Keysight Technologies

Spectrum Analysis Basics

Application Note
Keysight Technologies. Inc. dedicates this application note to Blake Peterson.

Blake’s outstanding service in technical support reached customers in all corners of

the world during and after his 45-year career with Hewlett-Packard and Keysight.
For many years, Blake trained new marketing and sales engineers in the “ABCs” of
spectrum analyzer technology, which provided the basis for understanding more
advanced technology. He is warmly regarded as a mentor and technical contributor
in spectrum analysis.

Blake’s many accomplishments include:

–– Authored the original edition of the Spectrum Analysis Basics application note
and contributed to subsequent editions
–– Helped launch the 8566/68 spectrum analyzers, marking the beginning of
modern spectrum analysis, and the PSA Series spectrum analyzers that set
new performance benchmarks in the industry when they were introduced
–– Inspired the creation of Blake Peterson University––required training for all
engineering hires at Keysight

As a testament to his accomplishments and contributions, Blake was honored with

Microwaves & RF magazine’s first Living Legend Award in 2013.

2
Table of Contents

Chapter 1 – Introduction................................................................................................................. 5
Frequency domain versus time domain ....................................................................................... 5
What is a spectrum? ....................................................................................................................... 6
Why measure spectra? ................................................................................................................... 6
Types of signal analyzers ................................................................................................................ 8

Chapter 2 – Spectrum Analyzer Fundamentals ........................................................................... 9

RF attenuator..................................................................................................................................10
Low-pass filter or preselector.......................................................................................................10
Tuning the analyzer .......................................................................................................................11
IF gain..............................................................................................................................................12
Resolving signals............................................................................................................................13
Residual FM....................................................................................................................................15
Phase noise.....................................................................................................................................16
Sweep time.....................................................................................................................................18
Envelope detector .........................................................................................................................20
Displays...........................................................................................................................................21
Detector types................................................................................................................................22
Sample detection...........................................................................................................................23
Peak (positive) detection...............................................................................................................24
Negative peak detection...............................................................................................................24
Normal detection...........................................................................................................................24
Average detection..........................................................................................................................27
EMI detectors: average and quasi-peak detection....................................................................27
Averaging processes......................................................................................................................28
Time gating.....................................................................................................................................31

Chapter 3 – Digital IF Overview ...................................................................................................36

Digital filters....................................................................................................................................36
All-digital IF....................................................................................................................................37
Custom digital signal processing.................................................................................................38
Additional video processing features ..........................................................................................38
Frequency counting ......................................................................................................................38
More advantages of all-digital IF..................................................................................................39

Chapter 4 – Amplitude and Frequency Accuracy.......................................................................40

Relative uncertainty ......................................................................................................................42
Absolute amplitude accuracy.......................................................................................................42
Improving overall uncertainty.......................................................................................................43
Specifications, typical performance and nominal values..........................................................43
Digital IF architecture and uncertainties.....................................................................................43
Amplitude uncertainty examples..................................................................................................44
Frequency accuracy.......................................................................................................................44

3
Table of Contents
continued
Chapter 5 – Sensitivity and Noise................................................................................................46
Sensitivity........................................................................................................................................46
Noise floor extension.....................................................................................................................48
Noise figure.....................................................................................................................................49
Preamplifiers...................................................................................................................................50
Noise as a signal............................................................................................................................53
Preamplifier for noise measurements..........................................................................................54

Chapter 6 – Dynamic Range.........................................................................................................55

Dynamic range versus internal distortion ...................................................................................55
Attenuator test................................................................................................................................56
Noise................................................................................................................................................57
Dynamic range versus measurement uncertainty......................................................................58
Gain compression..........................................................................................................................60
Display range and measurement range.......................................................................................60
Adjacent channel power measurements.....................................................................................61

Chapter 7 – Extending the Frequency Range.............................................................................62

Internal harmonic mixing..............................................................................................................62
Preselection....................................................................................................................................66
Amplitude calibration....................................................................................................................68
Phase noise ....................................................................................................................................68
Improved dynamic range..............................................................................................................69
Pluses and minuses of preselection............................................................................................70
External harmonic mixing..............................................................................................................71
Signal identification.......................................................................................................................73

Chapter 8 – Modern Signal Analyzers.........................................................................................76

Application-specific measurements............................................................................................76
The need for phase information.................................................................................................. 77
Digital modulation analysis...........................................................................................................79
Real-time spectrum analysis........................................................................................................80

Chapter 9 – Control and Data Transfer........................................................................................81

Saving and printing data...............................................................................................................81
Data transfer and remote instrument control ............................................................................81
Firmware updates..........................................................................................................................82
Calibration, troubleshooting, diagnostics and repair.................................................................82

Summary.........................................................................................................................................82

Glossary of Terms...........................................................................................................................83

4
Chapter 1. Introduction
This application note explains the Fourier1 theory tells us any time- Some measurements require that we
fundamentals of swept-tuned, domain electrical phenomenon is preserve complete information about
superheterodyne spectrum analyzers made up of one or more sine waves the signal frequency, amplitude and
and discusses the latest advances in of appropriate frequency, amplitude, phase. However, another large group
spectrum analyzer capabilities. and phase. In other words, we can of measurements can be made without
transform a time-domain signal into knowing the phase relationships among
At the most basic level, a spectrum its frequency-domain equivalent. the sinusoidal components. This type
analyzer can be described as a Measurements in the frequency domain of signal analysis is called spectrum
frequency-selective, peak-responding tell us how much energy is present analysis. Because spectrum analysis is
voltmeter calibrated to display the rms at each particular frequency. With simpler to understand, yet extremely
value of a sine wave. It is important to proper filtering, a waveform such as useful, we begin by looking first at how
understand that the spectrum analyzer the one shown in Figure 1-1 can be spectrum analyzers perform spectrum
is not a power meter, even though it can decomposed into separate sinusoidal analysis measurements, starting in
be used to display power directly. As waves, or spectral components, which Chapter 2.
long as we know some value of a sine we can then evaluate independently.
wave (for example, peak or average) and Each sine wave is characterized by Theoretically, to make the
know the resistance across which we its amplitude and phase. If the signal transformation from the time domain
measure this value, we can calibrate our we wish to analyze is periodic, as in to the frequency domain, the signal
voltmeter to indicate power. With the our case here, Fourier says that the must be evaluated over all time, that is,
advent of digital technology, modern constituent sine waves are separated in over infinity. However, in practice, we
spectrum analyzers have been given the frequency domain by 1/T, where T always use a finite time period when
many more capabilities. In this note, we is the period of the signal 2. making a measurement. You also can
describe the basic spectrum analyzer make Fourier transformations from
as well as additional capabilities made
possible using digital technology and
digital signal processing.

Frequency domain versus

time domain
Before we get into the details of
describing a spectrum analyzer, we
might first ask ourselves: “Just what is
a spectrum and why would we want
to analyze it?” Our normal frame
of reference is time. We note when
certain events occur. This includes
electrical events. We can use an
oscilloscope to view the instantaneous
value of a particular electrical event (or
some other event converted to volts
through an appropriate transducer)
as a function of time. In other words,
we use the oscilloscope to view
the waveform of a signal in the time
domain.

Figure 1-1. Complex time-domain signal

1.  Jean Baptiste Joseph Fourier, 1768-1830. A French mathematician and physicist who discovered that periodic functions can be expanded into a
series of sines and cosines.
2.  If the time signal occurs only once, then T is infinite, and the frequency representation is a continuum of sine waves.

5
the frequency to the time domain. This not give us a definitive indication of of a signal. People involved in wireless
case also theoretically requires the the reason why. Figure 1-2 shows our communications are extremely
evaluation of all spectral components complex signal in both the time and interested in out-of-band and spurious
over frequencies to infinity. In reality, frequency domains. The frequency- emissions. For example, cellular
making measurements in a finite domain display plots the amplitude radio systems must be checked for
bandwidth that captures most of the versus the frequency of each sine harmonics of the carrier signal that
signal energy produces acceptable wave in the spectrum. As shown, the might interfere with other systems
results. When you perform a Fourier spectrum in this case comprises just operating at the same frequencies
transformation on frequency domain two sine waves. We now know why our as the harmonics. Engineers and
data, the phase of the individual original waveform was not a pure sine technicians are also very concerned
components is indeed critical. For wave. It contained a second sine wave, about distortion of the message
example, a square wave transformed to the second harmonic in this case. Does modulated onto a carrier.
the frequency domain and back again this mean we have no need to perform
could turn into a sawtooth wave if you time-domain measurements? Not at Third-order intermodulation (two tones
do not preserve phase. all. The time domain is better for many of a complex signal modulating each
measurements, and some can be made other) can be particularly troublesome
What is a spectrum? only in the time domain. For example, because the distortion components can
pure time-domain measurements fall within the band of interest, which
So what is a spectrum in the context include pulse rise and fall times, means they cannot be filtered away.
of this discussion? A spectrum is a overshoot and ringing.
collection of sine waves that, when Spectrum monitoring is another
combined properly, produce the time- important frequency-domain
domain signal under examination.
Why measure spectra?
measurement activity. Government
Figure 1-1 shows the waveform of a The frequency domain also has its regulatory agencies allocate different
complex signal. Suppose that we were measurement strengths. We have frequencies for various radio services,
hoping to see a sine wave. Although already seen in Figures 1-1 and 1-2 such as broadcast television and
the waveform certainly shows us that that the frequency domain is better radio, mobile phone systems, police
the signal is not a pure sinusoid, it does for determining the harmonic content and emergency communications,
and a host of other applications. It
is critical that each of these services
operates at the assigned frequency
and stays within the allocated channel
bandwidth. Transmitters and other
intentional radiators often must
operate at closely spaced adjacent
frequencies. A key performance
measure for the power amplifiers
and other components used in these
systems is the amount of signal energy
that spills over into adjacent channels
and causes interference.

Time domain Frequency domain Electromagnetic interference (EMI) is

measurements measurements a term applied to unwanted emissions
from both intentional and unintentional
Figure 1-2. Relationship between time and frequency domain radiators. These unwanted emissions,
either radiated or conducted
(through the power lines or other
interconnecting wires), might impair
the operation of other systems. Almost
anyone designing or manufacturing
electrical or electronic products
must test for emission levels versus
frequency according to regulations
Figure 1-2.
set by various government agencies or
industry-standard bodies.

6
Figure 1-3. Harmonic distortion test of a transmitter Figure 1-4. GSM radio signal and spectral mask showing limits of
unwanted emissions

Figure 1- 5. Two-tone test on an RF power amplifier Figure 1-6. Radiated emissions plotted against CISPR11 limits as part of an
EMI test

7
Noise is often the signal you want to grade bench-top spectrum analyzers
measure. Any active circuit or device within 10ths of a dB. More information
will generate excess noise. Tests such
For additional information on
as noise figure and signal-to-noise ratio In this application note, we concentrate
vector measurements, see Vector
(SNR) are important for characterizing on swept amplitude measurements,
Signal Analysis Basics–Application Note,
the performance of a device and only briefly touching on measurements
literature number 5989-1121EN. For
its contribution to overall system involving phase–see Chapter 8.
information on FFT analyzers that
performance.
tune to 0 Hz, see the Web page for
Note: When computers became
the Keysight 35670A at
Figures 1-3 through 1-6 show some of Hewlett-Packard’s dominant business,
www.keysight.com/find/35670A.
these measurements on an X-Series it created and spun off Keysight
signal analyzer. Technologies in the late 1990’s to
continue the test and measurement
Types of signal analyzers business. Many older spectrum
analyzers carry the Hewlett-Packard
The first swept-tuned superheterodyne name but are supported by Keysight.
analyzers measured only amplitude.
However, as technology advanced This application note will give you
and communication systems grew insight into your particular spectrum or
more complex, phase became a more signal analyzer and help you use this
important part of the measurement. versatile instrument to its maximum
Spectrum analyzers, now often labeled potential.
signal analyzers, have kept pace. By
digitizing the signal, after one or more
stages of frequency conversion, phase
as well as amplitude is preserved
and can be included as part of the
information displayed. So today’s signal
analyzers such as the Keysight X-Series
combine the attributes of analog,
vector and FFT (fast Fourier transform)
analyzers. To further improve
capabilities, Keysight’s X-Series signal
analyzers incorporate a computer,
complete with a removable disk drive
that allows sensitive data to remain in a
controlled area should the analyzer be
removed.

Advanced technology also has allowed

circuits to be miniaturized. As a result,
rugged portable spectrum analyzers
such as the Keysight FieldFox simplify
tasks such as characterizing sites
for transmitters or antenna farms.
Zero warm-up time eliminates delays
in situations involving brief stops
for quick measurements. Due to
advanced calibration techniques,
field measurements made with these
handheld analyzers correlate with lab-

8
Chapter 2. Spectrum Analyzer Fundamentals
This chapter focuses on the it mixes with a signal from the local oscillator is tuned electronically rather
fundamental theory of how a spectrum oscillator (LO). Because the mixer is a than by a front-panel knob.
analyzer works. While today’s non-linear device, its output includes
technology makes it possible to replace not only the two original signals, but The output of a spectrum analyzer is
many analog circuits with modern also their harmonics and the sums and an X-Y trace on a display, so let’s see
digital implementations, it is useful to differences of the original frequencies what information we get from it. The
understand classic spectrum analyzer and their harmonics. If any of the mixed display is mapped on a grid (graticule)
architecture as a starting point in our signals falls within the pass band of with 10 major horizontal divisions and
discussion. the intermediate-frequency (IF) filter, generally 10 major vertical divisions.
it is further processed (amplified and The horizontal axis is linearly calibrated
In later chapters, we will look at the perhaps compressed on a logarithmic in frequency that increases from left
capabilities and advantages that scale). It is essentially rectified by the to right. Setting the frequency is a
digital circuitry brings to spectrum envelope detector, filtered through two-step process. First we adjust
analysis. Chapter 3 discusses digital the low-pass filter and displayed. A the frequency at the centerline of the
architectures used in spectrum ramp generator creates the horizontal graticule with the center frequency
analyzers available today. movement across the display from left control. Then we adjust the frequency
to right. The ramp also tunes the LO so range (span) across the full 10 divisions
Figure 2-1 is a simplified block diagram its frequency change is in proportion to with the frequency span control. These
of a superheterodyne spectrum the ramp voltage. controls are independent, so if we
analyzer. Heterodyne means to mix; change the center frequency, we do not
that is, to translate frequency. And If you are familiar with superheterodyne alter the frequency span. Alternatively,
super refers to superaudio frequencies, AM radios, the type that receive we can set the start and stop
or frequencies above the audio range. ordinary AM broadcast signals, you will frequencies instead of setting center
In the Figure 2-1 block diagram, note a strong similarity between them frequency and span. In either case, we
we see that an input signal passes and the block diagram shown in Figure can determine the absolute frequency
through an attenuator, then through 2-1. The differences are that the output of any signal displayed and the relative
a low-pass filter (later we will see why of a spectrum analyzer is a display frequency difference between any two
the filter is here) to a mixer, where instead of a speaker, and the local signals.

RF input
attenuator Log Envelope
Mixer IF gain IF filter amp detector

Input
signal

Pre-selector, or Video
low-pass filter filter
Local
oscillator

Reference
oscillator

Sweep
generator Display

Figure 2-1. Block diagram of a classic superheterodyne spectrum analyzer

9
The vertical axis is calibrated being damaged by a DC signal or a Low-pass filter or preselector
in amplitude. You can choose a DC offset of the signal being viewed.
linear scale calibrated in volts or a Unfortunately, it also attenuates low- The low-pass filter blocks high-
logarithmic scale calibrated in dB. The frequency signals and increases the frequency signals from reaching the
log scale is used far more often than minimum useable start frequency of mixer. This filtering prevents out-of-
the linear scale because it has a much the analyzer to 9 kHz, 100 kHz or 10 band signals from mixing with the
wider usable range. The log scale MHz, depending on the analyzer. local oscillator and creating unwanted
allows signals as far apart in amplitude responses on the display. Microwave
as 70 to 100 dB (voltage ratios of In some analyzers, an amplitude spectrum analyzers replace the
3200 to 100,000 and power ratios reference signal can be connected low-pass filter with a preselector,
of 10,000,000 to 10,000,000,000) to as shown in Figure 2-3. It provides which is a tunable filter that rejects all
be displayed simultaneously. On the a precise frequency and amplitude frequencies except those we currently
other hand, the linear scale is usable signal, used by the analyzer to wish to view. In Chapter 7, we go into
for signals differing by no more than periodically self-calibrate. more detail about the operation and
20 to 30 dB (voltage ratios of 10 to purpose of the preselector.
32). In either case, we give the top line
of the graticule, the reference level,
an absolute value through calibration
techniques1 and use the scaling per
division to assign values to other
locations on the graticule. Therefore,
we can measure either the absolute
value of a signal or the relative
amplitude difference between any two
signals.

Scale calibration, both frequency and

amplitude, is shown by annotations
written onto the display. Figure 2-2
shows the display of a typical analyzer.

Now, let’s turn our attention back to

the spectrum analyzer components
diagramed in Figure 2-1.

RF attenuator
The first part of our analyzer is the
RF input attenuator. Its purpose is to
ensure the signal enters the mixer at
the optimum level to prevent overload,
gain compression and distortion.
Because attenuation is a protective Figure 2-2. Typical spectrum analyzer display with control settings
circuit for the analyzer, it is usually set
automatically, based on the reference
level. However, manual selection of
0 to 70 dB, 2 dB steps
attenuation is also available in steps
RF input
of 10, 5, 2, or even 1 dB. The diagram
in Figure 2-3 is an example of an
attenuator circuit with a maximum
attenuation of 70 dB in increments Amplitude
of 2 dB. The blocking capacitor is reference
signal
used to prevent the analyzer from

Figure 2-3. RF input attenuator circuitry

figure 2-3

1.  See Chapter 4, “Amplitude and Frequency Accuracy.”

10
Tuning the analyzer tune from 0 Hz to 3.6 GHz (actually Figure 2-4 illustrates analyzer tuning.
from some low frequency because we In this figure, f LO is not quite high
We need to know how to tune our cannot view a 0-Hz signal with this enough to cause the f LO – fsig mixing
spectrum analyzer to the desired architecture). product to fall in the IF pass band, so
frequency range. Tuning is a function there is no response on the display. If
of the center frequency of the IF filter, If we start the LO at the IF (LO minus IF we adjust the ramp generator to tune
the frequency range of the LO and = 0 Hz) and tune it upward from there the LO higher, however, this mixing
the range of frequencies allowed to to 3.6 GHz above the IF, we can cover product will fall in the IF pass band at
reach the mixer from the outside world the tuning range with the LO minus IF some point on the ramp (sweep), and
(allowed to pass through the low- mixing product. Using this information, we will see a response on the display.
pass filter). Of all the mixing products we can generate a tuning equation:
emerging from the mixer, the two with The ramp generator controls both the
the greatest amplitudes, and therefore fsig = f LO - f IF horizontal position of the trace on the
the most desirable, are those created display and the LO frequency, so we
from the sum of the LO and input signal where f sig = signal frequency can now calibrate the horizontal axis of
and from the difference between the f LO = local oscillator frequency, the display in terms of the input signal
LO and input signal. If we can arrange and frequency.
things so that the signal we wish to f IF = intermediate frequency (IF)
examine is either above or below the We are not quite through with the
LO frequency by the IF, then only one If we wanted to determine the LO tuning yet. What happens if the
of the desired mixing products will fall frequency needed to tune the analyzer frequency of the input signal is 9.0
within the pass-band of the IF filter and to a low-, mid-, or high-frequency GHz? As the LO tunes through its
be detected to create an amplitude signal (say, 1 kHz, 1.5  GHz, or 3 3.8- to 8.7-GHz range, it reaches a
response on the display. GHz), we would first restate the tuning frequency (3.9 GHz) at which it is the
equation in terms of f LO: IF away from the 9.0-GHz input signal.
We need to pick an LO frequency and At this frequency we have a mixing
an IF that will create an analyzer with f LO = fsig + f IF product that is equal to the IF, creating
the desired tuning range. Let’s assume a response on the display. In other
that we want a tuning range from 0 to Then we would apply the numbers words, the tuning equation could just
3.6 GHz. We then need to choose the for the signal and IF in the tuning as easily have been:
IF. Let’s try a 1-GHz IF. Because this equation2:
frequency is within our desired tuning f sig = f LO + f IF
range, we could have an input signal f LO = 1 kHz + 5.1 GHz = 5.100001 GHz
at 1 GHz. The output of a mixer also This equation says that the architecture
includes the original input signals, so f LO = 1.5 GHz + 5.1 GHz = 6.6 GHz or
of Figure 2-1 could also result in a
an input signal at 1 GHz would give f LO = 3 GHz + 5.1 GHz = 8.1 GHz.
us a constant output from the mixer
at the IF. The 1-GHz signal would thus
Freq range
pass through the system and give us A of analyzer IF
a constant amplitude response on
the display regardless of the tuning
of the LO. The result would be a hole
in the frequency range at which we
could not properly examine signals fsig f f
fLO
because the amplitude response would fLO – f sig fLO + f sig
be independent of the LO frequency.
Therefore, a 1-GHz IF will not work. Freq range
of analyzer Freq range of LO
A
Instead, we choose an IF that is above
the highest frequency to which we wish
to tune. In the Keysight X-Series signal
analyzers that can tune to 3.6 GHz,
the first LO frequency range is 3.8 to fLO f
8.7 GHz, and the IF chosen is about
5.1 GHz. Remember that we want to
Figure 2-4. The LO must be tuned to f IF + f sig to produce a response on the display

2.  In the text, we round off some of the frequency values for simplicity, although the exact values are shown in the figures.

11
tuning range from 8.9 to 13.8 GHz, based on the architecture of a typical response on the display, as if it were an
but only if we allow signals in that spectrum analyzer. The full tuning input signal at 0 Hz. This response, the
range to reach the mixer. The job of equation for this analyzer is: LO feedthrough, can mask very low-
the input low-pass filter in Figure 2-1 frequency signals, so not all analyzers
is to prevent these higher frequencies fsig = f LO1 – (f LO2 + f LO3 + f final IF) allow the display range to include 0 Hz.
from getting to the mixer. We also want
to keep signals at the intermediate However, IF gain
frequency itself from reaching the
mixer, as previously described, so the f LO2 + f LO3 + f final IF Referring back to Figure 2-1, we see
low-pass filter must do a good job of the next component of the block
= 4.8 GHz + 300 MHz + 22.5 MHz
attenuating signals at 5.1 GHz as well diagram is a variable gain amplifier.
= 5.1225 GHz, the first IF. It is used to adjust the vertical
as in the range from 8.9 to 13.8 GHz.
position of signals on the display
Simplifying the tuning equation by without affecting the signal level at
In summary, we can say that for a
using just the first IF leads us to the the input mixer. When the IF gain is
single-band RF spectrum analyzer, we
same answers. Although only passive changed, the value of the reference
would choose an IF above the highest
filters level is changed accordingly to retain
frequency of the tuning range. We
are shown in figure 2-5, the actual the correct indicated value for the
would make the LO tunable from the
implementation includes amplification displayed signals. Generally, we do not
IF to the IF plus the upper limit of the
in the narrower IF stages. The want the reference level to change
tuning range and include a low-pass
final IF section contains additional when we change the input attenuator,
filter in front of the mixer that cuts off
components, such as logarithmic so the settings of the input attenuator
below the IF.
amplifiers or analog -to-digital and the IF gain are coupled together.
converters, depending on the design of
To separate closely spaced signals
the particular analyzer. A change in input attenuation will
(see “Resolving signals” later in this
chapter), some spectrum analyzers automatically change the IF gain to
Most RF spectrum analyzers allow offset the effect of the change in input
have IF bandwidths as narrow as 1 kHz;
an LO frequency as low as, and even attenuation, thereby keeping the signal
others, 10 Hz; still others, 1 Hz. Such
below, the first IF. Because there is at a constant position on the display.
narrow filters are difficult to achieve
finite isolation between the LO and IF
at a center frequency of 5.1 GHz, so
ports of the mixer, the LO appears at
we must add additional mixing stages,
the mixer output. When the LO equals
typically two to four stages, to down-
the IF, the LO signal itself is processed
convert from the first to the final IF.
by the system and appears as a
Figure 2-5 shows a possible IF chain

3.6 GHz 5.1225 GHz 322.5 MHz 22.5 MHz Envelope

detector

3.8 to 8.7 GHz

4.8 GHz 300 MHz

Sweep
generator
Display

Figure 2-5. Most spectrum analyzers use two to four mixing steps to reach the final IF.

12
Resolving signals have a definite width on the display. Two signals must be far enough apart
The output of a mixer includes the or the traces they make will fall on
After the IF gain amplifier, we find the sum and difference products plus top of each other and look like only
IF section, which consists of the analog the two original signals (input and one response. Fortunately, spectrum
or digital resolution bandwidth (RBW) LO). A bandpass filter determines the analyzers have selectable resolution
filters, or both. intermediate frequency, and this filter (IF) filters, so it is usually possible to
selects the desired mixing product select one narrow enough to resolve
Analog filters and rejects all other signals. Because closely spaced signals.
the input signal is fixed and the local
Frequency resolution is the ability of
oscillator is swept, the products from Keysight data sheets describe the
a spectrum analyzer to separate two
the mixer are also swept. If a mixing ability to resolve signals by listing the
input sinusoids into distinct responses.
product happens to sweep past the 3-dB bandwidths of the available IF
Fourier tells us that a sine-wave signal
IF, the characteristic shape of the filters. This number tells us how close
only has energy at one frequency, so
bandpass filter is traced on the display. together equal-amplitude sinusoids can
we should not have any resolution
See Figure 2-6. The narrowest filter be and still be resolved. In this case,
problems. Two signals, no matter how
in the chain determines the overall there will be about a 3-dB dip between
close in frequency, should appear
displayed bandwidth, and in the the two peaks traced out by these
as two lines on the display. But a
architecture of Figure 2-5, this filter is signals. See Figure 2-7. The signals
closer look at our superheterodyne
in the 22.5- MHz IF. can be closer together before their
receiver shows why signal responses

Figure 2-6. As a mixing product sweeps past the IF filter, the filter shape is traced on the display

3.  If you experiment with resolu-

tion on a spectrum analyzer using
the normal (rosenfell) detector
mode (See “Detector types” later
in this chapter) use enough video
filtering to create a smooth trace.
Otherwise, you will see smearing as
the two signals interact. While the
smeared trace certainly indicates
the presence of more than one
signal, it is difficult to determine the
amplitudes of the individual signals.
Spectrum analyzers with positive
peak as their default detector mode
may not show the smearing effect.
You can observe the smearing
by selecting the sample detector
mode.
Figure 2-7. Two equal-amplitude sinusoids separated by the 3-dB BW of the selected IF filter
can be resolved.

13
traces merge completely, but the 3-dB
bandwidth is a good rule of thumb for
resolution of equal-amplitude signals 3.

More often than not, we are dealing

with sinusoids that are not equal in
amplitude. The smaller sinusoid can
actually be lost under the skirt of the
response traced out by the larger. This
effect is illustrated in Figure 2-8. The
top trace looks like a single signal, but
in fact represents two signals: one
at 300 MHz (0 dBm) and another at
300.005 MHz (–30 dBm). The lower
trace shows the display after the 300-
MHz signal is removed.

Another specification is listed for the

resolution filters: bandwidth selectivity
(or selectivity or shape factor).
Bandwidth selectivity helps determine
Figure 2-8. A low-level signal can be lost under the skirt of the response to a larger signal
the resolving power for unequal
sinusoids. For Keysight analyzers,
bandwidth selectivity is generally
specified as the ratio of the 60-dB
bandwidth to the 3-dB bandwidth, as
shown in Figure 2-9. The analog filters
in Keysight analyzers are a four-pole,
synchronously tuned design, with a
nearly Gaussian shape4. This type of
filter exhibits a bandwidth selectivity of
about 12.7:1.

For example, what resolution

bandwidth must we choose to resolve

Figure 2-9. Bandwidth selectivity, ratio of 60-dB to 3-dB bandwidths

4.  Some older spectrum analyzer models used five-pole filters for the narrowest resolution bandwidths to provide improved selectivity of about 10:1.
Modern designs achieve even better bandwidth selectivity using digital IF filters.

14
signals that differ by 4 kHz and 30 dB,
assuming 12.7:1 bandwidth selectivity?
Because we are concerned with
rejection of the larger signal when
the analyzer is tuned to the smaller
signal, we need to consider not the
full bandwidth, but the frequency
difference from the filter center
frequency to the skirt. To determine
how far down the filter skirt is at a
given offset, we use the following
equation:

H(∆f) = –10(N) log10 [(∆f/f 0)2 + 1]

Where
H(∆f) is the filter skirt rejection in dB,
N is the number of filter poles,
∆f is the frequency offset from the
center in Hz, and

RBW Figure 2-10. The 3-kHz filter (top trace) does not resolve the smaller signal; reducing the resolution band-
f 0 is given by width to 1 kHz (bottom trace) does
2 √ 21/N –1

For our example, N=4 and ∆f = 4000.

Let’s begin by trying the 3-kHz RBW Digital filters transferred to any mixing product
filter. involving the LO, there was no point
Some spectrum analyzers use digital in having resolution bandwidths
First, we compute f 0:
techniques to realize their resolution narrower than 1 kHz because it was
3000 bandwidth filters. Digital filters can
f 0 = 2 √ 2¼ –1 = 3448.44 impossible to determine the cause of
provide important benefits, such as any instability on the display.
dramatically improved bandwidth
Now we can determine the filter
selectivity. The Keysight PSA and However, modern analyzers have
rejection at a 4-kHz offset:
X-Series signal analyzers implement all dramatically improved residual FM. For
resolution bandwidths digitally. Other example, residual FM in Keysight PXA
H(4000) = –10(4) log10 [(4000/3448.44)2 + 1]
analyzers, such as the Keysight ESA-E Series analyzers is nominally 0.25 Hz;
= −14.8 dB Series, take a hybrid approach, using in PSA Series analyzers, 1 to 4 Hz; and
analog filters for the wider bandwidths in ESA Series analyzers, 2 to 8 Hz. This
This is not enough to allow us to see and digital filters for bandwidths of 300 allows bandwidths as low as 1 Hz in
the smaller signal. Let’s determine Hz and below. Refer to Chapter 3 for many analyzers, and any instability we
H(∆f) again using a 1-kHz filter: more information on digital filters. see on a spectrum analyzer today is
1000 due to the incoming signal.
f0 = = 1149.48
2 √ 2¼ –1 Residual FM
This allows us to calculate the filter The instability and residual FM of the
rejection: LOs in an analyzer, particularly the
first LO, often determine the minimum
H(4000) = –10(4) log10[(4000/1149.48)2 + 1] usable resolution bandwidth. The
= −44.7 dB unstable YIG (yttrium iron garnet)
oscillator used in early analyzers
Thus, the 1-kHz resolution bandwidth typically had a residual FM of about
filter does resolve the smaller signal, as 1 kHz. Because this instability was
illustrated in Figure 2-10.

15
Phase noise
No oscillator is perfectly stable.
Even though we may not be able to
see the actual frequency jitter of a
spectrum analyzer LO system, there
is still a manifestation of the LO
frequency or phase instability that can
be observed. This is known as phase
noise (sometimes called sideband
noise).

All are frequency or phase modulated

by random noise to some extent. As
previously noted, any instability in
the LO is transferred to any mixing
products resulting from the LO and
input signals. So the LO phase noise
modulation sidebands appear around
any spectral component on the display
that is far enough above the broadband
noise floor of the system (Figure 2-11).
The amplitude difference between a Figure 2-11. Phase noise is displayed only when a signal is displayed far enough above the system noise floor
displayed spectral component and the
phase noise is a function of the stability Generally, we can see the inherent –– Optimize phase noise for
of the LO. The more stable the LO, the phase noise of a spectrum analyzer frequency offsets > 160 kHz from
lower the phase noise. The amplitude only in the narrower resolution filters, the carrier This mode optimizes
difference is also a function of the when it obscures the lower skirts of phase noise for offsets above 160
resolution bandwidth. If we reduce the these filters. The use of the digital kHz away from the carrier.
resolution bandwidth by a factor of 10, filters previously described does not
the level of the displayed phase noise change this effect. For wider filters, the –– Optimize LO for fast tuning
decreases by 10 dB5. phase noise is hidden under the filter When this mode is selected, LO
skirt, just as in the case of two unequal behavior compromises phase noise
The shape of the phase noise spectrum sinusoids discussed earlier. at all offsets from the carrier below
is a function of analyzer design, in
approximately 2 MHz. This mode
particular, the sophistication of the Today’s spectrum or signal analyzers, minimizes measurement time and
phase-lock loops employed to stabilize such as Keysight’s X-Series, allow allows the maximum measurement
the LO. In some analyzers, the phase you to select different LO stabilization throughput when changing the
noise is a relatively flat pedestal out modes to optimize the phase noise for center frequency or span.
to the bandwidth of the stabilizing different measurement conditions. For
loop. In others, the phase noise may example, the PXA signal analyzer offers
fall away as a function of frequency three different modes:
offset from the signal. Phase noise is
specified in terms of dBc (dB relative –– Optimize phase noise for
to a carrier) and normalized to a 1-Hz frequency offsets < 140 kHz from
noise power bandwidth. It is sometimes the carrier In this mode, the LO
specified at specific frequency offsets. phase noise is optimized for the
At other times, a curve is given to show area close in to the carrier at the
the phase noise characteristics over a expense of phase noise beyond
range of offsets. 140-kHz offset.

5.  The effect is the same for the broadband noise floor (or any broadband noise signal). See Chapter 5, “Sensitivity and Noise.”

16
Figure 2-12a. Phase noise performance can be optimized for different mea- Figure 2-12b. Detail of the 140-kHz carrier offset region
surement conditions

The PXA signal analyzers phase noise

optimization can also be set to auto
mode, which automatically sets the
instrument’s behavior to optimize
speed or dynamic range for various
operating conditions. When the span is
> 44.44 MHz or the RBW is > 1.9 MHz,
the PXA selects Fast Tuning mode.
Otherwise, the PXA automatically
chooses Best Close-In Phase Noise
when center frequency < 195 kHz, or
when center frequency ≥ 1 MHz and
span ≤ 1.3 MHz and RBW ≤ 75 kHz. If
these conditions are not met, the PXA
automatically chooses Best Wide-
Offset Phase Noise. Figure 2-13. Phase noise can prevent resolution of unequal signals

In any case, phase noise becomes the

ultimate limitation in an analyzer’s
ability to resolve signals of unequal
amplitude. As shown in Figure 2-13,
we may have determined that we can
resolve two signals based on the 3-dB
bandwidth and selectivity, only to find
that the phase noise covers up the
smaller signal.

17
Sweep time
Analog resolution filters
If resolution were the only criterion on
which we judged a spectrum analyzer,
we might design our analyzer with the
narrowest possible resolution (IF) filter
and let it go at that. But resolution
affects sweep time, and we care very
much about sweep time. Sweep time
directly affects how long it takes to
complete a measurement.

Resolution comes into play because

the IF filters are band-limited circuits
that require finite times to charge and
discharge. If the mixing products are
swept through them too quickly, there
will be a loss of displayed amplitude, as
shown in Figure 2-14. (See “Envelope
Figure 2-14. Sweeping an analyzer too fast causes a drop in displayed amplitude and a shift in indicated
detector,” later in this chapter, for
frequency
another approach to IF response time.)
If we think about how long a mixing
product stays in the pass band of the IF
filter, that time is directly proportional to
bandwidth and inversely proportional to The important message here is that
the sweep in Hz per unit time, or: a change in resolution has a dramatic
effect on sweep time. Older analog
Time in pass band = analyzers typically provided values in a
1, 3, 10 sequence or in ratios roughly
RBW
= (RBW)(ST) equaling the square root of 10. So
Span/ST Span sweep time was affected by a factor of
Where about 10 with each step in resolution.
RBW = resolution bandwidth and Keysight X-Series s igna l analyzers
ST = sweep time. offer bandwidth steps of just 10% for an
even better compromise among span,
On the other hand, the rise time of resolution and sweep time.
a filter is inversely proportional to its
bandwidth, and if we include a constant Spectrum analyzers automatically couple
of proportionality, k, then: sweep time to the span and resolution
bandwidth settings. Sweep time is
k adjusted to maintain a calibrated display.
Rise time =
RBW
If the need arises, we can override
If we make the terms equal and solve for the automatic setting and set sweep
sweep time, we have: time manually. If you set a sweep time
shorter than the maximum available,
k (RBW)(ST)
RBW = the analyzer indicates that the display
Span
is uncalibrated with a “Meas Uncal”
k (Span) message in the upper-right part of the
or ST =
RBW2 graticule.
For the synchronously-tuned, near-
Gaussian filters used in many analog
analyzers, the value of k is in the 2 to 3
range .

18
Digital resolution filters
The digital resolution filters used in
Keysight spectrum analyzers have an
effect on sweep time that is different
from the effects we’ve just discussed
for analog filters. For swept analysis,
the speed of digitally implemented
filters, with no further processing, can
show a two to four times improvement.

However, the X-Series signal analyzers

with Option FS1 are programmed to
correct for the effect of sweeping too
fast for resolution bandwidths between
about 3 kHz and 100 kHz. As a result,
sweep times that would otherwise
be many seconds may be reduced
to milliseconds, depending upon the
particular settings. See Figure 2-14a.
The sweep time without the correction
would be 79.8 seconds. Figure 2-14b
shows a sweep time of 1.506 s with Figure 2-14a. Full span sweep speed, RBW of 20 kHz, without Option FS1
Option FS1 installed. For the widest
resolution bandwidths, sweep times are
already very short. For example, using
the formula with k = 2 on a span of
1 GHz and a RBW of 1 MHz, the sweep
time calculates to just 2 msec.

For narrower resolution bandwidths,

analyzers such as the Keysight
X-Series use fast Fourier transforms
(FFTs) to process the data, also
producing shorter sweep times than
the formula predicts. The difference
occurs because the signal being
analyzed is processed in frequency
blocks, depending upon the particular
analyzer. For example, if the frequency
block was 1 kHz, then when we select
a 10-Hz resolution bandwidth, the
analyzer is in effect simultaneously
processing the data in each 1-kHz
block through 100 contiguous 10-Hz
filters. If the digital processing were
instantaneous, we would expect sweep Figure 2-14b. Full span sweep speed, RBW of 20 kHz, with Option FS1
time to be reduced by a factor of 100.
In practice, the reduction factor is
less, but is still significant. For more More information
information on the advantages of
digital processing, refer to Chapter 3. A more detailed discussion about fast sweep measurements can be found in
Using Fast-Sweep Techniques to Accelerate Spur Searches – Application Note, litera-
ture number 5991-3739EN

19
Envelope detector6
Older analyzers typically converted
the IF signal to video with an envelope t t
detector7. In its simplest form, an
envelope detector consists of a diode, IF signal
resistive load and low-pass filter, as
shown in Figure 2-15. The output of the
IF chain in this example, an amplitude Figure 2-15. Envelope detector
modulated sine wave, is applied to the
detector. The response of the detector
follows the changes in the envelope of
the IF signal, but not the instantaneous
value of the IF sine wave itself.

For most measurements, we choose a

resolution bandwidth narrow enough
to resolve the individual spectral
components of the input signal. If we
fix the frequency of the LO so that our
analyzer is tuned to one of the spectral
components of the signal, the output
of the IF is a steady sine wave with a
constant peak value. The output of Figure 2-16. Output of the envelope detector follows the peaks of the IF signal
the envelope detector will then be a
constant (DC) voltage, and there is no enough so that, with the analyzer
The width of the resolution (IF) filter tuned half-way between them, there
variation for the detector to follow.
determines the maximum rate at is negligible attenuation due to the
which the envelope of the IF signal can roll-off of the filter 8. The analyzer
However, there are times when we
change. This bandwidth determines display will vary between a value that
deliberately choose a resolution
how far apart two input sinusoids can is twice the voltage of either (6 dB
bandwidth wide enough to include
be so that after the mixing process greater) and zero (minus infinity on the
two or more spectral components.
they will both be within the filter at the log scale). We must remember that the
At other times, we have no choice.
same time. Let’s assume a 22.5-MHz two signals are sine waves (vectors)
The spectral components are closer
final IF and a 100-kHz bandwidth. Two at different frequencies, and so they
in frequency than our narrowest
input signals separated by 100 kHz continually change in phase with
bandwidth. Assuming only two spectral
would produce mixing products of 22.45 respect to each other. At some time
components within the pass band,
and 22.55 MHz and would meet the they add exactly in phase; at another,
we have two sine waves interacting to
criterion. See Figure 2-16. The detector exactly out of phase.
create a beat note, and the envelope of
must be able to follow the changes
the IF signal varies, as shown in Figure
in the envelope created by these two So the envelope detector follows the
2-16, as the phase between the two
signals but not the 22.5-MHz IF signal changing amplitude values of the peaks
sine waves varies.
itself. of the signal from the IF chain but not
the instantaneous values, resulting
The envelope detector is what makes in the loss of phase information.
More information the spectrum analyzer a voltmeter. This gives the analyzer its voltmeter
Let’s duplicate the situation above and characteristics.
Additional information on envelope
have two equal-amplitude signals in
detectors can be found in Spectrum
the pass band of the IF at the same Digitally implemented resolution
and Signal Analyzer Measurements and
time. A power meter would indicate a bandwidths do not have an analog
Noise–Application Note, literature
power level 3 dB above either signal, envelope detector. Instead, the digital
number 5966-4008E.
that is, the total power of the two. processing computes the root sum
Assume that the two signals are close of the squares of the I and Q data,

6.  The envelope detector should not be confused with the display detectors. See “Detector types” later in this chapter.
7.  A signal whose frequency range extends from zero (DC) to some upper frequency determined by the circuit elements. Historically, spectrum analyz-
ers with analog displays used this signal to drive the vertical deflection plates of the CRT directly. Hence it was known as the video signal.
8.  For this discussion, we assume the filter is perfectly rectangular.

20
which is mathematically equivalent moved slowly across the cathode ray rate had to be readjusted for each
to an envelope detector. For more tube (CRT), with no real trace on the new measurement situation. When
information on digital architecture, display. So a meaningful display was digital circuitry became affordable
refer to Chapter 3. not possible with the longer sweep in the mid-1970s, it was quickly put
times. to use in spectrum analyzers. Once a
Displays trace had been digitized and put into
Keysight (part of Hewlett-Packard memory, it was permanently available
Up until the mid-1970s, spectrum at the time) pioneered a variable- for display. It became an easy matter
analyzers were purely analog. The persistence storage CRT in which to update the display at a flicker-free
displayed trace presented a continuous we could adjust the fade rate of the rate without blooming or fading. The
indication of the signal envelope, and display. When properly adjusted, data in memory was updated at the
no information was lost. However, the old trace would just fade out at sweep rate, and since the contents of
analog displays had drawbacks. The the point where the new trace was memory were written to the display at
major problem was in handling the updating the display. This display was a flicker-free rate, we could follow the
long sweep times required for narrow continuous, had no flicker and avoided updating as the analyzer swept through
resolution bandwidths. In the extreme confusing overwrites. It worked quite its selected frequency span just as we
case, the display became a spot that well, but the intensity and the fade could with analog systems.

Figure 2-17. When digitizing an analog signal, what value should be displayed at each point?

21
Detector types
With digital displays, we had to decide
what value should be displayed for
each display data point. No matter
how many data points we use across
the display, each point must represent
what has occurred over some
frequency range and time interval
(although we usually do not think in
terms of time when dealing with a
spectrum analyzer).

It is as if the data for each interval is

thrown into a bucket and we apply
whatever math is necessary to extract
the desired bit of information from our
input signal. This datum is put into
memory and written to the display. This
process provides great flexibility.

Here we will discuss six different

detector types. Figure 2-18. Each of the 1001 trace points (buckets) covers a 100-kHz frequency span and a 0.01-mil-
lisecond time span

In Figure 2-18, each bucket contains

data from a span and timeframe that
is determined by these equations: One bucket

Frequency: Positive peak

bucket width = span/(trace points – 1)
Time:
bucket width = sweep time/(trace points – 1)

The sampling rates are different for

various instruments, but greater
Sample
accuracy is obtained from decreasing
the span or increasing the sweep time
because the number of samples per
bucket will increase in either case. Even
in analyzers with digital IFs, sample Negative peak
rates and interpolation behaviors
are designed to be the equivalent of
continuous-time processing.
Figure 2-19. The trace point saved in memory is based on the detector type algorithm

The “bucket” concept is important,

as it will help us differentiate the six The first three detectors, sample, Let’s return to the question of how to
detector types: peak, and negative peak are easy to display an analog system as faithfully
–– Sample understand and are visually represented as possible using digital techniques.
–– Positive peak (also simply called in Figure 2-19. Normal, average, and Let’s imagine the situation illustrated
peak) quasipeak are more complex and will be in Figure 2-17. We have a display that
–– Negative peak discussed later. contains only noise and a single CW
–– Normal signal.
–– Average
–– Quasipeak

22
Sample detection
As a first method, let us simply select
the data point as the instantaneous
level at the center of each bucket
(see Figure 2-19). This is the sample
detection mode. To give the trace a
continuous look, we design a system
that draws vectors between the points. Figure 2-20. Sample display mode using 10 points Figure 2-21. More points produce a display closer
Comparing Figure 2-17 with 2-20, it to display the signal shown in Figure 2-17 to an analog display

appears that we get a fairly reasonable

display. Of course, the more points
there are in the trace, the better the
replication of the analog signal will be.
The number of available display points
can vary for different analyzers. On
X-Series signal analyzers, the number
of display points for frequency domain
traces can be set from a minimum of 1
point to a maximum of 40,001 points.
As shown in Figure 2-21, more points
do indeed get us closer to the analog
signal.

While the sample detection mode

does a good job of indicating the
randomness of noise, it is not a good
mode for analyzing sinusoidal signals.
If we were to look at a 100-MHz comb
on a Keysight PXA, we might set it to
span from 0 to 26.5 GHz. Even with
1,001 display points, each display
point represents a span (bucket) of
Figure 2-22a. A 10-MHz span of a 250-kHz comb in the sample display mode
26.5 MHz. This is far wider than the
maximum 8-MHz resolution bandwidth.

As a result, the true amplitude of a

comb tooth is shown only if its mixing
product happens to fall at the center
of the IF when the sample is taken.
Figure 2-22a shows a 10-MHz span
with a 750-Hz bandwidth using sample
detection. The comb teeth should
be relatively equal in amplitude, as
shown in Figure 2-22b (using peak
detection). Therefore, sample detection
does not catch all the signals, nor
does it necessarily reflect the true
peak values of the displayed signals.
When resolution bandwidth is more
narrow than the sample interval (the
bucket width), sample mode can give
erroneous results.

Figure 2-22b. The actual comb over a 10-MHz span using peak (positive) detection

23
Peak (positive) detection Negative peak detection Normal detection
One way to insure that all sinusoids Negative peak detection displays the To provide a better visual display of
are reported at their true amplitudes minimum value encountered in each random noise than offered by peak
is to display the maximum value bucket. It is generally available in mode and yet avoid the missed-signal
encountered in each bucket. This is most spectrum analyzers, though it problem of the sample mode, the
the positive peak detection mode, is not used as often as other types of normal detection mode (informally
or peak. This mode is illustrated in detection. Differentiating CW from known as rosenfell 9 mode) is offered on
Figure 2-22b. Peak is the default mode impulsive signals in EMC testing is many spectrum analyzers. Should the
offered on many spectrum analyzers one application where negative peak signal both rise and fall, as determined
because it ensures that no sinusoid is detection is valuable. Later in this by the positive peak and negative peak
missed, regardless of the ratio between application note, we will see how detectors, the algorithm classifies the
resolution bandwidth and bucket width. negative peak detection is also used in signal as noise.
However, unlike sample mode, peak signal identification routines when you
does not give a good representation of use external mixers for high-frequency In that case, an odd-numbered data
random noise because it only displays measurements. point displays the maximum value
the maximum value in each bucket and encountered during its bucket. And an
ignores the true randomness of the even-numbered data point displays the
noise. So spectrum analyzers that use minimum value encountered during its
peak detection as their primary mode bucket. See Figure 2-25. Normal and
generally also offer sample mode as an sample modes are compared in Figures
alternative. 2-23a and 2-23b.10

Figure 2-23a. Normal mode Figure 2-23b. Sample mode

9.  Rosenfell is not a person’s name but rather a description of the algorithm that tests to see if the signal rose and fell within the bucket represented
by a given data point. It is also sometimes written as “rose’n’fell.”
10.  Because of its usefulness in measuring noise, the sample detector is usually used in “noise marker” applications. Similarly, the measurement of
channel power and adjacent-channel power requires a detector type that gives results unbiased by peak detection. For analyzers without
averaging detectors, sample detection is the best choice.

24
What happens when a sinusoidal
signal is encountered? We know that
as a mixing product is swept past
the IF filter, an analyzer traces out
the shape of the filter on the display.
If the filter shape is spread over
many display points, we encounter
a situation in which the displayed
signal only rises as the mixing product
approaches the center frequency of
the filter and only falls as the mixing
product moves away from the filter
center frequency. In either of these
cases, the positive-peak and negative-
peak detectors sense an amplitude
change in only one direction, and,
according to the normal detection
algorithm, the maximum value in each
bucket is displayed. See Figure 2-24.

What happens when the resolution

bandwidth is narrow, relative to a
Figure 2-24. Normal detection displays maximum values in buckets where the signal only rises or only falls
bucket? The signal will both rise and
fall during the bucket. If the bucket
happens to be an odd-numbered
one, all is well. The maximum
value encountered in the bucket
is simply plotted as the next data
point. However, if the bucket is even-
numbered, then the minimum value
in the bucket is plotted. Depending
on the ratio of resolution bandwidth
to bucket width, the minimum value
can differ from the true peak value
(the one we want displayed) by a
little or a lot. In the extreme, when
the bucket is much wider than the
resolution bandwidth, the difference
between the maximum and minimum
values encountered in the bucket is
the full difference between the peak
signal value and the noise. This is
true for the example in Figure 2-25.
See bucket 6. The peak value of the
previous bucket is always compared
to that of the current bucket. The
greater of the two values is displayed
if the bucket number is odd, as
depicted in bucket 7. The signal peak
actually occurs in bucket 6 but is not
displayed until bucket 7.

25
The normal detection
algorithm:
If the signal rises and falls within a
bucket:

Even-numbered buckets display the

minimum (negative peak) value in the
bucket. The maximum is remembered.
Odd-numbered buckets display the
maxi(positive peak) value determined
by comparing the current bucket peak
with the previous (remembered) bucket
peak. If the signal only rises or only falls
within a bucket, the peak is displayed.
See Figure 2-25.

This process may cause a maximum

value to be displayed one data point
too far to the right, but the offset is
usually only a small percentage of the
span. Some spectrum analyzers, such
as the Keysight PXA signal analyzer,
compensate for this potential effect
by moving the LO start and stop Figure 2-25. Trace points selected by the normal detection algorithm
frequencies.

Another type of error occurs when two

peaks are displayed when only one
actually exists. Figure 2-26 shows this
error. The outline of the two peaks is
displayed using peak detection with a
wider RBW.

So peak detection is best for locating

CW signals well out of the noise.
Sample is best for looking at noise, and
normal is best for viewing signals and
noise.

Figure 2-26. Normal detection can show two peaks when only one peak actually exists

26
Average detection the true average power, and is best measurements changes from being
for measuring the power of complex a summation over display buckets
Although modern digital modulation signals. to being a summation over FFT bins.
schemes have noise-like In both swept and FFT cases, the
characteristics, sample detection Voltage averaging averages the linear integration captures all the power
does not always provide us with the voltage data of the envelope signal information available, rather than just
information we need. For instance, measured during the bucket interval. that which is sampled by the sample
when taking a channel power It is often used in EMI testing for detector. As a result, the average
measurement on a W-CDMA signal, measuring narrowband signals (this detector has a lower variance result for
integration of the rms values is topic will be discussed further in the the same measurement time. In swept
required. This measurement involves next section). Voltage averaging is analysis, it also allows the convenience
summing power across a range of also useful for observing rise and fall of reducing variance simply by
analyzer frequency buckets. Sample behavior of AM or pulse-modulated extending the sweep time.
detection does not provide this signals such as radar and TDMA
capability. transmitters. EMI detectors: average and
While spectrum analyzers typically Log-power (video) averaging averages
quasipeak detection
collect amplitude data many times in the logarithmic amplitude values An important application of average
each bucket, sample detection keeps (dB) of the envelope signal measured detection is for characterizing devices
only one of those values and throws during the bucket interval. Log for electromagnetic interference
away the rest. On the other hand, an power averaging is best for observing (EMI). In this case, voltage averaging,
averaging detector uses all the data sinusoidal signals, especially those as described in the previous section,
values collected within the time (and near noise.11 is used for measuring narrowband
frequency) interval of a bucket. Once we signals that might be masked by the
have digitized the data, and knowing the Thus, using the average detector presence of broadband impulsive
circumstances under which they were with the averaging type set to power noise. The average detection used in
digitized, we can manipulate the data in provides true average power based EMI instruments takes an envelope-
a variety of ways to achieve the desired upon rms voltage, while the average detected signal and passes it through a
results. detector with the averaging type set low-pass filter with a bandwidth much
to voltage acts as a general-purpose less than the RBW. The filter integrates
Some spectrum analyzers refer to the average detector. The average detector (averages) the higher-frequency
averaging detector as an rms detector with the averaging type set to log has components such as noise. To perform
when it averages power (based on the no other equivalent. this type of detection in an older
root mean square of voltage). Keysight spectrum analyzer that doesn’t have
X-Series signal analyzers have an Average detection is an improvement a built-in voltage averaging detector
average detector that can average the over using sample detection for the function, set the analyzer in linear
power, voltage or log of the signal by determination of power. Sample mode and select a video filter with a
including a separate control to select detection requires multiple sweeps to cut-off frequency below the lowest PRF
the averaging type: collect enough data points to give us of the measured signal.
accurate average power information.
Power (rms) averaging computes Average detection changes channel Quasipeak detectors (QPD) are
rms levels, by taking the square power measurements from being a also used in EMI testing. QPD is a
root of the average of the squares of summation over a range of buckets weighted form of peak detection. The
the voltage data measured during into integration over the time interval measured value of the QPD drops as
the bucket interval. This computed representing a range of frequencies the repetition rate of the measured
voltage is squared and divided by in a swept analyzer. In a fast Fourier signal decreases. Thus, an impulsive
the characteristic input impedance transfer (FFT) analyzer 12, the signal with a given peak amplitude
of the spectrum analyzer, normally summation used for channel power and a 10-Hz pulse repetition rate will
50 ohms. Power averaging calculates have a lower quasipeak value than a
11.  See Chapter 5, “Sensitivity and Noise.” signal with the same peak amplitude
12.  Refer to Chapter 3 for more information on the FFT analyzers. They perform math computa- but having a 1-kHz repetition rate. This
tions on many buckets simultaneously, which improves measurement speed. signal weighting is accomplished by
13.  CISPR, the International Special Committee on Radio Interference, was established in 1934
by a group of international organizations to address radio interference. CISPR is a non- circuitry with specific charge, discharge
governmental group composed of National Committees of the International Electrotechnical and display time constants defined by
Commission (IEC), as well as numerous international organizations. CISPR’s recommended CISPR13.
standards generally form the basis for statutory EMC requirements adopted by governmental
regulatory agencies around the world.

27
QPD is a way of measuring and
quantifying the “annoyance factor”
of a signal. Imagine listening to a radio
station suffering from interference. If you
hear an occasional “pop” caused by noise
once every few seconds, you can still
listen to the program without too much
trouble. However, if that same amplitude
pop occurs 60 times per second, it
becomes extremely annoying, making
the radio program intolerable to listen to.

Averaging processes
There are several processes in a
spectrum analyzer that smooth the
variations in envelope-detected
amplitude. The first method, average
detection, was discussed previously. Figure 2-27. Spectrum analyzers display signal plus noise

Two other methods, video filtering and

trace averaging, are discussed next.14

Video filtering
Discerning signals close to the noise
is not just a problem when performing
EMC tests. Spectrum analyzers display
signals plus their own internal noise,
as shown in Figure  2-27. To reduce the
effect of noise on the displayed signal
amplitude, we often smooth or average
the display, as shown in Figure 2-28.
Spectrum analyzers include a variable
video filter for this purpose. The video
filter is a low-pass filter that comes after
the envelope detector and determines
the bandwidth of the video signal that
will later be digitized to yield amplitude
Figure 2-28. Display of Figure 2-27 after full smoothing
data. The cutoff frequency of the video
filter can be reduced to the point where
it becomes smaller than the bandwidth
of the selected resolution bandwidth
(IF) filter. When this occurs, the video
system can no longer follow the more
rapid variations of the envelope of the

More information
A more detailed discussion about
noise markers can be found in
Spectrum and Signal Analyzer Measure-
ments and Noise – Application Note,
literature number 5966-4008E

14.  A fourth method, called a noise marker,

is discussed in Chapter 5, “Sensitivity and
Noise.”
Figure 2-29. Smoothing effect of VBW-to-RBW ratios of 3:1, 1:10, and 1:100

28
signal(s) passing through the IF chain.
The result is an averaging or smoothing
of the displayed signal.

The effect is most noticeable in

measuring noise, particularly when you
use a wide-resolution bandwidth. As
we reduce the video bandwidth, the
peak-to-peak variations of the noise
are reduced. As Figure 2-29 shows,
the degree of reduction (degree of
averaging or smoothing) is a function
of the ratio of the video to resolution
bandwidths. At ratios of 0.01 or less,
the smoothing is very good. At higher
ratios, the smoothing is not as good.
The video filter does not affect any part
of the trace that is already smooth (for
example, a sinusoid displayed well out
of the noise).

If we set the analyzer to positive peak

Figure 2-30a. Positive peak detection mode: reducing video bandwidth lowers peak noise but not
detection mode, we notice two things: average noise
First, if VBW > RBW, then changing
the resolution bandwidth does not
make much difference in the peak-
to-peak fluctuations of the noise.
Second, if VBW < RBW, changing the
video bandwidth seems to affect the
noise level. The fluctuations do not
change much because the analyzer is
displaying only the peak values of the
noise. However, the noise level appears
to change with video bandwidth
because the averaging (smoothing)
changes, thereby changing the peak
values of the smoothed noise envelope.
See Figure 2-30a. When we select
average detection, we see the average
noise level remains constant. See
Figure 2-30b.

Because the video filter has its

own response time, the sweep time
increases approximately inversely with
video bandwidth when the VBW is
less than the resolution bandwidth.
Figure 2-30b. Average detection mode: noise level remains constant, regardless of VBW-to-RBW ratios
The sweep time (ST) can therefore be (3:1, 1:10 and 1:100)
described by this equation:
k(Span)
ST ≈ (RBW)(VBW)

The analyzer sets the sweep time

automatically to account for video
bandwidth as well as span and
resolution bandwidth.

29
Trace averaging
Digital displays offer another choice for
smoothing the display: trace averaging.
Trace averaging uses a completely
different process from the smoothing
performed using the average detector.
In this case, averaging is accomplished
over two or more sweeps on a point-
by-point basis. At each display point,
the new value is averaged in with the
previously averaged data:
n– 1
( )
A avg = n A prior avg +
1
( )
n An

where
A avg = new average value
A prior avg = average from prior sweep
A n= measured value on current sweep
n = number of current sweep

Thus, the display gradually converges

to an average over a number of sweeps. Figure 2-31. Trace averaging for 1, 5, 20 and 100 sweeps, top to bottom (trace position offset for each
set of sweeps
As with video filtering, we can select
the degree of averaging or smoothing. filtering or trace averaging. However, As a result, we can get significantly
We do this by setting the number of there is a distinct difference between different results from the two averaging
sweeps over which the averaging occurs. the two. Video filtering performs methods on certain signals. For
Figure 2-31 shows trace averaging for averaging in real time. That is, we example, a signal with a spectrum that
different numbers of sweeps. While see the full effect of the averaging or changes with time can yield a different
trace averaging has no effect on sweep smoothing at each point on the display average on each sweep when we use
time, the time to reach a given degree as the sweep progresses. Each point is video filtering. However, if we choose
of averaging is about the same as with averaged only once, for a time of about trace averaging over many sweeps, we
video filtering because of the number of 1/VBW on each sweep. Trace averaging, will get a value much closer to the true
sweeps required. on the other hand, requires multiple average. See Figures 2-32a and 2-32b.
sweeps to achieve the full degree of
In many cases, it does not matter averaging, and the averaging at each Figures 2-32a and 2-32b show how
which form of display smoothing we point takes place over the full time video filtering and trace averaging yield
pick. If the signal is noise or a low-level period needed to complete the multiple different results on an FM broadcast
sinusoid very close to the noise, we sweeps. signal.
get the same results with either video

Figure 2-32a. Video filtering Figure 2-32b. Trace averaging

30
Time gating Measuring time division signal shown, as in Figure 2-33c.
Time gating can be achieved using
Time-gated spectrum analysis allows duplex signals
three different methods we will discuss
you to obtain spectral information To illustrate using time-gating below. However, there are certain basic
about signals occupying the same part capability to perform difficult concepts of time gating that apply
of the frequency spectrum that are measurements, consider Figure 2-33a, to any implementation. In particular,
separated in the time domain. Using which shows a simplified digital you must have, or be able to set, the
an external trigger signal to coordinate mobile-radio signal in which two following four items:
the separation of these signals, you can radios, #1 and #2, are time-sharing
perform the following operations: a single frequency channel. Each –– An externally supplied gate trigger
radio transmits a single 1-ms burst, signal
–– Measure any one of several signals then shuts off while the other radio –– The gate control or trigger mode
separated in time (For example, transmits for 1 ms. The challenge (edge or level) (The X-Series signal
you can separate the spectra of is to measure the unique frequency analyzers can be set to gate-
two radios time-sharing a single spectrum of each transmitter. trigger holdoff to ignore potential
frequency.) false triggers.)
–– Measure the spectrum of a signal Unfortunately, a traditional spectrum –– The gate delay setting, which
in one time slot of a TDMA system analyzer cannot do that. It simply determines how long after the
–– Exclude the spectrum of interfering shows the combined spectrum, as seen trigger signal the gate actually
signals, such as periodic pulse in Figure 2-33b. Using the time-gating becomes active and the signal is
edge transients that exist for only a capability and an external trigger observed
limited time signal, you can see the spectrum of just –– The gate length setting, which
radio #1 (or radio #2 if you wish) and determines how long the gate is on
Why time gating is needed identify it as the source of the spurious and the signal is observed

Traditional frequency-domain spectrum

analysis provides only limited
information for certain difficult-to-
analyze signals. Examples include the
following signal types:

–– Pulsed RF
–– Time multiplexed
–– Time domain multiple access
(TDMA)
–– Interleaved or intermittent
–– Burst modulated

In some cases, time-gating capability

enables you to perform measurements
that would otherwise be very difficult,
if not impossible to make.
Figure 2-33a. Simplified digital mobile-radio signal in the time domain

Figure 2-33b. Frequency spectrum of combined Figure 2-33c. The time-gated spectrum of signal Figure 2-33d. The time-gated spectrum of signal
signals. Which radio produces the spurious emis- #1 identifies it as the source of spurious emission #2 shows it is free of spurious emissions
sions?

31
Controlling these parameters will allow
us to look at the spectrum of the signal
during a desired portion of the time.
If you are fortunate enough to have a
gating signal that is only true during
the period of interest, you can use
level gating, as shown in Figure 2-34.
However, in many cases the gating
signal will not perfectly coincide with
the time we want to measure the
spectrum. Therefore, a more flexible
approach is to use edge triggering in
conjunction with a specified gate delay
and gate length to precisely define the
time period in which to measure the
signal. Figure 2-34. Level triggering: the spectrum analyzer only measures the frequency spectrum when the
gate trigger signal is above a certain level

Consider the GSM signal with eight

time slots in Figure 2-35. Each burst is
0.577 ms and the full frame is 4.615 ms.
We may be interested in the spectrum
of the signal during a specific time slot.
For the purposes of this example, let’s
assume we are using only two of the
eight available time slots (time slots 1
and 3), as shown in Figure 2-36. When
we look at this signal in the frequency
domain in Figure 2-37, we observe an
unwanted spurious signal present in
the spectrum. In order to troubleshoot
the problem and find the source of
this interfering signal, we need to
determine the time slot in which it is
occurring. If we wish to look at time
slot 3, we set up the gate to trigger
on the rising edge of the burst in time
slot 3, and, then specify a gate delay of
1.4577 ms and a gate length of 461.60
µs, as shown in Figure 2-38. The gate
delay assures that we only measure the
spectrum of time slot 3 while the burst Figure 2-35. A TDMA format signal (in this case, GSM) with 8 time slots, time slot zero is “off”.
is fully on. Note that the gate start
and stop value is carefully selected to
avoid the rising and falling edge of the
burst, as we want to allow time for the
RBW filtered signal to settle out before
we make a measurement. Figure 2-39
shows the spectrum of time slot 3,
which reveals that the spurious signal is
not caused by this burst.

Three methods are commonly used to

perform time gating:
–– Gated FFT
–– Gated LO
–– Gated video

32
Figure 2-36. A zero span (time domain) view of the GSM signal with only time Figure 2-37. Frequency domain view of the GSM signal with 2 time slots “on”
slots 1 and 3 “on”. showing an unwanted spurious signal present in the spectrum.

Figure 2-38. Time gating is used to look at the spectrum of the GSM time Figure 2-39. Spectrum of time slot 3 reveals that the spurious signal is not
slot 3. caused by this burst.

33
 RF IF resolution Envelope Video
step bandwidth IF log detector bandwidth Peak/sample Analog-digital
attenuator Mixer filter amplifier (IF to video) filter detector converter
RF
input

Display logic

Local Scan generator

oscillator
Gate
control
Display
Figure 2-40. In gated LO mode, the LO sweeps only during gate interval

Gated FFT Gated LO described earlier in this chapter. Using

an X-Series signal analyzer, a standard,
The Keysight X-Series signal analyzers LO gating, sometimes referred to as non-gated, spectrum sweep over a
have built-in FFT capabilities. In this gated sweep, is another technique 1-MHz span takes 14.6 ms, as shown
mode, the data is acquired for an FFT for performing time gating. With this in Figure 2-41. With a gate length of
starting at a chosen delay following method, we control the voltage ramp 0.3 ms, the spectrum analyzer sweep
a trigger. The IF signal is digitized produced by the scan generator to must be built up in 49 gate intervals
and captured for a time period of sweep the LO, as shown in Figure (14.6 divided by 0.3). Or, if the full
1.83 divided by resolution bandwidth. 2-40. When the gate is active, the frame of the GSM signal is 4.615 ms,
An FFT is computed based on this LO ramps up in frequency like any the total measurement time is 49
data acquisition and the results are spectrum analyzer. When the gate is intervals times 4.615 ms = 226 ms. This
displayed as the spectrum. Thus, the blocked, the voltage out of the scan represents a significant improvement
spectrum is that which existed at a generator is frozen, and the LO stops in speed compared to the gated video
particular time of known duration. This rising in frequency. This technique technique, which will be described
is the fastest gating technique when can be much faster than gated video in the following section. LO gating is
the span is not wider than the FFT because multiple buckets can be available on X-Series signal analyzers
maximum width. measured during each burst. As an and PSA Series spectrum analyzers.
example, let’s use the same GSM signal
To get the maximum possible
frequency resolution, choose the
narrowest available RBW with a
capture time that fits within the time
period of interest. You may not always
need that much resolution, however,
and you could choose a wider RBW
with a corresponding narrower gate
length. The minimum usable RBW in
gated FFT applications is always lower
than the minimum usable RBW in other
gating techniques, because the IF must
fully settle during the burst in other
techniques, which takes longer than
1.83 divided by RBW.

Figure 2-41. Spectrum of the GSM signal

34
Gated video or bucket, so the peak detector is times 4.615 ms or 1.85 s. Some TDMA
able to see real data during that time formats have cycle times as large as 90
Gated video is the analysis technique interval. Otherwise, there will be trace ms, resulting in long sweep times using
used in a number of spectrum points with no data, resulting in an the gated video technique.
analyzers, including the Keysight 8560, incomplete spectrum. Therefore, the
8590 and ESA Series. In this case, minimum sweep time is N display Now that you’ve seen how a classic
the video voltage is switched off, or buckets times burst cycle time. For analog spectrum analyzer works and
to “negative infinity decibels,” during example, in GSM measurements, the how to use some of the important
the time the gate is supposed to be full frame lasts 4.615 ms. For an ESA features and capabilities, let’s take
in its “blocked” mode. The detector spectrum analyzer set to its default a look at how replacing some analog
is set to peak detection. The sweep value of 401 display points, the circuits with digital technology
time must be set so that the gates minimum sweep time for GSM gated improves spectrum analyzer
occur at least once per display point, video measurements would be 401 performance.

RF IF resolution Envelope Video

step bandwidth IF log detector bandwidth Peak/sample Analog-digital
attenuator Mixer filter amplifier (IF to video) filter detector converter

RF
input
– ∞ dB
Reset
Gate control

Display logic

Local Scan generator

oscillator

Display

Figure 2-42. Block diagram of a spectrum analyzer with gated video

35
Chapter 3. Digital IF Overview
Since the 1980s, one of the most bandwidths (1 Hz to 300 Hz) are A key benefit of the digital processing
profound changes in spectrum analysis realized using digital techniques. As done in these analyzers is a bandwidth
has been the application of digital shown in Figure 3-1, the linear analog selectivity of about 4:1. This selectivity
technology to replace portions of signal is mixed down to an 8.5-kHz is available on the narrowest filters, the
spectrum analyzers that had been IF and passed through a bandpass ones we would choose to separate the
implemented previously as analog filter only 1 kHz wide. This IF signal is most closely spaced signals.
circuits. With the availability of amplified, then sampled at an 11.3-kHz
high-performance analog-to-digital rate and digitized. In Chapter 2, we did a filter skirt
converters, the latest spectrum selectivity calculation for two
analyzers digitize incoming signals Once in digital form, the signal is signals spaced 4 kHz apart, using a
much earlier in the signal path put through a fast Fourier transform 3-kHz analog filter. Let’s repeat that
compared to spectrum analyzer algorithm. To transform the appropriate calculation using digital filters. A good
designs of just a few years ago. The signal, the analyzer must be fixed- model of the selectivity of digital filters
change has been most dramatic in the tuned (not sweeping). That is, the is a near-Gaussian model:
IF section of the spectrum analyzer. transform must be done on a time-
α
Digital IFs1 have had a great impact on domain signal. Thus the ESA-E Series H(∆f) = –3.01 dB x [ RBW/2
∆f
]
spectrum analyzer performance, with analyzers step in 900-Hz increments,
significant improvements in speed, instead of sweeping continuously, when where H(∆f) is the filter skirt rejection in dB.
accuracy and the ability to measure we select one of the digital resolution
complex signals using advanced DSP bandwidths. This stepped tuning ∆f is the frequency offset from the
techniques. can be seen on the display, which is center in Hz, and α is a parameter that
updated in 900-Hz increments as the controls selectivity. α = 2 for an ideal
Digital filters digital processing is completed. Gaussian filter. The swept RBW filters
used in Keysight spectrum analyzers
You will find a partial implementation As you will see in a moment, other are based on a near-Gaussian model
of digital IF circuitry in the Keysight spectrum and signal analyzers, such with an α value equal to 2.12, resulting
ESA-E Series spectrum analyzers. as the Keysight X-Series analyzers, in a selectivity ratio of 4.1:1.
While the 1-kHz and wider RBWs are use an all-digital IF, implementing all
implemented with traditional analog resolution bandwidth filters digitally.
LC and crystal filters, the narrowest

Log
21.4 MHz

Video ADC µC

Linear

3rd LO Sample and hold

at 11.3 kHz

8.5 kHz CF
1 kHz BW

Figure 3-1. Digital implementation of 1-, 3-, 10-, 30-, 100- and 300-Hz resolution filters in ESA-E Series spectrum analyzers

1.  Strictly speaking, once a signal has been digitized, it is no longer at an intermediate frequency, or IF. At that point, the signal is represented by digi-
tal data values. However, we use the term “digital IF” to describe the digital processing that replaces the analog IF processing found in traditional
spectrum analyzers.

36
Entering the values from our example the all-digital IF in the X-Series signal three cycles of the ADC clock (30
into the equation, we get: analyzer, as shown in Figure 3-2. MHz) through the anti-alias filter. The
H(4 kHz) = –3.01 dB x [
4000 2.12
3000/2 ] In this case, all 160 resolution
delay allows time for an impending
large signal to be recognized before it
= –24.1 dB
bandwidths are digitally implemented. overloads the ADC. The logic circuitry
At an offset of 4 kHz, the 3-kHz digital However, there is some analog circuitry controlling the autorange detector
filter is down –24.1 dB compared to the prior to the ADC, starting with several will decrease the gain in front of the
analog filter which was only down –14.8 stages of down conversion, followed ADC before a large signal reaches it,
dB. Because of its superior selectivity, by a pair of single-pole prefilters (one thus preventing clipping. If the signal
the digital filter can resolve more an LC filter, the other crystal-based). envelope remains small for a long time,
closely spaced signals. A prefilter helps prevent succeeding the autoranging circuit increases the
stages from contributing third-order gain, reducing the effective noise at the
distortion in the same way a prefilter input. The digital gain after the ADC
All-digital IF
would in an analog IF. In addition, it is also changed to compensate for the
Analyzers such as the Keysight enables dynamic range extension via analog gain in front of it. The result is
X-Series combine several digital autoranging. The output of the single- a “floating point” ADC with very wide
techniques to achieve the all-digital IF. pole prefilter is routed to the autorange dynamic range when autoranging is
The all-digital IF offers users a wealth detector and the anti-alias filter. enabled in swept mode.
of advantages. The combination of FFT
analysis for narrow spans and swept As with any FFT-based IF architecture,
analysis for wider spans optimizes the anti-alias filter is required to
sweeps for the fastest possible prevent aliasing (the folding of out-
measurements. Architecturally, the of-band signals into the ADC sampled
ADC is moved closer to the input data). This filter has many poles and
port, a move made possible by thus has substantial group delay.
improvements to the A-to-D converters Even a very fast-rising RF burst,
and other digital hardware. Let’s begin downconverted to the IF frequency,
by taking a look at the block diagram of will experience a delay of more than

Custom IC

Analog Digital Counter

Anti-alias gain gain I
filter
I, Q Display
ADC VBW det
Q r, log (r)

log pwr pwr log

Hilbert log v v log
Ranging –1 transform
Prefilter rules log log log log

Autoranging ADC system

Processor

Display
FFT Processing
log/lin dB/div Display

Figure 3-2. Block diagram of the all-digital IF in the Keysight X-Series signal analyzers

37
Figure 3-3 illustrates the sweeping
Amplitude
behavior of the X-Series analyzers. The (log)
single-pole prefilter allows the gain to
be turned up high when the analyzer
is tuned far from the carrier. As the ADC
carrier gets closer, the gain falls and clipping threshold
the ADC quantization noise rises. The Prefilter gain
noise level will depend on the signal Typical
analog IF Digital IF RBW response
level frequency separation from the response Noise floor after autoranging
carrier, so it looks like a step-shaped Typical LO phase noise
phase noise. However, phase noise
is different from this autoranging
noise. Phase noise cannot be avoided
in a spectrum analyzer. However,
Frequency or time
reducing the prefilter width can reduce
autoranging noise at most frequency
Figure 3-3. Autoranging keeps ADC noise close to the carrier and lower than LO noise or RBW filter response
offsets from the carrier. Since the
prefilter width is approximately 2.5
times the RBW, reducing the RBW compute the FFT is greatly reduced, need is total power in a channel
reduces the autoranging noise. speeding up computations. or across a frequency range. In a
measurement such as this, the display
Custom digital signal For swept analysis, the filtered I and points might represent the average
Q pairs are converted to magnitude power during the time the LO sweeps
processing
and phase pairs. For traditional swept through that point. The VBW filter can
Turning back to the block diagram of analysis, the magnitude signal is video- be reconfigured into an accumulator
the digital IF (Figure 3-2), after the ADC bandwidth (VBW) filtered and samples to perform averaging on either a log,
gain has been set with analog gain and are taken through the display detector voltage or power scale.
corrected with digital gain, a custom circuit. The log/linear display selection
IC begins processing the samples. and dB/division scaling occur in the Frequency counting
First, it splits the 30-MHz IF samples processor, so a trace can be displayed
into I and Q pairs at half the rate (15 on any scale without remeasuring. Swept spectrum analyzers usually
Mpairs/s). The I and Q pairs are given have a frequency counter. This counter
a high-frequency boost with a single- counts the zero crossings in the IF
Additional video processing signal and offsets that count by the
stage digital filter that has gain and
phase approximately opposite to that of
features known frequency offsets from LOs in
the single-pole analog prefilter. Next, The VBW filter normally smoothes the the rest of the conversion chain. If the
I and Q signals are low-pass filtered log of the magnitude of the signal, but count is allowed to run for a second,
with a linear-phase filter with nearly it has many additional features. It can you can achieve a resolution of 1 Hz.
ideal Gaussian response. Gaussian convert the log magnitude to a voltage
filters have always been used for envelope before filtering and convert Because of its digitally synthesized
swept spectrum analysis, because of it back for consistent behavior before LOs and all-digital RBWs, the native
their optimum compromise between display detection. frequency accuracy of the X-Series
frequency domain performance (shape signal analyzer is very good (0.1% of
factor) and time-domain performance Filtering the magnitude on a linear span). In addition, the X-Series signal
(response to rapid sweeps). With the voltage scale is desirable for observing analyzer includes a frequency counter
signal bandwidth now reduced, the I pulsed-RF envelope shapes in zero that observes not just zero crossings,
and Q pairs may be decimated and sent span. The log-magnitude signal but also the change in phase. Thus, it
to the processor for FFT processing also can be converted to a power can resolve frequency to the tens-of-
or demodulation. Although FFTs can (magnitude squared) signal before millihertz level in 0.1 second. With this
be performed to cover a segment of filtering, and then it can be converted design, the ability to resolve frequency
frequency span up to the 10-MHz back. Filtering the power allows the changes is not limited by the spectrum
bandwidth of the anti-alias filter, even analyzer to give the same average analyzer, but rather is determined
a narrower FFT span, such as 1 kHz, response to signals with noise- by the noisiness of the signal being
with a narrow RBW, such as 1 Hz, would like characteristics, such as digital counted.
require FFTs with 20 million data points. communications signals, as to CW
Using decimation for narrower spans, signals with the same rms voltage. An
the number of data points needed to increasingly common measurement

38
More advantages of all-digital IF implementations is specified at ± 0.07 prefilter. Again, the prefilter is highly
dB for any level up to –20 dBm at the stable and contributes only 20 percent
We have already discussed a number input mixer of the analyzer. The range of the error that would exist with an
of advantages of signal analyzers with of the log amp does not limit the log RBW made of five such stages. As a
all-digital IF: power/voltage/log video fidelity at low levels, as it would be in result, most RBWs are within 2 percent
filtering, high-resolution frequency an analog IF; the range is only limited of their stated bandwidth, compared
counting, log/linear switching of by noise around –155 dBm at the to 10 to 20 percent specifications in
stored traces, excellent shape factors, input mixer. Because of single-tone analog-IF analyzers.
an average-across-the display-point compression in upstream circuits at
detector mode, 160 RBWs, and of higher powers, the fidelity specification Bandwidth accuracy is important
course, FFT or swept processing. In degrades to ± 0.13 dB for signal levels for minimizing the inaccuracy of
spectrum analysis, the filtering action of down to –10 dBm at the input mixer. channel power measurements and
RBW filters causes errors in frequency By comparison, analog log amps are similar measurements. The noise
and amplitude measurements that are usually specified with tolerances in the bandwidth of the RBW filters is known
a function of the sweep rate. For a fixed ± 1 dB region. to much better specifications than
level of these errors, the all-digital the 2 percent setting tolerance, and
IF’s linear phase RBW filters allow Other IF-related accuracies are noise markers and channel-power
faster sweep rates than analog filters improved as well. The IF prefilter is measurements are corrected to a
permit. The digital implementation analog and must be aligned like an tolerance of ± 0.5 percent. Therefore,
also allows well-known compensations analog filter, so it is subject to alignment bandwidth uncertainties contribute
to frequency and amplitude readout, errors, but it is much better than most only ± 0.022 dB to the amplitude error
permitting sweep rates typically twice analog filters. With only one stage to of noise density and channel-power
as fast as older analyzers and excellent manufacture, that stage can be made measurements.
performance at even four times the much more stable than the 4- and
sweep speed. Keysight X-Series signal 5-stage filters of analog IF-based Finally, with no analog reference-level-
analyzers can achieve over 50 times spectrum analyzers. As a result, the dependent gain stages, there is no “IF
faster sweep speeds (see Chapter 2 - gain variations between RBW filters gain” error at all. The sum of all these
Digital resolution filters). is held to a specification of ± 0.03 dB improvements means that the all-digital
for general digital IF implementations, IF makes a quantum improvement
Digitally implemented logarithmic which is ten times better than all-analog in spectrum analyzer accuracy. It
amplification is very accurate. Typical designs. also allows you to change analyzer
errors of the entire analyzer are settings without significantly impacting
much smaller than the measurement The accuracy of the IF bandwidth is measurement uncertainty. We will cover
uncertainty with which the determined by settability limitations this topic in more detail in the next
manufacturer proves the log fidelity. in the digital part of the filtering and chapter.
The log fidelity on all digital IF calibration uncertainties in the analog

39
Chapter 4. Amplitude and Frequency Accuracy
Now let’s look at amplitude accuracy, Impedance mismatch is an important Spectrum analyzer data sheets
or perhaps better, amplitude factor in measurement uncertainty typically specify the input voltage
uncertainty. Most spectrum analyzers that is often overlooked. Analyzers do standing wave ratio (VSWR). Knowing
are specified in terms of both absolute not have perfect input impedances, the VSWR, we can calculate ρ with the
and relative accuracy. However, relative and signal sources do not have ideal following equation:
performance affects both, so let’s output impedances. When a mismatch
(VSWR–1)
look first at factors affecting relative exists, the incident and reflected signal ρ=
(VSWR+1)
measurement uncertainty. vectors may add constructively or
destructively. Thus the signal received As an example, consider a spectrum
Before we discuss these uncertainties, by the analyzer can be larger or analyzer with an input VSWR of 1.2 and
let’s look again at the block diagram smaller than the original signal. In most a device under test (DUT) with a VSWR
of an analog swept-tuned spectrum cases, uncertainty due to mismatch is of 1.4 at its output port. The resulting
analyzer, shown in Figure 4-1, and see relatively small. However, as spectrum mismatch error would be ±0.13 dB.
which components contribute to the analyzer amplitude accuracy has
uncertainties. Later in this chapter, we improved dramatically in recent
will see how a digital IF and various years, mismatch uncertainty now
correction and calibration techniques constitutes a more significant part of
can substantially reduce measurement total measurement uncertainty. In any More information
uncertainty. case, improving the match of either the For more information about how
source or analyzer reduces uncertainty. improving the match of either
Components that contribute to the source or analyzer reduces
uncertainty: The general expression used to calculate uncertainty, see the Keysight
–– Input connector (mismatch) the maximum mismatch error in dB is: PSA Performance Spectrum Analyzer
–– RF input attenuator Series Amplitude Accuracy – Technical
–– Mixer and input filter (flatness) Error (dB) = –20 log[1 ± |(ρanalyzer)(ρsource)|] Overview literature number 5980-
–– IF gain/attenuation (reference level) 3080EN.
–– RBW filters where ρ is the reflection coefficient.
–– Display scale fidelity
–– Calibrator (not shown)

RF input Log Envelope

attenuator Mixer IF gain IF filter amp detector

Input
signal

Pre-selector, or Video
low-pass filter filter
Local
oscillator

Reference
oscillator

Sweep
generator Display

Figure 4-1. Spectrum analyzer block diagram

40
Since the analyzer’s worst-case match uncertainty. Figure 4-2 illustrates what one frequency and, therefore, do not
occurs when its input attenuator is the frequency response might look contribute to frequency response.
set to 0 dB, we should avoid the 0 like in one frequency band. Frequency However, some amplitude uncertainty
dB setting if we can. Alternatively, response is usually specified as ± x dB is always introduced and it depends
we can attach a well-matched pad relative to the midpoint between the on how accurately the IF amplifier and
(attenuator) to the analyzer input and extremes. The frequency response of attenuator can be set to a desired
greatly reduce mismatch as a factor. a spectrum analyzer represents the value. This uncertainty is known as
Adding attenuation is a technique that overall system performance resulting reference level accuracy.
works well to reduce measurement from the flatness characteristics and
uncertainty when the signal we wish interactions of individual components Another parameter we might change
to measure is well above the noise. in the signal path up to and including during the course of a measurement
However, in cases where the signal- the first mixer. Microwave spectrum is resolution bandwidth. Different
to-noise ratio is small (typically ≤ 7 analyzers use more than one frequency filters have different insertion losses.
dB), adding attenuation will increase band to go above 3.6 GHz. This is done Generally, we see the greatest
measurement error because the by using a higher harmonic of the local difference when switching between
noise power adds to the signal power, oscillator, which will be discussed LC filters (typically used for the wider
resulting in an erroneously high in detail in Chapter 7. When making resolution bandwidths) and crystal
reading. relative measurements between signals filters (used for narrow bandwidths).
in different frequency bands, you must This results in resolution bandwidth
Let’s turn our attention to the add the frequency response of each switching uncertainty.
input attenuator. Some relative band to determine the overall frequency
measurements are made with different response uncertainty. In addition, The most common way to display
attenuator settings. In these cases, we some spectrum analyzers have a band signals on a spectrum analyzer is to use
must consider the input attenuation switching uncertainty which must be a logarithmic amplitude scale, such as
switching uncertainty. Because an RF added to the overall measurement 10 dB per div or 1 dB per div. Therefore,
input attenuator must operate over the uncertainty. the IF signal usually passes through a
entire frequency range of the analyzer, log amplifier. The gain characteristic
its step accuracy varies with frequency. After the input signal is converted to of the log amplifier approximates a
The attenuator also contributes to the an IF, it passes through the IF gain logarithmic curve. So any deviation
overall frequency response. At 1 GHz, amplifier and IF attenuator, which are from a perfect logarithmic response
we expect the attenuator performance adjusted to compensate for changes adds to the amplitude uncertainty.
to be quite good; at 26 GHz, not as in the RF attenuator setting and Similarly, when the spectrum analyzer
good. mixer conversion loss. Input signal is in linear mode, the linear amplifiers
amplitudes are thus referenced to the do not have a perfect linear response.
The next component in the signal path top line of the graticule on the display, This type of uncertainty is called
is the input filter. Spectrum analyzers known as the reference level. The IF display scale fidelity.
use a fixed low-pass filter in the low amplifier and attenuator work only at
band and a tunable bandpass filter
called a preselector (we will discuss the
preselector in more detail in Chapter Frequency response
7) in the higher frequency bands. The Signals in the same harmonic band
low-pass filter has a better frequency
+0.5 dB
response than the preselector and
adds a small amount of uncertainty
to the frequency response error. A
0
preselector, usually a YIG-tuned
filter, has a larger frequency response
variation, ranging from 1.5 dB to 3 dB
at millimeter-wave frequencies. - 0.5 dB
BAND 1
Following the input filter are the Specification: 0.5 dB
mixer and the local oscillator, both of
which add to the frequency response Figure 4-2. Relative frequency response in a single band

41
Relative uncertainty Absolute amplitude accuracy It is best to consider all known
uncertainties and then determine
When we make relative measurements Almost all spectrum analyzers have a which ones can be ignored when
on an incoming signal, we use either built-in calibration source that provides making a certain type of measurement.
some part of the same signal or a a known reference signal of specified The range of values shown in Table
different signal as a reference. For amplitude and frequency. We rely on 4-1 represents the specifications of a
example, when we make second the relative accuracy of the analyzer variety of spectrum analyzers.
harmonic distortion measurements, we to translate the absolute calibration of
use the fundamental of the signal as the reference to other frequencies and Some of the specifications, such as
our reference. Absolute values do not amplitudes. Spectrum analyzers often frequency response, are frequency-
come into play; we are interested only have an absolute frequency response range dependent. A 3-GHz RF analyzer
in how the second harmonic differs in specification, where the zero point might have a frequency response of
amplitude from the fundamental. on the flatness curve is referenced to ± 0.38 dB, while a microwave spectrum
this calibration signal. Many Keysight analyzer tuning in the 26-GHz range
In a worst-case relative measurement spectrum analyzers use a 50-MHz could have a frequency response of
scenario, the fundamental of the reference signal. At this frequency, the ± 2.5 dB or higher. On the other hand,
signal may occur at a point where specified absolute amplitude accuracy other sources of uncertainty, such as
the frequency response is highest, is extremely good: ± 0.24 dB for the changing resolution bandwidths, apply
while the harmonic we wish to X-Series PXA signal analyzer. equally to all frequencies.
measure occurs at the point where
the frequency response is the lowest.
The opposite scenario is equally likely.
Therefore, if our relative frequency Table 4-1. Representative values of amplitude uncertainty for common spectrum analyzers

response specification is ± 0.5 dB, as

shown in Figure 4-2, then the total Amplitude uncertainties (± dB)
uncertainty would be twice that value, Relative
or ± 1.0 dB. RF attenuator switching uncertainty 0.18 to 0.7
Frequency response 0.38 to 2.5
Perhaps the two signals under test
are in different frequency bands of Reference level accuracy (IF attenuator/gain change) 0.0 to 0.7
the spectrum analyzer. In that case, Resolution bandwidth switching uncertainty 0.03 to 1.0
a rigorous analysis of the overall Display scale fidelity 0.07 to 1.15
uncertainty must include the sum of Absolute
the flatness uncertainties of the two
Calibrator accuracy 0.24 to 0.34
frequency bands.

Other uncertainties might be irrelevant

in a relative measurement, like RBW
switching uncertainty or reference level
accuracy, which apply to both signals at
the same time.

42
Improving overall uncertainty in calibrator. Finally, many analyzers During manufacture, all instruments
available today have self-calibration are tested for typical performance
When we look at total measurement routines. These routines generate error parameters.
uncertainty for the first time, we may coefficients (for example, amplitude
well be concerned as we add up the changes versus resolution bandwidth) Nominal values indicate expected
uncertainty figures. The worst-case view that the analyzer later uses to correct performance or describe product
assumes each source of uncertainty measured data. As a result, these self- performance that is useful in the
for your spectrum analyzer is at the calibration routines allow us to make application of the product, but is not
maximum specified value, and all are good amplitude measurements with a covered by the product warranty.
biased in the same direction at the same spectrum analyzer and give us more Nominal parameters generally are
time. The sources of uncertainty can be freedom to change controls during the not tested during the manufacturing
considered independent variables, so it course of a measurement. process.
is likely that some errors will be positive
while others will be negative. Therefore,
a common practice is to calculate the
Specifications, typical Digital IF architecture and
root sum of squares (RSS) error. performance and nominal uncertainties
values As described in the previous chapter,
Regardless of whether we calculate the a digital IF architecture eliminates or
When evaluating spectrum analyzer
worst-case or RSS error, we can take minimizes many of the uncertainties
accuracy, it is important to have a clear
steps to improve the situation. First of experienced in analog spectrum
understanding of the many different
all, we should know the specifications analyzers. These include:
values found on an analyzer data sheet.
for our particular spectrum analyzer.
Keysight defines three classes of
These specifications may be good
enough over the range in which we are
instrument performance data: Reference level accuracy (IF
making our measurement. If not, Table gain uncertainty)
Specifications describe the
4-1 suggests some opportunities to Spectrum analyzers with an all-digital
performance of parameters covered
improve accuracy. IF, such as the Keysight X-Series, do
by the product warranty over a
temperature range of 0 to 55 °C (unless not have IF gain that changes with
Before taking any data, we can step reference level. Therefore, there is no
otherwise noted). Each instrument
through a measurement to see if any IF gain uncertainty.
is tested to verify it meets the
controls can be left unchanged. We
specification and takes into account
might find that the measurement
can be made without changing the
the measurement uncertainty of the Display scale fidelity
equipment used to test the instrument.
RF attenuator setting, resolution A digital IF architecture does not
All of the units tested will meet the
bandwidth or reference level. If so, include a log amplifier. Instead, the log
specification.
all uncertainties associated with function is performed mathematically,
changing these controls drop out. We and traditional log fidelity uncertainty
Some test equipment manufacturers
may be able to trade off reference does not exist. However, other factors,
use a “2 sigma” or 95% confidence
level accuracy against display fidelity, such as RF compression (especially
value for certain instrument
using whichever is more accurate and for input signals above –20 dBm),
specifications. When evaluating data
eliminating the other as an uncertainty ADC range gain alignment accuracy
sheet specifications for instruments
factor. We can even get around and ADC linearity (or quantization
from different manufacturers, it
frequency response if we are willing error) contribute to display scale
is important to make sure you are
to go to the trouble of characterizing uncertainty. The quantization error
comparing like numbers in order to
our particular analyzer 2. You can can be improved by the addition of
make an accurate comparison.
accomplish this by using a power noise, which smoothes the average of
meter and comparing the reading of the ADC transfer function. This added
Typical performance describes
the spectrum analyzer at the desired noise is called dither. While the dither
additional product performance
frequencies with the reading of the improves linearity, it does slightly
information that is not covered by the
power meter. degrade the displayed average noise
product warranty. It is performance
level. In the X-Series signal analyzers,
beyond specification that 80% of the
The same applies to the calibrator. If we we generally recommend you use
units exhibit with a 95% confidence
have a more accurate calibrator, or one dither when the measured signal has
level over the temperature range 20
closer to the frequency of interest, we a signal-to-noise ratio of greater than
to 30 °C. Typical performance does
may wish to use that in lieu of the built- or equal to 10 dB. When the signal-
not include measurement uncertainty.
to-noise ratio is less than 10 dB, the
2.  Should we do so, then mismatch may become a more significant error. degradations to accuracy of any single

43
measurement (in other words, without Table 4-2. Amplitude uncertainties when measuring a 1-GHz signal
averaging) that come from a higher
noise floor are worse than the linearity Source of uncertainty Absolute uncertainty of 1-GHz, –20-dBm signal
problems solved by adding dither, so   N9030A PXA N9020A MXA N9010A EXA
dither is best turned off. Absolute amplitude accuracy ± 0.24 dB ± 0.33 dB ± 0.40 dB
Frequency response ± 0.35 dB ± 0.45 dB ± 0.60 dB
RBW switching uncertainty
Total worst-case uncertainty ± 0.59 dB ± 0.78 dB ± 1.00 dB
The digital IF in the X-Series signal Total RSS uncertainty ± 0.42 dB ± 0.56 dB ± 0.72 dB
analyzers includes an analog prefilter
set to 2.5 times the desired resolution
bandwidth. This prefilter has some
Table 4-3. Absolute and relative amplitude accuracy comparison (8563EC and N9030A PXA)
uncertainty in bandwidth, gain and
center frequency as a function of the
Source of uncertainty Measurement of a 10-GHz signal at –10 dBm    
RBW setting. The rest of the RBW
filtering is done digitally in an ASIC in   Absolute uncertainty of Relative uncertainty of second
the digital IF section. Though the digital fundamental at 10 GHz harmonic at 20 GHz
filters are not perfect, they are very   8563EC N9030A PXA 8563EC N9030A PXA
repeatable, and some compensation Calibrator ± 0.3 dB N/A N/A N/A
is applied to minimize the error.
Absolute amplitude N/A ± 0.24 dB N/A N/A
This results in a tremendous overall
accuracy
improvement to the RBW switching
uncertainty compared to analog Attenuator N/A N/A N/A N/A
implementations. Frequency response ± 2.9 dB ± 2.0 dB ± (2.2 + 2.5) dB ± (2.0 + 2.0) dB
Band switching uncer- N/A N/A ± 1.0 dB N/A
Amplitude uncertainty tainty
examples IF gain N/A N/A N/A N/A

Let’s look at some amplitude RBW switching N/A ± 0.03 dB N/A N/A
uncertainty examples for various Display scale fidelity N/A ± 0.07 dB ± 0.85 dB ± 0.07 dB
measurements. Suppose we want to Total worst-case ± 3.20 dB ± 2.34 dB ± 6.55 dB ± 4.07 dB
measure a 1-GHz RF signal with an uncertainty
amplitude of –20 dBm. If we use a
Total RSS uncertainty ± 2.91 dB ± 2.02 dB ± 3.17 dB ± 2.83 dB
Keysight PXA X-Series signal analyzer
with Atten = 10 dB, RBW = 1 kHz, VBW
= 1 kHz, Span = 20 kHz, Ref level = 4-3, we compare the absolute and hand, are useful for discovering the
–20 dBm, log scale, and coupled sweep relative amplitude uncertainty of two distance between spectral components
time, and an ambient temperature of different Keysight spectrum and signal or the modulation frequency.
20 to 30 °C, the specifications tell us analyzers, an 8563EC (with analog IF)
that the absolute uncertainty equals and N9030A PXA (with digital IF). Up until the late 1970s, absolute
± 0.24 dB plus the absolute frequency frequency uncertainty was measured
response. The MXA X-Series signal Frequency accuracy in megahertz because the first LO was
analyzer measuring the same signal a high-frequency oscillator operating
using the same settings would have a So far, we have focused almost above the RF range of the analyzer, and
specified uncertainty of ± 0.33 plus the exclusively on amplitude there was no attempt to tie the LO to
absolute frequency response. These measurements. What about frequency a more accurate reference oscillator.
values are summarized in Table 4-2. measurements? Again, we can classify Today’s LOs are synthesized to provide
two broad categories, absolute and better accuracy. Absolute frequency
At higher frequencies, the uncertainties relative frequency measurements. uncertainty is often described under
get larger. In this example, we want Absolute measurements are used to the frequency readout accuracy
to measure a 10-GHz signal with an measure the frequencies of specific specification and refers to center
amplitude of –10 dBm. In addition, signals. For example, we might want frequency, start, stop and marker
we also want to measure its second to measure a radio broadcast signal frequencies.
harmonic at 20 GHz. Assume the to verify it is operating at its assigned
following measurement conditions: frequency. Absolute measurements are With the introduction of the Keysight
0 to 55 °C, RBW = 300 kHz, Atten = also used to analyze undesired signals, 8568A in 1977, counter-like frequency
10 dB, Ref level = –10 dBm. In Table such as when you search for spurs. accuracy became available in a general-
Relative measurements, on the other

44
purpose spectrum analyzer, and are a number of factors to consider on the display and the number of
ovenized oscillators were used to reduce before we can determine frequency display points selected. Also, to get the
drift. Over the years, crystal reference uncertainty. best frequency accuracy, we must be
oscillators with various forms of indirect careful to place the marker exactly at
synthesis have been added to analyzers In a factory setting, there is often an the peak of the response to a spectral
in all cost ranges. The broadest in-house frequency standard available component. If we place the marker at
definition of indirect synthesis is that the that is traceable to a national standard. some other point on the response, we
frequency of the oscillator in question is Most analyzers with internal reference will get a different frequency reading.
in some way determined by a reference oscillators allow you to use an external For the best accuracy, we may narrow
oscillator. This includes techniques such reference. The frequency reference the span and resolution bandwidth to
as phase lock, frequency discrimination error in the foregoing expression then minimize their effects and to make it
and counter lock. becomes the error of the in-house easier to place the marker at the peak
standard. of the response.
What we care about is the effect
these changes have had on frequency When you make relative Many analyzers have marker modes
accuracy (and drift). A typical readout measurements, span accuracy comes that include internal counter schemes
accuracy might be stated: into play. For Keysight analyzers, to eliminate the effects of span and
span accuracy generally means the resolution bandwidth on frequency
± [(freq readout x freq ref error) + A% of uncertainty in the indicated separation accuracy. The counter does not count
span + B% of RBW + C Hz] of any two spectral components on the the input signal directly, but instead
display. For example, suppose span counts the IF signal and perhaps one
Note that we cannot determine an accuracy is 0.5% of span and we have or more of the LOs, and the processor
exact frequency error unless we two signals separated by two divisions computes the frequency of the input
know something about the frequency in a 1-MHz span (100 kHz per division). signal. A minimum signal-to-noise
reference. In most cases, we are The uncertainty of the signal separation ratio is required to eliminate noise as
given an annual aging rate, such as ± would be 5 kHz. The uncertainty would a factor in the count. Counting the
1 x 10 –7 per year, though sometimes be the same if we used delta markers signal in the IF also eliminates the need
aging is given over a shorter period and the delta reading was 200 kHz. So to place the marker at the exact peak
(for example, ± 5 x 10 –10 per day). we would measure 200 kHz ± 5 kHz. of the signal response on the display.
In addition, we need to know when If you are using this marker counter
the oscillator was last adjusted and When making measurements in the function, placement anywhere near the
how close it was set to its nominal field, we typically want to turn our peak of the signal sufficiently out of the
frequency (usually 10 MHz). Other analyzer on, complete our task, and noise will do. Marker count accuracy
factors that we often overlook when we move on as quickly as possible. It is might be stated as:
think about frequency accuracy include helpful to know how the reference
how long the reference oscillator has in our analyzer behaves under short ± [(marker freq x freq ref error)
been operating. Many oscillators take warm-up conditions. For example, + counter resolution]
24 to 72 hours to reach their specified the Keysight ESA-E Series portable
drift rate. To minimize this effect, some spectrum analyzers will meet published We must still deal with the frequency
spectrum analyzers continue to provide specifications after a 5-minute warm reference error, as we previously
power to the reference oscillator as up. discussed. Counter resolution refers to
long as the instrument is plugged the least-significant digit in the counter
into the AC power line. In this case, Most analyzers offer markers you can readout, a factor here just as with any
the instrument is not really turned put on a signal to see amplitude and simple digital counter. Some analyzers
“off.” It is more accurate to say it is on absolute frequency. However, the allow you to use the counter mode with
“standby.” We also need to consider indicated frequency of the marker is a delta markers. In that case, the effects
the temperature stability, as it can be function of the frequency calibration of of counter resolution and the fixed
worse than the drift rate. In short, there the display, the location of the marker frequency would be doubled.

45
Chapter 5. Sensitivity and Noise
Sensitivity typically high enough in amplitude that when it is well above the effective
the noise generated in subsequent (displayed) noise floor. The effective
One of the primary ways engineers use gain stages adds only a small amount input noise floor includes the losses
spectrum analyzers is for searching to the total noise power. The input caused by the input attenuator, mixer
out and measuring low-level signals. attenuator and one or more mixers may conversion loss, and other circuit
The limitation in these measurements be between the input connector of a elements prior to the first gain stage.
is the noise generated within the spectrum analyzer and the first stage We cannot do anything about the
spectrum analyzer itself. This noise, of gain, and all of these components conversion loss of the mixers, but we
generated by the random electron generate noise. However, the noise can change the RF input attenuator.
motion in various circuit elements, is they generate is at or near the absolute This enables us to control the input
amplified by multiple gain stages in the minimum of –174 dBm/Hz, so they do signal power to the first mixer and thus
analyzer and appears on the display as not significantly affect the noise level change the displayed signal-to-noise
a noise signal. On a spectrum analyzer, input to the first gain stage, and its floor ratio. Clearly, we get the lowest
this noise is commonly referred to as amplification is typically insignificant. DANL by selecting minimum (zero) RF
the displayed average noise level, or attenuation.
DANL1. The noise power observed in While the input attenuator, mixer
the DANL is a combination of thermal and other circuit elements between Because the input attenuator has no
noise and the noise figure of the the input connector and first gain effect on the actual noise generated
spectrum analyzer. While there are stage have little effect on the actual in the system, some early spectrum
techniques to measure signals slightly system noise, they do have a marked analyzers simply left the displayed
below the DANL, this noise power effect on the ability of an analyzer to noise at the same position on the
ultimately limits our ability to make display low-level signals because they display regardless of the input
measurements of low-level signals. attenuate the input signal. That is, they attenuator setting. That is, the IF gain
reduce the signal-to-noise ratio and so remained constant. In this case, the
Let’s assume a 50-ohm termination degrade sensitivity. input attenuator affected the location
is attached to the spectrum analyzer of a true input signal on the display.
input to prevent any unwanted signals We can determine the DANL simply by As input attenuation was increased,
from entering the analyzer. This passive noting the noise level indicated on the further attenuating the input signal,
termination generates a small amount display when the spectrum analyzer the location of the signal on the display
of noise energy equal to kTB, where: input is terminated with a 50-ohm load. went down while the noise remained
This level is the spectrum analyzer’s stationary.
k = Boltzmann’s constant own noise floor. Signals below this
(1.38 x 10 –23 joule/K) level are masked by the noise and
T = temperature, in Kelvin cannot be seen. However, the DANL is
B = bandwidth in which the noise is not the actual noise level at the input,
measured, in Hertz but rather the effective noise level. An
analyzer display is calibrated to reflect
The total noise power is a function the level of a signal at the analyzer
of measurement bandwidth, so the input, so the displayed noise floor
value is typically normalized to a represents a fictitious or effective noise
1-Hz bandwidth. Therefore, at room floor at the input.
temperature, the noise power density is
–174 dBm/Hz. When this noise reaches The actual noise level at the input is
the first gain stage in the analyzer, the a function of the input signal. Indeed,
amplifier boosts the noise, plus adds noise is sometimes the signal of
some of its own. As the noise signal interest. Like any discrete signal, a
passes on through the system, it is noise signal is much easier to measure

1.  Displayed average noise level is sometimes confused with the term “sensitivity.” While related, these terms have different meanings. Sensitivity is a
measure of the minimum signal level that yields a defined signal-to-noise ratio (SNR) or bit error rate (BER). It is a common metric of radio receiver
performance. Spectrum analyzer specifications are always given in terms of the DANL.

46
Beginning in the late 1970s, spectrum
analyzer designers took a different
approach. In newer analyzers, an
internal microprocessor changes the
IF gain to offset changes in the input
attenuator. Thus, signals present at
the analyzer’s input remain stationary
on the display as we change the input
attenuator, while the displayed noise
moves up and down. In this case, the
reference level remains unchanged, as
shown in Figure 5-1. As the attenuation
increases from 5 to 15 to 25 dB, the
displayed noise rises while the –30-
dBm signal remains constant. In either
case, we get the best signal-to-noise
ratio by selecting minimum input
attenuation.

Resolution bandwidth also affects

signal-to-noise ratio, or sensitivity.
The noise generated in the analyzer is
Figure 5-1. In modern signal analyzers, reference levels remain constant when you change input attenuation
random and has a constant amplitude
over a wide frequency range. Since
the resolution, or IF, bandwidth filters
come after the first gain stage, the
total noise power that passes through
the filters is determined by the width of
the filters. This noise signal is detected
and ultimately reaches the display.
The random nature of the noise signal
causes the displayed level to vary as:

10 log (BW2 /BW1)

where
BW1 = starting resolution bandwidth
BW = ending resolution bandwidth
2

So if we change the resolution

bandwidth by a factor of 10, the
displayed noise level changes by 10 dB,
as shown in Figure 5-2. For continuous
wave (CW) signals, we get best signal-
to-noise ratio, or best sensitivity, using
the minimum resolution bandwidth
available in our spectrum analyzer 2. Figure 5-2. Displayed noise level changes as 10 log (BW2 /BW1 )

2.  Broadband, pulsed signals can exhibit the opposite behavior, where the SNR increases as the bandwidth gets larger.

47
A spectrum analyzer displays signal
plus noise, and a low signal-to-noise
ratio makes the signal difficult to
distinguish. We noted previously that
the video filter can be used to reduce
the amplitude fluctuations of noisy
signals without affecting constant
signals. Figure 5-3 shows how the
video filter can improve our ability to
discern low-level signals. The video
filter does not affect the average
noise level and so does not, by this
definition, affect the sensitivity of an
analyzer.

In summary, we get best sensitivity for

narrowband signals by selecting the
minimum resolution bandwidth and
minimum input attenuation. These
settings give us the best signal-to-noise
ratio. We can also select minimum video
bandwidth to help us see a signal at
or close to the noise level3. Of course,
selecting narrow resolution and video
bandwidths does lengthen the sweep
time.

Noise floor extension

While lowering an analyzer’s inherent
noise floor through hardware design
and component choices is obviously
beneficial for dynamic range, there are
practical limits, and another approach
offers significant improvement.
With sufficient processing and other
technical innovations, the noise power
in a signal analyzer can be modeled
and subtracted from measurement
results to reduce the effective noise
level. In the Keysight PXA signal
analyzer this operation is called noise
floor extension (NFE).

Generally, if you can accurately

Figure 5-3. Video filtering makes low-level signals more discernible
identify the noise power contribution
of an analyzer, you can subtract this
time-domain displays of signals. available in the Keysight X-Series signal
power from various kinds of spectrum
analyzers.) This capability is effective,
measurements. Examples include
Keysight has been demonstrating though somewhat inconvenient. It
signal power or band power, ACPR,
noise subtraction capability for some involves disconnecting the signal from
spurious, phase noise, harmonic and
time, using trace math in vector the analyzer, measuring analyzer noise
intermodulation distortion. Noise
signal analyzers to remove analyzer level with a large amount of averaging,
subtraction techniques do not improve
noise from spectrum and band power reconnecting the signal and using trace
the performance of vector analysis
measurements. (Similar trace math is math to display a corrected result. It is
operations such as demodulation or
3.  For the effect of noise on accuracy, see “Dynamic range versus measurement uncertainty” in Chapter 6.

48
necessary to re-measure the analyzer
noise power every time the analyzer
configuration (frequency center/span,
attenuator/input range, resolution
bandwidth) changed.

The Keysight PXA analyzers

dramatically improve this measurement
technique for many measurement
situations. Critical parameters that
determine the analyzer’s noise floor
are measured when it is calibrated,
and these parameters are used (with
current measurement information such
as analyzer temperature) to fully model
the analyzer’s noise floor, including
changes in analyzer configuration
and operating conditions. The
analyzer’s noise power contribution
is then automatically subtracted from
spectrum and power measurements.
This process in the PXA is called noise
Figure 5-4. Noise floor extension view of harmonics
floor extension and is enabled with a
keystroke in the Mode Setup menu. An
VBW, and it also works with any type of the gain of the analyzer. Second, the
example is shown in Figure 5-4.
detector or averaging. gain of our analyzer is unity because
the signal level at the output (indicated
The effectiveness of NFE can be
Noise figure on the display) is the same as the
expressed in several ways. Average
level at the input (input connector).
noise power in the display (DANL) is Many receiver manufacturers specify So our expression, after substitution,
usually reduced by 10 to 12 dB in the the performance of their receivers cancellation and rearrangement,
analyzer’s low band (below 3.6 GHz) in terms of noise figure, rather than becomes:
and about 8 dB in its high band sensitivity. We will show you how
(above 3.6 GHz). While the apparent the two can be equated. A spectrum F = No/Ni
noise level will be reduced, only analyzer is a receiver, and we will
the analyzer’s noise power is being examine noise figure on the basis of a This expression tells us that all we need
subtracted. Therefore, the apparent sinusoidal input. to do to determine the noise figure is
power of signals in the display will be
compare the noise level as read on the
reduced if the analyzer’s noise power Noise figure can be defined as the display to the true (not the effective)
is a significant part of their power, and degradation of signal-to-noise ratio noise level at the input connector.
not otherwise. as a signal passes through a device, a Noise figure is usually expressed in
spectrum analyzer in our case. We can terms of dB, or:
Thus measurements of both discrete express noise figure as:
signals and the noise floor of signal
Si /Ni NF = 10 log(F) = 10 log(No) – 10 log(Ni).
sources connected to the PXA are F =
more accurately measured with NFE So/No
We use the true noise level at the input,
enabled. NFE works with all spectrum where rather than the effective noise level,
measurements regardless of RBW or F = noise figure as power ratio (also because our input signal-to-noise ratio
known as noise factor) was based on the true noise. As we saw
Si = input signal power earlier, when the input is terminated in
More information Ni = true input noise power 50 ohms, the kTB noise level at room
So = output signal power temperature in a 1-Hz bandwidth is
For more information on using
No = output noise power –174 dBm.
noise floor extension, please refer
to, Using Noise Floor Extension in the
We can simplify this expression for We know the displayed level of noise on
PXA Signal Analyzer – Application
our spectrum analyzer. First of all, the the analyzer changes with bandwidth.
Note, literature number 5990-
output signal is the input signal times So all we need to do to determine the
5340EN.

49
noise figure of our spectrum analyzer lower the noise figure, we also improve Testing by experiment means we must
is to measure the noise power in some the system sensitivity. have the equipment at hand. We do
bandwidth, calculate the noise power not need to worry about numbers. We
that we would have measured in a 1-Hz When we introduced noise figure in the simply connect the preamplifier to the
bandwidth using 10 log(BW2 /BW1), and previous discussion, we did so on analyzer, note the average displayed
compare that to –174 dBm. the basis of a sinusoidal input signal. noise level and subtract the gain of
We can examine the benefits of the preamplifier. Then we have the
For example, if we measured –110 dBm a preamplifier on the same basis. sensitivity of the system.
in a 10-kHz resolution bandwidth, we However, a preamplifier also amplifies
would get: noise, and this output noise can be However, we really want to know ahead
higher than the effective input noise of of time what a preamplifier will do for
NF = [measured noise in dBm] – the analyzer. In the “Noise as a signal” us. We can state the two cases above
10 log(RBW/1) – kTBB=1 Hz section later in this chapter, you will as follows:
–110 dBm –10 log(10,000/1) – (–174 dBm) see how a spectrum analyzer using log
power averaging displays a random If NFpre + Gpre ≥ NFSA + 15 dB,
–110 – 40 + 174
= 24 dB noise signal 2.5 dB below its actual Then NFsys = NFpre – 2.5 dB
value. As we explore preamplifiers,
Noise figure is independent of we shall account for this 2.5 dB factor And
bandwidth4. Had we selected a where appropriate.
different resolution bandwidth, our If NFpre + Gpre ≤ NFSA – 10 dB,
results would have been exactly the Rather than develop a lot of formulas Then NFsys = NFSA – Gpre
same. For example, had we chosen to see what benefit we get from a
a 1-kHz resolution bandwidth, the preamplifier, let us look at two extreme Using these expressions, we’ll see how
measured noise would have been –120 cases and see when each might apply. a preamplifier affects our sensitivity.
dBm and 10 log(RBW/1) would have First, if the noise power out of the Assume that our spectrum analyzer
been 30. Combining all terms would preamplifier (in a bandwidth equal to has a noise figure of 24 dB and the
have given –120 – 30 + 174 = 24 dB, the that of the spectrum analyzer) is at preamplifier has a gain of 36 dB and a
same noise figure as above. least 15 dB higher than the DANL of the noise figure of 8 dB. All we need to do
spectrum analyzer, then the sensitivity is to compare the gain plus noise figure
The 24- dB noise figure in our example of the system is approximately that of the preamplifier to the noise figure
tells us that a sinusoidal signal must of the preamplifier, less 2.5 dB. How of the spectrum analyzer. The gain plus
be 24  dB above kTB to be equal to the can we tell if this is the case? Simply noise figure of the preamplifier is 44
displayed average noise level on this connect the preamplifier to the dB, more than 15 dB higher than the
particular analyzer. Thus we can use analyzer and note what happens to noise figure of the spectrum analyzer,
noise figure to determine the DANL the noise on the display. If it goes up so the sensitivity of the preamplifier/
for a given bandwidth or to compare 15 dB or more, we have fulfilled this spectrum-analyzer combination is
DANLs of different analyzers with the requirement. that of the preamplifier, less 2.5 dB.
same bandwidth.5 In a 10 kHz resolution bandwidth, our
On the other hand, if the noise power preamplifier/analyzer system has a
out of the preamplifier (again, in sensitivity (displayed average noise
Preamplifiers level, DANL) of:
the same bandwidth as that of the
One reason for introducing noise figure spectrum analyzer) is 10 dB or more
is that it helps us determine how much lower than the displayed average kTBB=1 + 10log(NBW/1Hz) + NFSYS +
benefit we can derive from the use of noise level on the analyzer, the noise LogCorrectionFactor
a preamplifier. A 24-dB noise figure, figure of the system is that of the
while good for a spectrum analyzer, is spectrum analyzer less the gain of In this expression, kTB = −174 dBm/
not so good for a dedicated receiver. the preamplifier. Again we can test by Hz, so kTBB=1 is −174 dBm. The
However, by placing an appropriate inspection. Connect the preamplifier noise bandwidth (NBW) for typical
preamplifier in front of the spectrum to the analyzer; if the displayed noise digital RBW’s is 0.2 dB wider than
analyzer, we can obtain a system does not change, we have fulfilled the the RBW, thus 40.2 dB. The noise
(preamplifier/spectrum analyzer) noise requirement. figure of the system is 8 dB. The
figure lower than that of the spectrum LogCorrectionFactor is −2.5 dB. So the
analyzer alone. To the extent that we sensitivity is −128.3 dBm.

4.  This may not always be precisely true for a given analyzer because of the way resolution bandwidth filter sections and gain are distributed in the
IF chain.
5.  The noise figure computed in this manner cannot be directly compared to that of a receiver because the “measured noise” term in the equation
understates the actual noise by 2.5 dB. See the section titled “Noise as a signal” later in this chapter.

50
This is an improvement of 18.3 dB the analyzer’s own noise floor, even after To choose the correct preamplifier, we
over the –110 dBm noise floor without accounting for the 2.5 dB factor. It is must look at our measurement needs.
the preamplifier. from this higher noise level that we now If we want absolutely the best
subtract the gain of the preamplifier. sensitivity and are not concerned about
However, there might be a drawback to With the preamplifier in place, our measurement range, we would choose a
using this preamplifier, depending upon measurement range is 92.5 dB, 17.5 dB high-gain, low-noise-figure preamplifier
our ultimate measurement objective. If less than without the preamplifier. The so that our system would take on the
we want the best sensitivity but no loss loss in measurement range equals the noise figure of the preamplifier, less
of measurement range, this preamplifier change in the displayed noise when the 2.5 dB. If we want better sensitivity
is not the right choice. Figure 5-5 preamplifier is connected. but cannot afford to give up any
illustrates this point. A spectrum measurement range, we must choose a
analyzer with a 24-dB noise figure will Finding a preamplifier that will give us lower-gain preamplifier.
have an average displayed noise level better sensitivity without costing us
of –110 dBm in a 10-kHz resolution measurement range dictates that we Interestingly enough, we can use the
bandwidth. If the 1-dB compression must meet the second of the above input attenuator of the spectrum analyzer
point 6 for that analyzer is 0 dBm, the criteria; that is, the sum of its gain and to effectively degrade the noise figure
measurement range is 110 dB. When noise figure must be at least 10 dB less (or reduce the gain of the preamplifier,
we connect the preamplifier, we must than the noise figure of the spectrum if you prefer). For example, if we need
reduce the maximum input to the analyzer. In this case, the displayed slightly better sensitivity but cannot
system by the gain of the preamplifier noise floor will not change noticeably afford to give up any measurement
to –36 dBm. However, when we connect when we connect the preamplifier, range, we can use the above
the preamplifier, the displayed average so although we shift the whole preamplifier with 30 dB of RF input
noise level will rise by about 17.5 dB measurement range down by the gain attenuation on the spectrum analyzer.
because the noise power out of the of the preamplifier, we end up with the This attenuation increases the noise
preamplifier is that much higher than same overall range we started with. figure of the analyzer from 24 to 54

Spectrum analyzer Spectrum analyzer and preamplifier

1 dB compression
0 dBm
Gpre
System 1 dB compression
–36 dBm

110 dB spectrum
analyzer range
92.5 dB
system
range DANL
–92.5 dBm
DANL
–110 dBm Gpre
System sensitivity
–128.5 dBm

Figure 5-5. If displayed noise goes up when a preamplifier is connected, measurement range is diminished by the amount the noise changes

6.  See the section titled “Mixer compression” in Chapter 6.

51
dB. Now the gain plus noise figure of
the preamplifier (36 + 8) is 10 dB less
than the noise figure of the analyzer,
and we have met the conditions of the NFSA – Gpre + 3 dB NFpre + 3 dB
second criterion above.
NFSA – Gpre + 2 dB NFpre + 2 dB
The noise figure of the system is now: System noise
NFSA – Gpre + 1 dB NFpre + 1 dB
Figure (dB)
NFsys = NFSA – GPRE
NFSA – Gpre NFpre
= 54 dB – 36  dB
= 18 dB
NFpre – 1 dB
This represents a 6-dB improvement
NFpre – 2 dB
over the noise figure of the analyzer
alone with 0 dB of input attenuation. So NFpre – 2.5 dB
–10 –5 0 +5 +10
we have improved sensitivity by 6 dB NFpre + Gpre – NFSA (dB)
and given up virtually no measurement
range.
Figure 5-6. System noise figure for sinusoidal signals
Of course, there are preamplifiers that
fall in between the extremes. Figure would the system noise figure be if spectrum analyzer combination is
5-6 enables us to determine system we add a Keysight 8447D amplifier, calibrated as a system, and amplitude
noise figure from a knowledge of the a preamplifier with a noise figure of values displayed on screen are already
noise figures of the spectrum analyzer about 8 dB and a gain of 26 dB? First, corrected for proper readout. With
and preamplifier and the gain of the NFPRE + GPRE – NFSA is +10 dB. From an external preamplifier, you must
amplifier. We enter the graph of Figure the graph of Figure 5-6 we find a correct the spectrum analyzer reading
5-6 by determining NFPRE + GPRE – system noise figure of about NFPRE – with a reference level offset equal
NFSA . If the value is less than zero, we 1.8 dB, or about 8 – 1.8 = 6.2 dB. The to the preamp gain. Most modern
find the corresponding point on the graph accounts for the 2.5-dB factor. spectrum analyzers allow you to
dashed curve and read system noise On the other hand, if the gain of the enter the gain value of the external
figure as the left ordinate in terms preamplifier is just 10 dB, then NFPRE preamplifier from the front panel. The
of dB above NFSA – GPRE. If NFPRE + + GPRE – NFSA is –6 dB. This time the analyzer then applies this gain offset
GPRE – NFSA is a positive value, we find graph indicates a system noise figure to the displayed reference level value,
the corresponding point on the solid of NFSA – GPRE + 0.6 dB, or 24 – 10 + so you can directly view corrected
curve and read system noise figure as 0.6 = 14.6 dB. (We did not introduce measurements on the display.
the right ordinate in terms of dB above the 2.5-dB factor previously when
NFPRE. we determined the noise figure of
the analyzer alone because we read
Let’s first test the two previous extreme the measured noise directly from the
cases.
More information
display. The displayed noise included
the 2.5-dB factor.) For more details on noise
As NFPRE + GPRE – NFSA becomes less figure, see Fundamentals of
than –10 dB, we find that system noise Many modern spectrum analyzers RF and Microwave Noise Figure
figure asymptotically approaches have optional built-in preamplifiers Measurements – Application Note,
NFSA – GPRE. As the value becomes available. Compared to external literature number
greater than +15 dB, system noise preamplifiers, built-in preamplifiers 5952-8255E.
figure asymptotically approaches simplify measurement setups and
NFPRE less 2.5 dB. eliminate the need for additional
cabling. Measuring signal amplitude is
Next, let’s try two numerical examples. much more convenient with a built-in
Above, we determined that the noise preamplifier, because the preamplifier/
figure of our analyzer is 24 dB. What

52
Noise as a signal Figure 5-7.
Random noise has
So far, we have focused on the noise a Gaussian ampli-
tude distribution
generated within the measurement
system (analyzer or analyzer/
preamplifier). We described how the
measurement system’s displayed
average noise level limits the overall
sensitivity. However, random noise
is sometimes the signal we want to
measure. Because of the nature of
noise, the superheterodyne spectrum Figure 5-8. The
analyzer indicates a value that is lower envelope of band-
than the actual value of the noise. Let’s limited Gaussian
noise has a Ray-
see why this is so and how we can leigh distribution
correct for it.

By random noise, we mean a signal

whose instantaneous amplitude has
a Gaussian distribution versus time,
as shown in Figure 5-7. For example,
thermal or Johnson noise has this
characteristic. Such a signal has no
discrete spectral components, so the mean value, we use video filtering or of signal amplitude, so the higher
we cannot select some particular averaging. The mean value of a Rayleigh noise values are not amplified as much
component and measure it to get an distribution is 1.253 σ. as the lower values. As a result, the
indication of signal strength. In fact, we output of the envelope detector is a
must define what we mean by signal However, our analyzer is a peak- skewed Rayleigh distribution, and the
strength. If we sample the signal at an responding voltmeter calibrated to mean value that we get from video
arbitrary instant, we could theoretically indicate the rms value of a sine wave. filtering or averaging is another 1.45
get any amplitude value. We need To convert from peak to rms, our dB lower. In the log mode, then, the
some measure that expresses the analyzer scales its readout by 0.707 (–3 mean or average noise is displayed 2.5
noise level averaged over time. Power, dB). The mean value of the Rayleigh- dB too low. Again, this error is not an
which is of course proportionate to rms distributed noise is scaled by the same ambiguity, and we can correct for it 7.
voltage, satisfies that requirement. factor, giving us a reading of 0.886 σ
(l.05 dB below σ). To equate the mean This is the 2.5-dB factor we accounted
We have already seen that both video value displayed by the analyzer to the for in the previous preamplifier
filtering and video averaging reduce rms voltage of the input noise signal, discussion, when the noise power out
the peak-to-peak fluctuations of a we must account for the error in the of the preamplifier was approximately
signal and can give us a steady value. displayed value. Note, however, that equal to or greater than the analyzer’s
We must equate this value to either the error is not an ambiguity; it is a own noise.
power or rms voltage. The rms value constant error that we can correct for
of a Gaussian distribution equals its by adding 1.05 dB to the displayed Another factor that affects noise
standard deviation, σ. value. measurements is the bandwidth in
which the measurement is made. We
Let’s start with our analyzer in the linear In most spectrum analyzers, the have seen how changing resolution
display mode. The Gaussian noise at the display scale (log or linear in voltage) bandwidth affects the displayed level
input is band limited as it passes through controls the scale on which the noise of the analyzer’s internally generated
the IF chain, and its envelope takes on distribution is averaged with either noise. Bandwidth affects external
a Rayleigh distribution (Figure 5-8). The the VBW filter or with trace averaging. noise signals in the same way. To
noise we see on our analyzer display, Normally, we use our analyzer in the compare measurements made on
the output of the envelope detector, is log display mode, and this mode adds different analyzers, we must know the
the Rayleigh-distributed envelope of the to the error in our noise measurement. bandwidths used in each case.
input noise signal. To get a steady value, The gain of a log amplifier is a function

7.  In X-Series analyzers, the averaging can be set to video, voltage or power (rms), independent of display scale. When using power averaging, no cor-
rection is needed, since the average rms level is determined by the square of the magnitude of the signal, not by the log or envelope of the voltage.

53
Not only does the 3-dB (or 6-dB)
bandwidth of the analyzer affect the
measured noise level, the shape of the
NF SA – G pre + 3 dB NF pre + 3 dB
resolution filter also plays a role. To
make comparisons possible, we define System noise
a standard noise-power bandwidth: the Figure (dB) NF SA – G pre + 2 dB NF pre + 2 dB
width of a rectangular filter that passes
the same noise power as our analyzer’s NF SA – G pre + 1 dB NF pre + 1 dB
filter. For the near-Gaussian filters
in Keysight analyzers, the equivalent NF SA – G pre NF pre
noise-power bandwidth is about 1.05 –10 –5 0 +5 +10
to 1.13 times the 3-dB bandwidth, NF pre + G pre – NF SA (dB)
depending on bandwidth selectivity.
For example, a 10-kHz resolution Figure 5-9. System noise figure for noise signals
bandwidth filter has a noise-power
bandwidth in the range of 10.5 to
11.3 kHz. Power (rms voltage) averaging: But noise is displayed 2.5 dB too low,
Power distribution: 0.00 dB so an input noise signal must be 2.5 dB
If we use 10 log(BW2 /BW1) to adjust 3-dB/noise power bandwidths: –0.50 dB above the analyzer’s displayed noise
the displayed noise level to what we Total correction: –0.50 dB floor to be at the same level by the time
would have measured in a noise-power Many of today’s microprocessor- it reaches the display. The input and
bandwidth of the same numeric value controlled analyzers allow us to activate internal noise signals add to raise the
as our 3-dB bandwidth, we find that a noise marker. When we do so, the displayed noise by 3 dB, a factor of two
the adjustment varies from: microprocessor switches the analyzer in power. So we can define the noise
into the power (rms) averaging mode, figure of our analyzer for a noise signal
10 log(10,000/10,500) = –0.21 dB computes the mean value of a number as:
to of display points about the marker 9,
10 log(10,000/11,300) = –0.53 dB normalizes and corrects the value to NFSA(N) = (noise floor)dBm/RBW –
a 1-Hz noise-power bandwidth and 10 log(RBW/1) – kTBB=1 + 2.5 dB
In other words, if we subtract displays the normalized value.
something between 0.21 and 0.53 If we use the same noise floor we used
dB from the indicated noise level, The analyzer does the hard part. It is previously, –110 dBm in a 10-kHz
we have the noise level in a noise- easy to convert the noise-marker value resolution bandwidth, we get:
power bandwidth that is convenient to other bandwidths. For example, if we
for computations. For the following want to know the total noise in a 4-MHz NFSA(N) = –110 dBm – 10 log(10,000/1)
examples, we will use 0.5 dB as communication channel, we add 10 – (–174 dBm) + 2.5 dB = 26.5 dB
a reasonable compromise for the log(4,000,000/1), or 66 dB to the noise-
bandwidth correction8. marker value10. As was the case for a sinusoidal signal,
NFSA(N) is independent of resolution
Let’s consider the various correction Preamplifier for noise bandwidth and tells us how far above
factors to calculate the total correction kTB a noise signal must be to be equal
measurements to the noise floor of our analyzer.
for each averaging mode:
Noise signals are typically low-level
Linear (voltage) averaging: signals, so we often need a preamplifier When we add a preamplifier to our
Rayleigh distribution (linear mode): 1.05 dB to have sufficient sensitivity to measure analyzer, the system noise figure and
3-dB/noise power bandwidths: –0.50 dB them. However, we must recalculate sensitivity improve. However, we have
Total correction: 0.55 dB sensitivity of our analyzer first. We accounted for the 2.5-dB factor in our
previously defined sensitivity as the definition of NFSA(N), so the graph of
Log averaging: level of a sinusoidal signal that is equal system noise figure becomes that of
Logged Rayleigh distribution: 2.50 dB to the displayed average noise floor. Figure 5-9. We determine system noise
3-dB/noise power bandwidths: –0.50 dB Since the analyzer is calibrated to show figure for noise the same way that we
Total correction: 2.00 dB the proper amplitude of a sinusoid, no did previously for a sinusoidal signal.
correction for the signal was needed.

8.  The X-Series analyzers specify noise power bandwidth accuracy to within 0.5% (± 0.022 dB).
9.  For example, the X-Series analyzers compute the mean over half a division, regardless of the number of display points.
10.  Most modern spectrum analyzers make this calculation even easier with the channel power function. You enter the integration bandwidth of the
channel and center the signal on the analyzer display. The channel power function then calculates the total signal power in the channel.

54
Chapter 6. Dynamic Range
Dynamic range is generally thought of We can expand this expression into a that for every 1 dB we drop the level of
as the ability of an analyzer to measure power series: the fundamental at the input mixer, the
harmonically related signals and the internally generated second harmonic
interaction of two or more signals, i = IS(k1v + k 2 v 2 + k 3 v 3 +...) drops by 2 dB. See Figure 6-1. The
for example, to measure second- or second term includes 3ω1, the third
third-harmonic distortion or third-order where k1 = q/kT
harmonic, and the cube of the input-
intermodulation. In dealing with such k 2 = k12 /2!
signal voltage, V13. So a 1-dB change
k 3 = k13 /3!, etc.
measurements, remember that the in the fundamental at the input mixer
input mixer of a spectrum analyzer is a changes the internally generated third
Let’s now apply two signals to the
nonlinear device, so it always generates harmonic by 3 dB.
mixer. One will be the input signal we
distortion of its own. The mixer is
wish to analyze; the other, the local
nonlinear for a reason. It must be Distortion is often described by its
oscillator signal necessary to create
nonlinear to translate an input signal order. The order can be determined by
the IF:
to the desired IF. But the unwanted noting the coefficient associated with
distortion products generated in the the signal frequency or the exponent
v = VLO sin(ωLO t) + V1 sin(ω1t)
mixer fall at the same frequencies as associated with the signal amplitude.
the distortion products we wish to Thus second-harmonic distortion
If we go through the mathematics, we
measure on the input signal. is second order and third harmonic
arrive at the desired mixing product
that, with the correct LO frequency, distortion is third order. The order
So we might define dynamic range in also indicates the change in internally
equals the IF:
this way: it is the ratio, expressed in dB, generated distortion relative to the
of the largest to the smallest signals change in the fundamental tone that
k2VLOV1 cos[(ωLO – ω1)t]
simultaneously present at the input of created it.
the spectrum analyzer that allows
A k 2VLOV1 cos[(ωLO + ω1)t] term is also
measurement of the smaller signal to a Now let us add a second input signal:
generated, but in our discussion of the
given degree of uncertainty.
tuning equation, we found that we want
the LO to be above the IF, so (ωLO + ω1) v = VLO sin(ωLO t) + V1 sin(ω1t) + V2 sin(ω2t)
Notice that accuracy of the
is also always above the IF.
measurement is part of the definition. This time, when we go through the
In the following examples, you will see math to find internally generated
With a constant LO level, the mixer
how both internally generated noise distortion, in addition to harmonic
output is linearly related to the input
and distortion affect accuracy. distortion, we get:
signal level. For all practical purposes,
this is true as long as the input signal
Dynamic range versus is more than 15 to 20 dB below the (k4/8)VLOV12V2cos[ωLO – (2ω1 – ω2)]t,
internal distortion level of the LO. There are also terms (k4/8)VLOV1V22 cos[ωLO – (2ω2 – ω1)]t, etc.
involving harmonics of the input signal:
To determine dynamic range versus These equations represent
distortion, we must first determine just intermodulation distortion, the
(3k3/4)VLOV12 sin(ωLO – 2 ω1)t,
how our input mixer behaves. Most interaction of the two input signals
(k4/8)VLOV13 sin(ωLO – 3ω1)t, etc.
analyzers, particularly those using with each other. The lower distortion
harmonic mixing to extend their tuning product, 2ω1 – ω2, falls below ω1 by
These terms tell us that dynamic range
range1, use diode mixers. (Other types a frequency equal to the difference
due to internal distortion is a function of
of mixers would behave similarly.) The between the two fundamental tones,
the input signal level at the input mixer.
current through an ideal diode can be ω2 – ω1. The higher distortion product,
Let’s see how this works, using as our
expressed as: 2ω2 – ω1, falls above ω2 by the same
definition of dynamic range, the difference
in dB between the fundamental tone and frequency. See Figure 6-1.
i = Is(eqv/kT–1)
the internally generated distortion.
The argument of the sine in the first Once again, dynamic range is a
where IS = the diode’s saturation current
term includes 2ω1, so it represents the function of the level at the input mixer.
q = electron charge (1.60 x 10–19 C)
second harmonic of the input signal. The internally generated distortion
v = instantaneous voltage
k = Boltzmann’s constant The level of this second harmonic is a changes as the product of V12 and
(1.38 x 10 –23 joule/K) function of the square of the voltage of V2 in the first case, of V1 and V22 in
T= temperature in Kelvin the fundamental, V12. This fact tells us the second. If V1 and V2 have the

1.  See Chapter 7, “Extending the Frequency Range.”

55
same amplitude, the usual case when
testing for distortion, we can treat D dB D dB D dB
their products as cubed terms (V13
or V23). Thus, for every dB that we 2D dB
3D dB
simultaneously change the level of 3D dB 3D dB
the two input signals, there is a 3-dB
change in the distortion components,
as shown in Figure 6-1.

w 2w 3w 2w 1 – w 2 w1 w2 2w 2 – w 1
This is the same degree of change that
we see for third harmonic distortion
in Figure 6-1. And in fact, this too, is Figure 6-1. Changing the level of fundamental tones at the mixer

third-order distortion. In this case, we

can determine the degree of distortion measurement range. In this case, for internally generated third-order
by summing the coefficients of ω1 and every 1 dB the fundamental changes distortion would be equal to the
ω2 (e.g., 2ω1 – 1ω2 yields 2 + 1 = 3) or at the mixer, our measurement range fundamental(s), or 0 dBc. This situation
the exponents of V1 and V2. also changes by 1 dB. In our second- cannot be realized in practice because
harmonic example, then, when the level the mixer would be well into saturation.
All this says that dynamic range at the mixer changes from –40 to –50 However, from a mathematical
depends upon the signal level at the dBm, the internal distortion, and thus standpoint, TOI is a perfectly good data
mixer. How do we know what level our measurement range, changes from point because we know the slope of
we need at the mixer for a particular –75 to –85 dBc. In fact, these points the line. So even with TOI as a starting
measurement? Most analyzer data fall on a line with a slope of 1 that point, we can still determine the degree
sheets include graphs to tell us how describes the dynamic range for any of internally generated distortion at a
dynamic range varies. However, if no input level at the mixer. given mixer level.
graph is provided, we can draw our
own2. We can construct a similar line for We can calculate TOI from data
third-order distortion. For example, sheet information. Because third-
We do need a starting point, and this a data sheet might say third-order order dynamic range changes 2 dB
we must get from the data sheet. distortion is –85 dBc for a level of for every 1 dB change in the level of
Let’s look at second-order distortion –30 dBm at this mixer. Again, this is the fundamental tone(s) at the mixer,
first. Let’s assume the data sheet our starting point, and we would plot we get TOI by subtracting half of the
says second-harmonic distortion is 75 the point shown in Figure 6-2. If we specified dynamic range in dBc from
dB down for a signal –40 dBm at the now drop the level at the mixer to the level of the fundamental(s):
mixer. Because distortion is a relative –40 dBm, what happens? Referring
measurement, and, at least for the again to Figure 6-1, we see that both TOI = Afund – d/2
moment, we are calling our dynamic third-harmonic distortion and third-
range the difference in dB between order intermodulation distortion where Afund = level of the fundamental
fundamental tone or tones and the fall by 3 dB for every 1 dB that the in dBm
internally generated distortion, we have fundamental tone or tones fall. Again, d = difference in dBc (a negative
our starting point. Internally generated it is the difference that is important. If value) between fundamental and
second-order distortion is 75 dB down, the level at the mixer changes from –30 distortion
so we can measure distortion down to –40 dBm, the difference between
75 dB. We plot that point on a graph fundamental tone or tones and Using the values from the previous
whose axes are labeled distortion internally generated distortion changes discussion:
(dBc) versus level at the mixer (level by 20 dB. So the internal distortion
at the input connector minus the is –105 dBc. These two points fall on TOI = –30 dBm – (–85 dBc)/2 = +12.5 dBm
input-attenuator setting). See Figure a line with a slope of 2, giving us the
6-2. What happens if the level at the third-order performance for any level Attenuator test
mixer drops to –50 dBm? As noted in at the mixer.
Understanding the distortion graph
Figure 6-1, for every 1 dB change in
Sometimes third-order performance is important, but we can use a simple
the level of the fundamental at the
is given as TOI (third-order intercept). test to determine whether displayed
mixer there is a 2 dB change in the
This is the mixer level at which the distortion components are true input
internally generated second harmonic.
But for measurement purposes, we are
2.  For more information on how to construct a dynamic range chart, see Optimizing Dynamic
interested only in the relative change, Range for Distortion Measurements – Keysight PSA Performance Spectrum Analyzer Series
that is, in what happened to our Product Note, literature number 5980-3079EN.

56
signals or internally generated signals. Figure 6-2.
Dynamic range
Change the input attenuator. If the 0
TOI SHI
versus distortion
displayed value of the distortion and noise
components remains the same, the –10
components are part of the input
–20
signal. If the displayed value changes,

2n order

er
the distortion components are

rd
–30

do
3rd
generated internally or are the sum
of external and internally generated –40

signals. We continue changing the –50 Nois

e(

(dBc)
attenuator until the displayed distortion 10
kH
does not change and then complete –60 z BW
the measurement. )
–70
Maximum 2nd order
dynamic range
Noise –80
Maximum 3rd order
dynamic range
Another constraint on dynamic range –90 Optimum
mixer levels
is the noise floor of our spectrum
–100
analyzer. Going back to our definition –60 –50 –40 –30 –20 –10 0 +10
of dynamic range as the ratio of the Mixer level (dBm)
largest to the smallest signal we can
measure, the average noise of our
Figure 6-3.
spectrum analyzer puts the limit on Reducing resolu-
the smaller signal. So dynamic range 0 TOI SHI tion bandwidth
versus noise becomes signal-to- improves dy-
–10 namic range
noise ratio in which the signal is the
fundamental whose distortion we wish –20

er
to measure.
ord

er
rd
–30
3rd
do
2n
We can easily plot noise on our
–40
dynamic range chart. For example,
suppose the data sheet for our –50 No
ise
(dBc)

spectrum analyzer specifies a displayed (1

N 0k
average noise level of –110 dBm in a –60 ois
e ( Hz B
1k W
10-kHz resolution bandwidth. If our Hz )
–70 BW
signal fundamental has a level of –40 ) 2nd order
dynamic range improvement
dBm at the mixer, it is 70 dB above the –80
average noise, so we have a 70-dB 3rd order
dynamic range improvement
–90
signal-to-noise ratio. For every 1 dB we
reduce the signal level at the mixer, we
lose 1 dB of signal-to-noise ratio. Our –60 –50 –40 –30 –20 –10 0 +10
Mixer level (dBm)
noise curve is a straight line having a
slope of –1, as shown in Figure 6-2.
because noise adds to the CW-like noise floor and the improvement in
If we ignore measurement accuracy distortion products, reducing dynamic dynamic range. For second-order
considerations for a moment, the range by 2 dB when you use the log distortion, the improvement is one half
best dynamic range will occur at power scale with log scale averaging. the change in the noise floor; for third-
the intersection of the appropriate order distortion, two-thirds the change
distortion curve and the noise curve. Figure 6-2 shows the dynamic range in the noise floor. See Figure 6-3.
Figure 6-2 tells us that our maximum for one resolution bandwidth. We
dynamic range for second-order certainly can improve dynamic range The final factor in dynamic range is the
distortion is 72.5 dB; for third-order by narrowing the resolution bandwidth, phase noise on our spectrum analyzer
distortion, 81.7 dB. In practice, the but there is not a one-to-one LO, and this affects only third-order
intersection of the noise and distortion correspondence between the lowered distortion measurements. For example,
graphs is not a sharply defined point,

57
suppose we are making a two-tone,
third-order distortion measurement
on an amplifier, and our test tones are –10
separated by 10 kHz. The third-order
distortion components will also be
separated from the test tones by 10 –20
kHz. For this measurement, we might
find ourselves using a 1-kHz resolution –30
bandwidth. Referring to Figure 6-3
and allowing for a 10-dB decrease
in the noise curve, we would find a –40
maximum dynamic range of about
88 dB. Suppose however, that our
–50
phase noise at a 10-kHz offset is only
–80 dBc. Then 80 dB becomes the
ultimate limit of dynamic range for this –60
measurement, as shown in Figure 6-4. (dBc)

–70
In summary, the dynamic range of a
spectrum analyzer is limited by three
–80 Phase noise
factors: the distortion performance of
Dynamic range (10 kHz offset)
the input mixer, the broadband noise reduction due
floor (sensitivity) of the system and the to phase noise
–90
phase noise of the local oscillator.

Dynamic range versus –100

measurement uncertainty
–110
In our previous discussion of amplitude –60 –50 –40 –30 –20 –10 0 +10
accuracy, we included only those items Mixer level (dBm)
listed in Table 4-1, plus mismatch.
We did not cover the possibility of
Figure 6-4. Phase noise can limit third-order intermodulation tests
an internally generated distortion
product (a sinusoid) being at the same
frequency as an external signal we
wished to measure. However, internally 6
5
generated distortion components 4
fall at exactly the same frequencies 3
as the distortion components we 2
wish to measure on external signals. 1
The problem is that we have no way 0 Maximum
–30 –25 –20 –15 –10 –5 –1 error (dB)
of knowing the phase relationship
–2
between the external and internal –3
signals. So we can determine only a –4
potential range of uncertainty: –5
–6
Uncertainty (in dB) = 20 log(l ± 10d/20) –7
Delta (dBc) –8

where d = difference in dB between

the larger and smaller sinusoid Figure 6-5. Uncertainty versus difference in amplitude between two sinusoids at the
same frequency
(a negative number)

See Figure 6-5. For example, if we

set up conditions such that the
internally generated distortion is
equal in amplitude to the distortion on
the incoming signal, the error in the

58
measurement could range from +6
dB (the two signals exactly in phase)
to negative infinity (the two signals 0
exactly out of phase and so canceling).
Such uncertainty is unacceptable
–10
in most cases. If we put a limit of ±1
dB on the measurement uncertainty,
Figure 6-5 shows us that the internally –20
generated distortion product must
be about 18 dB below the distortion
–30
product we wish to measure. To draw r
rde
dynamic range curves for second- and do

er
2n

ord
third-order measurements with no –40

3rd
more than 1 dB of measurement error,

No
we must then offset the curves of

ise
–50
Figure 6-2 by 18 dB as shown in Figure 5 dB
(dBc)
6-6.
–60
Next, let’s look at uncertainty due to
low signal-to-noise ratio. The distortion –70
components we wish to measure are,
we hope, low-level signals, and often
–80 18 dB
they are at–or very close to–the noise
level of our spectrum analyzer. In such
cases, we often use the video filter –90 18 dB
to make these low-level signals more
discernible. Figure 6-7 shows the error
–100
in displayed signal level as a function of
–60 –50 –40 –30 –20 –10 0 +10
displayed signal-to-noise for a typical
Mixer level (dBm)
spectrum analyzer. The error is only in
one direction, so we could correct for
it. However, we usually do not. So for Figure 6-6. Dynamic range for 1.3-dB maximum error
our dynamic range measurement, let’s
accept a 0.3-dB error due to noise and
offset the noise curve in our dynamic
7
range chart by 5 dB, as shown in Figure
6-6. Where the distortion and noise
curves intersect, the maximum error 6
Error in displayed signal level (dB)

possible would be less than 1.3 dB.

5
Let’s see what happened to our
dynamic range as a result of our 4
concern with measurement error.
As Figure 6-6 shows, second-order-
3
distortion dynamic range changes from
72.5 to 61 dB, a change of 11.5 dB. This
is one half the total offsets for the two 2
curves (18 dB for distortion; 5 dB for
noise). Third-order distortion changes 1
from 81.7 dB to about 72.7 dB for a
change of about 9 dB. In this case, the 0
change is one third of the 18-dB offset 0 1 2 3 4 5 6 7 8
for the distortion curve plus two thirds Displayed S/N (dB)
of the 5-dB offset for the noise curve.

Figure 6-7. Error in displayed signal amplitude due to noise

59
Gain compression A third method, called pulse The range of the log amplifier can
compression, measures the change in be another limitation for spectrum
In our discussion of dynamic range, we system gain to a narrow (broadband) analyzers with analog IF circuitry.
did not concern ourselves with how RF pulse while the power of the pulse For example, ESA-L Series spectrum
accurately the larger tone is displayed, is swept upward. When measuring analyzers use an 85-dB log amplifier.
even on a relative basis. As we raise pulses, we often use a resolution Thus, only measurements that are
the level of a sinusoidal input signal, bandwidth much narrower than the within 85 dB below the reference level
eventually the level at the input mixer bandwidth of the pulse, so our analyzer are calibrated.
becomes so high that the desired displays the signal level well below
output mixing product no longer the peak pulse power. As a result, we The question is, can the full display
changes linearly with respect to the could be unaware of the fact that range be used? From the previous
input signal. The mixer is in saturation, the total signal power is above the discussion of dynamic range, we know
and the displayed signal amplitude is mixer compression threshold. A high the answer is generally yes. In fact,
too low. Saturation is gradual rather threshold improves signal-to-noise dynamic range often exceeds display
than sudden. To help us stay away ratio for high-power, ultranarrow or range or log amplifier range. To bring
from the saturation condition, the 1-dB widely-chirped pulses. The threshold is the smaller signals into the calibrated
compression point is normally specified. about 12 dB higher than for two-tone area of the display, we must increase
Typically, this gain compression occurs compression in the Keysight X-Series IF gain. But in so doing, we may move
at a mixer level in the range of –5 to +5 signal analyzers. Nevertheless, the larger signals off the top of the
dBm. Thus we can determine what input because different compression display, above the reference level.
attenuator setting to use for accurate mechanisms affect CW, two-tone and Some Keysight analyzers, such as
measurement of high-level signals3. pulse compression differently, any of the X-Series, allow measurements
Spectrum analyzers with a digital IF will the compression thresholds can be of signals above the reference level
indicate that ADC is over-ranged. lower than any other. without affecting the accuracy with
which the smaller signals are displayed,
Actually, there are three different as shown in Figure 6-8 (see page 61).
methods of evaluating compression.
Display range and
So we can indeed take advantage of
A traditional method, called CW measurement range the full dynamic range of an analyzer
compression, measures the change Two additional ranges are often even when the dynamic range exceeds
in gain of a device (amplifier or mixer confused with dynamic range: display the display range. In Figure 6-8, even
or system) as the input signal power range and measurement range. though the reference level has changed
is swept upward. This method is the Display range, often called display from –20 dBm to –50 dBm, driving
one just described. Note that the CW dynamic range, refers to the calibrated the signal far above the top of the
compression point is considerably amplitude range of the spectrum screen, the marker readout remains
higher than the levels for the analyzer display. For example, a unchanged.
fundamentals indicated previously display with ten divisions would seem
for even moderate dynamic range. to have a 100-dB display range when Measurement range is the ratio of the
So we were correct in not concerning we select 10 dB per division. This is largest to the smallest signal that can
ourselves with the possibility of certainly true for today’s analyzers be measured under any circumstances.
compression of the larger signal(s). with digital IF circuitry, such as the The maximum safe input level, typically
Keysight X-Series. It is also true for the +30 dBm (1 watt) for most analyzers,
A second method, called two-tone Keysight ESA-E Series analyzers when determines the upper limit. These
compression, measures the change you use the narrow (10- to 300-Hz) analyzers have input attenuators you can
in system gain for a small signal digital resolution bandwidths. However, set to 60 or 70 dB, so you can reduce
while the power of a larger signal is spectrum analyzers with analog IF +30 dBm signals to levels well below the
swept upward. Two-tone compression sections typically are calibrated only compression point of the input mixer
applies to the measurement of multiple for the first 85 or 90 dB below the and measure them accurately. The
CW signals, such as sidebands and reference level. In this case, the bottom displayed average noise level sets the
independent signals. The threshold of line of the graticule represents other end of the range. Depending on
compression of this method is usually signal amplitudes of zero, so the the minimum resolution bandwidth of
a few dB lower than that of the CW bottom portion of the display covers the particular analyzer and whether or
method. This is the method used the range from –85 or –90 dB to not you are using a preamplifier, DANL
by Keysight Technologies to specify infinity, relative to the reference level. typically ranges from –115 to –170 dBm.
spectrum analyzer gain compression. Measurement range, then, can vary from

3.  Many analyzers internally control the combined settings of the input attenuator and IF gain so that a CW signal as high as the compression level at
the input mixer creates a response above the top line of the graticule. This feature keeps us from making incorrect measurements on CW signals
inadvertently.
60
145 to 200 dB. Of course, we cannot
view a –170-dBm signal while a +30-dBm
signal is also present at the input.

Adjacent channel power

measurements
TOI, SOI, 1- dB gain compression,
and DANL are all classic measures
of spectrum analyzer performance.
However, with the tremendous growth
of digital communication systems,
other measures of dynamic range
have become increasingly important.
For example, adjacent channel power
(ACP) measurements are often done in
CDMA-based communication systems
to determine how much signal energy
leaks or “spills over” into adjacent
or alternate channels located above
and below a carrier. An example ACP
measurement is shown in Figure 6-9.

Note the relative amplitude difference

between the channel power and the
adjacent and alternate channels. You
can measure up to six channels on
either side of the carrier at a time.

Typically, we are most interested in

the relative difference between the
signal power in the main channel
and the signal power in the adjacent
or alternate channel. Depending
on the particular communication
standard, these measurements are Figure 6-8. Display range and measurement range on the PXA spectrum analyzer
often described as “adjacent channel
power ratio” (ACPR) or “adjacent
channel leakage ratio” (ACLR) tests.
Because digitally modulated signals
and the distortion they generate are
very noise-like in nature, the industry
standards typically define a channel
bandwidth over which the signal power
is integrated.

To accurately measure ACP

performance of a device under
test such as a power amplifier, the
spectrum analyzer must have better
ACP performance than the device
being tested. Therefore, spectrum
analyzer ACPR dynamic range has
become a key performance measure for
digital communication systems.
Figure 6-9. Adjacent channel power measurement using a PXA spectrum analyzer

61
Chapter 7. Extending the Frequency Range
As more wireless services continue But let us take one step at a time. examples. In summary, for the low band,
to be introduced and deployed, the In developing our tuning equation in up to 3.6 GHz, our first IF is 5.1 GHz. For
available spectrum has become Chapter 2, we found that we needed the upper frequency bands, we switch to
more and more crowded. As a result, the low-pass filter shown in Figure 2-1 a first IF of 322.5 MHz. In Figure 7-1, the
there has been an ongoing trend to prevent higher-frequency signals second IF is already 322.5 MHz, so all we
toward developing new products from reaching the mixer. The result was need to do when we want to tune to the
and services at higher frequencies. In a uniquely responding, single-band higher ranges is bypass the first IF.
addition, new microwave technologies analyzer that tuned to 3.6 GHz. To
continue to evolve, driving the need for observe and measure higher-frequency In Chapter 2, we used a mathematical
more measurement capability in the signals, we must remove the low-pass approach to conclude that we needed
microwave bands. Spectrum analyzer filter. a low-pass filter. The math becomes
designers have responded by developing more complex in the situation here, so
instruments capable of directly tuning Other factors that we explored in we will use a graphical approach to see
up to 50 GHz using a coaxial input. Even developing the tuning equation were what is happening. The low band is the
higher frequencies can be measured the choice of LO and intermediate simpler case, so we’ll start with that.
using external mixing techniques. This frequencies. We decided that the IF In all of our graphs, we will plot the LO
chapter describes the techniques used to should not be within the band of interest frequency along the horizontal axis and
enable tuning the spectrum analyzer to because it created a hole in our tuning signal frequency along the vertical axis,
such high frequencies. range in which we could not make as shown in Figure 7-2. We know we get
measurements. So we chose 5.1 GHz, a mixing product equal to the IF (and
Internal harmonic mixing moving the IF above the highest tuning therefore a response on the display)
frequency of interest (3.6 GHz). Our new whenever the input signal differs from
In Chapter 2, we described a single- tuning range will be above 3.6 GHz, so the LO by the IF. Therefore, we can
range spectrum analyzer that tunes to it seems logical to move the new IF to a determine the frequency to which the
3.6 GHz. Now we want to tune higher frequency below 3.6 GHz. A typical first analyzer is tuned simply by adding the
in frequency. The most practical way IF for these higher frequency ranges in IF to, or subtracting it from, the LO
to achieve an extended range is to use Keysight spectrum analyzers is 322.5 frequency. To determine our tuning
harmonic mixing. MHz. We will use this frequency in our range, we start by plotting the LO

Low Analog or
band path digital IF
3.6 GHz 5.1225 GHz 322.5 MHz 22.5 MHz

Input
signal

High
band path 3.8 to 8.7 GHz

To external
mixer
4.8 GHz 300 MHz

322.5 MHz

Preselector

Sweep generator
Display

Figure 7-1. Switching arrangement for low band and high bands
62
frequency against the signal frequency
axis, as shown by the dashed line in
Figure 7-2. Subtracting the IF from the 13

Signal frequency (GHz)

dashed line gives us a tuning range of
0 to 3.6 GHz, the range we developed 9
in Chapter 2. Note that this line in +IF LO frequency,
Figure 7-2 is labeled “1−” to indicate GHz
5 1-
fundamental mixing and the use of the
–IF
minus sign in the tuning equation. We 1+
can use the graph to determine what 1
LO frequency is required to receive
a particular signal or to what signal -3
the analyzer is tuned for a given LO 3 4 5 6 7 8 9
frequency. To display a 1-GHz signal, LO frequency (GHz)
the LO must be tuned to 6.1 GHz.
For an LO frequency of 8 GHz, the Figure 7-2. Tuning curves for fundamental mixing in the low band, high IF case
spectrum analyzer is tuned to receive a
signal frequency of 2.9 GHz. In our text,
we round off the first IF to one decimal 25
place; the true IF, 5.1225 GHz, is shown
on the block diagram.
20
Now let’s add the other fundamental-
mixing band by adding the IF to the LO
Signal frequency (GHz)

15 1-
line in Figure 7-2. This gives us the solid
1+
upper line, labeled 1+, that indicates
a tuning range from 8.9 to 13.8 GHz. 10 2-
Note that for a given LO frequency, the 2+
two frequencies to which the analyzer
5 LO
is tuned are separated by twice the IF.
Assuming we have a low-pass filter at 2xLO
the input while measuring signals in the
0
low band, we will not be bothered by 3 4 5 6 7 8 9
signals in the 1+ frequency range.
-5
LO frequency (GHz)
Next let’s see to what extent harmonic
mixing complicates the situation.
Harmonic mixing comes about because Figure 7-3. Signals in the “1 minus” frequency range produce single, unambiguous responses in the
the LO provides a high-level drive low-band, high-IF case

signal to the mixer for efficient mixing,

and because the mixer is a non-linear Let’s add second-harmonic mixing signals in the 1− tuning range produce
device, it generates harmonics of the to our graph in Figure 7-3 and see unique, unambiguous responses on our
LO signal. Incoming signals can mix to what extent this complicates our analyzer display. The same low-pass
against LO harmonics just as well measurement procedure. As before, we filter used in the fundamental mixing
as the fundamental, and any mixing first plot the LO frequency against the case works equally well for eliminating
product that equals the IF produces signal frequency axis. Multiplying the responses created in the harmonic
a response on the display. In other LO frequency by two yields the upper mixing case.
words, our tuning (mixing) equation dashed line of Figure 7-3. As we did for
now becomes: fundamental mixing, we simply subtract The situation is considerably different
the IF (5.1 GHz) from and add it to the for the high-band, low-IF case. As
fsig = nf LO ± f IF LO second-harmonic curve to produce before, we start by plotting the LO
the 2− and 2+ tuning ranges. Since fundamental against the signal-
where n = LO harmonic neither of these overlap the desired frequency axis and then add and
(Other parameters remain the 1− tuning range, we can again argue subtract the IF, producing the results
same as previously discussed.) that they do not really complicate the shown in Figure 7-4. Note that the
measurement process. In other words, 1− and 1+ tuning ranges are much

63
closer together, and in fact overlap,
because the IF is a much lower 10
frequency, 322.5 MHz in this case.
Does the close spacing of the tuning Image Frequencies

Signal frequency (GHz)

ranges complicate the measurement
process? Yes and no. First of all, our
system can be calibrated for only one 5.3
5 1-
tuning range at a time. In this case, we
4.7
would choose the 1− tuning to give us 1+
a low-end frequency of about 3.5 GHz, LO Frequency
so we have some overlap with the 3.6-
GHz upper end of our low-band tuning
range. So what are we likely to see on
the display? If we enter the graph at an 0 4.4 5.6
3 4 5 6 7 8 9
LO frequency of 5 GHz, we find there LO frequency (GHz)
are two possible signal frequencies
that would give us responses at the
Figure 7-4. Tuning curves for fundamental mixing in the high-band, low-IF case
same point on the display: 4.7 and 5.3
GHz (rounding the numbers again). On
the other hand, if we enter the signal in a displayed signal that appears to be
and those produced on the 1+ tuning
frequency axis at 5.3 GHz, we find that at 13.0 GHz.
curve. However, before we look at
in addition to the 1+ response at an
signal identification solutions, let’s
LO frequency of 5 GHz, we could also The displayed signals created by the
add harmonic-mixing curves to 26.5
get a 1− response. This would occur responses to the 3+ and 3− tuning
GHz and see if there are any additional
if we allowed the LO to sweep as high curves are known as in-band multiple
factors we must consider in the signal
as 5.6 GHz, twice the IF above 5 GHz. responses. Because they occur when
identification process. Figure 7-5 shows
Also, if we entered the signal frequency the LO is tuned to 4.63 GHz and
tuning curves up to the fourth harmonic
graph at 4.7 GHz, we would find a 1+ 4.43 GHz, they will produce false
of the LO.
response at an LO frequency of about responses on the display that appear
4.4 GHz (twice the IF below 5 GHz) in to be genuine signals at 8.96 GHz and
In examining Figure 7-5, we find
addition to the 1− response at an LO 8.56 GHz.
some additional complications. The
frequency of 5 GHz. Thus, for every
spectrum analyzer is set up to operate
desired response on the 1− tuning Other situations can create out-of-
in several tuning bands. Depending on
line, there will be a second response band multiple responses. For example,
the frequency to which the analyzer is
located twice the IF below it. These suppose we are looking at a 5-GHz
tuned, the analyzer display is frequency
pairs of responses are known as image signal in band 1 that has a significant
calibrated for a specific LO harmonic.
responses. third harmonic at 15 GHz (band 3). In
For example, in the 8.3- to 13.6-GHz
input frequency range, the spectrum addition to the expected multiple pair
With this type of mixing arrangement, caused by the 5-GHz signal on the
analyzer is calibrated for the 2− tuning
it is possible for signals at different 1+ and 1− tuning curves, we also get
curve. Suppose we have an 13.6-GHz
frequencies to produce responses at responses generated by the 15-GHz
signal present at the input. As the
the same point on the display, that is, signal on the 4+, 4−, 3+, and 3− tuning
LO sweeps, the signal will produce IF
at the same LO frequency. As we can curves. Since these responses occur
responses with the 3+, 3-, 2+ and 2−
see from Figure 7-4, input signals at 4.7 when the LO is tuned to 3.7, 3.8, 4.9,
tuning curves. The desired response
and 5.3 GHz both produce a response and 5.1 GHz respectively, the display
of the 2− tuning curve occurs when
at the IF when the LO frequency is set will show signals that appear to be
the LO frequency satisfies the tuning
to 5 GHz. These signals are known as located at 3.4, 3.5, 4.6, and 4.8 GHz.
equation:
image frequencies, and they are also This is shown in Figure 7-6.
separated by twice the IF frequency.
13.6 GHz = 2 f LO – 0.3
f LO = 6.95GHz Multiple responses generally always
Clearly, we need some mechanism come in pairs1, with a “plus” mixing
to differentiate between responses product and a “minus” mixing product.
Similarly, we can calculate that the
generated on the 1− tuning curve When we use the correct harmonic
response from the 2+ tuning curve
for which our analyzer is calibrated mixing number for a given tuning band,
occurs when f LO = 6.65 GHz, resulting

1.  Often referred to as an “image pair.” This is inaccurate terminology, since images are actually two or more real signals present at the spectrum ana-
lyzer input that produce an IF response at the same LO frequency. The numbers for your analyzer may differ.

64
 25.00 4+ 3+
Apparent location of an input signal 3-
resulting from the response to the 4-
Band 4
2- tuning curve
20.00
In-band multiple 2+
Signal frequency (GHz)

responses 2-
15.00 Band 3
13.6
13.0
Band 2
10.00 1+
8.96
8.56 1-
Band 1
5.00

Apparent locations of in-band Band 0

multiples of a 13.6 GHz input signal (lowband)
0.00
4.43 4.63 6.65 6.95
3 4 5 6 7 8 9
LO frequency (GHz)

Figure 7-5. Tuning curves up to 4th harmonic of LO showing in-band multiple responses to a 13.6-GHz input signal

25.00 4+ 3+

4- 3-
Band 4
20.00 Out-of-band
multiple responses 2+
Signal frequency (GHz)

2-
15.00 Band 3

Band 2
10.00 1+
1-
Band 1
5.00

Band 0
(lowband)
0.00
3.7 3.8 4.9 5.1 5.3
3 4 5 6 7 8 9
LO frequency (GHz)

Figure 7-6. Out-of-band multiple responses in band 1 as a result of a signal in-band 3

the responses will be separated by 2 2f IF (Nc /NA)

times f IF. Because the slope of each pair
of tuning curves increases linearly with where Nc = the correct harmonic
the harmonic number N, the multiple number for the desired tuning
pairs caused by any other harmonic band
mixing number appear to be separated NA = the actual harmonic
by: number generating the multiple
pair

65
 25.00 2+ LO Doubled

Band 4
20.00 2- LO Doubled
Signal frequency (GHz)

2+
2-
15.00 1+ LO Doubled Band 3

1- LO Doubled Band 2
10.00 1+
1-
Band 1
5.00

Band 0
(lowband)
0.00
3 4 5 6 7 8 9
LO frequency (GHz)

Figure 7-7. X-Series analyzer harmonic bands using LO doubling

In X-Series analyzers, the LO is doubled Fortunately, there is a way to tracking preselector. Signals beyond
to produce a new, higher-frequency essentially eliminate image and the dashed lines are rejected. Let’s
LO for harmonic mixing. As a result, the multiple responses through a process continue with our previous example of
LO harmonics are twice as far apart as of prefiltering the signal. This technique 4.7- and 5.3- GHz signals present at
they would otherwise be and likelihood is called preselection. the analyzer input. If we set a center
of multiple responses is significantly frequency of 5 GHz and a span of 2
reduced. Compare Figure 7-6 and Figure Preselection GHz, let’s see what happens as the
7-7. analyzer tunes across this range. As the
What form must our preselection take? LO sweeps past 4.4 GHz (the frequency
Can we conclude from this discussion Referring back to Figure 7-4, assume at which it could mix with the 4.7-GHz
that a harmonic mixing spectrum we have two signals at 4.7 and 5.3 GHz input signal on its 1+ mixing mode),
analyzer is not practical? Not present at the input of our analyzer. the preselector is tuned to 4.1 GHz
necessarily. In cases where the signal If we were particularly interested in and therefore rejects the 4.7-GHz
frequency is known, we can tune to one, we could use a band-pass filter to signal. The input signal does not reach
the signal directly, knowing that the allow that signal into the analyzer and the mixer, so no mixing occurs, and
analyzer will select the appropriate reject the other. However, the fixed filter no response appears on the display.
mixing mode for which it is calibrated. does not eliminate multiple responses; As the LO sweeps past 5 GHz, the
In controlled environments with only so if the spectrum is crowded, there preselector allows the 4.7-GHz signal
one or two signals, it is usually easy is still potential for confusion. More to reach the mixer, and we see the
to distinguish the real signal from important, perhaps, is the restriction appropriate response on the display.
the image and multiple responses. that a fixed filter puts on the flexibility of The 5.3-GHz image signal is rejected,
However, there are many cases in the analyzer. If we are doing broadband so it creates no mixing product to
which we have no idea how many testing, we certainly do not want to be interact with the mixing product from
signals are involved or what their continually forced to change bandpass the 4.7-GHz signal and cause a false
frequencies might be. For example, filters. display. Finally, as the LO sweeps past
we could be searching for unknown 5.6 GHz, the preselector allows the
spurious signals, conducting site The solution is a tunable filter 5.3-GHz signal to reach the mixer, and
surveillance tests as part of a configured such that it automatically we see it properly displayed. Note in
frequency-monitoring program or tracks the frequency of the appropriate Figure 7-8 that nowhere do the various
performing EMI tests to measure mixing mode. Figure 7-8 shows the mixing modes intersect. So as long as
unwanted device emissions. In all effect of such a preselector. Here we the preselector bandwidth is narrow
these cases, we could be looking for take advantage of the fact that our enough (it typically varies from about
totally unknown signals in a potentially superheterodyne spectrum analyzer 35 MHz at low frequencies to 80 MHz
crowded spectral environment. Having is not a real-time analyzer; that is, at high frequencies) it will greatly
to perform some form of identification it tunes to only one frequency at attenuate all image and multiple
routine on each and every response a time. The dashed lines in Figure responses.
would make measurement time 7-8 represent the bandwidth of the
intolerably long.
66
The word “eliminate” may be a little
strong. Preselectors do not have
infinite rejection. Rejection in the 70- to
1+
80-dB range is more likely. So if we are
looking for very low-level signals in the 6
presence of very high-level signals, we
might see low-level images or multiples
5.3
of the high-level signals. What 1–
about the low band? Most tracking

Signal frequency (GHz)

4.7
preselectors use YIG technology, and
YIG filters do not operate well at low
frequencies. Fortunately, there is a 4
simple solution. Figure 7-3 shows that
no other mixing mode overlaps the
1− mixing mode in the low-frequency,
3 Preselector
high-IF case. So a simple low-pass
filter attenuates both image and bandwidth
multiple responses. Figure 7-9 shows
the input architecture of a typical 2
microwave spectrum analyzer. 4.4 5.6
3 4 5 6
LO frequency (GHz)

Figure 7-8. Preselection; dashed gray lines represent bandwidth of tracking preselector

Low Analog or
band path digital IF
3.6 GHz 5.1225 GHz 322.5 MHz 22.5 MHz

Input
signal

High
band path 3.8 to 8.7 GHz

To external
mixer
4.8 GHz 300 MHz

322.5 MHz

Preselector

Sweep generator
Display

Figure 7-9. Front-end architecture of a typical preselected spectrum analyzer

67
Amplitude calibration
So far, we have looked at how a
harmonic mixing spectrum analyzer
responds to various input frequencies.
What about amplitude?

The conversion loss of a mixer is a

function of harmonic number, and the
loss goes up as the harmonic number
goes up. This means that signals of
equal amplitude would appear at
different levels on the display if they
involved different mixing modes.
To preserve amplitude calibration,
something must be done. In Keysight
spectrum analyzers, the IF gain is
changed. The increased conversion
loss at higher LO harmonics causes Figure 7-10. Rise in noise floor indicates changes in sensitivity with changes in LO harmonic used
a loss of sensitivity just as if we had
increased the input attenuator. And
since the IF gain change occurs after
the conversion loss, the gain change is
reflected by a corresponding change
in the displayed noise level. So we can
determine analyzer sensitivity on the
harmonic-mixing ranges by noting the
average displayed noise level just as we
did on fundamental mixing.

In older spectrum analyzers, the

increase in displayed average noise
level with each harmonic band was
very noticeable. More recent models
of Keysight spectrum analyzers use
a double-balanced, image-enhanced
harmonic mixer to minimize the
increased conversion loss when using
higher harmonics. Thus, the “stair step” Figure 7-11. Phase noise levels for fundamental and 4th harmonic mixing
effect on DANL has been replaced by
a gentle sloping increase with higher
mixing, phase noise (in decibels) degree of modulation is very small, as
frequencies, as shown in Figure 7-10.
increases by: in the situation here, the amplitude of
the modulation side bands is directly
Phase noise 20 log(N), proportional to the deviation of the
In Chapter 2, we noted that instability of where N = harmonic of the LO carrier (LO). If the deviation doubles, the
an analyzer LO appears as phase noise level of the side bands must also double
around signals that are displayed far For example, suppose that the LO in voltage; that is, increase by 6 dB or
enough above the noise floor. We also fundamental has a peak-to-peak 20 log(2). As a result, the ability of our
noted that this phase noise can impose deviation of 10 Hz. The second analyzer to measure closely spaced
a limit on our ability to measure closely harmonic then has a 20-Hz peak-to- signals that are unequal in amplitude
spaced signals that differ in amplitude. peak deviation; the third harmonic, 30 decreases as higher harmonics of the
The level of the phase noise indicates Hz; and so on. Since the phase noise LO are used for mixing. Figure 7-11
the angular, or frequency, deviation of indicates the signal (noise in this case) shows the difference in phase noise
the LO. What happens to phase noise producing the modulation, the level between fundamental mixing of a 5-GHz
when a harmonic of the LO is used in the of the phase noise must be higher to signal and fourth-harmonic mixing of a
mixing process? Relative to fundamental produce greater deviation. When the 20-GHz signal.

68
Improved dynamic range
A preselector improves dynamic range –45
if the signals in question have sufficient
–50
frequency separation. The discussion of
dynamic range in Chapter 6 assumed
that both the large and small signals –60
were always present at the mixer and
–70

Internal distortion (dBc)

their amplitudes did not change during
the course of the measurement. But as
we have seen, if signals are far enough –80
apart, a preselector allows one to reach
the mixer while rejecting the others. For –90
example, if we were to test a microwave
oscillator for harmonics, a preselector
–100
would reject the fundamental when
we tuned the analyzer to one of the
harmonics. –110
–115
Let’s look at the dynamic range of –120
a second-harmonic test of a 3-GHz –90 –80 –70 –60 –50 –40 –30 –20 –10 0
oscillator. Using the example from Mixed level (dBm)
Chapter 6, suppose that a –40-dBm
signal at the mixer produces a second
Figure 7-12. Second-order distortion graph
harmonic product of –75 dBc. We also
know, from our discussion, that for
every 1 dB the level of the fundamental
changes at the mixer, measurement distortion, dynamic range is limited 35 MHz at the low end and 80 MHz
range also changes by 1 dB. The on the low-level (harmonic) end only at the high end. As a conservative
second-harmonic distortion curve is by the noise floor (sensitivity) of the figure, we might use 18 dB per octave
shown in Figure 7-12. For this example, analyzer. of bandwidth roll-off of a typical YIG
we assume plenty of power from the preselector filter beyond the 3 dB point.
oscillator and set the input attenuator What about the upper, high-level So to determine the improvement in
so that when we measure the oscillator end? When measuring the oscillator dynamic range, we must determine to
fundamental, the level at the mixer is fundamental, we must limit power at what extent each of the fundamental
–10 dBm, below the 1-dB compression the mixer to get an accurate reading tones is attenuated and how that
point. of the level. We can use either affects internally generated distortion.
internal or external attenuation to From the expressions in Chapter 6 for
From the graph, we see that a –10-dBm limit the level of the fundamental at third-order intermodulation, we have:
signal at the mixer produces a second- the mixer to something less than the
harmonic distortion component of 1-dB compression point. However, (k4 /8)VLOV12V2 cos[ωLO – (2ω1 – ω2)]t
–45 dBc. Now we tune the analyzer the preselector highly attenuates the and
to the 6-GHz second harmonic. If the fundamental when we are tuned to the
(k4 /8)VLOV1V22cos[ωLO – (2ω2 – ω1)]t
preselector has 70-dB rejection, the second harmonic, so we can remove
fundamental at the mixer has dropped some attenuation if we need better
to –80 dBm. Figure 7-12 indicates that sensitivity to measure the harmonic. Looking at these expressions, we
for a signal of –80 dBm at the mixer, A fundamental level of +20 dBm at the see that the amplitude of the lower
the internally generated distortion preselector should not affect our ability distortion component (2ω1 – ω2) varies
is –115 dBc, meaning 115 dB below to measure the harmonic. as the square of V1 and linearly with
the new fundamental level of –80 V2. On the other side, the amplitude of
dBm. This puts the absolute level of Any improvement in dynamic range the upper distortion component (2ω2
the harmonic at –195 dBm. So the for third-order intermodulation – ω1) varies linearly with V1 and as the
difference between the fundamental measurements depends upon square of V2. However, depending on
we tuned to and the internally separation of the test tones versus the signal frequencies and separation,
generated second harmonic we tuned preselector bandwidth. As we noted, the preselector may not attenuate the
to is 185 dB! Clearly, for harmonic typical preselector bandwidth is about two fundamental tones equally.

69
Consider the situation shown in Figure
7-13 in which we are tuned to the 3 dB
lower distortion component, and the 21 dB
two fundamental tones are separated
by half the preselector bandwidth. In
this case, the lower-frequency test
tone lies at the edge of the preselector
pass band and is attenuated 3 dB. The
upper test tone lies above the lower
distortion component by an amount
equal to the full preselector bandwidth. 27 dB
It is attenuated approximately 21
dB. Since we are tuned to the lower
distortion component, internally
generated distortion at this frequency
drops by a factor of two relative to the
attenuation of V1 (2 times 3 dB = 6 dB) Figure 7-13. Improved third-order intermodulation distortion; test tone separation is significant relative to
preselector bandwidth
and equally as fast as the attenuation
of V2 (21 dB). The improvement in mixer, the interaction of the mismatch a certain amount of ripple. In most
dynamic range is the sum of 6 dB + 21 of the preselector with that of the configurations, the tuning ramp for
dB, or 27 dB. As in the case of second input mixer can cause a degradation the preselector and local oscillator
harmonic distortion, the noise floor of of frequency response. You must come from the same source, but
the analyzer must be considered, too. use proper calibration techniques to there is no feedback mechanism
For very closely spaced test tones, the compensate for this ripple. Another to ensure the preselector exactly
preselector provides no improvement, approach to minimize this interaction tracks the tuning of the analyzer.
and we determine dynamic range as if would be to insert a matching pad Another source of post-tuning drift
the preselector was not there. (fixed attenuator) or isolator between is the self-heating caused by current
the preselector and mixer. In this case, flowing in the preselector circuitry.
The discussion of dynamic range in sensitivity would be degraded by the The center of the preselector pass
Chapter 6 applies to the low-pass- full value of the pad or isolator. band will depend on its temperature
filtered low band. The only exceptions and temperature gradients, which
occur when a particular harmonic Some spectrum analyzer architectures depend on the history of the
of a low-band signal falls within the eliminate the need for the matching preselector tuning. As a result, you
preselected range. For example, if pad or isolator. As the electrical length obtain the best flatness by centering
we measure the second harmonic of between the preselector and mixer the preselector at each signal. The
a 2.5- GHz fundamental, we get the increases, the rate of change of phase centering function is typically built into
benefit of the preselector when we of the reflected and re-reflected the spectrum analyzer firmware and is
tune to the 5- GHz harmonic. signals becomes more rapid for a given selected either by a front-panel key in
change in input frequency. The result manual measurement applications or
Pluses and minuses is a more exaggerated ripple effect on programmatically in automated test
of preselection flatness. Architectures such as those systems. When activated, the centering
used in PSA Series analyzers include function adjusts the preselector tuning
We have seen the pluses of the mixer diodes as an integral part of DAC to center the preselector pass
preselection: simpler analyzer the preselector/mixer assembly. In such band on the signal. The frequency
operation, uncluttered displays, an assembly, there is minimal electrical response specification for most
improved dynamic range and length between the preselector and microwave analyzers applies only after
wide spans. But there are also mixer. This architecture thus removes centering the preselector, and it is
some disadvantages relative to an the ripple effect on frequency response generally a best practice to perform
unpreselected analyzer. and improves sensitivity by eliminating this function (to mitigate the effects
the matching pad or isolator. of post-tuning drift) before making
First of all, the preselector has amplitude measurements of microwave
insertion loss, typically 6 to 8 dB. This Even aside from its interaction with signals.
loss comes prior to the first stage of the mixer, a preselector causes some
gain, so system sensitivity is degraded degradation of frequency response. In our discussion of sweep time, we
by the full loss. In addition, when a The preselector filter pass band is found that analyzers such as the PXA
preselector is connected directly to a never perfectly flat, but rather exhibits use FFTs when the narrower resolution

70
bandwidths are selected. Because the second harmonic (N=2–) to tune to to the external mixer, which uses the
LO is stepped and fixed for each FFT 17.1GHz and the second harmonic higher harmonics to mix with the high-
segment, the preseletor must be stepped (N=2–), with the LO doubled, to tune to frequency signals. The external mixer’s
and fixed as well. Since the preselector 26.5 GHz. However, what if you want IF output connects to the analyzer’s IF
takes several milliseconds to tune and to test outside the upper frequency “in” port. The latest analyzers have only
stabilize, sweep time may be negatively range of the signal analyzer? Some one front-panel port, and this is possible
impacted relative to similar settings analyzers provide the ability to use because the LO frequency supplied from
in the low band. The X-Series signal an external mixer to make high- the analyzer is between 3 and 14 GHz,
analyzers allow you to select the width frequency measurements, where the while the IF output frequency from the
of each step to minimize the number of external mixer becomes the front end external mixer to the analyzer is 322.5
steps. (For details, see the operating of the analyzer, bypassing the input MHz. Because of the wide frequency
manual for your particular analyzer.) If attenuator, the preselector and the first difference between the LO and IF
your analyzer has Option MPB, you may mixers. The external mixer uses higher signals, both signals can exist on the
bypass the preselector to eliminate its harmonics of the analyzer’s first LO, and same coaxial interconnect cable that
impact on sweep time. However, be sure in some cases, the first LO frequency attaches the analyzer and the mixer.
your signal is such that no images or is doubled before being sent to the As long as the external mixer uses the
multiples can cause confusion. external mixer. Higher fundamental same IF as the spectrum analyzer, the
LO frequencies allow for lower mixer signal can be processed and displayed
External harmonic mixing conversion loss. Typically, a spectrum internally, just like any signal that came
analyzer that supports external mixing from the internal first mixer. Figure 7-14
We have discussed tuning to higher has one or two additional connectors illustrates the block diagram of an
frequencies within the signal analyzer. on the front panel. Early analyzers had external mixer used in conjunction with
For internal harmonic mixing, the two connectors. An LO “out” port routes a spectrum analyzer.
X-Series signal analyzers use the the analyzer’s internal first LO signal

External mixer

Waveguide IF out
input IF in

Low
band path Analog or
3.6 GHz 5.1225 GHz 322.5 MHz 22.5 MHz digital IF

Analyzer
input

LO
3.8 to 8.7 GHz
High
band path To external
mixer
4.8 GHz 300 MHz

322.5 MHz

Preselector

Sweep generator
Display

Figure 7-14. Spectrum analyzer and external mixer block diagram

71
Table 7-1. Harmonic mixing modes used by X-Series analyzers with external mixers

Band Keysight 11970 Series Keysight M1970 Series Other Manufacturer’s Other Manufacturer’s
Mixers Mixers mixers mixers
(LO range 3–7 GHz) (LO range 6–14 GHz) (LO range 3–7 GHz) (LO range 6–14 GHz)
A (26.5 to 40.0 GHz) 6− and 8−
Q (33.0 to 50.0 GHz) 8− and 10−
U (40.0 to 60.0 GHz) 10−
V (50.0 to 75.0 GHz) 12− and 14− 6−
E (60.0 to 90.0 GHz) N.A. 6− and 8−
W (75.0 to 110.0 GHz) 18− 8−
F (90.0 to 140.0 GHz) 16− 10−
D (110.0 to 170.0 GHz) 20− 14−
G (140.0 to 220.0 GHz) 26− 18−
Y (170.0 to 260.0 GHz ) 30− 20−
J (220.0 to 325.0 GHz) 38− 24−
(325.0 to 500.0 GHz) 58− 36−
(500.0 to 750.0 GHz) 86− 54−
(750.0 to 1,100.0 GHz) 80−

Table 7-1 shows the harmonic mixing frequency. Some external mixers from that may be present at the input. This
modes used by the X-Series analyzers other manufacturers require a bias produces mixing products that can be
at various millimeter-wave bands current to set the mixer diodes to the processed through the IF just like any
for both the Keysight M1970 Series proper operating point. The X-Series other valid signals.
and the earlier 11970 Series external analyzers can provide up to ± 10 mA
mixers. For ease of use and low of DC current through the front-panel A tunable filter that performs
conversion loss, the M1970 Series external mixer port to provide this bias preselection of the signals reaching the
mixers provide a USB connection that and keep the measurement setup as first mixer in the internal signal path
is used to automatically identify the simple as possible. is common in most signal analyzers.
mixer model number and serial number, External mixers that are unpreselected
perform an LO adjustment to optimize Whether you perform harmonic mixing will produce unwanted responses on
performance, and download the mixer with an internal or an external mixer, screen that are not true signals. A way
conversion loss table into the analyzer the issues are similar. The LO and its to deal with these unwanted signals
memory. You also can use external harmonics mix not only with the desired has been designed into the signal
mixers from other manufactures if you input signal, but also with any other analyzer. This function is called “signal
know the mixer’s conversion loss with signal, including out-of-band signals, identification.”

72
Signal identification
It is quite possible that the particular
response we have tuned onto the
display has been generated on an
LO harmonic or mixing mode other
than the one for which the display is
calibrated. So our analyzer must have
some way to tell us whether or not
the display is calibrated for the signal
response in question. For this example,
assume we are using a Keysight
M1970V 50- to 75-GHz unpreselected
mixer, which uses the 6− mixing mode.
The full V-Band measurement can be
seen in Figure 7-15.

Keysight X-Series signal analyzers offer

two different identification methods:
image shift and image suppress. Let’s
first explore the image shift method.
Looking at Figure 7-15, let’s assume we
have tuned the analyzer to a frequency Figure 7-15. Which ones are the real signals?
of 62.50 GHz. The 6th harmonic of
the LO produces a pair of responses, frequency of 61.85 GHz, which is 2
where the 6− mixing product appears times f IF below the real response. The
on screen at the correct frequency of X-Series analyzer has an IF frequency
62.50 GHz, while the 6+ mixing product of 322.5 MHz, so the pair of responses
produces a response with an indicated is separated by 645 MHz.

Harmonic mixing tuning lines

80
8+

IF Frequency = 322.5 MHz 8-

75
Signal Frequency (GHz)

70

Input Image RF 6-

RF 6+ RF 8- RF 8+ 6+
65

6-
62.500
61.855

60

55

10.36

10.47
50
7 7.5 8 8.5 9 9.5 10 10.5 11
LO Frequency (GHz)

Figure 7-16 Harmonic tuning lines for the X-Series analyzers using the M1970 Series mixers

73
Let’s assume we have some idea of the
characteristics of our signal, but we do
not know its exact frequency. How do
we determine which is the real signal?
The image-shift process retunes the LO
fundamental frequency by an amount
equal to 2f IF/N. This causes the Nth
harmonic to shift by 2f IF.

If we are tuned to a real signal, its

corresponding pair will now appear at
the same position on screen that the
real signal occupied in the first sweep.
If we are tuned to another multiple
pair created by some other incorrect
harmonic, the signal will appear to
shift in frequency on the display. The
X-Series signal analyzer shifts the LO
on alternate sweeps, creating the two
displays show in Figures 7-17a and
7-17b. In Figure 7-17a, the real signal
(the 6 − mixing product) is tuned to
Figure 7-17a: 6− centered (yellow trace)
the center of the screen. Figure 7-17b
shows how the image shift function
moves the corresponding pair (the 6+
mixing product) to the center of the
screen.

Figures 7-17a and 7-17b display

alternate sweeps taken with the image
shift function.

Figure 7-17b: 6+ centered (blue trace)

74
Let’s examine the second method
of signal identification, image
suppression. In this mode, two sweeps
are taken using the minimum hold
function, which saves the smaller value
of each display point, or bucket, from
the two sweeps. The first sweep is
done using normal LO tuning values.
The second sweep offsets the LO
fundamental frequency by 2f IF/N. As
we saw in the first signal ID method,
the image product generated by the
correct harmonic will land at the same
point on the display as the real signal
did on the first sweep. Therefore,
the trace retains a high amplitude
value. Any false response that shifts
in frequency will have its trace data
replaced by a lower value. Thus, all
image and incorrect multiple responses
will appear as noise, as shown in Figure
7-18.
Figure 7-18. The image suppress function displays only real signals

Note that both signal identification

methods are used for identifying
correct frequencies only. You should
not attempt to make amplitude
measurements while the signal
identification function is turned on.
Once we have identified the real signal
of interest, we turn off the signal ID
function and zoom in on it by reducing
the span. We can then measure the
signal’s amplitude and frequency. See
Figure 7-19.

To make an accurate amplitude

measurement, it is important that you
first enter the calibration data for your
external mixer. This data is normally
supplied by the mixer manufacturer,
and it is typically presented as a
table of mixer conversion loss, in
dB, at a number of frequency points
across the band. This data is entered
into a correction table on the signal
Figure 7-19. Measurement of a positively identified signal
analyzer, and the analyzer uses this
data to compensate for the mixer
conversion loss. If you are using the
M1970 Series harmonic mixers, the
mixer conversion loss is automatically
transferred from the mixer memory to
the X-Series signal analyzer memory,
which eliminates manual entry into a
correction file. The spectrum analyzer
reference level is now calibrated for
signals at the input to the external
mixer.
75
Chapter 8. Modern Signal Analyzers
In the previous chapters of this Application-specific time the instantaneous power of the
application note, we have looked at the signal exceeds the average power by a
measurements
fundamental architecture of spectrum certain number of dB. This information
analyzers and basic considerations In addition to measuring general signal is important in power amplifier design,
for making frequency-domain characteristics like frequency and for example, where it is important to
measurements. On a practical level, amplitude, you often need to make handle instantaneous signal peaks with
modern spectrum or signal analyzers specific measurements of certain signal minimum distortion while minimizing
must also handle many other tasks parameters. Examples include channel cost, weight and power consumption of
to help you meet your measurement power measurements and adjacent the device.
requirements. These tasks include: channel power (ACP) measurements,
which we described in Chapter 6. Other examples of built-in
–– Providing application-specific Many signal analyzers now have these measurement functions include
measurements, such as adjacent built-in functions available. You simply occupied bandwidth, TOI, harmonic
channel power (ACP), noise figure, specify the channel bandwidth and distortion, and spurious emissions
and phase noise spacing, then press a button to activate measurements. The instrument
–– Providing digital modulation the automatic measurement. settings – such as center frequency,
analysis measurements defined by span and resolution bandwidth – for
industry or regulatory standards, The complementary cumulative these measurements depend on the
such as LTE, GSM, cdma2000 ®, distribution function (CCDF), which specific radio standard to which the
802.11, or Bluetooth® shows power statistics, is another device is being tested. Most modern
–– Performing vector signal analysis measurement capability increasingly signal analyzers have these instrument
–– Saving, printing or transferring found in modern signal analyzers, settings stored in memory so you
data as you can see in Figure 8-1. CCDF can select the desired radio standard
–– Offering remote control and measurements provide statistical (LTE, MSR, GSM/EDGE, cdma2000,
operation over GPIB, LAN or the information showing the percent of W-CDMA, 802.11a/b/g/n/ac and so on)
Internet
–– Allowing you to update instrument
firmware to add new features and
capabilities, as well as to repair
defects
–– Making provisions for self-
calibration, troubleshooting,
diagnostics and repair
–– Recognizing and operating with
optional hardware or firmware to
add new capabilities
–– Allowing you to make
measurements in the field with
a rugged, battery-powered
handheld spectrum analyzer that
correlate with data taken with
high-performance bench-top
equipment

Figure 8-1. CCDF measurement

76
to make the measurements properly.
RF designers are often concerned
with the noise figure of their devices,
as noise figure directly affects the
sensitivity of receivers and other
systems. Some s ign a l analyzers,
such as the X-Series, have optional
noise figure measurement capabilities
available. This option provides control
for the noise source needed to drive
the input of the device under test
(DUT) as well as firmware to automate
the measurement process and display
the results. Figure 8-2 shows a typical
measurement result, with DUT noise
figure (upper trace) and gain (lower
trace) displayed as a function of
frequency.

The need for phase

information
Figure 8-2. Noise figure measurement
Phase noise is a common measure
of oscillator performance. In digitally
modulated communication systems,
phase noise can negatively impact
bit error rates. Phase noise can also
degrade the ability of Doppler radar
systems to capture the return pulses
from targets. The X-Series signal
analyzers offer optional phase noise
measurement capabilities. These
options provide firmware to control the
measurement and display the phase
noise as a function of frequency offset
from the carrier, as shown in Figure
8-3.

Figure 8-3. Phase noise measurement

77
Today’s digital modulation techniques
use amplitude and phase information
to carry more baseband data in limited
spectrum and time. Therefore, it is
essential for a signal analyzer to handle
amplitude and phase in an appropriate
manner. QPSK (Quadrature Phase-Shift
Keying) is one of the simple examples
of digital modulation techniques,
with which two bits of digital data are
transmitted at once, or two bits per
symbol. Figure 8-4 shows an example
of QPSK demodulation analysis with
Agilent 89601B option AYA. Remember
you need four (22) states to transmit 2
bits at once.

As an easy and intuitive way to

understand what’s going on the digital
radio transmission, use an I/Q plane, Figure 8-4. Modulation analysis of a QPSK signal measured with Keysight’s 89600 VSA software
which is a two dimensional chart
comprising in-phase and quadrature
components of the demodulated
signal on the horizontal axis and the
vertical axis, respectively. An example
of the chart is again shown on the top
left window of Figure 8-4. The yellow
trace called trajectory shows a vector
combining phase and amplitude moves
around as time goes while red points
indicates the instantaneous position
of trajectory at the time of decision
when a receiver actually judges the
symbol value. Essentially, for digital
radios, vectors at these decision points
are most important for modulation
quality. As you can see on the bottom
left window of Figure 8-4, a “scalar”
analyzer meaning traditional spectrum
analyzer may be able to show the
modulated signal in frequency domain Figure 8-5. Modulation analysis of WLAN 802.11ac signal using Keysight 89600 VSA software
so that you can see whether the signal
is properly modulated in power wise to
some extent, and you can also make
sure that there is no unwanted emission maximum power is limited, so the
or leakage power to the adjacent data points are much closer in both
channels. You need, however, some phase and magnitude than for QPSK.
sort of “vector” analyzer to perform The analyzer you use to evaluate the
meaningful analysis of modulation transmitted signal must be sufficiently
quality for digital data transmission accurate that it does not lead you to
where phase information is involved. a false conclusion about the quality
of the transmission. Pure amplitude
A newer and much more complicated measurements are also required to
system is 802.11ac, which uses determine signal attributes such as
256QAM (quadrature-amplitude flatness, adjacent-channel power levels
modulation). See Figure 8-5. The and distortion.

78
Digital modulation analysis
The common wireless communication
systems used throughout the world
today all have prescribed measurement
techniques defined by standards-
development organizations and
governmental regulatory bodies.
Optional measurement personalities
are available on the X-Series signal
analyzers to perform the key tests
defined for a particular wireless
communication format. For example,
if we need to test a transmitter to the
Bluetooth wireless communication
standard, we must measure
parameters such as:

–– Average/peak output power

–– Modulation characteristics
–– Initial carrier frequency tolerance
–– Carrier frequency drift
–– Monitor band/channel Figure 8-6. EVM measurement of LTE FDD downlink signal
–– Modulation overview
–– Output spectrum Not all digital communication systems
–– 20-dB bandwidth are based on well-defined industry More information
–– Adjacent channel power standards. If you are working on
Additional information is
nonstandard proprietary systems or
available on the following:
These measurements are available on the early stages of proposed industry-
the Keysight X-Series signal analyzers standard formats, you need more
Noise figure measurements, see
with appropriate options. flexibility to analyze vector-modulated
Keysight Noise Figure Measurements of
signals under varying conditions. You
Frequency Converting Devices Using the
Other optional measurement can achieve that flexibility two ways.
Keysight NFA Series Noise Figure Ana-
capabilities for a wide variety of First, modulation analysis personalities
lyzer − Application Note, literature
wireless communications standards are available on the X-Series signal
number 5989-0400EN.
that are available on the X-Series analyzers. Alternatively, you can
signal analyzers: perform more extensive analysis
Measurements involving phase,
with software running on an external
see Vector Signal Analysis Basics –
–– LTE/LTE-Advanced computer. For example, you can use
Application Note, literature num-
–– WLAN Keysight 89600 VSA software with
ber 5989-1121EN.
–– Multi-standard radio (MSR) X-Series signal analyzers to provide
–– GSM/EDGE flexible vector signal analysis. In this
Bluetooth measurements, see
–– W-CDMA case, the signal analyzer acts as
Performing Bluetooth RF Measurements
–– HSDPA an RF downconverter and digitizer.
Today – Application Note, literature
–– cdma2000 The software can run internally on
number 5968-7746E.
–– 1xEV-DO the signal analyzer or communicate
–– 1xEV-DV with the analyzer over a GPIB or LAN
–– cdmaOne connection. IQ data is transferred to
–– NADC and PDC the computer, where it performs the
–– TD-SCDMA vector signal analysis. Measurement
settings, such as modulation type,
Figure 8-6 illustrates an error vector symbol rate, filtering, triggering
magnitude (EVM) measurement and record length, can be varied as
performed on a LTE FDD downlink necessary for the particular signal you
signal. This test helps you diagnose are analyzing.
modulation or amplification distortions
that lead to bit errors in the receiver.

79
Real-time spectrum analysis
For the capable RF engineer, continuous-
wave (CW) and predictably-repeating
signals are no great challenge – but
today’s complex and agile signals and
multi-signal environments are proving
to be another matter. To keep up with
evolving analysis needs, new types
of signal analyzers and application
software have emerged in recent years.
The Keysight PXA and MXA signal
analyzers now offer a combination of
swept spectrum, real-time and vector
signal analysis capability – all in one
instrument.

Design and troubleshooting tasks are

much more difficult when dealing with
agile signals, and the challenges are Figure 8-7 . Even when you use fast sweeps and max hold over a period of many seconds, the swept
often made more difficult when these spectrum analyzer view of the radar signal is not very informative
signals are in an environment of other
agile signals. Even the analysis of a
single signal can be a challenge when
that signal is very agile or complex. You
can use the Keysight PXA and MXA
real-time spectrum analysis capability
to capture the behavior of dynamic
and elusive signals with true gap-free
spectrum analysis.

An example of a single complex signal

is the agile S-band acquisition radar
signal. The signal at the receiver varies
widely in amplitude over a period of
several seconds, and this long-duration
characteristic, combined with the
short-duration characteristics of
its pulse length and repetition interval
(and therefore short duty cycle) make
it agile and difficult to measure well. A Figure 8-8. Real-time capture of S-band acquisition radar signal
basic spectrum analysis of this signal
with a swept spectrum analyzer shows
the measurement difficulty it poses, data into a single display that shows measure (or even to rapidly and reliably
as illustrated in Figure 8-7. Even after both rare and frequent events, with find) this signal with a swept spectrum
many sweeps and the use of an indication of relative frequency of analyzer.
a max hold function, the signal is not occurrence.
clearly represented.
The PXA’s real-time analyzer mode
The Keysight PXA real-time spectrum and density display provide a fast and More information
analyzer screen shown in Figure 8-8, insight-producing representation of this For additional information on
in contrast with the swept spectrum wideband, dynamic and agile signal. measurements involving real-time
screen, readily shows the main The blue color of all but the noise spectrum analysis, see Measur-
characteristics of the signal using floor indicates that the pulses, while ing Agile Signals and Dynamic Signal
the density or histogram display. The prominent, have a very low frequency- Environments – Application Note,
density or histogram display collects of-occurrence. This is the principal literature number 5991-2119EN.
a large amount of real-time spectrum characteristic that makes it difficult to

80
Chapter 9. Control and Data Transfer
Saving and printing data Data transfer and remote A variety of commercial software
products are available to control
After making a measurement, we instrument control
spectrum analyzers remotely over
normally want to keep a record of the In 1977, Keysight Technologies (part an I/O bus. Also, you can write your
test data. We might simply want to of Hewlett-Packard at that time) own software to control spectrum
make a quick printout of the instrument introduced the world’s first GPIB- analyzers in a number of different
display. Depending on the particular controllable spectrum analyzer, the ways. One method is to directly send
analyzer and printer model, we might 8568A. The GPIB interface (also known programming commands to the
use the USB or LAN ports to connect as HP-IB or IEEE-488) made it possible instrument. Older spectrum analyzers
the two units. to control all major functions of the typically used proprietary command
analyzer from an external computer sets, but newer instruments, such as
Very often, we may want to save and transfer trace data to an external Keysight’s X-Series signal analyzers,
measurement data as a file, either computer. This innovation paved the use industry-standard SCPI (standard
in the spectrum analyzer’s internal way for a wide variety of automated commands for programmable
memory or on a USB mass-storage spectrum analyzer measurements that instrumentation) commands. A more
device. There are several different were faster and more repeatable than common method is to use standard
kinds of data we can save this way: manual measurements. By transferring software drivers, such as VXI plug&play
the raw data to a computer, it could be drivers, which enable higher-level
–– An image of the display – saved on disk, analyzed, corrected and functional commands to the instrument
Preferably in a popular file format, operated on in a variety of ways. without the need for detailed
such as bitmap, GIF, PNG or knowledge of the SCPI commands.
Windows metafile. Today, automated test and Most recently, a new generation of
–– Trace data – Saved as X-Y data measurement equipment has become language-independent instrument
pairs representing frequency and the norm, and nearly all modern drivers, known as “interchangeable
amplitude points on the screen. spectrum analyzers come with a variety virtual instrument,” or IVI-COM drivers,
The number of data pairs can vary. of standard interfaces, including LAN, has become available for the X-Series
Modern spectrum analyzers such as USB 2.0 and GPIB. LAN connectivity is signal analyzers. The IVI-COM drivers
the X-Series allow you to select the the most commonly used interface, as are based on the Microsoft Component
desired display resolution by setting it can provide high data-transfer rates Object Model standard and work in a
a minimum of 1 up to a maximum over long distances and integrates variety of PC application development
of 40,001 frequency sweep points easily into networked environments environments, such as the Keysight
(buckets) on the screen. This data such as a factory floor. Other standard T&M Programmers Toolkit and
format is well suited for transfer interfaces used widely in the computer Microsoft’s Visual Studio .NET.
to a spreadsheet program on a industry are likely to become available
computer. on spectrum analyzers in the future Some applications require you to
–– Instrument state – To keep a record to simplify connectivity between control the spectrum analyzer and
of the spectrum analyzer settings, instruments and computers. collect measurement data from a very
such as center frequency, span, long distance. For example, you may
reference level and so on, used in Keysight’s X-Series signal analyzers want to monitor satellite signals from
the measurement. This information literally have computer firmware a central control room, collecting
is useful for documenting running USB ports and a Windows data from remote tracking stations
test setups used for making operating system. These features located hundreds or even thousands
measurements. Consistent test greatly simplify control and data of kilometers from the central site. The
setups are essential for maintaining transfer. In addition, the X-Series X-Series signal analyzers have software
repeatable measurements over analyzers can be operated remotely, available to control these units, capture
time. and the analyzer’s display appears screen images and transfer trace data
on the remote computer. Details are over the Internet using a standard Web
beyond the scope of this application browser.
note; see the operating manual for your
particular analyzer.

81
Firmware updates Calibration, troubleshooting, Summary
Modern spectrum analyzers have much diagnostics and repair This application note has provided
more software inside them than do a broad survey of basic spectrum
Spectrum analyzers must be
instruments from just a few years ago. As analyzer concepts. However, you may
periodically calibrated to insure the
new features are added to the software wish to learn more about many other
instrument performance meets all
and defects repaired, it becomes highly topics related to spectrum analysis. An
published specifications. Typically, this
desirable to update the spectrum excellent place to start is to visit the
is done once a year. However, between
analyzer’s firmware to take advantage of Keysight Technologies Web site at
these annual calibrations, the spectrum
the improved performance. www.keysight.com and search for
analyzer must be aligned periodically to
signal or spectrum analyzer.
compensate for thermal drift and aging
The latest revisions of spectrum and effects. Modern spectrum analyzers
signal analyzer firmware can be found such as the X-Series have built-in
on the Keysight Technologies website. alignment routines that operate when
You can download this firmware to a the instrument is first turned on and
file on your local computer. A common during retrace (between sweeps) at
method to transfer new firmware into predetermined intervals. The alignment
the spectrum analyzer is to copy the routines also operate if the internal
firmware onto a USB drive and then temperature of the instrument changes.
insert it into one of the spectrum These alignment routines continuously
analyzer’s USB ports. Some models, adjust the instrument to maintain
such as the X-Series, allow you to specified performance.
transfer the new firmware directly
into the spectrum analyzer using the Modern spectrum analyzers usually
instrument’s Ethernet LAN port. have a service menu available. In this
area, you can perform useful diagnostic
It is a good practice to periodically functions, such as a test of the front-
check your spectrum analyzer model’s panel keys. You also can display more
Web page to see if updated firmware is details of the alignment process, as
available. well as a list of all optional hardware
and measurement personalities
installed in the instrument. When you
upgrade a spectrum analyzer with a
new measurement personality, Keysight
provides a unique license key tied to
the serial number of the instrument.
You install this license key through the
USB port or enter it on the front-panel
keypad to activate the measurement
capabilities of the personality.

82
Glossary of Terms
Absolute amplitude accuracy: Average detection: A method of Delta marker: A mode in which a fixed,
The uncertainty of an amplitude detection that sums power across a reference marker has been established
measurement in absolute terms, frequency interval. It is often used for and a second, active marker is available
either volts or power. Includes relative measuring complex, digitally modulated that we can place anywhere on the
uncertainties (see Relative amplitude signals and other types of signals with displayed trace. A read-out indicates
accuracy) plus calibrator uncertainty. noise-like characteristics. Modern the relative frequency separation and
For improved accuracy, some spectrum Keysight spectrum analyzers typically amplitude difference between the
analyzers have frequency response offer three types of average detection: reference marker and the active marker.
specified relative to the calibrator as power (rms) averaging, which measures
well as relative to the midpoint between the true average power over a bucket Digital display: A display technology
peak-to-peak extremes. interval; voltage averaging, which where digitized trace information,
measures the average voltage data stored in memory, is displayed on an
ACPR: Adjacent channel power ratio is over a bucket interval; and log-power instrument’s screen. The displayed trace
a measure of how much signal energy (video) averaging, which measures is a series of points designed to present
from one communication channel spills the logarithmic amplitude in dB of the a continuous-looking trace. While
over or leaks into an adjacent channel. envelope of the signal during the bucket the default number of display points
This is an important metric in digital interval. varies between different models, most
communication components and modern spectrum analyzers allow the
systems, as too much leakage will cause Average noise level: See Displayed user to choose the desired resolution
interference on adjacent channels. It is average noise level. by controlling the number of points
sometimes also described as ACLR, or displayed. The display is refreshed
adjacent channel leakage ratio. Bandwidth selectivity: A measure of (rewritten from data in memory) at a
an analyzer’s ability to resolve signals flicker-free rate; the data in memory is
Amplitude accuracy: The uncertainty unequal in amplitude. Also called shape updated at the sweep rate. Nearly all
of an amplitude measurement. It can be factor, bandwidth selectivity is the ratio modern spectrum analyzers have digital
expressed either as an absolute term or of the 60-dB bandwidth to the 3-dB flat-panel LCD displays, rather than
relative to another reference point. bandwidth for a given resolution (IF) CRT-based analog displays that were
filter. For some analyzers, the 6-dB used in earlier analyzers.
Amplitude reference signal: A signal of bandwidth is used in lieu of the 3-dB
precise frequency and amplitude that bandwidth. In either case, bandwidth Display detector mode: The manner
the analyzer uses for self-calibration. selectivity tells us how steep the filter in which the signal information is
skirts are. processed prior to being displayed on
Analog display: A display technology screen. See Neg peak, Pos peak,
where analog signal information (from Blocking capacitor: A filter that keeps Normal, Average and Sample.
the envelope detector) is written directly unwanted low-frequency signals
to an instrument’s display, typically (including DC) from damaging circuitry. Digital IF: An architecture found in
implemented on a cathode ray tube A blocking capacitor limits the lowest modern spectrum analyzers in which the
(CRT). Analog displays were once frequency that can be measured signal is digitized soon after it has been
the standard method of displaying accurately. downconverted from an RF frequency
information on spectrum analyzers. to an intermediate frequency (IF). At
However, modern spectrum analyzers no CDMA: Code division multiple access that point, all further signal processing
longer use this technology; instead, they is a method of digital communication in is done using digital signal processing
now use digital displays. which multiple communication streams (DSP) techniques.
are orthogonally coded, enabling them
to share a common frequency channel. It Display dynamic range: The maximum
is a popular technique used in a number dynamic range for which both the
of widely used mobile communication larger and smaller signal may be viewed
systems. simultaneously on the spectrum analyzer
display. For analyzers with a maximum
Constellation diagram: A display type logarithmic display of 10 dB/div, the
commonly used when analyzing digitally actual dynamic range (see Dynamic
modulated signals in which the detected range) may be greater than the display
symbol points are plotted on an IQ dynamic range.
graph.

83
Display scale fidelity: The degree input signal, and we can observe the them individually. Resolution of equal
of uncertainty in measuring relative modulating signal as a function of time amplitude components is determined
differences in amplitude on a spectrum on the display. by resolution bandwidth. The ability to
analyzer. The logarithmic and linear resolve unequal amplitude signals is a
IF amplifiers found in analyzers with Error vector magnitude (EVM): A quality function of both resolution bandwidth
analog IF sections never have perfect metric in digital communication systems. and bandwidth selectivity.
logarithmic or linear responses, and EVM is the magnitude of the vector
therefore they introduce uncertainty. difference at a given instant in time Frequency response: Variation in the
Modern analyzers with digital IF sections between the ideal reference signal and displayed amplitude of a signal as a
have significantly better display scale the measured signal. function of frequency (flatness). Typically
fidelity. specified in terms of ± dB relative to the
External mixer: An independent mixer, value midway between the extremes.
Display range: The calibrated range usually with a waveguide input port, Also may be specified relative to the
of the display for the particular display used to extend the frequency range of calibrator signal.
mode and scale factor. See Linear and spectrum analyzers that use external
Log display and Scale factor. mixers. The analyzer provides the LO Frequency span: The frequency range
signal and, if needed, mixer bias. Mixing represented by the horizontal axis of
Displayed average noise level: The products are returned to the analyzer’s the display. Generally, frequency span
noise level as seen on the analyzer’s IF input. is given as the total span across the full
display after setting the video bandwidth display. Some earlier analyzers indicate
narrow enough to reduce the peak- FFT (fast Fourier transform): A frequency span (scan width) on a per-
to-peak noise fluctuations such that mathematical operation performed division basis.
the displayed noise is essentially seen on a time-domain signal to yield the
as a straight line. Usually refers to the individual spectral components that Frequency stability: A general phrase
analyzer’s own internally generated constitute the signal. See Spectrum. that covers both short- and long-term
noise as a measure of sensitivity and LO instability. The sweep ramp that
is typically specified in dBm under Fast sweep: A digital signal processing tunes the LO also determines where
conditions of minimum resolution technique that implements complex- a signal should appear on the display.
bandwidth and minimum input valued resolution bandwidth filtering for Any long term variation in LO frequency
attenuation. a sweeping spectrum analyzer, allowing (drift) with respect to the sweep
faster sweep rates than a traditional ramp causes a signal to slowly shift
Drift: The very slow (relative to sweep analog or digital resolution bandwidth its horizontal position on the display.
time) change of signal position on filter would allow. Shorter-term LO instability can appear
the display as a result of a change in as random FM or phase noise on an
LO frequency versus sweep voltage. Flatness: See Frequency response. otherwise stable signal.
The primary sources of drift are the
temperature stability and aging rate of Frequency accuracy: The degree of Full span: For most modern spectrum
the frequency reference in the spectrum uncertainty with which the frequency analyzers, full span means a frequency
analyzer. of a signal or spectral component is span that covers the entire tuning
indicated, either in an absolute sense or range of the analyzer. These analyzers
Dynamic range: The ratio, in dB, relative to some other signal or spectral include single -band RF analyzers and
between the largest and smallest signals component. Absolute and relative microwave analyzers such as the ESA,
simultaneously present at the spectrum frequency accuracies are specified PSA and X- Series that use a solid-state
analyzer input that can be measured to independently. switch to switch between the low and
a given degree of accuracy. Dynamic preselected ranges.
range generally refers to measurement of Frequency range: The minimum to
distortion or intermodulation products. maximum frequencies over which a NOTE: On some earlier spectrum
spectrum analyzer can tune. While the analyzers, full span referred to a sub-
Envelope detector: A circuit element maximum frequency is generally thought range. For example, with the Keysight
whose output follows the envelope, of in terms of an analyzer’s coaxial input, 8566B, a microwave spectrum analyzer
but not the instantaneous variation, of the range of many microwave analyzers that used a mechanical switch to switch
its input signal. In a superheterodyne can be extended through use of external between the low and preselected
spectrum analyzer, the input to the waveguide mixers. ranges, full span referred to either the
envelope detector comes from the final low, non-preselected range or the high,
IF, and the output is a video signal. Frequency resolution: The ability of a preselected range.
When we put our analyzer in zero span, spectrum analyzer to separate closely
the envelope detector demodulates the spaced spectral components and display

84
Gain compression: That signal level at frequency indicated by the spectrum Linear display: The display mode in
the input mixer of a spectrum analyzer analyzer. For each harmonic of the LO, which vertical deflection on the display
at which the displayed amplitude of there is an image pair, one below and is directly proportional to the voltage of
the signal is a specified number of dB one above the LO frequency by the IF. the input signal. The bottom line of the
too low due just to mixer saturation. Images usually appear only on non- graticule represents 0 V, and the top
The signal level is generally specified preselected spectrum analyzers. line, the reference level, some nonzero
for 1-dB compression and is usually value that depends upon the particular
between +3 and –10 dBm, depending on Incidental FM: Unwanted frequency spectrum analyzer. On most modern
the model of spectrum analyzer. modulation on the output of a device analyzers, we select the reference
(signal source, amplifier) caused by level, and the scale factor becomes
GSM: The global system for mobile (incidental to) some other form of the reference level value divided by the
communication is a widely used digital modulation, e.g., amplitude modulation. number of graticule divisions. Although
standard for mobile communication. It is the display is linear, modern analyzers
a TDMA-based system in which multiple Input attenuator: A step attenuator allow reference level and marker values
communication streams are interleaved between the input connector and first to be indicated in dBm, dBmV, dBuV, and
in time, enabling them to share a mixer of a spectrum analyzer. Also in some cases, watts as well as volts.
common frequency channel. called the RF attenuator. The input
attenuator is used to adjust level of LO emission or feedout: The emergence
Harmonic distortion: Unwanted the signal incident upon the first mixer. of the LO signal from the input of a
frequency components added to a signal The attenuator is used to prevent spectrum analyzer. The level can be
as the result of the nonlinear behavior gain compression due to high-level or greater than 0 dBm on non-preselected
of the device (e.g., mixer, amplifier) broadband signals and to set dynamic spectrum analyzers but is usually less
through which the signal passes. These range by controlling the degree of than –70 dBm on preselected analyzers.
unwanted components are harmonically internally generated distortion. In
related to the original signal. some analyzers, the vertical position of LO feedthrough: The response on the
displayed signals is changed when the display when a spectrum analyzer is
Harmonic mixing: Using the LO
input attenuator setting is changed, tuned to 0 Hz, i.e., when the LO is tuned
harmonics generated in a mixer to
so the reference level is also changed to the IF. The LO feedthrough can be
extend the tuning range of a spectrum
accordingly. In modern Keysight used as a 0-Hz marker, and there is no
analyzer beyond the range achievable
analyzers, the IF gain is changed frequency error.
using just the LO fundamental.
to compensate for input attenuator
IF gain/IF attenuation: Adjusts the changes, so signals remain stationary on Log display: The display mode in which
vertical position of signals on the display the display, and the reference level is not vertical deflection on the display is a
without affecting the signal level at changed. logarithmic function of the voltage of
the input mixer. When changed, the the input signal. We set the display
value of the reference level is changed Input impedance: The terminating calibration by selecting the value of the
accordingly. impedance that the analyzer presents top line of the graticule, the reference
to the signal source. The nominal level and scale factor in dB/div. On
IF feedthrough: A raising of the baseline impedance for RF and microwave Keysight analyzers, the bottom line of
trace on the display due to an input analyzers is usually 50 ohms. For some the graticule represents zero volts for
signal at the intermediate frequency systems, e.g., cable TV, 75 ohms is scale factors of 10 dB/div or more, so the
passing through the input mixer. standard. The degree of mismatch bottom division is not calibrated in these
Generally, this is a potential problem between the nominal and actual input cases. Modern analyzers allow reference
only on non-preselected spectrum impedance is given in terms of VSWR level and marker values to be indicated
analyzers. The entire trace is raised (voltage standing wave ratio). in dBm, dBmV, dBuV, volts, and in some
because the signal is always at the IF; cases, watts. Earlier analyzers usually
mixing with the LO is not required. Intermodulation distortion: Unwanted offered only one choice of units, and
frequency components resulting from dBm was the usual choice.
Image frequencies: Two or more real the interaction of two or more spectral
signals present at the spectrum analyzer components passing through a device Marker: A visible indicator we can
input that produce an IF response at with nonlinear behavior (e.g., mixer, place anywhere along the displayed
the same LO frequency. Because the amplifier). The unwanted components signal trace. A read out indicates the
mixing products all occur at the same are related to the fundamental absolute value of both the frequency and
LO and IF frequencies, it is impossible to components by sums and differences amplitude of the trace at the marked
distinguish between them. of the fundamentals and various point. The amplitude value is given in the
harmonics, e.g. f1 ± f 2, 2f1 ± f 2, 2f 2 ± f1, currently selected units. Also see Delta
Image response: A displayed signal that 3f1 ± 2f 2, and so forth. marker and Noise marker.
is actually twice the IF away from the
85
Measurement range: The ratio, bandwidth. The normalization process Quasi-peak detector (QPD): A type of
expressed in dB, of the maximum signal accounts for detection and bandwidth detector whose output is a function of
level that can be measured (usually plus the effect of the log amplifier when both signal amplitude as well as pulse
the maximum safe input level) to the we select the log display mode. repetition rate. The QPD gives higher
displayed average noise level (DANL). weighting to signals with higher pulse
This ratio is almost always much Noise power bandwidth: A fictitious repetition rates. In the limit, a QPD will
greater than can be realized in a single filter that would pass the same exhibit the same amplitude as a peak
measurement. See Dynamic range. noise power as the analyzer’s actual detector when measuring a signal with a
filter, making comparisons of noise constant amplitude (CW) signal.
Mixing mode: A description of the measurements among different
particular circumstance that creates a analyzers possible. Raster display: A TV-like display in
given response on a spectrum analyzer. which the image is formed by scanning
The mixing mode, e.g., 1+, indicates the Noise sidebands: Modulation sidebands the electron beam rapidly across and
harmonic of the LO used in the mixing that indicate the short-term instability slowly down the display face and
process and whether the input signal is of the LO (primarily the first LO) system gating the beam on as appropriate.
above (+) or below (–) that harmonic. of a spectrum analyzer. The modulating The scanning rates are fast enough to
signal is noise, in the LO circuit itself produce a flicker-free display. Also see
Multiple responses: Two or more or in the LO stabilizing circuit, and the Vector display and Sweep time.
responses on a spectrum analyzer sidebands comprise a noise spectrum.
display from a single input signal. The mixing process transfers any LO Real-time spectrum analyzer: A method
Multiple responses occur only when instability to the mixing products, so the of signal analysis in which all signal
mixing modes overlap and the LO is noise sidebands appear on any spectral samples are processed for some sort
swept over a wide enough range to allow component displayed on the analyzer far of measurement result or triggering
the input signal to mix on more than one enough above the broadband noise floor. operation. There are no gaps between
mixing mode. Normally not encountered Because the sidebands are noise, their time acquisitions while nonreal-time
in analyzers with preselectors. level relative to a spectral component operations leave gaps.
is a function of resolution bandwidth.
Negative peak: The display detection Noise sidebands are typically specified Reference level: The calibrated vertical
mode in which each displayed point in terms of dBc/Hz (amplitude in a 1-Hz position on the display used as a
indicates the minimum value of the video bandwidth relative to the carrier) at a reference for amplitude measurements.
signal for that part of the frequency span given offset from the carrier, the carrier The reference level position is normally
or time interval represented by the point. being a spectral component viewed on the top line of the graticule.
the display.
Noise floor extension: Developed by Relative amplitude accuracy:
Keysight Technologies, Inc., a modeling Phase noise: See Noise sidebands. The uncertainty of an amplitude
algorithm of the noise power in a signal measurement in which the amplitude
analyzer which can be subtracted from Positive peak: The display detection of one signal is compared to the
the measurement results to reduce the mode in which each displayed point amplitude of another regardless of
effective noise level. indicates the maximum value of the absolute amplitude of either.
the video signal for that part of the Distortion measurements are relative
Noise figure: The ratio, usually frequency span or time interval measurements. Contributors to
expressed in dB, of the signal-to-noise represented by the point. uncertainty include frequency response
ratio at the input of a device (mixer, Preamplifier: An external, low-noise- and display fidelity and changes of input
amplifier) to the signal-to-noise ratio at figure amplifier that improves system attenuation, IF gain, scale factor and
the output of the device. (preamplifier/spectrum analyzer) resolution bandwidth.
sensitivity over that of the analyzer itself.
Noise marker: A marker whose value Residual FM: The inherent short-term
indicates the noise level in a 1-Hz noise Preselector: A tunable bandpass frequency instability of an oscillator in
power bandwidth. When the noise filter that precedes the input mixer of the absence of any other modulation.
marker is selected, the sample display a spectrum analyzer and tracks the In the case of a spectrum analyzer, we
detection mode is activated, the values appropriate mixing mode. Preselectors usually expand the definition to include
of a number of consecutive trace points are typically used only above 2 GHz. the case in which the LO is swept.
(the number depends upon the analyzer) They essentially eliminate multiple and Residual FM is usually specified in peak-
about the marker are averaged, and image responses and, for certain signal to-peak values because they are most
this average value is normalized to an conditions, improve dynamic range. easily measured on the display, if visible
equivalent value in a 1-Hz noise power at all.

86
Residual responses: Discrete responses Shape factor: See Bandwidth selectivity. and the start of the next. In zero span,
seen on a spectrum analyzer display the spectrum analyzer’s LO is fixed,
with no input signal present. Signal analyzer: A spectrum analyzer so the horizontal axis of the display
that also uses digital signal processing is calibrated in time only. In nonzero
Resolution: See Frequency resolution. to perform other more complex spans, the horizontal axis is calibrated
measurements such as vector signal in both frequency and time, and sweep
Resolution bandwidth: The width of analysis. time is usually a function of frequency
the resolution bandwidth (IF) filter of span, resolution bandwidth and video
a spectrum analyzer at some level Signal identification: A routine, either bandwidth.
below the minimum insertion loss point manual or automatic, that indicates
(maximum deflection point on the whether or not a particular response on Time gating: A method of controlling
display). For Keysight analyzers, the the spectrum analyzer’s display is from the frequency sweep of the spectrum
3-dB bandwidth is specified; for some the mixing mode for which the display is analyzer based on the characteristics
others, it is the 6-dB bandwidth. calibrated. If automatic, the routine may of the signal being measured. It is often
change the analyzer’s tuning to show useful when analyzing pulsed or burst
Rosenfell: The display detection mode the signal on the correct mixing mode, modulated signals’ time-multiplexed
in which the value displayed at each or it may tell us the signal’s frequency signals and intermittent signals.
point is based upon whether or not the and give us the option of ignoring the
video signal both rose and fell during the signal or having the analyzer tune itself TDMA: Time division multiple access
frequency or time interval represented properly for the signal. Generally not is a digital communication method in
by the point. If the video signal only needed on preselected analyzers. which multiple communication streams
rose or only fell, the maximum value is are interleaved in time, enabling them to
displayed. If the video signal did both Span accuracy: The uncertainty of the share a common frequency channel.
rise and fall, then the maximum value indicated frequency separation of any
during the interval is displayed by odd- two signals on the display. Units: Dimensions of the measured
numbered points, the minimum value, quantities. Units usually refer to
by even-numbered points. To prevent Spectral purity: See Noise sidebands. amplitude quantities because they
the loss of a signal that occurs only in an can be changed. In modern spectrum
even-numbered interval, the maximum Spectral component: One of the sine analyzers, available units are dBm
value during this interval is preserved, waves comprising a spectrum. (dB relative to 1 milliwatt dissipated
and in the next (odd-numbered) interval, in the nominal input impedance of
the displayed value is the greater of Spectrum: An array of sine waves of the analyzer), dBmV (dB relative to
either the value carried over or the differing frequencies and amplitudes and 1 millivolt), dBuV (dB relative to 1
maximum that occurs in the current properly related with respect to phase microvolt), volts, and in some analyzers,
interval. that, taken as a whole, constitute a watts. In Keysight analyzers, we can
particular time-domain signal. specify any units in both log and linear
Sample: The display detection mode in displays.
which the value displayed at each point Spectrum analyzer: A device that
is the instantaneous value of the video effectively performs a Fourier transform Vector diagram: A display type
signal at the end of the frequency span and displays the individual spectral commonly used when analyzing digitally
or time interval represented by the point. components (sine waves) that constitute modulated signals. It is similar to a
a time-domain signal. Phase may or may constellation display, except that in
Scale factor: The per-division calibration not be preserved, depending upon the addition to the detected symbol points,
of the vertical axis of the display. analyzer type and design. the instantaneous power levels during
Sensitivity: The level of the smallest state transitions are also plotted on an
sinusoid that can be observed on Spurious responses: The improper IQ graph.
a spectrum analyzer, usually under responses that appear on a spectrum
optimized conditions of minimum analyzer display as a result of the input Vector display: A display type used in
resolution bandwidth, 0-dB RF input signal. Internally generated distortion earlier spectrum analyzer designs, in
attenuation and minimum video products are spurious responses, as are which the electron beam was directed
bandwidth. Keysight defines sensitivity image and multiple responses. so that the image (trace, graticule,
as the displayed average noise level. A annotation) was written directly on the
sinusoid at that level will appear to be Sweep time: The time to tune the CRT face, not created from a series of
about 2 dB above the noise. LO across the selected span. Sweep dots as in the raster displays commonly
time does not include the dead time used today.
between the completion of one sweep

87
Video: In a spectrum analyzer, a term Video bandwidth: The cutoff frequency
describing the output of the envelope (3-dB point) of an adjustable low-
detector. The frequency range extends pass filter in the video circuit. When
from 0 Hz to a frequency typically well the video bandwidth is equal to or
beyond the widest resolution bandwidth less than the resolution bandwidth,
available in the analyzer. However, the the video circuit cannot fully respond
ultimate bandwidth of the video chain is to the more rapid fluctuations of the
determined by the setting of the video output of the envelope detector. The
filter. result is a smoothing of the trace,
i.e., a reduction in the peak-to-peak
Video amplifier: A post-detection, DC- excursion of broadband signals such
coupled amplifier that drives the vertical as noise and pulsed RF when viewed
deflection plates of the CRT. See Video in the broadband mode. The degree of
bandwidth and Video filter. averaging or smoothing is a function of
the ratio of the video bandwidth to the
Video average: A digital averaging of a resolution bandwidth.
spectrum analyzer’s trace information.
The averaging is done at each point Video filter: A post-detection, low-pass
of the display independently and is filter that determines the bandwidth of
completed over the number of sweeps the video amplifier. Used to average or
selected by the user. The averaging smooth a trace. See Video bandwidth.
algorithm applies a weighting factor
(1/n, where n is the number of the Zero span: That case in which a
current sweep) to the amplitude value spectrum analyzer’s LO remains fixed
of a given point on the current sweep, at a given frequency so the analyzer
applies another weighting factor [(n – becomes a fixed-tuned receiver. The
1)/n] to the previously stored average, bandwidth of the receiver is that of
and combines the two for a current the resolution (IF) bandwidth. Signal
average. After the designated number amplitude variations are displayed as
of sweeps are completed, the weighting a function of time. To avoid any loss
factors remain constant, and the display of signal information, the resolution
becomes a running average. bandwidth must be as wide as the signal
bandwidth. To avoid any smoothing, the
video bandwidth must be set wider than
the resolution bandwidth.

88
89 | Keysight | Spectrum Analysis Basics - Application Note

myKeysight For more information on Keysight

Technologies’ products, applications or
www.keysight.com/find/mykeysight
services, please contact your local Keysight
A personalized view into the information most relevant to you.
office. The complete list is available at:
www.lxistandard.org www.keysight.com/find/contactus
LAN eXtensions for Instruments puts the power of Ethernet and the
Web inside your test systems. Keysight is a founding member of the LXI Americas
consortium. Canada (877) 894 4414
Brazil 55 11 3351 7010
Three-Year Warranty Mexico 001 800 254 2440
www.keysight.com/find/ThreeYearWarranty United States (800) 829 4444
Keysight’s commitment to superior product quality and lower total cost
of ownership. The only test and measurement company with three-year Asia Pacific
warranty standard on all instruments, worldwide. Australia 1 800 629 485
China 800 810 0189
Keysight Assurance Plans Hong Kong 800 938 693
www.keysight.com/find/AssurancePlans India 1 800 112 929
Up to five years of protection and no budgetary surprises to ensure your Japan 0120 (421) 345
instruments are operating to specification so you can rely on accurate Korea 080 769 0800
measurements. Malaysia 1 800 888 848
Singapore 1 800 375 8100
www.keysight.com/go/quality Taiwan 0800 047 866
Keysight Technologies, Inc. Other AP Countries (65) 6375 8100
DEKRA Certified ISO 9001:2008
Europe & Middle East
Quality Management System
Austria 0800 001122
Keysight Channel Partners Belgium 0800 58580
www.keysight.com/find/channelpartners Finland 0800 523252
Get the best of both worlds: Keysight’s measurement expertise and product France 0805 980333
Germany 0800 6270999
breadth, combined with channel partner convenience.
Ireland 1800 832700
Israel 1 809 343051
Italy 800 599100
Luxembourg +32 800 58580
Bluetooth and the Bluetooth logos are trademarks owned by Bluetooth SIG, Inc., U.S.A. Netherlands 0800 0233200
and licensed to Keysight Technologies, Inc. Russia 8800 5009286
cdma2000 is a US registered certification mark of the Telecommunications Industry Spain 800 000154
Association.
Sweden 0200 882255
www.keysight.com/find/SA Switzerland 0800 805353
Opt. 1 (DE)
Opt. 2 (FR)
Opt. 3 (IT)
United Kingdom 0800 0260637

For other unlisted countries:

www.keysight.com/find/contactus
(BP-09-23-14)

This information is subject to change without notice.

© Keysight Technologies, 1971 - 2015
Published in USA, February 13, 2015
5952-0292
www.keysight.com