Spectrum-Analyzer Basics; Spectrum-Analyzer Specifi Cations; Frequency Measurements - Hameg HM5530 Handbuch

S p e c t r u m a n a l y z e r b a s i c s

Spectrum-Analyzer basics

Introduction to spectrum analysis, advantages of
spectrum analyzers.
The analysis of eletrical signals is a fundamental task for many
engineers and scientists. Even if the parameters to be measured
are basically non-electrical, in many cases they are converted
to electrical signals. Such transducers are available for me-
chanical parameters like pressure or acceleration as well as
for chemical and biological ones. The conversion allows to use
the many electrical and electronic measuring instruments for
analysis in the time and frequency domains.
Traditionally, electrical signals are observed and measured
in the amplitude – time – domain, e.g. with an oscilloscope in
the Y/t mode. This yields information about waveforms, ampli-
tudes and time relationships. However, not all signals can be
adequately characterized that way. An oscilloscope displays
the waveform, but not the individual components of which this
is composed. So to speak the oscilloscope shows the sum of
the components, but it can not measure the frequencies and
amplitudes of them.
A spectrum analyzer displays the amplitudes of the spectral
components of a signal with respect to frequency (Y/f). The
signal resp. its components must repeat periodically. There
are oscilloscopes which calculate and display a mathemati-
cally derived Fourier spectrum, but even with this feature an
oscilloscope will not become a spectrum analyzer by far! There
remain fundamental differences, although such oscilloscope
Fourier spectra may suffi ce for many applications. In general,
one needs both types of instruments.
1. The sensitivity of spectrum analyzers is several orders of
magnitude higher than that of any oscilloscope. This fact,
also in conjunction with the following item, allows the ana-
lysis of signals which can not be displayed on a scope.
2. The dynamic range of spectrum analyzers is several orders
of magnitude larger than that of any oscilloscope.
3. Spectrum analyzers excel also and especially in the analysis
of distortions of sine waves, the detection of weak amplitude
or frequency modulation of signals, in measurements of
AM, FM such as carrier frequency, modulation frequency,
modulation depth etc. Also frequency converters can be
characterized with respect to losses and distortions.
4. An oscilloscope amplifi es the whole signal in a wide-band
amplifi er up to its crt (in analog scopes) or up to the a/d
converter (in DSO's). Large signal components or inter-
ference dictate the setting of the input attenuator i.e. the
sensitivity, consequently weak signals or components can
not be seen any more. Increasing the sensitivity in order to
detect small signal components is not possible, because
this would cause overdrive and hence distortions. (There
is an exception: a true difference amplifi er with offset is
able to give a microscopic display of small signal waveform
portions, but not of spectral components.)
A spectrum analyzer is a high performance narrow bandpass
tunable receiver with high quality input preselection fi lters and
multiple superheterodyning with its known advantages. It is
able to detect and measure very small signal components even
in the presence of very much larger amplitudes nearby.
36
Subject to change without notice
5. A spectum analyzer can display simultaneously a wide fre-
quency band and also a 80 dB (HM 5530) amplitude range
due to its logarithmic scaling. This is an enormous advantage
in many important applications such as emi measurements,
because the results of circuit modifi cations will be evident
immediately over a wide frequency range. In emi work there
is the socalled „water bed effect" which means that a certain
measure to suppress a portion of the frequency spectrum
may cause an increase of amplitudes in another portion with
the net result of no improvement at all.
Spectrum analyzers operate according to two predominant
principles: tuned or real time analyzers. Real time analyzers
conforming to the principles of the discrete Fourier transform
consist of the parallel connection of a multitude of frequency
selective indicators. Only that many discrete frequencies can
be detected and measured as there are fi lters. Depending on
the number and quality of such fi lters, the increase in cost sets
limits to their practical application.
Almost all modern spectrum analyzers use the superheterody-
ne principle. One method is the use of a bandpass fi lter which
can be tuned over the interesting frequency range. A detector
generates the Y signal while a sweep generator tunes the fi lter
synchronously with the X defl ection. This simple principle is
low cost, but suffers from serious drawbacks with respect to
selectivity and sensitivity, one reason is the change of bandwidth
with tuning.
Practical spectrum analyzers function quite like a high perfor-
mance radio receiver and use one or several bandpass fi lters
with fi xed center frequencies. The disadvantages of tunable
bandpass fi lters are avoided by frequency conversion of the input
signal to a fi xed if. The if fi lter(s) allow such input frequencies to
pass which conform to the equation: fi nput = f LO +- f if.
Circuit design and layout of the input stage determine to a
large extent the frequency range as well as the sensitivity of a
spectrum analyzer. The hf input stage consists of the attenuator,
the input fi lter, and the 1st local oscillator.
Spectrum-Analyzer specifi cations
The many applications of spectrum analyzers require a variety
of properties which may partly exclude each other or which
can only be combined with great effort. The main application
areas are those where the accuracy, the resolution in time resp.
frequency and the low dynamic range of oscilloscopes limit the
analysis of signals.
A wide frequency tuning range, requirements on the fi lters from
extremely narrow to „full span" as well as high sensitivity need
not exclude each other; but their combination with also high
resolution, high stability, fl at frequency response, low distor-
tions mostly requires indeed high effort and cost.
Frequency measurement
Spectrum analyzers allow the measurement of frequencies
in SPAN (frequency sweep) mode as well as in the Zero Span
(SF = 0) mode. In SPAN mode, the whole frequency range of