Wireless systems have some of the toughest requirements in the electronics industry. Operating in an environment that may have atmospheric and electronic disturbances, wireless systems are often portable, necessitating compact size, low weight, and low power consumption to stretch battery endurance. DSP is fast becoming a standard feature on wireless systems too, adding computing power to the system requirements. Careful system design is necessary to create wireless designs that live up to their potential.

Electronic-design-automation (EDA) tools can help you develop wireless systems by letting you assemble the system in software before you actually build the hardware. Using a software simulation lets you perform a detailed analysis of the system's performance before you build it. Although analysis itself doesn't make a better system, it gives you a better understanding of the system so that you can determine what, if anything, needs improvement.

EDA tools at their most basic level can automate functions you once performed manually, such as schematic capture and physical layout. Although EDA tools that automate these tasks are more efficient than using manual design, the tools' greatest benefit is that they allow you to start a system design at a high level and work with a top-down design methodology. You model the system at its highest level and then simulate how the system behaves to inputs. Your system specification should guide you as to what to simulate and the level of detail you require.

The sooner you can get a handle on how your system will perform, the sooner you can get on a predictable development path. Of course, no design that advances the state of the art is completely predictable. You can't even be sure it's possible until a system is operating. Simulation, however, can go a long way toward verifying that a system will work. Furthermore, until you determine that your system design is satisfactory, designing details will be a waste of time.

Design with fewer iterations

The time you spend developing a system-level simulation is time taken away from other design tasks. In addition, EDA tools cost money, and training takes time away from productive design work. Even after training, it takes time to become really proficient with the tools. How do you know all this time and money is well-spent?

One way is to look at historical results within your company. If it takes several iterations of a design to get the final product right, then you can improve your current product-development methodology. Even removing one iteration could save time and money. A top-down design methodology could potentially save one or more iterations in developing a design.

If your current design methodology doesn't result in multiple iterations to reach your desired product, your method may be satisfactory. However, you should still consider the impact system simulation may have on your product. For example, you may design a better product with simulation because it allows you to experiment with design changes to see how they affect the system's overall performance. One of the major benefits of simulation is that it lets you quickly test many design changes.

Complex considerations

Almost any electronic design can benefit from system simulation, and wireless systems have a variety of complicating considerations that make system simulation particularly helpful. For example, many wireless systems, especially those for communication, use complex modulating schemes to transmit as much data as possible through a given portion of the electromagnetic spectrum. You cannot use simple calculations to approximate the performance of these complex modulation schemes. Furthermore, the wireless environment is a demanding one in which transmissions may be subject to interfering signals, reflections, weather-related effects, and other disturbances that make analyzing performance difficult.

Another aspect of wireless design is a trend toward the increasing use of DSP, which adds another variable to the design process. A typical DSP design requires you to consider cost and performance tradeoffs on the level of DSP vs analog signal processing.

DSP can make the performance of a system less sensitive to manufacturing variations. The performance of an analog design, however, varies from unit to unit, primarily due to component variations in resistors, inductors, and capacitors (although a properly designed analog product can accommodate analog-component variations and still meet performance specifications). Digital systems, on the other hand, perform identically on all units. In addition, in many cases, especially at the lower-frequency portions of wireless systems, DSP provides performance superior to that of analog and at a lower cost.

Portions of wireless systems will remain analog for a long time to come. For example, the RF portion is currently always analog, and DSP techinques are just starting to become viable for some IF applications. You can best find the proper mix of analog signal processing and DSP by simulating the system as a whole to make sure the performance is up to the specification. You can then evaluate which approach provides the lowest total system cost.

Experiment with tradeoffs

Creating a system simulation can be relatively quick if you have a library of models that covers the system components you need; acquiring or developing new models, however, takes time. Once you develop a simulation, it becomes relatively easy to make small design changes and run the many system tradeoffs necessary to analyze and settle on a design.

Once you settle on a design, you can move on to designing the details of the system. One of the great benefits of simulating and iterating a design until the simulation indicates that a system will meet your requirements is predictability. If you start with a system simulation and develop a top-level design that works, then you have a predictable design process as you work to change each block in the simulation into a detailed circuit performing the functions required.

Choosing the proper system simulator for wireless designs means finding one that fits your design process. One of the most important considerations in choosing a simulator is the product's model library.

The importance of models

Models are the currency for simulators: Without models, a simulator is useless. When you shop for a simulator, spend plenty of time examining the model libraries to make sure the model goes into the level of detail you require. For most wireless work, you need a library of models specifically dedicated to telecomm design.

The model library needs to include models not only for system components, such as amplifiers and mixers, but also for modulation schemes, atmospheric degradation, and other parameters. If you cannot obtain the models you need in a library, you can obtain them from other sources. For example, the simulator vendor may develop models for a fee, and manufacturers of wireless-system components sometimes develop models. Alternatively, you can develop the models yourself.

You also need to evaluate how the simulator works with the rest of your design tools as you go from a top-level system design down to a detailed design. If you are doing only the top-level system design, and someone else or another company is designing the modules that go into the system, then you don't necessarily need the simulator to work with lower-level design tools. However, if you are designing the modules or function blocks that comprise the system, you might want to be able to simulate each block and drop them back into the system simulation to see if the block still performs as expected.

As your design progresses from a block-level system design into a detailed design, you need to decide what level of fidelity you need from the simulator in representing the real system. High-level system simulators shouldn't have difficulty representing both digital and analog portions of wireless systems. As you move down into detailed design, you may require more sophisticated mixed analog-digital simulation.

Systems with DSP blocks require special consideration. For example, you may simulate your system at the top level with a floating-point representation. As you work out the detail design, you may actually be using fixed-point processing for higher speed and lower cost. The floating-point representation now needs to change to represent the quantization effects of the fixed-point computations. The ability to represent the actual resolution of the hardware is called "bit-true simulation."

You may place other requirements on a wireless-system simulator. It may take time to find a simulator, acquire or develop the models, and learn to use the tool effectively. In the long run, though, it should help you design the system you want with fewer iterations.