You are on page 1of 49


A comprehensive primer from
the editors of Microwaves & RF
LOU FRENZEL, Contributing Editor
CHRIS DeMARTINO, Technology Editor
JOHN BLYLER, Contributing Editor

Sponsored by

Copyright 2016 by Penton Media,Inc.

All rights reserved.

CHAPTER 1: SPECTRUM ANALYZERSThe main test instrument of wireless. The frequency domain vs.
time domain. Basic types, operation, and specifications..................................................4
CHAPTER 2: OSCILLOSCOPESA time domain instrument that is still useful at RF. The digital
oscilloscope, types, and how it works. Specifications and RF use cases.............................8
CHAPTER 3: NETWORK ANALYZERSNetwork analysis defined. Characterization of
components and circuits. S-parameters. Vector network analyzer (VNA) organization,
operation, and specification.....................................................................................12
modulation such as ASK, FSK, BPSK, QPSK, mQAM, OFDM, and variations.
Emphasis on higher-level modulation methods. ............................................................17
CHAPTER 5: VECTOR SIGNAL ANALYZERSHow VSAs are organized and how they work.
Common uses. Specifications...................................................................................24
CHAPTER 6: POWER METERSMethods for measuring RF/microwave power. Types of power meters
and their specifications. ..........................................................................................28
CHAPTER 7: THE BASICS OF SIGNAL GENERATIONTypes of signal synthesizers. Signal generator
types including vector signal generators and arbitrary function generators. Specifications....32
CHAPTER 8: PULSED MEASUREMENTSInstruments for making pulsed power measurements.
Radar system testing. Pulsed signal measurement parameters such as peak power,
rise/fall times, etc. Time- and frequency-domain analysis...............................................36
importance of instrument connection
cables and probes. Types and limitations. Accessories such as attenuators, isolators,
couplers, etc.........................................................................................................39
CHAPTER 10: EMI/EMCThe importance of electromagnetic interference and electromagnetic

compatibility. Instruments for EMI tests, probes, and antennas. Solutions: grounding,
shielding, and filtering.............................................................................................44

VISIT: for more T&M content | 1


LOU FRENZEL, Contributing Editor



RF and microwave instrument concepts and applications.
his book addresses the knowledge needs of engineers, manag-
ers, and other technical professionals already working in or new
to the wireless field who have not learned the basics of test and
measurement (T&M), either through experience or other means.
With fundamental T&M knowledge foundation, engineers will
be better prepared to dig deeper as required by their work with
specific equipment.
The book focuses on test instruments used for making mea-
surements on RF and microwave signals, components, equip-
ment, and systems. Its goal is to provide a baseline knowledge
of the most widely used test instruments in design, R&D, pro-
duction, test, and repair and maintenance. Each chapter is an
introduction to a particular type of RF/microwave test instrument,
what it is used to measure, and the most important specifications to be aware of. Some chap-
ters just concentrate on related fundamentals.
T&M is not a subject generally taught in colleges and universities as part of an engineering
curriculum. If a student is lucky, he or she may get to use multimeters, oscilloscopes, and sig-
nal generators in the context of lab experiments. Rarely is any RF/microwave test equipment
available. Any test and measurement information is typically taught as an aside, if it is taught
at all. Therefore, college graduates come to work with little knowledge of the equipment that
will play a major role in their work.
This book is designed to provide a base amount of knowledge on test and measurement to
make graduates aware of the types of equipment in general use, how they work, and how they
are used. The coverage is generic and fundamental, rather than addressing specific makes
and models of equipment, their features, and their operation.
Test and measurement is critical to RF and microwave work. Advanced instruments that
operate at microwave and millimeter-wave frequencies are essential to the design of next-
generation cellular systems, along with radar, satellite, and automotive equipment. Fifth-

VISIT: for more T&M content | 2


Generation (5G) cellular and the Internet of Things (IoT) are connect to a laptop or other PC by way of a USB interface.
examples of new technologies that rely on the availability of The processor and VI software then handle the signal pro-
instruments that can test and measure in an R&D environ- cessing, measurement, and display.
ment. Luckily, the test and measurement industry is alive Another major trend is the movement toward modular
and competitive, making plenty of high-quality instruments instruments. Instead of the traditional larger rack-and-stack
available. bench instruments, the use of smaller instrument modules
There are several key trends in T&M equipment. The first plugged into a chassis with a processor allows engineers
of these is the movement toward digital processing. Signals to customize their equipment, focusing on the application
to be measured are digitized by a high-sampling-rate ADC while lowering overall cost. Modules using the PXI standard
and stored in a memory. Next, a processor running special with a PCIe bus are now common.
instrument software analyzes the signal data and makes One more key trend is softwarenow the heart of most
the measurement. DSP filtering and processing are com- test instruments. Modern instruments digitize inputs, then
mon. The signal is then formatted and displayed on an LCD process the digital version to yield the measurement and
screen. While many analog instruments are still available, display. To update an instrument, vendors simply revise
most instruments incorporate digital techniques, even at the software or create new software as needed. Engineers
the higher microwave and millimeter-wave frequencies. purchase a base instrument, subsequently adding the soft-
One key aspect of the digital trend is the rise of virtual ware products as needed for their application. n
instruments (VIs). These small devices feature front-end
signal conditioning, digitizing, and memory circuits. They From the Editors of Microwaves & RF

VISIT: for more T&M content | 3


LOU FRENZEL, Contributing Editor


pectrum analyzers are the most widely used measurement capability to analyze the complex modulated
test instruments in the RF and microwave signals in modern equipment.
field. Their purpose is to display input signals
in the frequency domain. An oscilloscope What Spectrum Analyzers Do
displays signals in the time domain, voltages Spectrum analyzers are used to test transmitters, trans-
versus time. A spectrum analyzer displays ceivers, oscillators, phase-locked loops (PLLs), frequency
signal power versus frequency. synthesizers, and other signal-generating circuits and
Spectrum analyzers allow engineers to see signal equipment. The main display shows signal power on a dB
characteristics not present in a time domain display. While scale versus frequency. The best feature is the ability to see
oscilloscopes are still useful at RF and microwave frequen- what a Fourier analysis of a time domain signal indicates.
cies, it is the spectrum analyzer that provides the best Figure 1-1 shows how time and frequency domains
relate. The time domain signal is a distorted square wave.
Fourier components

Amplitude 20

Power (dBm)


domain 70
Time 80
0 10 20 30 40 50 60 70 80
Time domain Frequency (MHz)
display 1-2. Displayed is the frequency domain display of a
1-1. This figure shows the relationship between the time distorted square wave showing fundamental sine wave
and frequency domains. and third and fifth harmonics.

VISIT: for more T&M content | 4


display capabilities. There are two basic types: the analog

swept-tuned analyzer and the digital vector signal analyzer
(VSA). The swept-tuned type is discussed here. The VSA is
covered in a later chapter.
Figure 1-4 shows a general block diagram of the swept-
tuned analyzer. The input signal is applied to a variable
Power (dBm)

attenuator that is used to adjust the height of the signal on

the display and to accommodate the amplitude range of
the internal circuits. An amplifier may also be a part of the
input circuitry so that small signals can be displayed. An
input lowpass or bandpass filter is also used to select the
general frequency range to be analyzed.
The resulting signal is then fed to a mixer along with the
signal from a local oscillator (LO). The LO is a voltage-
0.998 1 1.002 controlled oscillator (VCO) that is controlled by the sweep
Frequency (MHz) signal. The LO signal is mixed with the input signal, and the
1-3. Shown is the spectrum display of an amplitude- sum and difference frequencies are generated. The differ-
modulated carrier with sidebands. ence signal is selected by a filter.
The input signal and all the surrounding spectrum is
Recall that a Fourier analysis of a square wave indicates therefore downconverted to a lower frequency. This is
that the signal is made up of a fundamental frequency sine called the intermediate frequency (IF) signal. Some IF am-
wave plus an infinite number of odd harmonics. This cannot plification may also be present. In some analyzers a sec-
be seen on the time domain display that shows a distorted ond or even third downconversion is used to further lower
square wave. the IF. Such dual- or triple-conversion architecture permits
However, a spectrum analyzer identifies these compo- narrower filters, and avoids a problem with conversion im-
nents and displays them as shown in Fig. 1-1. The funda- ages that are detrimental to any superhetrodyne receiver.
mental frequency sine wave is displayed along with the A key circuit in the analyzer is the IF filter. This is a
third and fifth harmonics. The actual display would look variable bandwidth bandpass filter. Called the resolution
more like that in Fig. 1-2. bandwidth (RBW) filter, it sets the bandwidth in which it
The benefit of the spectrum analyzer is that it shows will display signals. The RBW filter is usually set by a front
all of the frequency components of a signal along with panel control. It provides a window through which passes
other related or unrelated signals. The spectrum analyzer the segment of spectrum to be displayed.
will show the Fourier analysis of the signal, signal power, To allow the input to pass through the window, the LO
harmonics, sidebands produced by modulation, intermod- frequency is varied or swept over a frequency range similar
ulation components, spurs, noise, carrier-to-noise ratio, and to that of the IF window. A sawtooth ramp signal modulates
interfering signals. Fig. 1-3 shows the spectrum of a 1-MHz the LO frequency so that it varies linearly from a lower-
carrier amplitude modulated with a 2-kHz tone. The related frequency limit to an upper-frequency limit. The ramp
sidebands are clearly shown. repeats the sweep at a high rate. That signal is also used to
The spectrum analyzer is also useful in spectrum mon- trigger the display presentation.
itoring, detecting spurious noise and other emissions, and As the LO is swept, the IF varies with a lower-frequency
seeking sources of electromagnetic interference.
There is no faster way to evaluate a signal and Variable
make critical measurements. These are essential attenuator Amp Mixer detector
to meeting design specifications, passing stan- Input
Filter RBW Video
dards tests, complying with relevant regulations, signal filter filter
IF Log
and avoiding interference to nearby electronic amp amp
equipment. Local LCD
oscillator display
How Spectrum Analyzers Work Sweep
A spectrum analyzer is essentially a very generator
sensitive RF superhetrodyne receiver with special 1-4. This block diagram shows a swept-tuned spectrum analyzer.

VISIT: for more T&M content | 5


Wide RBW Narrow Here is a short list of specifications you should look for
filter RBW filter when buying or evaluating a spectrum.
Frequency range. This is the main specification. Obvi-
ously, the analyzer should cover the frequency range of the
devices you are testing. However, you should pinpoint the
maximum frequencies you will be seeing along with any
significant harmonics or sidebands. A good rule of thumb
(as with oscilloscopes) is to be sure the analyzer can re-
solve frequencies three to five times higher than the fun-

damental carrier frequencies of your application. Modern

spectrum analyzers come is a variety of frequency ranges
from roughly 10 Hz to 100 GHz. Accessory down con-
verters are sometimes used when higher millimeter-wave
signals are to be analyzed.
Input sensitivity. This is the range of input power levels
the analyzer can accommodate. The range varies widely
Frequency from model to model, but it usually falls between a low
1-5. Wideband RBW filters obscure signal details. of 160 dBm and a maximum of +30 dBm. The low-end
Narrow RBW filters show more detail. sensitivity is typically given as the displayed average noise
level (DANL) measured in dBm/Hz of BW. DANL defines
version of the original input. The signal is then processed the analyzers noise floor at a given IF bandwidth. This is
by a logarithmic amplifier. The log amplifier allows the an- the best sensitivity rating of the instrument, as input signals
alyzer to deal with the wide dynamic range of the circuitry lower than the noise floor will not be displayed.
that can accommodate signals ranging from a few micro- Care should be taken not to overload the analyzer front-
volts to many volts. The log amp compresses the signal end, as permanent damage can result. For signals exceed-
range for display and provides an output signal level in dB ing the 20-to-30 dBm range, an external attenuator should
measuring units. be used. A front panel control allows you to set the input
That signal is then applied to a detector, which outputs range and the dBm/division scale on the display grid.
a signal with amplitude proportional to the envelope of the Resolution bandwidth. Resolution bandwidth (RBW) is
IF output. The detector is usually a diode power detector. the bandwidth of the IF filter. It is a variable bandpass filter
Its output is then filtered into the envelope signal by a video that is usually made up of multiple fixed-bandwidth filters
lowpass filter. That is the displayed signal. Older spectrum that can be switched in as necessary to the measurement.
analyzers use a cathode ray tube (CRT) display, but today The narrower the bandwidth is, the better the resolution. If
a liquid crystal display (LCD) is standard with most equip- the bandwidth is set too wide, signal components like side-
ment. bands, harmonics, or interfering signals may be obscured
Most of the newer spectrum analyzers use digital tech- during the sweep. Figure1-5 shows the effect of both wide
niques to generate the spectrum display. One approach is and narrowband RBW filters. The wide BW obscures the
to digitize the IF output in a fast-sampling ADC, and then true detail, but the narrow BW filter shows this detail. RBW
store the result in a local memory. Next, a fast processor ranges are typically from about 1 kHz to 1 MHz.
analyzes the data using the discrete Fourier transform Accuracy. There are usually specifications for both am-
(DFT) or fast Fourier transform (FFT). plitude and frequency. Amplitude accuracy is governed by
The DFT and FFT are mathematical algorithms that take the analyzers gain precision and filter responses. A typical
time domain sampled data and converts it into a frequen- value is 0.3 dB and is the result of all the various errors
cy domain output. VSAs also use the FFT method, but in involved. Frequency measurement accuracy is similar in
addition use special IQ demodulation to allow complex that it is based primarily upon errors in the LOs precise
modulated signals to be disassembled into different forms frequency and filter-center frequencies.
of output displays. See Chapter 5 for details. Spurious free dynamic range. Spurious free dynamic
range (SFDR) is the ratio of the maximum signal level to
Spectrum Analyzer Specifications the largest spur that occurs during measurement as stated
There are several key characteristics of spectrum an- in dB. The spur is a spurious unwanted signal that may be
alyzers that will determine their fitness for an application. a harmonic, intermodulation product, or noise spike that

VISIT: for more T&M content | 6


originates within the analyzer (the larger, the better). A

typical specification is 80 dB or more.
Phase noise. Phase noise consists of unwanted
sidebands around a signal, which results from varia-
tions in the LO frequency. The LO is frequency modu-
lated by internal short-term instabilities. This results in
low-level sidebands that appear above and below the
LO signal or carrier. These minor variations are translat-
ed into noise variations by the mixer, and so will appear
on the spectrum display. Phase noise is measured in
dBc/Hz at some offset frequency from the LO carrier. A
typical level is 140 dBc at 10 MHz.
This concludes the introduction to spectrum analyz-
ers, but please note that this chapter is the prerequisite
to Chapter 5 on vector signal analyzers (VSAs). You
may also want to review your knowledge of modulation
Learn the basics of the architecture and
in Chapter 4 before attempting Chapter 5. n
proper uses of the super-heterodyne signal

VISIT: for more T&M content | 7


JOHN BLYLER, Contributing Editor


n oscilloscope displays electronic signals in Applications: What Oscilloscopes Do
the time domain. The displayed waveforms An oscilloscope, or scope, is a test and measurement
can be analyzed for a variety of measure- instrument that displays electronic signals. Most microwave
ments and to determine the proper behavior and RF signals are comprised of a collection of regularly
patterns of electronic subsystems. Most repeating sinusoidal waveforms. Viewing the wave shape
engineers equate time-based measure- of a signal allows you to determine if a component of an
ments with oscilloscopes and frequency measurements electronic system is behaving properly. To make these
with spectrum analyzers. But this perception is changing determinations, you must understand four basic waveform
thanks to a number of improvements in oscilloscopes. parameters: amplitude, phase shift, period, and frequency.
Amplitude refers to the peak amplitude or maximum
Oscilloscopes for the RF/Microwave Space displacement of a signal; alternately, it can refer to the root-
It wasnt so long ago that microwave engineers shied mean-square (RMS) amplitude value of the signal, such as
away from using oscilloscopes to measure voltage and a sine wave. The RMS value of a sine wave is calculated
current. The direct measurement of microwave voltages by squaring the waveform, finding its average voltage, and
was usually accompanied by considerable noise fluctua- taking the square root. For a sine wave, the RMS amplitude
tions, resulting in poor sensitivity, low dynamic range, and is equal to 0.707 times the peak amplitude. Most scope
significant errors. Back then, microwave measurements measurements are peak-to-peak (pp), the distance be-
usually involved the accurate determination of impedance tween the maximum and minimum voltage excursions. The
and power, rather than voltage and current. peak value for a sine wave is one-half the pp value.
Thanks to advancements in semiconductor technology, Phase shift refers to the amount of horizontal translation
sophisticated software algorithms, and improved micro- between two otherwise-identical waves. It is measured in
wave and RF measurement techniques, it is now common- degrees or radians. For a sine wave, one cycle is repre-
place to use time-domain equipment to measure voltages sented by 360 deg. Therefore, if two sine waves differ by
and currents operating in the microwave range. In addi- half of a cycle, their relative phase shift is 180 deg.
tion, modern oscilloscopes come with features that enable The amount of time it takes for a wave to repeat itself
the additional functionality of spectrum and logic analysis. is known as the period of the wave (t). It is the distance
The capability to measure higher frequencies in the between adjacent peaks, or sequential positive or negative
time-domain is the result of several improvements, ranging voltage transitions.
from faster semiconductor front-end amplifiers and ana- The frequency (f) is simply the number of times a period
log-digital converters (ADC) to greater stability of sample wave repeats itself within one second, measured in Hertz.
clock generators (less jitter) and triggering. The frequency is also the reciprocal of the period (f = 1/t).
To understand how todays RF/microwave engineers An oscilloscope displays one variable as a function of
use these time-based instruments, we first must under- another, typically voltage (y-axis) versus time (x-axis). A
stand the basics of oscilloscope measurements, types, major use for this capability is to determine if a specific
and specifications. electronic component or series of components are behav-

VISIT: for more T&M content | 8


One period 16.67 ns can be divided into several functional areas: vertical and
horizontal controls, triggers, and display.
The vertical area controls the vertical positioning, chan-
nel coupling (dc or ac), bandwidth limit, and attenuation of
the input signals. Note that there are two input channels,
1 and 2. BNC connectors are used to connect the probes
and cables. The math controls enable selection of add,
Peak-to- subtract, multiply, and FFT (fast Fourier transform) function-
ality on specific input channels.
Horizontal controls are used for selecting the time per
division in the scale readout of the screen. These controls
change the horizontal scale and position of the waveforms.
The trigger control determines when the oscilloscope
starts to acquire data and display a waveform. Most os-
cilloscopes provide three trigger types: edge, video, and
0.5 V/div. pulse. An edge trigger can be used with analog and digital
circuits; it occurs when the trigger input passes through a
specified voltage level with the specified slope.
5 ns/div.
Some scopes allow you to store and recall waveforms or
2-1. This is a generalized view of an oscilloscope configuration setups within the oscilloscope, or else to an
screen showing grid and measurement for a sine wave. external device like a USB memory stick.
The LCD screen of a digital oscilloscope is used both
ing properly. Figure 2-1 is a typical oscilloscope display. to display the waveforms and serve as the menu screen
The screen implements a grid pattern to facilitate voltage for the instrument. The menu is used to set many of the
and time measurements. Each large division represents a controls listed above and many others into a unique config-
calibrated voltage or time increment. uration in memory. Also, the menu can be used to display
In the figure, each large division vertically represents 0.5 specific measured calculation on the LCD screen.
volts and each horizontal division represents 5 nanosec- There is a rather application-specific version of an
onds (ns). Using these calibrations, the sine wave shown oscilloscope that incorporates aspects of analog and
has an amplitude of 5 0.5 = 2.5 volts peak-to-peak (pp). digital instruments. Its called a mixed-signal oscilloscope
Most amplitude measurements on a scope are pp. The
period is measured between two adjacent peaks of the
waveform. The period is approximately 3.3 divisions 5 ns
= 16.67 ns. This gives a frequency of 60 MHz.
An oscilloscope with multiple inputs or channels can
be used to compare and contrast the signals from different Rise-and-fall
sources. A common application is observing both the input times of pulses
and output of a circuit.
Another common signal shape is the eye pattern
or eye diagram, so called because the shapes on the 1-bit interval
oscilloscope display look like a series of eyes between a
pair of rails (Fig. 2-2). This shape appears when a digital
signal from a receiver is repetitively sampled and applied
to the vertical input while the data rate is used as a trigger (b)
for the horizontal sweep. Eye patterns are used in testing
high-speed serial communications buses to determine the
integrity of baseband-related signals.

Instrument Controls and Display 2-2. This eye diagram shows (a) a normal waveform;
The display and controls of a typical digital oscilloscope and (b) a compromised waveform caused by lack of
are shown in Fig. 2-3. The front panel of most oscilloscopes bandwidth.

VISIT: for more T&M content | 9


and many have four. Each input has its own ADC; the latter
Math Horizontal
sample the inputs and convert the amplified analog input
signal into digital words (usually 8 bits). The digitized sam-
Vertical Trigger
ples are stored in an acquisition memory and reconstructed
screen in a MCU processing subsystem prior to display.

1 2
USB Advanced Oscilloscope Features
Modern oscilloscopes include features that cross over
into the domains of other test equipmente.g., spectrum
and logic analyzers. For example, oscilloscopes can
Vertical inputs Trigger input
BNCs BNC include capabilities for RF applications, such as viewing
signals in the frequency domain, by performing the FFT.
2-3. Shown is a typical digital oscilloscope front panel For embedded design, some oscilloscopes have the
layout. same capabilities as logic analyzers. Such oscilloscopes
can be used for debugging mixed analog and digital
(MSO). Analog-mixed signal technology forms the basis signals, in addition to series interface technologies such as
of todays modern communication systems, such as the automotive CAN/LIN, FlexRay, MOST, SPI, RS-232, I2C, and
mobile phone; hence, it has become increasingly neces- other serial bus structures. Furthermore, modern high-end
sary to monitor analog and digital signals simultaneously. oscilloscopes often include signal integrity and jitter mea-
Mixed-signal oscilloscopes can trigger on and display both surement accessories.
analog and digital signals simultaneously, all correlated to
the same time base. Oscilloscope Specifications
Here is a short list of specifications you should consider
How Oscilloscopes Work when buying or evaluating an oscilloscope.
An oscilloscope is basically a graph-displaying de- Bandwidth. This is one of the most critical specifications
vice that draws a trace of an electrical signal. Around the for an oscilloscope. The bandwidth tells you the frequency
mid-1990s, most analog oscilloscopes were based on range which the oscilloscope can accurately measure. You
cathode-ray tube (CRT) technology for the display. CRTs should compare this range to the signal frequencies of your
were relatively bulky, heavy, power-hungry, unreliable, and applications.
expensive. Most entry-level oscilloscopes have a maximum band-
Today, CRT displays have been almost completely width of 100 MHz. They can accurately (within 2%) show
replaced by more convenient liquid crystal display (LC- the amplitudes of sinusoidal signals up to 20 MHz. For
D)-type screens. Instead of using moving electron beams digital signals, the bandwidth of the oscilloscope plus the
to draw traces, LCD oscilloscopes use digital electronics to probe should similarly be at least 5 the maximum signal
draw a signal trace. Most scopes offer a color screen today. bandwidth for better than 2% measurement error.
The most commonly used scope is a digital (storage) High-speed digital, serial communications, video, and
oscilloscope, or DSO, that can measure both digital and other complex signals thus require oscilloscope band-
analog signals. To measure the later, the analog input sig- widths of 500 MHz or more. Oscilloscopes with bandwidths
nal voltage is first converted to a digital form after passing up to 70 GHz are available, making them useful for some
through an analog-to-digital converter (ADC).
The input analog signal becomes a series of Amp
digital samples that are stored in a memory
ADC Display LCD
and reassembled as the input waveform on Acquisition
Inputs Amp MCU memory display
the oscilloscopes screen. Atten memory
A simplified diagram of a basic digital ADC
oscilloscope is shown in Fig. 2-4. The inputs
of the digital oscilloscope are applied to cali- Clock &
brated attenuators followed by vertical ampli- Trigger timing
fiers. Vertical attenuation controls allow you
to adjust the amplitude range of the input 2-4. This is a simplified functional block diagram for a
signal. Most scopes today have at least two vertical inputs, typical digital storage oscilloscope.

VISIT: for more T&M content | 10


microwave and millimeter-wave measurements. can turn input voltages into digital values. Calculation
Rise time. Another way to describe the frequency range techniques via the microprocessor can be used to improve
of an oscilloscope is with the rise time, especially when the effective resolution. Most oscilloscopes have 8-bit ADC
measuring pulses and step function signals. The rise time resolution.
of these signals is the time required for a pulse to rise from Memory depth. This helps determine the sampling
10% to 90% of its steady-state value. rate of the input signal after it passes through the ADC.
An oscilloscope cannot accurately display pulses with The digitized data is then stored in the oscilloscopes
rise times faster than the specified rise time of the oscil- high-speed memory. Memory depth refers to exactly how
loscope. Thus, an oscilloscopes rise time should be less many samples or points can be stored in memory, which
than 1/5 times the fastest rise time of the signale.g., a directly affects the length of time sample of the signal being
4-nanosecond rise time needs an oscilloscope with faster measured. Insufficient memory depth will limit the type of
than 800-pico-second rise time. Rise time (tr) and upper signals that can be displayed. If long periods of time are
frequency bandwidth are related by the expression: needed with high resolution between points, you will need
deep memory.
BW = 0.35/tr For example, capturing enough detail to decode a USB
serial data stream requires a high-resolution sample (200
Vertical sensitivity. This specification characterizes micro-seconds). An oscilloscope with a long record length
how much the vertical amplifier can amplify a weak signal. (1Mpoints) is needed. An entry-level oscilloscope can store
It is usually measured in millivolts (mV) per division. The over 2,000 points per record length. Time captured equals
smallest voltage a general-purpose oscilloscope can de- the record length divided by the sample rate. That means
tect is typically about 2 mV per vertical screen division. that a record length of 1Mpoints and a sample rate of 250
Gain accuracy. This indicates the accuracy with which MS/s will capture a signal 4 ms in length.
the vertical system attenuates or amplifies a signal. It is (Please note that this chapter is the prerequisite to
typically listed as a percentage error. Chapter 9 on probes, cables, and attenuators.) n
Time base. This specification, also known as the hor-
izontal accuracy, indicates the accuracy with which the BACK TO TABLE OF CONTENTS
horizontal system displays the timing of a signal. This is
usually listed as a percentage error.
Sample rate. Applicable to digital oscilloscopes, the
sample rate indicates how many samples per second (S/s)
the oscilloscope can sample the signal. As with band-
width, the five-time rule should apply. Thus, oscilloscopes
should have a sample rate of at least 5 your circuits high-
est frequency component. To capture slower signals, the
oscilloscope must have a sufficiently low-end sample rate.
Typically, most entry-level oscilloscopes have a maxi-
mum sample rate of 1 to 2 GS/s, while mid-range scopes
are closer to 5 to 10 GS/s.
Caution should be exercised when evaluating the
oscilloscopes sample rate specification. Manufacturers
may specify the maximum sample rate an oscilloscope can
attain if only one or two input channels are used. This may
be a problem if more channels are used for simultaneous, If your organization values time, efficiency
high-bandwidth measurements. and productivity, theres no substitute for a
ADC resolution. Also known as the vertical resolution, PXI scope. Test 40 50X faster and gain
this specification indicates the resolution, in bits, of the up to 35 hours of test time every week.
ADC. It determines how precisely the digital oscilloscope

VISIT: for more T&M content | 11


CHRIS DeMARTINO, Technology Editor


etwork analyzers are indispens-
a1 a2
able test instruments in the RF/ 3-2. Shown is an
microwave industry. Their mea- DUT illustration of a
Port 1 Port 2
surement capabilities allow them b1 b2 two-port network.
to characterize a wide range
of devices, components, and
systems. Many componentsboth passive
and activeare commonly measured with a
network analyzer, including amplifiers, filters, a1 = Incident wave applied to port 1
attenuators, switches, and many others. b1 = Reflected wave from port 1
a2 = Incident wave applied to port 2
Network analyzers are relied upon heav-
b2 = Reflected wave from port 2
ily for manufacturing production testing. A ZS = Source impedance
components datasheet typically contains a ZL = Load impedance
significant amount of information obtained
from a network analyzer. In addition, network
analyzers are used for research and development purpos- What Network Analyzers Do
es. They can be used to measure engineering prototypes, Network analyzers can both measure and display a
thus allowing engineers to optimize performance character- device-under-tests (DUT) magnitude and phase information
istics like gain flatness and return loss. across a frequency range. A generic illustration of a network
analyzer plot is shown in Fig. 3-1. In essence, a network
analyzer characterizes a DUT in terms of
scattering parameters, or S-parameters.
0 3-1. This plot is S-parameters are used to characterize
5 a depiction of a performance at RF and microwave fre-
network analyzer quencies in terms of incident and reflected
measurement. waves. S-parameters are vector quantities,
15 meaning they contain both magnitude and

phase information. A scalar network analyz-

er (SNA) can only measure magnitude, while
25 a vector network analyzer (VNA) can mea-
sure both magnitude and phase. Each of
these will be discussed in more detail later.
5.0 5.5 6.0 6.5 Figure 3-2 shows an illustration of a two-
Frequency (GHz) port network. 50- impedances are typical
of most RF/microwave applications. Hence,

VISIT: for more T&M content | 12


S-parameter matrix of the two-port network that has been
Port 1 Port 2 discussed is shown below:
b1 b2
b1 S11 S12 a1
b2 S21 S22 a2

3-3. In this figure, an incident wave, a1, is applied to the

DUT. As stated earlier, S-parameters contain both magnitude
and phase information. Magnitude is typically expressed in
ZS and ZL in Fig. 3-2 would generally be 50 . Cable-tele- decibels (dB). This is mathematically defined below:
vision (CATV) applications are the main exception, as they
operate in a 75- environment. S11(dB): 20 log10 |S11|
Figure 3-3 is based on the network shown in Fig. 3-2.
Here, ZS and ZL are equal to 50 . Port 2 of the DUT is ter- S12(dB): 20 log10 |S12|
minated in a 50- load, thus setting a2 to be equal to zero.
In Fig. 3-3, a1 is applied to Port 1 of the DUT. A portion
S21(dB): 20 log10 |S21|
of this incident wave is transmitted through the DUT and
exits through Port 2, thereby resulting in b2. A portion of the
incident wave is also reflected back to the source, thereby S22(dB): 20 log10 |S22|
resulting in b1. Now, two S-parameters, S11 and S21, can be
mathematically defined as follows:
The dB representations of S11 and S22 are known as re-
a1 a2| 0 turn loss, which is the difference in dB between the reflected
signal and the incident signal. Thus, a return loss of 15 dB
b2 means that the reflected signal is 15 dB lower in magnitude
s21: a1 a2| 0 than the incident signal. Return loss is commonly expressed
as a positive value, so a return loss of 15 dB is often ex-

Figure 3-4 is also basedb on the network shown in Fig. pressed as just simply 15 dB.
3-2. Here, Port 1 of thes12:DUT 1
a 2isaterminated in a 50- load, The magnitude element of S21 is known as gain or inser-
1 0
thus setting a1 to be equal to zero. tion loss, depending on whether the DUT is active or pas-
In Fig. 3-4, a2 is now applied to Port 2 of the DUT. A sive. In other words, an active devicesuch as an amplifi-
portion of this incident a | is transmitted
a12 aa21 0 through the DUT erhas gain because it increases the magnitude of an input
and exits through Port 1, thereby resulting in b1. A portion of signal. A passive component like a filter does not have gain,
the incident wave is also reflected,
thereby resulting in b2. meaning the output signal is smaller in magnitude than the
Now, the two remaining s21: S-parameters,
a1 a2 0 S12 and S22, can be input signal. In this case, S21 is referred to as insertion loss.
mathematically defined as follows: S12 defines transmission in the reverse direction. Hence, it is
known as reverse gain or reverse transmission.
s12: a 2 a1 It is often important to characterize phase along with
magnitude. As discussed already, when an incident signal

a 2 a1 0
is applied to a DUT, a portion of that signal is transmitted.
The transmitted signal that exits the DUT differs not only in

A two-port network therefore has four 3-4. In this figure,
S-parameter elements: S11, S21, S12, and an incident wave, Port 1 Port 2
S22. S11 and S22 are known as reflection a2, is applied to b1 b2
coefficients. S21 and S12 are known as the DUT. 50
transmission coefficients. Furthermore,
a network can be defined in matrix form,
known as an S-parameter matrix. The

VISIT: for more T&M content | 13


Network analyzer
magnitude from the incident signal, but 3-6. Shown is an
also in phase. Consequently, S21 and illustration of a vector
S12 also describe the phase difference network analyzer (VNA)
in degrees between a transmitted signal measuring a two-port
and an incident signal. A linear phase component.
response over frequency is desirable
deviating from one can cause distortion.
Port 1 Port 2
Group delay is the transit time of a
signal as it passes through a DUT. It is
mathematically defined as the negative
of the derivative of phase response with
respect to frequency, meaning that it
quantifies phase linearity. Thus, a flat
group delay means the phase response
is linear. Network analyzers are com-
monly used to measure the group delay DUT
of components, such as filters.
The two-port model that has been
discussed represents a component
with two ports. However, many components have more than widely used in recent years. However, SNAs will be dis-
just two ports. Figure 3-5 shows an illustration of an N-port cussed here first.
network. The DUT here could possibly be a power divider or SNAs can only measure magnitude. They do not have
some other component with multiple ports. The S-parameter the capability to perform phase measurements. SNAs
matrix of this network is shown below: perform transmission measurements by means of a diode
detector, which converts an RF signal to a dc voltage. This
b1 S11 S1n a1 dc voltage is proportional to the magnitude of the RF signal.
Unfortunately, this process does not take phase information
into account. Return loss measurements can be achieved
bn Sn1 Snn an by means of a directional bridge.
VNAs are much more powerful than SNAs, as they
can measure both magnitude and phase. Figure 3-6 is
How Network Analyzers Work an illustration of a VNA being used to measure a two-port
The fundamental purpose of a network analyzer is to component, such as a filter or amplifier. The input of the
measure S-parameters. A network analyzer can be classi- DUT is connected to the end of a cable that is attached to
fied as either a VNA or an SNA. VNAs have become more Port 1 of the VNA. The DUTs output is connected to the end
of a cable that is attached to Port 2
3-5. This figure of the VNA. The DUT shown in Fig.
Port 1 shows a network 3-6 essentially represents the DUT
b1 with N ports. shown in Fig. 3-2.
a2 In Fig. 3-6, the VNA contains an
Port 2 RF source, which generates signals
b2 over the frequency range of inter-
Z aN est. These signals exit through Port
1 and are applied to the DUT. When
Port N
a signal at any frequency is applied
Z to the DUT, a portion of that signal
is transmitted through the DUT to
Port 2 of the VNA. Another portion
is reflected back to Port 1. The user
can then see S21 or S11 measure-
ments on the VNAs display.

VISIT: for more T&M content | 14


To measure S22 or S12, 3-7. Shown is a

the VNA generates sig- simplified block Reference Measurement
nals that exits through Port diagram of a vector receiver 2 receiver 2
2 and are applied to the network analyzer
DUT. A portion of a signal (VNA) with two test Test
Port 2
at any frequency applied ports. cable
to the DUT now transmits
Directional Directional
through the DUT to Port 1 RF source
device R2 device M2
of the VNA. Another portion Computer DUT
is reflected back to Port 2. Directional Directional
Now, the user can see S12 device R1 device M1
or S22 measurements on the Test
display. Port 1 cable
The VNAs architecture
can now be examined to Reference Measurement
understand how these mea- receiver 1 receiver 1
surements are achieved.
Figure 3-7 shows a general
block diagram of a VNA with
two test ports. (It should
be noted that Fig. 3-7 shows a general VNA architecture. then downconverted and processed. Now, the VNA can
Additional variations are possible, but the general concept compare the data obtained from measurement receiver 2
remains.) The signals generated by the VNAs RF source with the data obtained from reference receiver 1. This pro-
enter a switch, which routes the signal toward either Port 1 cess summarizes how S21 is measured.
or Port 2. To measure S22 or S12, the switch would route the signal
To measure S11 or S21, the switch would route the signal to Port 2. The same process occurs in the opposite direc-
toward Port 1. The signal first enters a directional device, tion: The signal first enters directional device R2 in Fig. 3-7,
shown as directional device R1 in Fig. 3-7. This directional which directs a portion of the signal to reference receiver 2.
device could be a directional coupler or bridge. A power After entering this reference receiver, the signal is downcon-
divider could also be used here. The purpose of this device verted and processed.
is to direct a portion of the input signal to a reference receiv- The signal that passes through directional device R2
er, shown as reference receiver 1. Once the signal enters then passes through directional device M2 on its way
this reference receiver, it is downconverted and ultimately toward the DUT. After arriving at the DUT, a portion of the
processed. signal is reflected back to Port 2 of the VNA and reenters
The signal that passes through directional device R1 directional device M2. Now, a portion of this reflected signal
then passes through another directional device, shown as is directed to measurement receiver 2. This signal is then
directional device M1 in Fig. 3-7, before arriving at the DUT. downconverted and processed. The data obtained from
When the signal does arrive at the DUT, a portion is reflect- measurement receiver 2 can then be compared with the
ed back to Port 1 of the VNA and reenters directional device data obtained from reference receiver 2, thus summarizing
M1. This device now directs a portion of the reflected signal the S22 measurement process.
to a measurement receiver, shown as measurement receiv- The signal that is transmitted through the DUT enters
er 1. Next, this signal is downconverted and processed. Port 1 of the VNA. This transmitted signal enters directional
The VNA can then compare the data obtained from mea- device M1 in Fig. 3-7, which directs a portion of the signal
surement receiver 1 with the data obtained from reference to measurement receiver 1. This signal is likewise down-
receiver 1. This process summarizes how S11 is measured. converted and processed. The data obtained from mea-
The signal applied to the DUT also results in a trans- surement receiver 1 can now be compared with the data
mitted signal, which exits the DUT and enters Port 2 of the obtained from reference receiver 2, thus summarizing the
VNA. This transmitted signal enters another directional S12 measurement process.
device, shown as directional device M2 in Fig. 3-7. This de- As previously stated, many components have more than
vice directs a portion of the signal to another measurement just two ports. Many VNAs also have more than two test
receiver, shown as measurement receiver 2. This signal is ports to conveniently measure such components. For exam-

VISIT: for more T&M content | 15


ple, four measurement receivers and four reference

receivers could be used to build a four-port VNA,
allowing a four-port component to be easily mea-

Network Analyzer Specifications

There are several important specifications that
define a network analyzer:
Frequency range. This is the main specification.
A network analyzers frequency range defines the
minimum and maximum frequencies it can measure.
Dynamic range. This defines the range of power
that the network analyzer can measure.
Number of test ports. A network analyzer can
have two, four, or more test ports.
Measurement speed. This is the time required to Learn key VNA concepts such as basic
perform measurements across a range of architectures, sources of error, and
frequencies. n calibration. Also explore best practices for
network analysis measurement.

VISIT: for more T&M content | 16


LOU FRENZEL, Contributing Editor


(Editors Note: This chapter is based on an article of the responding AM and FM signals resulting from modulation.
same name that previously appeared on electronicdesign. There are two types of AM signals: on-off keying (OOK)
com.) and amplitude shift keying (ASK). In Fig. 4-1a, the carrier

amplitude is shifted between two amplitude levels to pro-
undamental to all wireless communications is duce ASK. In Fig. 4-1b, the binary signal turns the carrier
modulation, the process of impressing the data off and on to create OOK.
to be transmitted on the radio carrier. Most wire- AM produces sidebands above and below the carrier
less transmissions today are digital, and with equal to the highest frequency content of the modulating
the limited spectrum available, the type of mod- signal. The bandwidth required is two times the highest fre-
ulation is more critical than it has ever been. quency content including any harmonics for binary pulse
The main goal of modulation today is to squeeze as modulating signals.
much data into the least amount of spectrum possible. Frequency shift keying (FSK) shifts the carrier between
That objective, known as spectral efficiency, measures two different frequencies called the mark and space fre-
how quickly data can be transmit-
ted in an assigned bandwidth. The 4-1. Three basic 1 1
unit of measurement is bits per sec- digital modulation data
ond per Hz (b/s/Hz). Multiple tech- formats are still very 0 0
niques have emerged to achieve popular with low-
and improve spectral efficiency. data-rate short-range
wireless applications:
Amplitude Shift Keying (ASK) amplitude shift
and Frequency Shift Keying keying (a), on-off (a) ASK

(FSK) keying (b), and

There are three basic ways to frequency shift
modulate a sine-wave radio carrier: keying (c). These sine
modifying the amplitude, frequen- waveforms are
cy, or phase. More sophisticated coherent as the (b) OOK
methods combine two or more of binary state change
these variations to improve spectral occurs at carrier zero Higher
efficiency. These basic modulation crossing points.
forms are still used today with digi-
(c) FSK
tal signals.
Figure 1 shows a basic serial
Lower frequency
digital signal of binary zeros and
ones to be transmitted and the cor-

VISIT: for more T&M content | 17


quencies, or fm and fs (Fig. 4-1c). FM 1 0 0 1 0 1 fall times and reducing the harmonic
produces multiple sideband frequen- Serial content. Special Gaussian and raised
cies above and below the carrier data cosine lowpass filters are used for
frequency. The bandwidth produced this purpose. GSM cell phones widely
is a function of the highest modulating BPSK use a popular combination, Gaussian
frequency including harmonics and filtered MSK (GMSK), which allows a
the modulation index, which is: Phase changes data rate of 270 kbits/s in a 200-kHz
m = f(T) channel.
f is the frequency deviation or 4-2. In binary phase shift keying, note
shift between the mark and space how a binary 0 is 0 deg. while a binary Binary Phase Shift Keying (BPSK)
0209EEmodulation-FIGURE 2
frequencies, or: 1 is 180 deg. The phase changes and Quadrature Phase Shift Keying
f = fs fm when the binary state switches so the (QPSK)
T is the bit time interval of the data signal is coherent. A very popular digital modulation
or the reciprocal of the data rate (1/ scheme, binary phase shift key-
bit/s). ing (BPSK), shifts the carrier sine
Smaller values of m produce fewer sidebands. A wave 180 deg. for each change in binary state (Fig. 4-2).
popular version of FSK called minimum shift keying (MSK) BPSK is coherent as the phase transitions occur at the
specifies m = 0.5. Smaller values are also used, such as m zero crossing points. The proper demodulation of BPSK
= 0.3. requires the signal to be compared to a sine carrier of
Here are two ways to further improve the spectral the same phase. This involves carrier recovery and other
efficiency for both ASK and FSK. First, select data rates, complex circuitry.
carrier frequencies, and shift frequencies so there are no A simpler version is differential BPSK or DPSK, where
discontinuities in the sine carrier when changing from one the received bit phase is compared to the phase of the
binary state to another. These discontinuities produce previous bit signal. BPSK is very spectrally efficient in that
glitches that increase the harmonic content and the band- you can transmit at a data rate equal to the bandwidth or 1
width. bit/Hz.
The idea is to synchronize the stop and start times of In a popular variation of BPSK, quadrature PSK (QPSK),
the binary data with when the sine carrier is transitioning the modulator produces two sine carriers 90 deg. apart.
in amplitude or frequency at the zero The binary data modulates each phase,
crossing points. This is called continuous Table 1: Carrier Phase producing four unique sine signals shift-
phase or coherent operation. Both co- Shift for Each Pair of Bits ed by 45 deg. from one another. The two
herent ASK/OOK and coherent FSK have Represented phases are added together to produce
Bit pairs Phase (degrees)
fewer harmonics and a narrower band- the final signal. Each unique pair of bits
00 45
width than non-coherent signals. generates a carrier with a different phase
01 135
A second technique is to filter the bi- (Table 1).
nary data prior to modulation. This rounds 11 225
Figure 4-3a illustrates QPSK with a
the signal off, lengthening the rise and 10 315
phasor diagram where the phasor rep-
resents the carrier sine amplitude peak
90 90 and its position indicates the phase. A
01 00 constellation diagram in Fig. 4-3b shows
the same information. QPSK is very
spectrally efficient since each carrier
180 0 180 0 phase represents two bits of data. The
spectral efficiency is 2 bits/Hz, meaning
twice the data rate can be achieved in
11 10
the same bandwidth as BPSK.
(a) 270 (b) 270

4-3. Modulation can be represented without time-domain waveforms. Data Rate and Baud Rate
For example, QPSK can be represented with a phasor diagram (a) or a The maximum theoretical data rate
constellation diagram (b), both of which indicate phase and amplitude or channel capacity (C) in bits/s is a
magnitudes. function of the channel bandwidth (B)
0209EEmodulation-FIGURE 3

VISIT: for more T&M content | 18


channel in Hz and the signal-to-noise 90 While multiple phase shift keying

ratio (SNR): (M-PSK) is much more spectrally effi-
C = B log2 (1 + SNR) cient, the greater the number of smaller
This is called the Shannon-Hartley phase shifts, the more difficult the sig-
Law. The maximum data rate is direct- nal is to demodulate in the presence of
180 0
ly proportional to the bandwidth and noise. The benefit of M-PSK is that the
logarithmically proportional to the SNR. constant carrier amplitude means that
Noise greatly diminishes the data rate more efficient nonlinear power amplifi-
for a given bit error rate (BER). cation can be used.
Another key factor is the baud rate, 270
or the number of modulation symbols 4-4. 16QAM uses a mix of Quadrature Amplitude Modulation
transmitted per second. The term sym- amplitudes and phases to achieve (QAM)
bol in modulation refers to one specific 4 bits/Hz. In this example, there The creation of symbols that are
state of a sine carrier signal. It can be are three amplitudes and 12 phase some combination of amplitude and
an amplitude, a frequency, a phase, shifts. phase can carry the concept of trans-
0209EEmodulation-FIGURE 4
or some combination of them. Basic 90 mitting more bits per symbol further.
binary transmission uses one bit per This method is called quadrature
symbol. A1 amplitude modulation (QAM). For ex-

In ASK, a binary 0 is one amplitude ample, 8QAM uses four carrier phases
and a binary 1 is another amplitude. In plus two amplitude levels to transmit
FSK, a binary 0 is one carrier frequen- 180
3 bits per symbol. Other popular
cy and a binary 1 is another frequency. variations are 16QAM, 64QAM, and
BPSK uses a 0-deg. shift for a binary 256QAM, which transmit 4, 6, and 8
0 and a 180-deg. shift for a binary 1. bits per symbol, respectively (Fig. 4-4).
In each of these cases there is one bit While QAM is enormously efficient
per symbol. of spectrum, it is more difficult to
Data rate in bits/s is calculated as demodulate in the presence of noise,
the reciprocal of the bit time (tb): 4-5. 16APSK uses two amplitude which is mostly random amplitude
bits/s = 1/tb levels, A1 and A2, plus 16 different variations. Linear power amplification is
With one symbol per bit, the baud phase positions with an offset
0209EEmodulation-FIGURE 5 of also required. QAM is very widely used
rate is the same as the bit rate. Howev- . This technique is widely used in in cable TV, Wi-Fi wireless local-area
er, if you transmit more bits per symbol, satellites. networks (LANs), satellites, and cellular
the baud rate is slower than the bit rate telephone systems to produce maxi-
by a factor equal to the number of bits mum data rate in limited bandwidths.
per symbol. For example, if 2 bits per symbol are transmit-
ted, the baud rate is the bit rate divided by 2. For instance, Amplitude Phase Shift Keying (APSK)
with QPSK a 70 Mb/s data stream is transmitted at a baud Amplitude phase shift keying (APSK), a variation of
rate of 35 symbols/second. both M-PSK and QAM, was created in response to the
need for an improved QAM. Higher levels of QAM, such as
Multiple Phase Shift Keying (M-PSK) 16QAM and above, have many different amplitude levels
QPSK produces two bits per symbol, making it very as well as phase shifts. These amplitude levels are more
spectrally efficient. QPSK can be referred to as 4-PSK susceptible to noise.
because there are four amplitude-phase combinations. By Furthermore, these multiple levels require linear power
using smaller phase shifts, more bits can be transmitted amplifiers (PAs) that are less efficient than nonlinear (e.g.,
per symbol. Some popular variations are 8-PSK and 16- class C). The fewer the number of amplitude levels or the
PSK. smaller the difference between the amplitude levels, the
8-PSK uses eight symbols with constant carrier ampli- greater the chance to operate in the nonlinear region of the
tude 45-deg. shifts between them, enabling three bits to PA to boost power level.
be transmitted for each symbol. 16-PSK uses 22.5-deg. APSK uses fewer amplitude levels. It essentially arrang-
shifts of constant amplitude carrier signals. This arrange- es the symbols into two or more concentric rings with a
ment results in a transmission of 4 bits per symbol. constant phase offset . For example, 16APSK uses a dou-

VISIT: for more T&M content | 19


56 subcarriers 4-6. In the OFDM wireless, Long Term Evolution

20-MHz channel
signal for the (LTE) 4G cellular systems, digital
IEEE 802.11n subscriber line (DSL) systems,
Wi-Fi standard, and in most
56 subcarriers power-line communications (PLC)
are spaced 312.5 applications. Some form of OFDM
kHz in a 20-MHz will be the modulation of choice
channel. Data rates for the forthcoming 5G cellular
to 300 Mbits/s can systems.
be achieved with
312.5-kHz 312.5-kHz Each subcarrier
64QAM.S Determining Spectral
subcarrier subcarrier modulated by Efficiency
spacing bandwidth BPSK, QPSK,
16QAM, or 64QAM Again, spectral efficiency is
a measure of how quickly data
ble-ring PSK format (Fig. 4-5). This is called 4-12 16APSK, can be transmitted in an assigned bandwidth, and the
with four symbols in the center ring and 12 in the outer ring. unit of measurement is bits/s/Hz (b/s/Hz). Each type of
Two close amplitude levels allow the amplifier to oper- modulation has a maximum theoretical spectral efficiency
ate closer to the nonlinear region, improving both efficien- measure (Table 2).
cy and power output. APSK is used primarily in satellites, SNR is another important factor that influences spectral
since it is a good fit with the popular traveling wave tube efficiency. It also can be expressed as the carrier-to-noise
(TWT) PAs. power ratio (CNR). The measure is the BER for a given
CNR value. BER is the percentage of errors that occur in
Orthogonal Frequency Division Multiplexing (OFDM) a given number of bits transmitted. As the noise becomes
Orthogonal frequency division multiplexing (OFDM) larger compared to the signal level, more errors occur.
combines modulation and multiplexing techniques to Some modulation methods are more immune to noise
improve spectral efficiency. A transmission channel is than others. Amplitude modulation methods like ASK/OOK
divided into many smaller subchannels or subcarriers. The and QAM are far more susceptible to noise, so they have
subcarrier frequencies and spacings are chosen so theyre a higher BER for a given modulation. Phase and frequency
orthogonal to one another. Their spectra wont interfere modulation (BPSK, FSK, etc.) fare better in a noisy environ-
with one another, then, so no guard bands are required ment, so they require less signal power for a given noise
(Fig. 4-6). level (Fig. 4-7).
0209EEmodulation-FIGURE 6
The serial digital data to be transmitted is subdivided
into parallel slower data rate channels. These lower data Other Factors Affecting Spectral Efficiency
rate signals are then used to modulate each subcarrier. While modulation plays a key role in the spectral effi-
The most common forms of modulation are BPSK, QPSK, ciency you can expect, other aspects in wireless design
and several levels of QAM. BPSK, QPSK, 16QAM, and
64QAM are defined with 802.11n. Data rates up to about
300 Mbits/s are possible with 64QAM. Table 2: Spectral Efficiency for Popular
The complex modulation process is only produced by Digital Modulation Methods
digital signal processing (DSP) techniques. An inverse fast Type of modulation Spectral efficiency (bits/s/Hz)
Fourier transform (IFFT) generates the signal to be trans- FSK <1 (depends on modulation index)
mitted. An FFT process recovers the signal at the receiver.
GMSK 1.35
OFDM is very spectrally efficient. That efficiency level
depends on the number of subcarriers and the type of BPSK 1

modulation, but it can be as high as 30 bits/s/Hz. Because QPSK 2

of the wide bandwidth it usually occupies and the large 8PSK 3
number of subcarriers, it also is less prone to signal loss 16QAM 4
due to fading, multipath reflections, and similar effects
64QAM 6
common in UHF and microwave radio signal propagation.
Currently, OFDM is the most popular form of digital >10 (depends on the type of modulation
and the number of subcarriers)
modulation. It is used in Wi-Fi LANs, WiMAX broadband

VISIT: for more T&M content | 20


influence it as well. For example, the use of forward error efficiency. Newer wireless LAN (WLAN) standards like
correction (FEC) techniques can greatly improve the BER. 802.11n and 802.11ac/ad, along with cellular standards
Such coding methods add extra bits so errors can be like LTE and WiMAX, use MIMO. 5G systems will use mas-
detected and corrected. sive MIMO to develop gigabit range data rates for mobile
These extra coding bits add overhead to the signal, devivces.
reducing the net bit rate of the data, but thats usually an
acceptable tradeoff for the single-digit dB improvement in Implementing Modulation and Demodulation
CNR. Such coding gain is common to almost all wireless In the past, unique circuits implemented modulation
systems today. and demodulation. Today, most modern radios are
Digital compression is another useful technique. The software-defined radios (SDR) where functions like mod-
digital data to be sent is subjected to a compression algo- ulation and demodulation are handled in software. DSP
rithm that greatly reduces the amount of information. This algorithms do the job previously assigned to modulator
allows digital signals to be reduced in content so they can and demodulator circuits.
be transmitted as shorter, slower data streams. The modulation process begins with the data to be
For example, voice signals are compressed for dig- transmitted being fed to a DSP device that generates two
ital cell phones and voice over Internet protocol (VoIP) digital outputs, which are needed to define the amplitude
phones. Music is compressed in MP3 or AAC files for and phase information required at the receiver to recover
faster transmission and less storage. Video is compressed the data. The DSP produces two baseband streams that
so high-resolution images can be transmitted faster or in are sent to digital-to-analog converters (DACs) that pro-
bandwidth-limited systems. duce the analog equivalents.
Another factor affecting spectral efficiency is the use of These modulation signals feed the mixers along with
multiple-input multiple-output (MIMO), which is the use of the carrier. There is a 90-deg. shift between the carrier sig-
multiple antennas and transceivers to transmit two or more nals to the mixers. The resulting quadrature output signals
bit streams. A single high-rate stream is divided into two from the mixers are summed to produce the signal to be
parallel streams and transmitted in the same bandwidth transmitted. If the carrier signal is at the final transmission
simultaneously. frequency, the composite signal is ready to be amplified
Because of their unique path characteristics, the sig- and sent to the antenna. This is called direct conversion.
nals can be identified and demodulated by the receivers Alternately, the carrier signal may be at a lower intermedi-
and reassembled into the original stream. MIMO, there- ate frequency (IF). The IF signal is upconverted to the final
fore, improves data rate, noise performance, and spectral carrier frequency by another mixer before being applied to
the transmitter PA.
103 4-7. This is a At the receiver, the signal
comparison of from the antenna is amplified and
several popular downconverted to IF or directly to
104 modulation methods the original baseband signals. The
and their spectral amplified signal from the antenna
efficiency expressed is applied to mixers along with the
105 in terms of BER carrier signal. Again, there is a
64QAM versus CNR. Note 90-deg. shift between the carrier
that for a given signals applied to the mixers.
106 BER, a greater CNR The mixers produce the original

BPSK 16QAM is needed for the baseband analog signals, which

higher QAM levels. are then digitized in a pair of an-
107 alog-to-digital converters (ADCs)
and sent to the DSP circuitry where
demodulation algorithms recover
the original digital data.
There are three important points
to consider. First, the modulation
and demodulation processes use
6 10 14 18 22 26 30 two signals in quadrature with one
CNR (dB)

VISIT: for more T&M content | 21


another. The DSP calculations call for two quadrature

signals if the phase and amplitude are to be preserved
and captured during modulation or demodulation.
Second, the DSP circuitry may be a conventional
programmable DSP chip or may be implemented by
fixed digital logic implementing the algorithm. Fixed
logic circuits are smaller and faster and are preferred
for their low latency in the modulation or demodulation
process. FPGAs are also widely used.
Third, the PA in the transmitter needs to be a linear
amplifier if the modulation is QPSK or QAM to faithfully
reproduce the amplitude and phase information. For
From wireless prototyping to test applications, NIs hard-
ASK, FSK, and BPSK, a more efficient non-linear ampli-
ware and software platform enables you to overcome the
fier may be used. n
challenges in your RF applications.


VISIT: for more T&M content | 22



The old approach to automated test isnt

scaling, but you already knew that. Look at your
balance sheet. To test smart devices, you need
a smarter test system built on a platform of NI
PXI, LabVIEW, and TestStand. More than 35,000
companies deploy NI technology to lower their
cost of testwhat are you waiting for?

Prepare for the future at

NI PXI, LabVIEW, and TestStand

2016 National Instruments. All rights reserved. LabVIEW, National Instruments, NI,, and NI TestStand are trademarks of National Instruments. Other product and company names listed are trademarks or trade names of their respective companies. 27634

LOU FRENZEL, Contributing Editor


ver the years, the analog swept-tuned chapter offers a hypothetical generic VSA as an example.
spectrum analyzers described in Chap-
ter 1 have morphed into digital signal What VSAs Do
analyzers (DSAs). Today these DSAs The main function of a VSA is to provide a frequency
have evolved into vector signal analyzers domain display so that engineers can do spectrum analy-
(VSAs). The DSA is essentially a swept- sis. VSAs show the frequency components of a signal as
tuned analyzer, but introduces digital techniques at the well as any accompanying interfering signals, harmonics,
IF output. An ADC samples the analog IF and digitizes it, intermodulation products, spurs, and noise. The main
and then stores the data in memory. A processor imple- difference between a DSA and a VSA is that the VSA can
menting the fast Fourier transform (FFT) converts the time demodulate the signal to provide a baseband output. The
domain digital data into a frequency domain output for demodulation process also allows other characteristics of
display. A VSA goes one step further and demodulates a signal to be measured.
the signals into base band outputs that can be measured. Some of the measurements that can now be accom-
One type of DSA is the digital sampling oscilloscope plished are error vector magnitude (EVM), IQ imbalance,
(DSO). Sometimes called a dynamic signal analyzer, crest factor, complementary cumulative distribution
DSOs incorporate a fast sampling ADC to digitize the (CCDF), adjacent channel leakage ratio (ACLR), and
input time domain signal. The resulting data is stored in phase error.
memory and then displayed. By adding an FFT processor A key feature of the VSA is that it has a much wider IF
to the DSO, it can now perform spectrum analysis. While bandwidth than swept-tuned analyzers. Digital modulation
lacking the dynamic range and flexibility of a real spec- schemes such as QAM and OFDM produce wideband
trum analyzer, the DSO DSA may be suitable for some signals with many sidebands. To make measurements
test applications. requiring demodulation, a wide IF bandwidth is needed to
The basic purpose of the newer generation of analyz- capture the full modulation bandwidth. Wider bandwidths
ers is still spectrum analysis, but thanks to modern digital let more noise through, thereby decreasing dynamic
processing techniques, new measurements are possible range, but this is a reasonable tradeoff.
on the more recent complex digitally modulated wireless
signals. This chapter describes the operation and appli- How VSAs Work
cation of the modern VSA. Figure 5-1 shows a generalized block diagram of a
One thing we should note before proceeding: There VSA. As with the swept-tuned analyzers, the VSA is really
are no formal names or designations for the various types a very sensitive superhetrodyne receiver. It uses hetero-
of analyzers. The generic term signal analyzer is often dyne mixers to downconvert the incoming signal into a
used. Furthermore, there are many different implemen- lower intermediate frequency (IF). Most VSAs use multiple
tations from multiple test equipment manufacturers. This conversions to eliminate the image problem that occurs

VISIT: for more T&M content | 24


when a wide input frequency range is to be bandwidth and wide frequency range are
covered. usually desired in a test instrument. An-
Keep in mind that a mixer receives the other solution is to choose the IF so that it
input RF signal and combines it with a permits input filter to reject most images.
local oscillator sine wave, then generates The ultimate solution is to use multiple
the sum and difference frequencies. The downconversionsor even upconver-
difference frequency is usually chosen as sionsto minimize the image problem.
the IF, but not always. IF filters are easier to Dual- and triple-conversion architectures
realize at lower frequencies and can have are common in VSAs. This simplifies the
better selectivity. input filtering requirements. A triple conver-
Furthermore, lower IFs make the sam- sion design is shown in Fig. 5-1. IF band-
pling speed of the ADC much lower. A sin- pass filters (BPF) filters set the bandwidth.
gle downconversion design is vulnerable The local oscillators are commonly
to image problems. The image is a signal tunable PLL synthesizers that are locked to
that is spaced from the desired input by an a stable crystal (XTAL) frequency reference
amount equal to twice the IF. For example, source. This reference source is often tem-
assume an input of 2 GHz and a LO of 2.5 perature-stabilized to reduce phase noise.
GHz. The IF is the difference, or 500 MHz. It is also common for the VSA to use
5-1. A generic block Should an input signal of 3 GHz appear internal circuit switching (SW1 and SW2)
diagram of a vector at the input, the mixer will also develop a to select a configuration that best fits the
signal analyzer. Different 500-MHz IF. The image interferes with the frequency range of interest. For the high-
conversion configurations desired 2-GHz signal. est frequency range, a single or double
are used to minimize One solution to the image problem is conversion may be used. An upconversion
image interference filtering the mixer input so that it passes may be used instead of a downconversion.
depending upon the input the desired signal but rejects the image. Lower frequency ranges will use two or
frequency range. This is often difficult to do since wide input three conversions as needed to minimize

8 GHz

Attenuator LNA
Mixer #1 Mixer #2
100 Hz-6 GHz
SW1 path
Input GHz
signal Filter 800 MHz


Mixer #3
High-frequency path High
Mixer #4
BPF BPF ADC Memory Other

800 MHz 100 MHz

LO Display

VISIT: for more T&M content | 25


the image problem. Another unique display is the spectrogram. This is a

In Fig. 5-1, lower-frequency signals from 100 Hz to 6 frequency domain display that varies over time to show
GHz pass through SW1 to mixer #1 and are upconverted changes. The frequency display is usually in color and
to 8 GHz; mixer #2 then downconverts this to 800 MHz. flows vertically to show signal changes with time. This
SW2 selects the low frequency path. Next, mixer #4 type of readout is sometimes called a waterfall display.
downconverts this 2nd IF to 100 MHz, where it is applied One variant on the VSA is the real-time (RT) VSA,
to the ADC. High-frequency inputs from 6 GHz to 40 GHz which offers some benefits for certain kinds of testing.
go through SW1 to mixer #3, then via SW2 to mixer #4. The RT VSA has the ability to perform all of the necessary
The final IF signal is applied to an anti-alias filter (not digital processing on the input as it occurs. Standard
shown), then to a fast ADC. Sampling rates of several VSAs digitize the incoming signal in blocks and store
hundred MHz are common. The digitized signal is stored these segments in memory. Then the digital calculations
in memory. This time domain signal does contain the operate on the data.
amplitude and phase information needed to demodulate For this reason, the output lags the input. What you
the signal. see hopefully is a relatively constant repeating input that
To develop the display and perform a measurement, is stable. If changes occur during the dead zone between
the digitized signal in memory is subjected to various signal capture and the end of computation, such as tran-
processes. IQ separation, filtering, downconversion, deci- sients or other short-term signals, the VSA misses them.
mation, and FFT are all employed to extract the frequency An RT VSA solves this problem.
content, as well as to demodulate the signal for further The RT VSA must have an ADC that meets the Nyquist
analysis and measurement. The digital signal processing requirement of being able to sample the input at more
(DSP) algorithms can be implemented in a microproces- than two times the VSA bandwidth. In addition, the RT
sor (MCU) or in a fast ASIC or FPGA. In most VSAs, the VSAs computation capability using a DSP or FPGA must
demodulation is done via software. But older designs be fast enough to keep up with the ADC. The process-
have used a hardware IQ demodulator to generate the ingFFT, filtering, demodulation, etc.must occur con-
baseband signals. The processed signals are then condi- currently with the input rate.
tioned for display. With this arrangement, the display keeps pace with
A unique capability of the VSA is its ability to out- any input signal changes. The processed data is then
put different display formats. In addition to a traditional stored in memory. RT VSAs are great for analyzing pulsed
frequency domain display, the VSA can display constel- RF, transients, and other signals that a standard VSA
lation, IQ vector, and eye diagrams thanks to the internal misses. Further processing occurs after storage to dis-
demodulation capability. Figure 5-2 shows a 16QAM play the various demodulated forms of baseband output.
constellation diagram as an example.
VSA Specifications
5-2. A constellation All of the specifications giv-
diagram output for en for the swept-tuned analyzer
16QAM. Each dot apply to the VSA. (see Chapter 1).
represents a unique Analysis bandwidth is one addi-
carrier amplitude tional specification of importance.
and phase symbol. Typical available bandwidths are
The dots are 40, 100, and 160 MHz. Newer VSA
derived from the models offer even wider band-
input signal through widthsup to several hundreds of
demodulation MHz.
computation. When selecting a VSA, keep in
mind that you can get a standard
dedicated bench instrument or
a modular version. The modular
version uses the PXI standard with
PCIe bus. Some PXI VSAs use
multiple modules. For example,
one VSA is made up of a frequen-

VISIT: for more T&M content | 26


cy reference module, multiple synthesiz-

er modules for LOs, a downconversion
module, and a digitizer module. An
internal PC manages the plug-in mod-
ules and runs the software.
A newer form of VSA is the vector
signal transceiver (VST). This is usual-
ly a PXI product that combines a VSA
and a vector signal generator (VSG) in
a single package. VSGs are covered in
Chapter 7.
One final, important consideration is
the software available to customize the
VSA to your measurement needs. Soft-
ware allows the VSA to make the need-
ed tests for cellular standards, Wi-Fi, or
Wireless devices require high performance
other common wireless technologies. n
test systems to characterize linearity.
Explore tips for getting the best performance
BACK TO TABLE OF CONTENTS from Vector Signal Analyzers.

VISIT: for more T&M content | 27


LOU FRENZEL, Contributing Editor


ost measurements made on microwave The relationship between power in
and RF equipment are power-related. watts and decibel measure
For example, measurements made with
Watts dBm dBW
a spectrum analyzer are power versus
1,000 +60 +30
frequency. While such measurements are
useful, there are some test needs that 100 +50 +20
require a more accurate power measurement. For such 10 +40 +10
use cases, a power meter is the desired instrument. This 1 +30 0
chapter concerns power meter types, their application, 0.1 +20 -10
and operation. 0.01 +10 -20
0.001 (one milliwatt) 0 -30
What Power Meters Do
0.0001 -10 -40
The basic electrical unit of power measurement is the
0.00001 -20 -50
watt. While power meters express their outputs in watts, a
more common unit of power measurement is the decibel 0.000001 (one microwatt) -30 -60
(dB). The dB is used to express the power gain or loss in a
device: Power is given in watts, but milliwatts (mW) can be
used in the calculation.
dB = 10log (Po/Pi) A power level of 500 mW is:

Po is the output power of the device and Pi is the input dBm = 10log (500/1) = 27 dBm
power. The dB is useful as its logarithmic nature allows
wide power level ranges to be compressed into a smaller Types of power. There are different types of power
range of values. measurements. These include continuous wave (CW), av-
The dB is also used to indicate power directly. Power is erage, peak, and peak envelope power (PEP). CW power
expressed in relationship with a fixed value. For instance, is a sine wave that dissipated into a load; this is an aver-
dBm is power referenced to one milliwatt. The unit dBW is age power. Another type of average power is that from a
power level referenced to one watt. The table shows the modulated carrier with amplitude variations. A power meter
relationships; the dBm is the more widely used unit. Micro- averages the instantaneous amplitude variations into an
wave and RF power measurements range from kilowatts average power over many cycles of the carrier.
(kW) to microwatts (W): Pulse or peak power is the maximum power value
averaged over the duration of a pulsed signal. This power
dBm = 10log (P/.001) value depends upon the peak value and the pulse width
and duty cycle. Others factors are pulse rise and fall times,

VISIT: for more T&M content | 28


6-1. A through- Output to coaxial Measuring power in complex
line power meter R1 modulation systems (FM,
cable transmission
samples the signal line and antenna CDMA, PSK, QAM, OFDM).
from a coaxial Power
Measuring power in pulsed
section and input from communications systems (ra-
produces an output transmitter C dar, navigation, etc.)
voltages in both the B Measuring power in wired
forward and reverse communications links like
directions, making 50 cable TV.
VSWR readings Copper LOAD
possible. Such microstrip D1 R2 How Power Meters Work
meters can handle on printed 50 The typical modern power
power levels up to circuit board meter usually consists of two
several kilowatts. D2 parts: a measurement head or
C1 R3
sensor that is connected to the
Reflected device under test (DUT) and
Forward a processing display unit. The
sensor reads the power signal
C2 and generates an analog signal
Meter A
proportional to power; performs
signal manipulation in certain
cases, some involving digiti-
R5 zation with an ADC; and then
sends the signal via a cable to
the main processing device.
The processing display
as well as any ringing or overshoot. device may be a bench instrument that per-
Peak envelope power (PEP) is the maxi- forms additional manipulation of the signal
mum value of the envelope of an RF signal. using DSP, then displays the measurement. A
This measurement is useful in some pulse conventional numerical readout may be used
measurements, as well as in power measure- showing the measured power in watts or dBm.
ment on amplitude modulated (AM) and single The processing unit may also be a laptop or
sideband (SSB) signals. other PC that uses the internal processor to
Applications. Power measurements are run software for DSP operations, conversions,
made as part of design and development, in and display presentation.
addition to manufacturing production testing. There are two basic types of power meters:
This testing may be evaluating a circuit to see through-line and absorptive meters. Here is
if it meets design specifications, testing a de- more detail on each type
vice to see if it meets standards or regulatory Through-line meter. This type of meter
requirements, or final production tests before passes the signal to be measured through
shipping. a circuit or component, in series with the
Here are just a few examples: coaxial feed line, on the way to the normal
Testing the output of oscillators and frequen- load. The feed line section samples the signal
cy synthesizers. by inductive and capacitive coupling; diodes
Testing power amplifiers used in cell rectify the samples into a DC voltage that is a
phones. measure of the power going to the load. The
Testing power amplifiers used in cellular sampling section may be a short segment of
base stations. air-coaxial line, a PC board pattern, or a toroi-
Testing power amplifiers in radio transmitters dal transformer.
(land mobile, aircraft, marine, broadcast, Figure 6-1 shows one type of through-

VISIT: for more T&M content | 29


line meter called a monomatch. It is made up of a short sensors are even more sensitive than thermistors. They are
segment of microstrip transmission line on a printed circuit stable and accurate and offer a fast response time.
board. The impedance of the line is usually 50 to match Thermocouple sensors for power measurements are
the input coax line, output coax line, and the external load, usually made on a chip of semiconductor materials. The
such as an antenna or other circuit. The center microstrip chip is coupled to the RF input that heats the junction pro-
section carries the signal while the segments on either side ducing the voltage. The very small output voltage requires
sample the signal. Segment A samples the forward power, amplification in a chopper stabilized preamplifier. A FET
and segment C samples any reverse power arising from chopper converts the dc into a low-frequency ac signal
a load and line impedance mismatch. Diodes rectify the that is amplified and further processed into a power value
samples creating a DC voltage that represents the power. by the processing software. Thermocouple sensors have
A meter calibrated in power units is used for the readout. a wide dynamic range and can handle power levels from
This type of meter provides both forward and reverse nanowatts to several watts. They can be used to frequen-
power measurements. With this information, the voltage cies up to 50 GHz.
standing wave ratio (VSWR) can be calculated. VSWR is The most widely used power sensor is the diode sen-
a measure of the mismatch between the line and load im- sor. The RF signal is applied across a load resistor, usually
pedances. A VSWR reading of 1:1 means an exact match, 50 , and a half wave diode rectifier converts the voltage
whereas any larger value indicates the degree of mismatch across the resistor into pulses that are then filtered with a
and loss of output power. capacitor into a dc voltage. That voltage is amplified and
Through-line power meters may be external units or then otherwise processed to a power value.
they may be built into other equipment, such as a transmit- The diode is typically a metal-semiconductor Schottky
ter. This type of meter is also the preferred way to measure diode that operates at very high frequencies and has a low
higher power levels. Absorptive power meters are usually forward voltage of about 0.3 volt. Some modern diode sen-
limited to a few watts input, but the through-line meter can sors use a GaAs diode for better high-frequency response.
easily handle hundreds or even thousands of watts. Error Typical operating frequency range is up to 30 GHz. These
of measurement is several percent or less. sensors can accommodate RF input voltages in the 70 to
Absorption meters. An absorption meter contains a +20 dBm range.
load that absorbs the input signal power and generates Most of these sensors work with low-level input signals.
a heat difference that can be measured as an average Anything more that about +20 dBm could damage the
power. There are three common types of absorption sensor sensor. The input level range varies with the instrument
heads: thermistor, thermocouple, and diode. vendor. If higher power levels are to be measured, it is
A thermistor sensor uses a negative temperature co- common practice to place a passive resistive attenuator
efficient thermistor in a bridge circuit that absorbs power between the power source and the sensor. This will protect
from the input. As the thermistor heats up, its resistance the sensor head.
decreases. This produces a current change in the bridge One final power measurement device is the IC power
that is in turn converted into a proportional voltage, which detector. This miniature chip is designed to be incorporat-
is sent to the processing unit that corrects for the non-lin- ed into specific products. One common example is a cell
earity of the thermistor and calculates the power value to phone. Some RF, modulation, and access methods require
be displayed. Thermistor sensors are very sensitive, and that the cell phone monitor its output power, increasing
as such produce excellent results with low-level signals. or decreasing it as needed to maintain contact with the
This type of sensor can be used with inputs from about basestation cell site. These devices measure input aver-
one microwatt to 20 milliwatts. The device is frequency age power levels from about 70 dBm to +3 dBm over
insensitive, meaning it can operate over a wide frequency a frequency range of 10 MHz to 10 GHz. The output is a
range from roughly 10 MHz to 20 GHz. DC voltage that varies linearly with dBm input power. The
An even better heat-based sensor is the thermocouple IC power detector is not widely used in measuring instru-
sensor. The thermocouple is made up of two dissimilar ments.
metals that form two junctions in a loop. When one junction
is heated and the other cooled, a small voltage is devel- Power Meter Specifications
oped within the loop. The loop is opened at the cooled There are only a few key specifications to remember in
junction end to access it. That dc voltage, in the microvolt selecting a power meter. These are type of meter, the kind
to millivolt range depending upon the metals used, is then of power it can measure, input dynamic range, frequency
translated into an equivalent power value. Thermocouple range, and accuracy.

VISIT: for more T&M content | 30


First select the kind of meter suited to the type of equipment

being tested and the kind of power to be measured. Absorption
meters are the most useful; they can usually measure average pow-
er and peak power and in some cases other forms of power. Diode
and thermocouple sensor heads are the most widely used.
Input dynamic range is another important specification. Total
range is usually about 70 to + 3 dBm.
Frequency range is also essential. Power meters are available
for measurement from a low of 10 MHz up to about 50 GHz or more
in the newer instruments.
Check the instrument vendors measurement error rating. It var- Complete your RF Test System, with a USB
ies with the sensor and the tests but is usually several percent. n RF Power Sensor, equipped to measure WiFi,
LTE, GSM and EDGE waveforms with a 80 dB

VISIT: for more T&M content | 31


LOU FRENZEL, Contributing Editor


ignal generators are among the key test Signal generators are also widely used for testing ampli-
instruments used in microwave and millime- fiers, filters, and some components. Amplifier tests include
ter-wave measurements. They provide the gain, frequency response, phase response, and distortion
signals needed to test circuits without over- products. Filter measurements include attenuation, fre-
the-air communications. These instruments quency response, and phase response. Components like
are widely used in research and develop- splitters and combiners, duplexers, isolators, and circula-
ment, design, maintenance/repair, and manufacturing tors can also be tested.
production testing. Multiple types are available depending There are several categories of signal generators. These
upon the application. This chapter summarizes the various are function generators, RF generators, vector signal gen-
types of signal generators, how they work, and where they erators (VSGs), and arbitrary waveform generators (AWGs).
are used. Function generators. Function generators are signal
sources that produce sine, square, and triangle waves over
Applications a frequency range from 0.001 Hz to 10 MHz. The upper-
A signal generator is a test instrument that produces a frequency limit may be higher and varies by manufacturer.
signal or carrier, usually sinusoidal, whose frequency and Most function generators can also be modulated using
amplitude may be varied to fit the device under test (DUT). AM or FM. The output is continuously variable from zero to
Most generators can also modulate the carrier with all the some upper level. The output is usually through a BNC con-
various forms of modulation, including the more advanced nector with a 50- coax cable to the DUT. Function genera-
types like QAM. Some generators are also capable of be- tors also generally incorporate a frequency sweep function
ing programmed to generate special, unusual, or non-stan- that uses an internal sawtooth wave to frequency modulate
dard signals. the sine wave. This produces an output that continuously
One common use is receiver testing. The generator sim- and linearly varies the output frequency over a selectable
ulates the normal signals to be received, taking the place of segment of frequencies. The sweep function is useful for
the normal transmitter. The signal is injected into the receiv- testing filters and amplifiers.
er front-end so that measurements on the receiver can be RF generators. These generators produce signals at
made. Some typical measurements may be receiver gain the higher wireless frequencies. A typical low-end frequen-
and sensitivity, frequency response and selectivity, reaction cy is 10 MHz, while upper-limit frequencies can vary from
to spurious signals, and output characteristics. a few GHz to nearly 100 GHz. The frequency is variable

VISIT: for more T&M content | 32


XTAL bench instruments. However, theyFigure

may7-2also be modular
REF Phase
detector VCO
units such as PXI modules thatDigital
into ainput
base chassis,
Phase increment register
along with other test instruments such as a VSA.
N LPF Output Clock
How Signal Generators Work Memory address register
All signal generators start with aMemory
main signal source
whose frequency can be varied. A highly LPF stable oscillator
divider was used in early generators, but the main signal source
Pattern input for custom waveforms
Digital in a modern generator is a frequency synthesizer. This is a
highly stable signal source whose frequency can be var-
7-1. Shown are the essential components of a phase- ied in increments. The two basic types used in commercial
Figure 7-1
locked loop (PLL) frequency synthesizer. generators are covered here.
Frequency synthesizers. The two types of synthesiz-
either continuously or in selected increments. Direct key- ers are the phase-locked loop (PLL) synthesizer and the
board entry of a desired frequency is a feature on some direct digital synthesizer (DDS). The basic PLL synthesizer
generators. The output amplitude is also variableeither is shown in Fig. 7-1. The stability and precision are derived
continuously or in attenuator steps. In addition, the RF gen- from a highly accurate temperature-stabilized or oven-sta-
erators feature extensive modulation capability, including bilized crystal (XTAL) reference (REF) oscillator (OSC).
AM, FM, PM, m-PSK, ASK, FSK, and m-QAM. Most of the It is followed by a frequency divider whose division
generators have a sweep function. factor (N) can be varied. The reference oscillator frequency
Vector signal generator. A special type of RF gen- and divider settings determine the increment of frequency
erator is the vector signal generator (VSG). It generates change in the output. For example, a reference frequency
the in-phase (I) and quadrature (Q) signals used in ad- of 10 MHz and a divider factor of 100 gives an output of
vanced modulation schemes like QAM and OFDM. These 100 kHz. This means that the output of the synthesizer can
more-complex signals are needed to test cell phones, Wi- change in 100-kHz steps.
Fi, and other wireless standards using complex modulation The reference signal goes to a phase detector whose
methods. output is a function of comparing the reference signal to the
Arbitrary waveform generators. An AWG is a gener- output of the synthesizer, divided down by a factor of M in
ator capable of being programmed to produce special or a variable frequency divider. The output of the synthesizer
unusual waveforms. Some examples are shaped pulses comes from a voltage-controlled oscillator (VCO) whose
and sine carrier bursts. The desired output wave is defined frequency is varied by a dc input from the phase detector
and stored in binary form in a memory. The binary pattern output by way of a lowpass loop filter (LPF).
is then sent to a digital-to-analog converter (DAC) that When the two phase detector inputs are at the same fre-
produces a stepped approximation of the desired signal. quency, the circuit is said to be locked and the VCO output
Frequency and amplitude are variable. The upper-frequen- has the stability of the reference. If the output frequency
cy limit is generally restricted since digital sampling tech- varies, the phase detector generates an error output that
niques are used. is converted into a dc voltage by the loop filter, thereby
Physically, signal generators are usually independent varying the VCO output frequency until the error is reduced
to zero.
To change frequency, you change the frequency
divider ratio (M). This is done from a keyboard or
Digital some other digital input. For instance, with a divider
tuning factor of 451 and a reference input to the phase de-
Memory DAC
Memory tector of 100 kHz, the VCO output will be 45.1 MHz.
address There are several variations of the PLL synthesiz-
er, but all are enhancements to the concepts just
Phase LPF
Output described.
register Pattern input for Direct digital synthesis (DDS) uses the method
custom waveforms shown in Fig. 7-2. A binary pattern representing
7-2. This is a basic block diagram of a direct digital synthesizer a sine wave is stored as a look-up table in a RAM
(DDS) and arbitrary waveform generator (AWG). or ROM. These binary samples are sent to a

Figure 7-2

VISIT: for more T&M content | 33


Sine Sine AM Coupler

shaper Step
AMP attenuator
Frequency Amplitude Out
Triangle Out
Synthesizer modulator

Square Attenuator

Integrator Comparator
7-4. A simplified RF signal-generator configuration is
Figure 7-4
7-3. Depicted is a low-cost function-generator circuit. shown.
DAC that produces a stepped approximation of the sine cation. The output is a sine wave. Generators are available
wave that is filtered by a lowpass filter to smooth it out. The in different ranges, from a low of 6 GHz to a high of 90 GHz.
frequency is changed by varying the rate of binary word The synthesizer output is sent to an amplitude modulator
output rate to the DAC. A stable crystal-controlled fixed that can vary the amplitude of the signal. This allows AM to
frequency clock updates the memory address register by be implemented.
adding a constant from a phase increment register to the Also, the modulator is used for automatic level control
current memory address. (ALC). Feedback from the output is rectified and filtered
The result is a different selection of memory sam- into dc to control the modulator, maintaining a constant out-
ples that changes the output frequency. The constant is put amplitude over the frequency range used. Step attenu-
changed by a front panel keyboard or other digital source. ators control the output amplitude.
By using many binary samples, large address words, and Vector signal generators. A VSG is essentially an RF
large phase words, frequency increments can be very pre- signal generator with added vector modulation capability.
ciseeven fractions of a Hz. One type of integrated circuit Today a wide range of wireless technologies (e.g., cellular
DDS uses 48-bit addresses and 17-bit phase increment and Wi-Fi) use complex digital modulation schemes like
constants. The upper-frequency range is limited by the M-PSK, M-QAM, and OFDM. Digital modulation involves
number of samples and DAC update rate. vector modulation methods that mix both amplitude and
Function generators. A typical low-cost function- phase changes of the carrier. A VSG uses an in-phase
generator block diagram is shown in Fig. 7-3. It does not (I) and quadrature (Q) modulator to generate these test
use a PLL or DDS source. Instead, it uses an op amp signals.
integrator and a comparator in a closed loop to form an os- Figure 7-5 shows a simplified version of a VSG. It begins
cillator. This circuit inherently generates triangle and square with a PLL or DDS synthesizer as the prime signal source.
waves. The frequency is varied by changing the integrator The synthesizer feeds the IQ modulator, the ALC circuits,
RC values. A shaping circuit converts the triangle wave into an amplifier, and the step attenuator. The digital data to be
a sine wave. Modern function generators
IQ modulator
typically use a DDS synthesizer instead AMP Coupler
of the integrator oscillator. Attenuators on Frequency Out
synthesizer ALC
the outputs vary the signal level. Function
90 Step
generators usually have a maximum output
frequency of 10 to 20 MHz; as such they Q I
are not used in microwave and millime-
Processor or Q
ter-wave testing. FPGA & RAM

Symbol I
RF Signal Generators mapping DAC LPF
A basic RF signal generator is illustrat-
ed in Fig. 7-4. It begins with a frequency Baseband generator
synthesizer (either PLL or DDS) that has an 7-5. Depicted are the primary components of a vector signal generator
upper-frequency limit suitable to the appli- (VSG).
Figure 7-5

VISIT: for more T&M content | 34


transmitted is developed in the baseband generator. variation caused by temperature and voltage instabilities
The data is symbol mapped according to the mod- or noise. These variations produce low-level sidebands
ulation, such as 16QAM, and the digital I and Q signals around the carrier. The level is measured at some offset
are produced. These are sent to DACs where they are from the carrier frequency (e.g., 100 kHz, 1 MHz, etc.). Typ-
converted into their analog versions and filtered before ical range should be less than about 120 dBm/Hz.
being applied to the IQ modulator. The key to a VSG is the Spectral purity. This refers to unwanted signals in the
baseband digital generator capability. Some VSGs can be output, such as harmonics and intermodulation products.
preprogrammed to exactly emulate commonly used test These should be more than 60 dBc, carrier referenced.
signals, like LTE cellular of some Wi-Fi version. Modulation capability. AM, FM, and PM, along with
Arbitrary waveform generator. An arbitrary waveform some pulse variants, are usually standard. For VSGs,
generator (AWG), sometimes called an ARB or arbitrary BPSK, QPSK, FSK, QAM, OFDM and others are available
function generator (AFG), is a special signal generator that as needed. A sweep mode is usually included in most
can produce any standard or unusual signal shape. For models.
example, the AWG can be used to simulate voice or video Configuration. Benchtop or modular (PXI).
signals, pulses with impairments like ringing and droop, or Interfaces. USB, SPI, I2C, etc. n
impulse ultrawideband (UWB) wavelets.
An AWG is based upon the concepts of a DDS, where BACK TO TABLE OF CONTENTS
stored samples of a waveform are fed to a DAC followed
by a lowpass filter to create an analog output (refer back
to Fig. 7-2). In an AWG, there is an input to the memory to
store the word samples of the desired custom waveform.
One source is special software running on a PC that can
be custom-programmed as needed. Another source is to
capture a desired signal using a digital oscilloscope. The
scope digitizes the waveform and stores the samples that
may then be transferred to the AWG. Most AWGs also have
pre-stored waveforms for sine, square, triangle, sawtooth,
and other common signals.
Learn about how signal generators generate
analog signals and other topics such as types
Generator Specifications of signal generators, bit resolution, bandwidth,
Keep in mind that there are many different commercial attenuation, digital gain, digital filtering, and
variations of the basic generator types described above. analog filtering.
Test instrument vendors usually offer several different
models of each type of generator with different features
and pricepoints to fit specific applications. However, the
specifications listed below generally apply to all types and
Frequency range. Both lower and upper limits are
given. Common upper limits are 6 GHz, 26.5 GHz, 40 GHz,
and 90 GHz. Higher frequency limits may sometimes be
achieved with external upconverters. Consider frequency
increment tuning ranges, as well.
Output level. Variability of signal level is a must. The
upper limit is typically less than +20 dBm, while the lower
level may be down to 160 dBm for receiver sensitivity
testing. Amplitude increment variation is a consideration (1
dB, 5 dB, etc.)
Output impedance. Usually 50 , but 75 may be
available. Coax cable connections via N-type, SMA, or
Phase noise. Phase noise is the minor output frequency

VISIT: for more T&M content | 35


CHRIS DeMARTINO, Technology Editor


ontinuous-wave (CW) signals are used power sensor can be used to measure pulsed RF signals,
very often in RF/microwave applications. To they may not always be the most effective option. Often-
measure the power level of such signals, times, measuring these signals can be performed more
a CW power sensor is commonly used. effectively by taking a different approach. It is therefore
Many diode-based CW power sensors have important to understand pulsed RF signals, as well as the
a wide dynamic range, allowing them to techniques that can be used to measure them.
be an excellent choice for measuring unmodulated CW
signals. However, some RF/microwave applications utilize Description of Pulsed RF Signals
pulsed signals rather than CW signals. Figure 8-1 shows an ideal pulse. Among the character-
For instance, radar applications often use pulsed RF istics that describe pulses are peak power, average power,
signals, which can be described as periodic bursts of an pulse repetition frequency (PRF), pulse repetition interval
RF carrier. In essence, these signals are on for a certain (PRI), pulse width, duty cycle, rise time, and fall time. PRF
amount of time. At the end of the on cycle, the signal is is the rate at which pulses are generated. PRI is inversely
then off for a period of time. While a CW or average


Peak power Modulating pulse

Pulse repetition

Pulse width
Duty cycle =
Pulse repetition interval
RF carrier
Average power = peak power duty cycle 8-2. Shown is a modulating pulse alongside an
8-1. In this figure, an ideal pulse is shown. RF carrier.

VISIT: for more T&M content | 36


RF signal can be calculated when the average power

value and the duty cycle are known quantities. Thus, an
average power meter can be used to indirectly measure
peak power. In other words, the average power is first
measured. If the duty cycle is known, the peak power can
be calculated mathematically as follows:

Peak Power = Average Power

Duty Cycle

Pulse-modulated CW signal However, this method of calculating peak power is

only accurate in the case of an ideal pulse, such as the
8-3. This figure shows a pulse-modulated RF signal. one shown in Fig. 8-1. Figure 8-4 shows a pulse that is not
purely rectangular. Additional factors, such as overshoot
related to the PRF. Pulse width is simply the duration of a and ringing, result in a non-ideal shape. In this scenario,
pulse. The peak power of a pulse can be seen clearly in calculating the peak power by the method described will
Fig. 8-1. The average power of a pulse is related to its duty produce an inaccurate result. Thus, to calculate an accu-
cycle and peak power level. rate peak power value based on the average power value,
A pulse-modulated RF signal is created when a pulse the pulse must be purely rectangular and have a constant
modulates an RF carrier. Fig. 8-2 is an illustration of a duty cycle. If these conditions are not met, the calculation
modulating pulse along with an RF carrier. When the pulse will not be accurate.
modulates the RF carrier, the result is a pulse-modulated Unlike an average power meter, a peak power me-
RF signal, shown in Fig. 8-3. ter can directly measure the peak power of a pulsed RF
signal. A peak power meter would be used with a peak
Options for Measuring Pulsed RF Signals power sensor, which is a diode-based detector with a fast
Pulsed RF signals can be measured by several different response. That response enables it to accurately measure
test instruments. Two options will be discussed here: peak the envelope of a pulsed signal. One drawback of peak
power meters and spectrum analyzers. Peak power meters power sensors is they generally have less dynamic range
are essential when it comes to measuring pulsed RF sig- than average power sensors.
nals, as they provide measurement capabilities that cannot Peak power meters often have trace display capabili-
be achieved by average power meters. They can easily ties, allowing one to see the envelope of a pulsed RF sig-
analyze pulsed RF signals, allowing users to measure nal on a display (Fig. 8-5). In essence, peak power meters
not only peak power, but also average and instantaneous display a response similar to an oscilloscope. It should be
power. noted that a peak power meters name may actually be
Spectrum analyzers can be used to analyze pulsed RF somewhat misleading. The reason why is that in addition
signals in the frequency domain. Viewing such signals on to peak power, peak power meters can measure both
a spectrum analyzer can reveal a average power over a defined time
number of potential problems. They Overshoot
interval and the instantaneous power
do have a disadvantage in terms at any point in time.
of cost, as spectrum analyzers are In addition to direct power mea-
typically more expensive than peak surements, peak power meters allow
power meters. users to analyze a variety of param-
eters, such as pulse width, PRI, and
Peak Power Meters rise/fall times. Many peak power
An average power meter is com- meters also have triggering capa-
monly used to measure CW signals. bilities. Some provide even more
Such power meters can also be used advanced capabilities, such as
to measure the average power of statistical analysis.
pulsed RF signals. However, average Video bandwidth is one of the
Non-ideal pulse
power meters cannot directly mea- most important parameters, as it
sure peak power levels. 8-4. In this figure, a pulse with non- must be sufficient to accurately track
The peak power level of a pulsed ideal characteristics is shown. the envelope of a pulsed signal.

VISIT: for more T&M content | 37




8-5. Shown is an illustration of the envelope of a pulsed

signal on a peak power meter display.

Video bandwidth and rise time are inversely related. In 3 1 1 3

fC fC fC fC + fC +
essence, the faster the rise time of the pulse being mea- 4 PW 2 PW PW 2 PW
fC +
fC fC fC +
sured, the larger the video bandwidth must be. Video PW PW PW PW

bandwidth is typically specified by manufacturers. 8-6. This illustration depicts a display seen on a
spectrum analyzer when measuring a pulsed RF signal.
Spectrum Analyzers
A traditional spectrum analyzer is another instrument the PRF, each individual spectral component cannot be
that can be used to measure pulsed RF signals. By using a displayed. However, the shape of the spectrum can still
spectrum analyzer, one can analyze a pulsed RF signal in be preserved. Therefore, this approach is still practical in
the frequency domain. Thus, a spectrum analyzer provides many instances. n
information that cannot be attained by a peak power meter.
Figure 8-6 is a generic illustration of a spectrum analyzer BACK TO TABLE OF CONTENTS
display of a pulsed RF signal.
The pulsed RF signals spectrum consists of a main
lobe accompanied by side lobes. The widths of the main
lobe and side lobes are inversely related to the pulse
width. Essentially, as the pulse width widens, the widths of
the main lobe and side lobes will decrease. The spacing
between each spectral component is determined by the
PRF. As the PRF increases, the spacing between spectral
components also increases.
When using a spectrum analyzer to measure pulsed
RF signals, one must use the proper resolution bandwidth
Examine the types of pulsed signals used in
(RBW). The spectrum analyzers RBW must be much
radar applications and learn best practices for
less than the PRF to differentiate between each spectral
accurate pulsed power measurements.
component. If the spectrum analyzers RBW is greater than

VISIT: for more T&M content | 38


JOHN BLYLER, Contributing Editor


probe is a device that connects a circuit of microwave signals, and sub-nanosecond rise times, signal
interest to a test instrument via a cable. In so paths must be considered as a transmission line that
doing, the probe and cable usually attenu- degrades the signal being measured. The probe, cables,
ate the desired measured signal. attenuators, and any adapters all contribute to the cor-
For relatively low-frequency signals, the ruption of the signal along the transmission line to the test
probe-cable-instrument connection does equipment.
not usually present a significant measurement obstacle.
However, the signal loss becomes more pronounced for Probe Details
higher-frequency microwave and RF signals. This chapter For DC signals (0 Hz), probes act as a low-value resis-
will examine the importance of instrument connection ca- tors. However, for microwave and RF alternating current
bles and probes. The discussion will cover types, limita- (AC) signals, probes have impedance with resistance,
tions, and accessories (including attenuators, isolators, and inductance, and capacitance (R, L, and C) as shown in Fig.
couplers). 9-2. The inductance reacts to AC signals by increasingly
impeding AC current flow with higher signal frequencies.
Importance of Probes and Connections The capacitance reacts to AC signals with decreasing im-
Before explaining the importance of probes and con- pedance to AC current flow as signal frequency increases.
nections, lets first define the terms. A probe is a device that The capacitance and inductance are distributed along the
makes a physical and electrical connection between a test
point on a device, known as the device under test (DUT), Probe
and an electronic test instrument (Fig. 9-1). In practice, Box
most probes consist of a probe head, a probe cable, and Oscilloscope
a compensation or signal condition box with a connector to
the instrument. The latter is needed to electrically match the
probe signal to the input instruments (more on that later).
Virtually all probes, oscilloscopes, and other test instru-
ments use coax cable. This type of cable is self-shielding,
and so helps reduce unwanted electromagnetic interfer- Circuit
under Test
ence (EMI) with the signal. These cables have a measur- Probe
able impedance and attenuate the signal being measured. Probe cable
So why should engineers worry about probes, cables,
and attenuators? When dealing with RF and (especially) 9-1. This is a probe-cable-equipment functional diagram.
Test point

VISIT: for more T&M content | 39


R L Probe and cable and cables designed for the

instrument, as it is essential to
get the desired measurement
C C Oscilloscope accuracy and results.

Probe Types
Ground lead
There are three common
9-2. In the presence of AC signals, probes and their cables act as distributed probe types: passive, active,
resistive, inductive, and capacitive circuits that form a transmission line. and differential. A passive
Figure 9-2 probe contains no active
cable forming a transmission line with a known impedance. components like transistors or
This transmission line acts like a lowpass filter that has an amplifiers. They tend to be the most rugged, economical,
upper cut-off frequency, as well as a finite amount of atten- and thus widely used probes. Passive voltage probes
uation. All this is factored into the measurement accuracy of are available with various attenuation factors1, 10,
the instrument. and 100for different voltage ranges. Of these, the 10
The effect of putting a probe in a circuit is known as cir- passive voltage probe is the most common, as it is typically
cuit loading. Circuit loading adds extra impedance (RLC) to supplied as a standard accessory with oscilloscopes.
the circuit being measured, which can change the desired To minimize circuit loading, attenuator probes (e.g.,
signals by a lot or a littledepending on the magnitude 10 probes) are used. A typical probe uses a 9-M series
of the loading. For signal frequencies under 5 kHz, circuit resistor shunted by a low-value capacitor to make an RC
loading is mostly resistive. To avoid significant circuit load- compensated divider with the cable capacitance and the 1
ing here, all you need is a probe with a resistance at least M test instrument input impedance. The RC time con-
two orders of magnitude greater than the circuit impedance stants are adjusted to match. At low frequencies where the
(100 M probes for 1 M sources, 1 M probes for IO k resistance of R is much less than the reactance of C, the
sources, etc.). circuit looks like a resistive divider. But at higher frequen-
For high-frequency RF and microwave signals, the cies, where the resistance is much greater than reactance,
inductance and capacitance become important. Its best to the circuit looks like a capacitive divider.
keep both as low as possible. One way to do that is to use A special category of passive probes provides much
an attenuator probe, which helps to reduce circuit load- higher bandwidths than typical passive probes. They are
ing. Instead of loading the circuit with capacitance from referred to as Zo probes. These probes are designed for
the probe tip plus the cable plus the test equipments own use in 50- environmentstypically, high-speed device
input, the 10 attenuator probe introduces about 10 times characterization, microwave communication, and time
less capacitance, as little as I0 to 14 picofarads (pF). The domain reflectometry (TDR). A typical 50- probe for such
penalty is the reduction in signal amplitude from the 10:1 applications has a bandwidth of several GHz and a rise
attenuation. time of 100 picoseconds or faster.
Another challenge when measuring high frequencies Active probes rely on active components, such as
is that the probes impedance varies with frequency. The a field-effect transistor (FET), for their operation. These
probes datasheet will contain a graphic to show this probes use a high-impedance high-frequency amplifier
change. Further, all probes and the connected test equip- mounted in the probe head, which serves as isolation or
ment should be well-grounded with as short a ground clip buffering between the circuit under test, the cable, and
as possible. This will reduce the effect of any grounding the test equipment. Active probes load the circuit with a
loops. low capacitance and high DC resistancefor example, a
In addition to loading the circuits, probes also impact capacitance of 1 picofarad or less in parallel with 1 M
the measureable signal bandwidth. Just like test equip- resistance. Such ultra-low capacitance has several desir-
ment, probes have a bandwidth limitation. The goal is to able effects.
stay within 3 db (or 0.707) of the original signals ampli- Since capacitive reactance is the primary input im-
tude value. The key is to consider the overall bandwidth of pedance element of a voltage probe, a low C results in a
both the probe and the test equipment to which it is con- high-input impedance over a broad band of frequencies.
nected. To get the full bandwidth from your test equipment, As a result, active probes will typically have specified
you need more bandwidth from the probe, or else to use a bandwidths ranging from 500 MHz to several GHz. The
probe designed for that instrument. Always use the probes most important advantage of active probes is the lower

VISIT: for more T&M content | 40


Inner conductor Coax cable types have expanded over the years from just
rigid, semi-rigid, and flexible coaxial varieties. Present-day
versions embody some characteristics of both flexible
Outer conductor and semi-rigid cables. Formable cables, for instance,
boast performance levels approaching those of semi-rigid
cables in terms of attenuation and phase stability, yet can
be formed into required shapes and forms without special
Inner conductor Insulation Outer conductor tools.
Two-conductor RF/microwave coax cables consist of
9-3. Two-conductor RF/microwave coax cables consist inner and outer conductors separated by some form of
Figure 9-3
of inner and outer conductors separated by some form dielectric material (Fig. 9-3) and often surrounded by an
of dielectric material. outer jacket. The outer conductor is typically some form
of braided metal shield. Since loss through the cable will
circuit loading. This means they can be used on high- come from a combination of conductor and dielectric loss-
impedance circuits that would be seriously loaded by pas- es, cable manufacturers usually select low-loss materials
sive probes. like copper for the conductors.
Active probes dont have the voltage range of pas- For some applications, such as test and measurement,
sive probes. The linear dynamic range of active probes is more flexible cable solutions are required because of the
generally anywhere from 0.6 V to 10 V, with a maximum number of flexures that take place during normal testing
voltage as low as 40 V (DC + peak AC). This means that operations. For this reason, companies providing cables
you cant measure from millivolts to tens of volts like you and cable assemblies aimed at test applications typically
can with a passive probe. Thats why active probes can be stress the durability of their connectors along with the elec-
more easily damaged than passive ones. trical performance.
Both passive and active probes are usually referenced Most RF measurements are accomplished with coaxial
to ground. Conversely, a differential probe measures cable. The following is a brief review of the more common
signals that are referenced to each other. To maximize the connector types:
common-mode rejection ratio (CMRR), differential probes APC-7 (7 mm) connector. The APC-7 (Amphenol Preci-
must provide two signal paths that are as nearly identical sion Connector-7 mm) offers the lowest reflection coefficient
as possible, matched in overall attenuation, frequency and most repeatable measurement for 18 GHz connectors.
response, and time delay. SMA connector. The SMA (Subminiature A) is one of
the most commonly used RF/microwave connectors. It is
Probe Tips intended for use on semi-rigid cables and in components
Weve covered the basic types of probes. Now lets that are connected infrequently. Most SMA connectors have
consider different probe tips. Most probes allow a variety of higher reflection coefficients than other connectors avail-
tips to be installed. The most common is a small, pointed able for use to 24 GHz because of the difficulty in securing
tip. To measure the leads on integrated circuits (ICs), a the dielectric support.
small plastic insulating foot with indentation is used. This BNC connector. The BNC (Bayonet Navy Connector)
makes it easier to probe very-fine-pitch integrated cir- was designed for military use and has gained wide accep-
cuits. The indentations align with the pitch of the IC leads, tance in video and RF applications to 2 GHz. Above 4 GHz,
stabilizing the probe and helping to maintain contact on the the slots may radiate signals. Both 50- and 75- versions
desired pin. are available. A threaded version (TNC) helps resolve leak-
Measuring todays monolithic microwave integrated age for common applications up to 12 GHz.
circuit (MMIC) devices on a wafer requires a different type SMC connector. The SMC (Subminiature C) is much
of RF probe tip, one that can measure up to 110 GHz. This smaller than an SMA connector, making it suitable for some
is typically done with a three-tipped ground-signal-ground applications with size constraints. It is often used up to 7
RF probe. Some probe tips for GHz signals are designed to GHz where low leakage and few connections are required.
be soldered to the test point. N-type connector. This larger connector is used on
some types of test equipment and can be used at frequen-
Cable Types cies up to about 18 GHz. It can handle higher power or
Cables and connectors for RF/microwave applications several hundred watts at gigahertz frequencies.
are designed for a characteristic impedance of 50 . To mitigate transmission losses, be sure to specify high

VISIT: for more T&M content | 41


frequency probes and cables. BNC connectors (and 50-

coaxial cable) are fine for signals with rise times greater
than 1 ns. However, for signals with sub-nanosecond rise
times, it is better to use cables terminated in SMA con-
nectors. SMA connectors are better shielded and maintain
impedance better in the microwave region than BNCs.
Additionally, cables should have a 50- impedance 9-4. Commercial terminators and attenuators are mostly
with a solid shield, a foam dielectric, high power handling, used with high power levels.
and much lower loss than ordinary coax cables. Every
cable requires a connector that fits it mechanically, and it or absorb power, respectively. In many ways, the two types
must maintain essentially constant impedance through the of components are similar, since they are both designed
connector. to stop RF/microwave power. Attenuators decrease some
Also, avoid mixing signal path hardware types via portion of the power, in fixed or variable amounts, while
adapters (e.g., BNC/SMA). Adapters introduce significant terminations absorb power applied to them.
parasitics, resulting in reflections, rise-time degradation, Both types of components are available in various
resonances, and other degrading behavior. forms, from miniature chips to higher-power coaxial com-
Fast linear signals for timing measurements typical- ponents, and the highest-power waveguide assemblies
ly have rise times of less than a few nanoseconds and with heat sinks (Fig. 9-4). Both types of components play
durations of less than 1 s. Because of the fast rise time, important roles in high-frequency circuits and systems,
interconnections between modules are always made with a especially when high-power signals must be managed.
50- coaxial cable, and the cable is always terminated with Attenuators and, in particular, high-power terminations
a 50- load at the receiving end. Modules intended for use are usually specified with size, weight, power-handling
with these signals normally have a 50- input impedance. capability, and frequency range as essential parameters for
Connector parasitic, especially capacitance, must be comparison. Power-handling capability is generally a func-
kept as low as possible. If probes must be used, their intro- tion of size, with the highest-power components occupying
duction to the sign path mandates attention to their connec- the greatest amount of volume in a design. The key specifi-
tion mechanism and high frequency compensation. Passive cations of an attenuator are its attenuation in dB, its pow-
Z0 types, commercial available in 500- (10) and 5-k er-handling capability, and its impedance (usually 50-).
(100) impedances, have input capacitance below 1pf. An isolator is a three-port passive electronic device that
Any such probe must be carefully frequency compensated helps direct the flow of microwave signals in RF equipment
before use or misrepresented measurement will result. and systems. A common use of an isolator is shown in Fig.
9-4. The isolator is connected between a signal generator
Attenuators, Isolators, and Couplers and some device under test (DUT). If all impedances are
Attenuators, isolators, and couplers are the oft-forgotten matched, the signal passes freely to the DUT. If there is a
part of the RF design. We shall consider each separately. mismatch at the DUT or if the DUT is disconnected, it cre-
Attenuators and terminations are commonly used ates a high-voltage standing wave ratio (VSWR), causing
components in high-frequency systems, used to adjust or a large reflected signal. The circulator absorbs this signal,
absorb power, respectively. In many ways, the two types protecting the usually expensive signal generator.
of components are similar, since they are both designed to The attenuation of an isolator in the reverse direction is
reduce RF/microwave power. Attenuators decrease some typically in the 20-dB range. If greater attenuation is need-
portion of the power, in fixed or variable ed, two isolators can be cascaded.
amounts, while terminations stop and The result is a four-port device that can
absorb power applied to them altogeth- boost attenuation to about 40 dB or so.
er. Attenuators are used to reduce the Such four-port units are available as a
power level to avoid damage to test single product rather than two individual
instruments. Terminators are used to isolators.
terminate a transmission line to prevent Couplers and adapters are nec-
reflections. essary links between components in
Attenuators and terminations 9-5. This isolator is connected to the test and operation of RF systems.
are commonly used components in protect a signal generator in a test These devices provide a way to modify
high-frequency systems, used to adjust setup. connections between test instruments,

VISIT: for more T&M content | 42


cables, probes, and DUTs. For example, a coupler may

provide a way to attached one cable to another while an
adapter may convert from one connector type to another.
As these components are add-ins to the RF signal stream,
they produce distortion, attenuation, and delay.
Minimizing these performance degraders is a high
priority for designers of adapters and couplers, who are
preparing their product lines to meet future demands for
higher-frequency and wider-band operation. The materials
used, quality of construction, finishing, size, and connec-
tor type all have significant impact on the performance
of adapters and couplers. For adapters in particular, they
impact frequency range, voltage-standing-wave ratio NI RF and wireless test solutions span
(VSWR), maximum power, and passive intermodulation applications from mobile device testing to
distortion (PIM). Couplers are affected in terms of frequen- high-performance measurements. NI systems
cy range, VSWR, PIM, coupling, accuracy, insertion loss, provide the fastest test times and simplify
directivity, and maximum power. n measurement challenges while offering
industry-leading performance.


VISIT: for more T&M content | 43


LOU FRENZEL, Contributing Editor


lectromagnetic interference (EMI) and elec- EMI Regulations
tromagnetic compatibility (EMC) testing have Governments regulate EMI. In the United States, the
become an essential part of the design of new regulatory body is the Federal Communications Commis-
and updated microwave and millimeter-wave sion (FCC). The EMI regulations are put forth in the Code
equipment. Such testing is necessary not only of Federal Regulations (CFR) Title 47, Parts 15 and 18. All
to avoid interference to and from other electrical engineers should have a copy (which can be obtained at
or electronic devices, but also to meet the strict rules of the
federal regulatory agencies. Another U.S. source of EMI regulations is the Depart-
Passing EMI and EMC tests is mandatory to ensure that ment of Defense (DoD); these regulations are designated
a product is legally certified for sales. RF and other engi- MIL-STD 461E. The American National Standards Institute
neers working on wireless products must be familiar with (ANSI) also has EMI standards C63.2.
the related rules and test procedures. This chapter pro- Other countries have their own regulatory bodies and
vides a general introduction to EMI and EMC definitions, standards institutes. One major source of EMI standards is
sources, equipment, and test processes. the Comite International Special des Pertubations Radi-
oelectriques (CISPR), an organization that is part of the
Defining EMI International Electrotechnical Commission (IEC). Each
EMI is any electrical or electronic signal that interferes country has its own ministry or regulatory body.
with a communications device, usually a receiver. EMI can
block reception completely, cause intermittent operation, Types and Sources of EMI
reduce overall reliability, or inflict some other disruptive There are two types of EMI: conducted and radiated.
malady. EMI used to be called radio frequency interfer- Conducted EMI is a signal that is transmitted over the AC
ence (RFI) and television interference (TVI) during the power line or other cable. Spurious signals originating in
earlier days of radio and TV. one piece of equipment can travel back through the power
EMC is a term that means that a piece of equipment will supply and ac line to a victim receiver. Any connecting
not generate EMI that will affect other equipment nearby, cables, such as unshielded twisted pair, can also carry
nor is it susceptible to EMI from other equipment in the EMI.
same environment. The goal of every wireless or other Radiated EMI comes from intentional radiators like
electronic device is to achieve this nirvana state. radios and unintentional radiators such as computers. Any
The presence of EMI and its detrimental effects have wireless device can cause interference to some other vic-
increased dramatically over the years due to the mas- tim. This is less likely if it is properly designed and tested
sive growth in computers and wireless products like cell to certify that it meets regulations. Yet, most radios have
phones, Wi-Fi, Bluetooth, and other radio sources. As the circuits that may radiate on frequencies other than the
Internet of Things (IoT) and machine-to-machine (M2M) operational frequency. Local oscillators in receivers and
products grow in volume, the EMI problem is expected to frequency synthesizers are examples. It is the unintentional
worsen. The solution is to design and build equipment that radiators that cause the most problems.
is EMI-free. Standards and regulations have been put in The ac power line is a major source of EMI caused by
place to define acceptable levels of EMI, along with how switching transients, motors, fluorescent lights, and CFLs.
devices are tested for EMI emissions or susceptibility. Automobile ignitions and other electrical transient-pro-

VISIT: for more T&M content | 44


ducing equipment are 10-1. Displayed Log periodic

other sources. Any device is a generic EMI Far field antenna
producing an electrostatic radiation testing distance
Coax Coax
discharge (ESD) is an EMI configuration DUT
source. conducted in an Spectrum
Optional analyzer
Any piece of digital isolation chamber. preamp
equipment with a high- Non-
frequency digital clock is a stand
major source. PCs and lap-
tops with processors operat-
ing at rates from 1 to 4 GHz
are major EMI generators.
Cell phones and tablets also contribute to the pollution. to run the related software and its screen serving as the
Then there are the switching power supplies, regulators, analyzer display. EMI receivers are also used for remote
dc-dc converters, and invertersall of which are produc- spectrum monitoring.
ers of harsh EMI. Inductive and capacitive coupling from The key specifications for a spectrum analyzer for EMI
cables can be yet another source. hunting are controllable resolution bandwidth (RBW), an
upper frequency limit three to five times the operation-
EMI Testing al test frequency, peak detection, and a spectragraph
Testing of all intended and unintended radiators is display. The spectragraph display is a frequency domain
mandatory before they can be legally sold. Known as display that occurs over time. On the display, the horizontal
compliance testing, this process incorporates all mandato- axis is frequency as usual, but the vertical axis is a mov-
ry measurements as dictated by the rules and regulations. ing time display that flows from top to bottom. This is also
Compliance testing requires expensive test instruments referred to as a waterfall display.
and facilities that few product development companies Signal strength is shown in colors, with red the highest
have or can afford. Therefore, most compliance testing is and blue to black the lowest. A persistence display that
conducted by professional companies specializing in this retains scans over time may be useful in finding some
process. It is an expensive and time-consuming process, intermittent interfering signals. Spectrum storage units can
but must be factored into the development of any wireless also be acquired to monitor and store spectrum data over
product or device that must meet radiation regulations. time so that non-continuous signals may be identified.
The goal of any design is to get it right the first time An antenna is another EMI testing device. Available
so that compliance testing does not have to be repeated. for different frequency ranges, it becomes the input to the
That is why pre-compliance testingdone by the design analyzer. Since antennas are usually resonant, they must
engineer or an internal testing departmentbecomes be made more broadband to accommodate a wide range
more important. This process can detect and correct of frequencies. This is done by enlarging dipole elements
potential problems before final compliance tests are per- with larger tubing or using large-area, conical-shaped
formed, saving significant time and money. elements. These are used mainly at frequencies below
300 MHz. Above that, a log periodic antenna is used. It is
EMI Testing Equipment smaller but covers a wide frequency span up to about 10
The primary test instrument is the spectrum analyzer GHz. At higher frequencies, a horn antenna is generally
or VSA (see Chapters 1 and 5). The oscilloscope can be used.
helpful in some cases, but the spectrum analyzer remains Also available are special probes that are used to sniff
the most useful piece of equipment. The objective is to go out EMI sources. Probes for both electric and magnetic
to the point where EMI was experienced and attempt to field detection allow you to pinpoint RF leakage for enclo-
experience it again, but using the spectrum analyzer. This sure seams, unintended PC board ground loops, or cable/
may be inconvenient as spectrum analyzers are bench connector faults.
instruments. A device called a linear impedance stabilization net-
To solve this problem, test equipment vendors now offer work (LISN) is needed for conducted EMI measurements.
portable handheld battery-powered analyzers to use in the The LISN is placed in the AC line between the AC input
field. Also available are EMI receivers. These are usually and the device under test (DUT). It isolates the DUT from
housed in a small enclosure, with a laptop being used line noise and transients and couples the signals from the

VISIT: for more T&M content | 45


Spectrum 10-2. This figure these basic solutions is the secret

analyzer shows a conducted to eliminating or controlling EMI so
EMI test setup. that you can meet the compliance
Grounding. Ground is that
common conductor that forms
AC the return path for most signals.
line The secret to a good RF ground is
AC LISN DUT ultra-low impedance. This means
filter a very large copper area on PC
boards where multiple connections
can be made. In some designs, a
DUT to the spectrum analyzer. metallic chassis or enclosure is used as the ground. Low
For compliance testing, a special shielded room known resistance and inductance of the ground is essential to
as an anechoic chamber is usually needed to prevent stray prevent ground loops that become effective EMI radiators.
signals from external sources from interfering with specific Another good design rule is to keep all component
tests. These rooms are very expensive, and usually only leads to ground as short as possible. As well, you should
commercial compliance testing companies have them. always use separate grounds for digital, analog, and I/O
interface circuits. In PC board design, using a multilayer
EMI Testing Procedure arrangement is a proven way to reduce EMI.
For intentional radiation tests the configuration shown in Separate power and ground layers offer good low
Fig. 10-1 is used. The DUT is placed on a non-metallic ta- impedance returns and inherent shielding. And separate
ble or stand. An antenna mounted on a tripod is placed at layers can isolate offending signals such as fast clocks.
a distance from the DUT such that the antenna is in the far Be sure to keep all clock lines as short as possible. Careful
field of the signal. The far field is the real EM signal, and it layout of the PCB is critical at RF, as some designs can
begins roughly 10 wavelengths from the radiating device. create long runs or loops that radiate.
Closer to the DUT is the near field that is mainly the Shielding. Shielding is the process of containing any
individual electric or magnetic fields. The antenna is then radiation or inductive or capacitive coupling by using a
connected to the spectrum analyzer. In some tests a metallic barrier. This means placing the offending circuits
preamplifier may be needed. Keep in mind that this test inside a metallic enclosure, or else wrapping a small
is conducted in an isolated chamber so that unwanted metallic enclosure around a component or circuit. Shields
signals are not picked up. may be steel, copper, aluminum, or even plastic coated
Once the radiated far field test is complete, the DUT with a metallic material. It must be thick enough to contain
can be further explored with the electric and magnetic field the radiation.
probes. These probes can detect near field leakage from Be careful to ensure that any seams in a multipart en-
enclosures and various mechanical ports. closure fit perfectly to seal in the radiation. Seams leakage
A conducted EMI test is shown in Fig. 10-2. The DUT is a major source of EMI. Special copper sealing gaskets
is placed on a metallic ground plane and connected to can be used to prevent such leakage.
the LISN. An additional ac line filter is optional to suppress Filtering. This is the most obvious way to get rid of
external noise and transients. The LISN output passes the offending EMI signals. For example, a lowpass filter on the
DUT EMI, if any, to the spectrum analyzer. A preamplifier output of a transmitter can greatly reduce any unwanted
may be needed. harmonics. Good bypassing of the dc power lines also
Remember that the procedures described here are minimizes EMI effectively. Use series inductors and shunt
general. Testing standards usually prescribe detailed bypass capacitors on all power lines. Input filters on a
physical implementations and electronic test equipment receiver can also effectively reduce some EMI.
and procedures. One mandatory filter in most designs is an ac line filter.
This is a lowpass filter that eliminates signals from being
Minimizing EMI conducted into the power supply of the equipment. This
There are three fundamental ways that EMI can be elim- filter also effectively eliminates any high-frequency EMI
inated or reduced to regulatory levels: grounding, shield- generated within the equipment from being passed along
ing, and filtering. A good design incorporating one or all of to the ac line. Such a line filter is mandated in most equip-

VISIT: for more T&M content | 46


ment today.
One not-so-obvious filter is a series resistor in line with
clock signals. This forms a lowpass filter along with any
stray shunt capacitance. Such a filter lengthens pulse rise-
and-fall times, and therefore decreases the level of har-
monic production. Another unsuspected filter is one that
reduces the level of common mode signals that affect a
cable. These signals can come from capacitive or induc-
tive coupling. A common practice is to wrap several turns
of the cable around a ferrite toroid or rod. This type of filter
is widely used on video cabling. n

NI RF and wireless test solutions span

BACK TO TABLE OF CONTENTS applications from mobile device testing to high-
performance measurements. NI systems provide
the fastest test times and simplify measurement
challenges while offering industry-leading

VISIT: for more T&M content | 47



The old approach to automated test isnt

scaling, but you already knew that. Look at your
balance sheet. To test smart devices, you need
a smarter test system built on a platform of NI
PXI, LabVIEW, and TestStand. More than 35,000
companies deploy NI technology to lower their
cost of testwhat are you waiting for?

Prepare for the future at

NI PXI, LabVIEW, and TestStand

2016 National Instruments. All rights reserved. LabVIEW, National Instruments, NI,, and NI TestStand are trademarks of National Instruments. Other product and company names listed are trademarks or trade names of their respective companies. 27634