Spectrum Analyzer BASICS
There are many ways to look at the world, and many ways to look at electronic signals. Many fundamental instruments, such as voltmeters and oscilloscopes, use a
time-oriented approach. They provide information on a signal's characteristics at a single point in time or show how a signal's characteristics change during some time interval. The job of the spectrum analyzer is to offer a different way of looking at signals, particularly communications signals.
One of the tenants of Fourier analysis is that any time-varying signal can be represented as a series of sine waves of differing amplitude at a variety of frequencies.
Repetitive signals are composed of sine waves that are harmonics of the repetition
frequency. Other signals are composed of a potentially infinite variety of unrelated
frequencies. Either way, Fourier analysis allows signals to be represented by their
frequency-power spectrum.
There are several advantages to examining the spectrum of a signal rather than its
instantaneous value. One is the ability to separate and identify signals that have been
added together, such as a Wi-Fi broadcast and EMI from a nearby computer. Added signals produce a time-varying pattern that cannot be easily separated into the desired
and interference components. The spectrums, however, are distinctive and readily identified by examination in the frequency domain. Once the characteristics of the interference
are known, tracking down the source becomes easier.
With modern wireless communications systems, the ability to see the power spectrum
becomes essential. The complex modulation schemes these systems use involve frequency
hopping that is impossible to interpret when looking only at the time history of signal
strength. The system behavior becomes readily apparent, however, when looking at the
power spectrum.
The essential concept of a spectrum analyzer is to pass the signal of interest through a
set of parallel narrow-band filters. The outputs of those filters are a measure of the signal's
strength within the filter bandwidth. The narrower the filters, the higher frequency resolution the resulting power spectrum will have. In practice, however, there are severe challenges to creating a matched set of narrowband filters that cover the frequency ranges of
interest, especially at RF frequencies. A variety of digital and analog approaches have
arisen to achieve the same functionality without requiring an array of filters.
Analog Vs. Digital Spectrum Analyzers
Several different technologies for spectrum analysis have arisen to produce an equivalent effect to a set of filters, including analog, digital, and hybrid analyzers. Each has its own approach to generating the equivalent of an array of narrow-band filters with center frequencies spanning the range of interest. Analog spectrum analyzers, for instance,
use a swept-frequency local signal to down-convert the input and pass the result to a single narrow-band filter (See Figure 1). Sweeping the local signal across a range of frequencies causes the filter output to represent a similar sweep across the input signal's frequency spectrum. In effect, the analyzer functions as a filter with a fixed bandwidth at a
tunable center frequency. The bandwidth range of an analog analyzer is set by the range
that the local signal can generate, while the resolution is set by the narrowband filter. A
sophisticated unit can have several selectable filters and local oscillators to be able to offer
users a range of parameters.
A digital spectrum analyzer takes advantage of the ability to convert a sampled
time series into a frequency spectrum directly by using a Fourier Transform. The analyzer samples the incoming signal at a high rate, saving the data into a block. The
unit then performs a digital Fast Fourier Transform (FFT) operation on the data block
to generate the power spectrum. The frequency range of the digital analyzer is set by
the sampling rate, and the resolution depends on the number of sample points in the
block.
One of the advantages of the digital spectrum analyzer is that the FFT preserves
phase information about the signal's frequency components as well as the amplitudes. This phase information is essential to analyzing complex digital modulation
schemes that shift both frequency and phase in order to encode data. In order to distinguish analyzers that work with phase information from those that simply display
amplitudes as on a traditional spectrum analyzer, the industry has coined the term
vector signal analyzer.
A hybrid spectrum analyzer combines elements of both approaches. It uses the
same type of signal down-conversion as the analog analyzer in order to translate the
signal of interest to a lower frequency, then samples the translated signal and performs the FFT. This approach allows the analyzer to work at a higher frequency
range than pure digital sampling could achieve while preserving the benefits of digital analysis.
Real-Time Analysis
The latest innovation in spectrum analysis is the advent of the real-time spectrum analyzer. A conventional analog analyzer sweeps across the range of interest,
with the result that the power spectrum of one resolution band is measured at a different time than another band. Similarly, a
conventional digital spectrum analyzer collects data for a time, then performs the
transform and displays the result before collecting data again.
Both conventional approaches work well
when the signals of interest are relatively stable, such as traditional television and radio
broadcast. As communications bandwidth
demands have increased, however, the situation has become more complex. Not only are complex phase and frequency modulation
schemes in use, but adaptive frequency-hopping has appeared. In addition, many communications methods share an unregulated frequency band and contend with one another.
The highly dynamic nature of this environment has created a need for spectrum analyzers to be able to watch the entire frequency
band of interest on a continual basis (See Figure 2). With the swept signal or intermittent
sampling of conventional analyzers, there is a
high probability that the unit will fail to capture
a transient signal. Yet such transients can be
the cause of communications system failures.
In order to reliably capture these transients, Tektronix developed its RTSA real-time
spectrum analyzer product family. This family uses a bank of digital processing engines
to ensure continuous signal sampling and
display of the input signals. In addition, the
instruments are able to augment the traditional two-dimensional display of power-versus-frequency with a color-coding that provides
a third "dimension" to show the temporal
variations in the spectrum. The result is that
stable signals form a background against
which transient signals will stand out (see
Figure 3). Such innovations are another step
in the continual capability expansion in
analysis tools for communications.
Conclusion
As communications have shifted from
simple modulation schemes on dedicated channels to complex, adaptive modulation on unregulated channels, spectrum analyzers have also evolved.
Moving from swept analog to phase-preserving digital and real-time methods, spectrum analyzers have remained
a vital instrument for understanding
communications system behavior.