increasing. Meeting these demands results in high energy consumption, which in turn drives up the acquisition and operating costs of base station systems. The base station power amplifier (PA), which transmits the wireless signal from the base station, accounts for up to 30% of the base station cost. Implementing crest factor reduction (CFR) and digital predistortion (DPD) techniques before the wireless signal reaches the base station PA can improve base station signal quality and coverage while reducing system acquisition and operating costs. The

need to improve spectrum efficiency

In recent years, the number of wireless users has increased dramatically. At the same time, the emergence of new services such as music downloads and Internet access through mobile phones has led to an increasing amount of data transmission in the wireless infrastructure. At the same time, the spectrum allocated for wireless communications has remained the same. As a result, the continuous increase in users and traffic has caused the wireless spectrum to become extremely congested.

This is similar to the situation where a highway is congested during rush hour. Imagine that the spectrum is a highway and the data (voice calls, music, or Internet content) is the vehicles. The width of the highway, or the number of lanes, represents the fixed amount of wireless spectrum available. Adding lanes to a wireless data highway is a huge undertaking, just like adding lanes to a real-world highway, which involves procurement, building demolition, and construction of lanes.

On a highway, all the cars are heading to their destinations, and drivers want to arrive on time. Each car represents part of a voice call or music download, and the on-time arrival of many cars represents the completion of the download or call. More and more users are using wireless devices to access digital content, just like more and more cars on the highway. When there are too many cars on the highway, traffic begins to slow down. The same is true for wireless networks.

To solve this problem, wireless providers have turned to wireless standards that improve spectrum efficiency. This is similar to stacking multiple cars going to the same destination on a truck and driving the truck down the highway to the destination. This approach allows more data traffic to be carried on the same highway without causing a decrease in traffic. To improve spectrum efficiency, we can deploy or define all wireless standards, including CDMA2000, W-CDMA, TD-SCDMA, MC-GSM, WiMAX, and LTE. Figure 1 shows the wireless standards that have been deployed or defined in recent years to improve spectrum efficiency.

To understand the importance of CFR and DPD to current and future wireless infrastructures, it is important to understand three basic characteristics of PAs. First, output power determines the range of a wireless signal. Higher output power will provide wider base station coverage. Second, the power efficiency of a PA increases as output power increases, reaching its highest value near its saturation level. As these two characteristics indicate, base stations provide the best coverage and consume the least power when they are operated at their highest power. In

the past, wireless standards enabled service providers to operate base stations near their maximum output power, thereby achieving the lowest capital cost (fewer base stations were needed to cover the same area) and the lowest operating cost (lower electricity costs). However, as the number of wireless users exploded and the amount of data transmitted through the infrastructure increased dramatically, this changed because of a third characteristic of PAs: distortion increases as output power increases, and distortion becomes particularly prominent when the PA begins to deviate from the ideal (linear) curve.

This third characteristic is important because current wireless standards are very sensitive to distortion, which means that base station PAs must be operated below saturation levels to ensure signal quality. Below saturation levels, more of the input energy is dissipated as heat rather than being consumed in the transmission of the output signal.

This is like owning a car with a powerful engine but not being able to achieve the desired speed. Because the gas mileage of a large engine is poor, the owner has to spend a lot of money on gas and still not get the most out of the engine. Ideally, if the engine is gas efficient, the owner can enjoy the power of the car while saving money on gas. Likewise, ideally, a PA operating close to its saturation level will output a strong signal with very low signal distortion. In addition, operating close to saturation levels can save energy. This is where DPD and CFR are valuable to PAs. They improve the linearity of the PA output power, thereby reducing energy consumption.

Linearizing Power Amplifier Performance

Adaptive DPD can extend the straight-line performance of wideband radio frequency (RF) PAs far beyond the normal range. As mentioned earlier, as the output power increases, the output power deviates from the ideal linear performance curve. This is where the PA's amplified output signal becomes distorted, resulting in signal degradation and interference. DPD compares the distorted output signal of the PA to the undistorted input signal. It then adds a signal from the output that is the exact opposite of the distortion to the input, effectively canceling the distortion. Because DPD extends the linear range of the PA, it is called a linearization technique (see Figure 4).

The process of comparing the output to the input signal is called feedback and is most efficiently implemented digitally. Since operating conditions such as temperature may change, digital feedback provides the information needed to adjust the predistortion to the changing conditions. Digital predistortion and feedback implementations are much more robust, flexible, and manufacturable than analog implementations. DPD can extend the PA linear performance by 2 or 3 decibels (dB), which is a significant extension of the PA's operating range.

By linearizing the PA output performance over a larger operating range, a lower-cost, lower-rate PA can still meet the system's performance requirements. In addition, PAs with more linear performance curves are more energy-efficient. This will not only reduce power costs, but also the system's acquisition cost due to reduced energy conversion to heat and reduced cooling requirements.

CFR is another digital predistortion technique that increases the operating output power of the PA by reducing the peak-to-average ratio (PAR). To understand how CFR works, refer to Figure 3 and the example of stacking cars on trucks to increase the amount of data or traffic on a wireless network as described above. As shown in Figure 3, if the input power is increased enough, the output power reaches saturation and the PA will not cause the output power to increase. This is like a truck that has to go through a tunnel. If too many cars are stacked on the truck and the stacking height exceeds the height of the tunnel entrance, the cars on top will be left on the highway, causing traffic to slow down. Moreover, when the truck reaches its destination, several cars have been lost. The lost cars represent the data lost in the phone call.

Let's continue with this metaphor. The CFR algorithm analyzes the height of each truck to determine whether it can pass through the tunnel. If necessary, it shrinks each car a little so that the truck can pass through the tunnel and all cars can reach their destination. CFR automatically reduces the signal peaks and allows the signal to pass through the PA without clipping or distortion.

TI's CFR algorithm has a unique feature that it cancels excessive signal peaks by adding a signal with an inverting characteristic, rather than simply clipping it when it reaches a peak. There is little benefit in clipping the input signal because it actually distorts the output signal instead of eliminating distortion.

TI's solution can reduce the PAR while maintaining the quality of the output signal. By reducing the PAR of the input signal, the output power of the PA can be scaled in a linear manner. For example, if the PAR is reduced by 3dB, the PA's operating point will be improved by 3dB, thereby improving the PA's operating efficiency. Improving the PA's operating point by 3dB means achieving twice the output power without increasing the power consumption. To put it in a more vivid way, it is like a car with the power and horsepower of a V8, but its energy efficiency is equivalent to that of a much smaller four-cylinder engine.

In general, base stations consume a lot of energy. Therefore, many wireless service providers have implemented strict energy-saving standards for their base stations, not only to reduce operating costs, but also for environmental considerations. By adopting some wireless standards, TI's DPD/CFR solution can not only improve the PA's operating point (output power) by 10 times, but also improve energy efficiency by 4 times. With this excellent performance, TI's solution will have a significant impact on the deployment and operating costs of wireless base stations.

Conclusion

TI has implemented innovative adaptive DPD and CFR algorithms in a single-chip application-specific semiconductor product (ASSP) for the wireless infrastructure market. By addressing and compensating for the inherent deficiencies in base station PAs, TI's GC5322 not only enables base station PAs to operate at higher output power, but also solves the two most critical issues in wireless base stations, namely cost and power consumption. In addition, the GC5322 also includes digital up-conversion (DUC) components and can meet the requirements of all existing and emerging wireless standards such as W-CDMA, WiMAX, LTE and MC-GSM to achieve the most integrated and complete transmit processor solution available today.