The Worldwide Interoperability for Microwave Access (WiMax) Forum is supported by over 520 corporate members from all segments of the industry. Most of the world's telecommunications carriers have indicated interest in WiMax, and many are conducting WiMax trials. Several are even now preparing to deploy it. The question is not if WiMax will be deployed, but when, and to what extent, and for what applications. This has prompted huge interest among component, subsystem and system manufacturers.
WiMax telecommunications technology aims to provide wireless data at high speeds over long distances in a variety of ways ranging from point-to-point links to full mobile cellular-type access. The complexity of the new generation of wireless devices using signals such as WiMax is driving a move to faster, more efficient design software in order to accurately predict performance and deliver communications products to market quickly and efficiently. Overdesigning, a common consequence of the use of traditional tools and methodologies, is no longer a competitive option. Older design methodologies address metrics specific to the radio portion only. Often, spreadsheets and calculators, along with disjointed simulation software tools, are used for cascaded budget analysis and to measure carrier-to-interference-plus-noise ratio (CNIR). Then "rules of thumb" are implemented to predict pertinent metrics, such as bit error rate (BER). This, however, can cause designers to overlook phase noise in an oscillator, impedance mismatch between components or the characteristics of the image noise rejection filter.
The performance of WiMax systems is evaluated by metrics similar to those applied to other existing wireless standards. However, WiMax is one of the most complex wireless standards ever deployed. WiMax components, subsystems and complete systems must meet very stringent requirements of performance that necessitate much more complex up-front system design, analysis, simulation and measurement procedures than preceding communication standards have dictated for design tools and methodologies.
Furthermore, traditional approaches may produce incomplete results if the baseband portion of the design is not included. Such approaches often require extra design margins to achieve the desired measured performance of the final integrated RF and baseband system, cauusing extra costs and delays in time-to-manufacture. The design of today's WiMax products requires a unified baseband and RF design methodology to help designers overcome the challenges of working with WiMax signals and take advantage of the market's windows of opportunity.
Such a design methodology uses realistic modulated signals and provides accurate performance predictions for RF components, eliminating overdesign. This expanded methodology enables the use of system-level measurements that are defined in the WiMax specifications and which are common to all other wireless standards:
* Error vector magnitude (EVM) is an important measurement used to characterize the performance of RF components. It measures the quality of the modulated signal passing through the RF components under test and indicates whether the signal-to-noise ratio (SNR) is maintained above a specified minimum value, ensuring reliable operation.
* Adjacent channel power ratio (ACPR) is a measure of the interference caused to adjacent channels and is used to characterize the distortion introduced by RF components.
* Bit error rate (BER) is the ultimate performance metric used to measure any wireless standard. System performance is measured by the capability to provide reliable data communications and maintain BER below a predefined value at a given signal level. All components of such systems should be designed to meet BER requirements.
To better illustrate, let's consider an example design methodology that accounts for joint circuit and system analysis (also known as cosimulation) and includes a WiMax signal generator, typical hardware impairments (amplitude/phase distortions, DC offset and phase noise), an RF device under test (DUT) that is being designed and analyzed and a WiMax receiver. There are several methods by which to characterize the RF components of a DUT. Many designers use behavioral models, allowing the definition of parameters specific to particular components. An amplifier may be characterized by defining its 1 dB compression point, its second- or third-order intercept points, its noise figure, its saturation level, and so on. Designers may also choose to characterize RF components as actual circuit designs embedded within a circuit design tool (a circuit design being a compilation of transistors and other discrete components). Designers can also always rely upon laboratory measurements or manufacturing data sheets to characterize RF components. Whether through a behavioral, circuit model or measurement approach, the resulting characterization of RF component can be readily employed within a system simulation platform.
Within a system simulation platform, several traditional RF measurements are used to evaluate RF components, and certain system-level measurements are used to ascertain the overall performance of the system, such as complementary cumulative distribution function (CCDF) and EVM. A CCDF measurement is performed before and after the DUT and is an indicator of the signal clipping introduced by nonlinear devices. EVM is measured for different levels of the signal power and shows the increase in distortion as the DUT is driven closer to saturation. The AM-to-AM characteristic of an amplifier and its current operating point enables designers to evaluate the level of backoff and the efficiency of the amplifier. Although it is a very commonly used measurement and a good indication of performance, in the case of complex signals such as WiMax, system-level measurements provided up front in a virtual design environment or system simulator platform are necessary to accurately predict system performance and avoid unnecessary design margins for RF components. Similarly, the IQ constellation of the undistorted transmitted signal and the constellation of the demodulated signal are plotted on the same graph. These simulated measurements visually illustrate the effect of the DUT and hardware impairments. The desired goal is to simulate reality as nearly as possible long before building the component, subsystem or system.
Spectral compliance is very important for any wireless system. Transmitted signals should comply with well-defined spectral masks under all conditions, avoiding spectral regrowth. ACPR is measured at several frequency offsets to evaluate the level of interference caused to adjacent channels. As expected, ACPR increases as the DUT is driven closer to saturation.
The overdesign of RF components lowers competitiveness, but the joint circuit and system design methodology discussed here helps designers avoid such extra margins, making possible interactions and tradeoffs between circuit and system implementations. It provides accurate systems simulations and helps reduce design margins, leading to diminished power consumption, decreased costs and more competitive products.
Gent Paparisto is a Senior Systems Engineer at AWR, Inc. He has extensive experience in systems engineering covering a broad range of communication technologies and signal processing algorithms for wireless and wireline applications. He received his PhD in electrical engineering from the University of Southern California, Los Angeles.