 |
|
| |
|
|
| |
Selecting an AWG for wideband signal simulation
( 01 Jul 2006 )
By Mike Flaherty, Agilent Technologies, Inc.
|
For modern radar and military communication systems, receivers must perform a multitude of tasks quickly and correctly. In the world of commercial communications, components and subsystems of the wireless infrastructure around the world must handle increasing levels of voice and data traffic more efficiently. Design cycles continue to shorten. Time-to-Market and Time-to-Money pressures are forcing designers to get it right the first time. To do this, better modeling and simulations must correctly predict product behavior and enable faster implementation.
Computers and software have enabled this revolution by providing better modeling and simulation tools to the design communities. Electronic Design Automation (EDA) tools from companies such as Agilent Technologies, OrCAD and Eagleware allow engineers to model their designs and produce sophisticated simulations of whole systems on the computer. Products like The MathWorks’ Matlab and Simulink enable this activity in the commercial communications, defense and academic communities. Once the models are built and the waveforms are designed, the engineer must have a way to transform the digital information into the analog signal domain.


CONCERNING BANDWIDTH AND DYNAMIC RANGE
The arbitrary waveform generator (AWG) has evolved over the years to meet this demand. Today, designers recognize that the AWG is the key piece of technology that bridges the world of mathematics and simulation to the world of physics and hardware. The more accurately the AWG translates the waveform, the more effective the model will be.
Engineers and scientists often use a mathematics-rich programming language such as Matlab or a simulation environment such as Simulink to build the necessary waveform vectors for their system. At some point, an AWG must be employed to transform the vectors into analog signals. Measurements are required in order to make any adjustments or corrections to the waveform algorithms or simulation. In today’s high-performance systems it may be difficult to determine if the measurement results reflect errors in the waveform mathematics or inadequacies in the generating equipment.
In choosing an AWG, the first choice to make is about bandwidth. The AWG must have enough bandwidth to generate the entire spectrum of waveforms. Bandwidth is made up of two major considerations. One is the available sampling rate of the digital-to-analog converter (DAC) used in the AWG. It must be at least twice the maximum frequency you wish to generate. The other consideration is the analog circuitry that supports the DAC. The components that follow the DAC determine how much degradation occurs before the signal is output. Any amplifiers, transformers, and filters must be of high quality and must adequately reject unwanted DAC output from the signal output. The analog circuitry in the AWG should include a lowpass filter that passes the highest frequency of interest and rejects frequencies above it. This filter, commonly referred to as a reconstruction filter, removes DAC harmonic output that can appear in the spectrum as unwanted signal components. This phenomenon is known as aliasing.
NUMBER OF EFFECTIVE BITS
Once bandwidth decisions are made, the next criterion is number of effective bits, or amplitude resolution. The number of effective DAC bits will determine the granularity with which the AWG transforms the signal. As the DAC samples the instantaneous voltage of the signal at any discrete time, it must try to provide a voltage output as close to the exact mathematical value of the signal. The more available bits of the DAC, the closer it can approximate the true value with the correct voltage. Any differences will show up as distortion in the signal, that is, imperfections in the transformation between the math and the physical signal at the output. One of the ways this distortion is seen in the signal spectrum is as spurious signals. As a rule of thumb, each bit in the DAC provides approximately 6dB of voltage resolution. A perfect 10-bit DAC will provide approximately 60 dB of voltage range. Unfortunately, AWGs are not perfect. Degradation due to the analog circuitry may reduce the number of usable DAC bits from 10 to 8 or even 7. This can reduce the AWG’s spurious-free dynamic range (SFDR) to 40dB or even less. If a 15-bit DAC is available, then the DAC voltage range is increased to approximately 90dB. After some degradation in the analog output circuitry, the SFDR can be 65dB or even greater than 70dB. The distortion products are attenuated far enough to have minimum impact on measurements. At this point, there is assurance that the measured results reflect the waveform mathematics, and not inadequate equipment.
CREATING WAVEFORMS WITH MORE CARRIER FREQUENCIES
New radar and emerging communication designs frequently process multi-emitter waveforms to locate targets of interest and to enhance channel capacity, respectively. Modern AWGs employ dedicated synchronizing circuitry that enables several AWGs to have their outputs precisely time aligned with a common sample clock. Such a high degree of alignment allows radar engineers to simulate antenna wave fronts used in antenna beam-forming and geolocation applications.
The last major consideration is the ease with which the mathematics can be realized as a physical signal. This refers to the simplicity of the interface between the waveform creation environment and the AWG. If Matlab or Simulink by The Math Works is used, it is helpful to have a direct interface between Matlab and the AWG. This interface should be provided by the AWG manufacturer and should be included with the product. Finally, if LabView by National Instruments, VEE by Agilent Technologies, or Microsoft Visual C++ or Visual .NET is employed, a complete interface that can be incorporated in software
is essential. An instrument driver that conforms to the IVI-C driver definition can ensure that a program will compile and link correctly with the AWG interface.
SUMMARY
This article has demonstrated the need for a high performance arbitrary waveform generator that has both the bandwidth and the dynamic range simultaneously to help today’s engineers successfully realize their complex signal environments and bring their designs quickly to market. It also emphasizes the need for a scalable AWG architecture that enables multiple emitter systems with precise time alignment between channels. Finally, it stresses the importance of a complete set of easy-to-use programming interfaces that allow efficient connectivity to any signal creation environment in which an engineer might choose to develop complex waveforms.
About the author
Mike Flaherty is a senior member of the technical staff with the Advanced Products Operation of Agilent Technologies, Inc. in
Santa Rosa, California. He can be reached at mflaherty@agilent.com.
|
|
|