Sie sind auf Seite 1von 32

Chapter 17

Receiver Tests and Measurements

In this chapter, we look at some tests and duce output over a very wide range of fre-
measurements used on radio receivers. But quencies. Add a modulator stage to these in-
first, we look at signal generators. These de- struments and a signal generator is created.
vices make the signals used in testing of ra- Figure 17.1 shows a typical commercial sig-
dio receivers. nal generator.

Grades of Instrument
SIGNAL SOURCES AND SIGNAL Signal generators and sources come in sev-
GENERATORS eral grades. Which to select depends on the
use. A service-grade instrument is used for
Signal generators and signal sources are in- troubleshooting common broadcast band re-
struments that generate controlled signals for ceivers. Such instruments often lack a cali-
use in testing and measurement. A distinction brated or metered output level control, and
is made by some people between signal the frequency accuracy usually is low. More
sources and signal generators. The former important for many sensitive measurements,
produce continuous wave (CW) output sig- at low output levels more signal will escape
nals without modulation, while the latter will around the flanges than comes through the
produce one or more forms of modulated output connector. Such instruments are use-
signal (AM, FM, SSB, PM) in addition to CW ful for simple troubleshooting but useless for
output. In many cases, however, the words accurate measurements.
source and generator are interchangeable in Laboratory-grade signal sources and
popular usage. generators are very high-quality instruments
Some signal sources produce a single- with accurate frequency readout and output
output frequency (or a discrete number of level controls. These instruments are used in
fixed output frequencies). These instruments making lab measurements of receivers and
sometimes are used for testing channelized other devices where high precision and ac-
receiver systems. Other signal sources pro- curacy is desired.


Fig. 17.1 Typical signal generators: (A) Digital readout synthesizer; (B) Analog readout signal generators;
(C) Sweep generator; (D) Service-grade signal generators.

High-quality service-grade instruments Output Level

fall somewhere between the two previous
The output of a calibrated signal generator is
grades. Several mainline manufacturers of
usually expressed in either microvolts (µV)
high-quality laboratory signal sources and
or dBm (decibels relative to 1 mW in 50 Ω)
generators also manufacture “economy” lines
or both µV and dBm. It is useful to have a
that fit this category. They are of consider-
feel for both forms of output level indication.
ably higher grade than the simple service in-
One microvolt (1 µV) is 10–6 V, so when ap-
struments but are not up to the lab grade.
plied across a 50 Ω resistive load, it produces
They often are used for troubleshooting
a power level of
high-quality telecommunications and land-
mobile systems, where the highly accurate
and precise measurements are not needed. V2/R = (10–6 V)2/(50 Ω) = 2 × 10–14 W (17.1)
Receiver Tests and Measurements 241

The 0 dBm reference level is one milli- units, then the required unit can be calcu-
watt (1 mW) dissipated in a 50 Ω resistive lated using these methods.
load. This represents an applied voltage of To find the output power level in watts
or milliwatts from dBm is similarly simple:
V = PR = (0.001 W )(50 Ω ) = 0.2236 V (17.2)
P = 10dBm / 10 mW (17.3)
If the output level set dial on a particu- To find the level in watts divide equa-
lar instrument is not calibrated in the correct tion 17.3 by 1000.


It would be nice if all signal sources and sig-
nal generators were ideal; that is, the output
frequency and output level were noiseless
and perfectly calibrated. This never occurs,
although the differences in these specifica-

tions involve a principal difference between
high- and low-quality instruments.


The important considerations regarding fre-

quency are the range, resolution, accuracy,
and (in automatic test equipment applica-
tions) the switching speed.
Range FSET

The frequency range is a specification of the
frequencies covered. In some cases, only one Fig. 17.2 Frequency response with an uncer-
or some small number of discrete frequen- tainty band.
cies are covered. In other cases, one or more
bands of frequencies are provided.

Resolution been six months (0.5 yr) since the last

Resolution is the statement of the smallest in-
crement of frequency that can be set. On Accuracy = ± FSet × τAging × τCal
analog instruments that have no counter, the
= ± (480 MHz) × (0.155 ppm/yr)
resolution is poor. The resolution may (but
× (0.5 yr)
not certainly) be improved by adding a digi-
tal frequency counter to measure the output = ± 37.2 Hz
frequency. On modern synthesizers, it is pos-
sible to set frequency with extremely good Also, some random variation may enter
resolution. the output frequency. Figure 17.2 shows the
uncertainty band around the set frequency.
Accuracy The actual output frequency, Fo, will be FSet ±
Accuracy. The general practice is to calibrate
Accuracy specifies to how nearly the set fre- a signal generator on a six-month or annual
quency matches the actual output frequency. schedule, depending on the use.
Accuracy is a function of the set frequency
(and how closely it can be set), FSet, long-
term aging (τAging), and the time since last cal- Switching Speed (Settling Time)
ibration (τCal). Mathematically, The switching speed is the length of time,
Accuracy = ± FSet × τAging × τCal (17.4) usually in milliseconds or microseconds, re-
quired for a synthesized signal source or
EXAMPLE generator to move to a new frequency when
A signal generator is set to 480 MHz and digitally commanded to change. It is calcu-
has an aging rate of 0.155 ppm/yr. It has lated as the length of time for the error of the
Receiver Tests and Measurements 243



Fig. 17.3 F
Amplitude response with an uncer- FO

tainty band. FREQUENCY

frequency or output level (or both) com- SPECTRAL PURITY

manded by the change to come into specifi-
cation range. The output signal is not always clean.
Although the purity of the output signal is a
distinguishing factor that differentiates lower-
Output Level
and higher-quality generators, all generators
The output level can be expressed in voltage, produce signals other than the one desired.
power, or dBm notation. All are equivalent, Figure 17.4A shows a typical spectrum out-
although one or another will be preferred in put that might be seen on a spectrum ana-
most cases. The most common method of lyzer. The main signal is a CW sine wave, so
describing the output level is dBm. A typical ideally we would expect only a single spike,
signal generator or CW signal source will with a height proportional to the output
produce output levels from −136 dBm to +10 level. But a lot of other signals are present.
dBm (some go to higher levels), with an out- First, note that the main signal is spread
put level accuracy of ±0.50 dBm and a reso- out by phase noise, random variation around
lution of 0.01–0.02 dB. the main frequency. When integrated over a
As with frequency, certain factors affect specified bandwidth, such as 300–3000 Hz,
the accuracy of the actual output vs. the set the phase noise is called residual FM (Figure
output. Figure 17.3 shows a zone of uncer- 17.4B).
tainty around the output level set. Any given Second, harmonics are present in Figure
setting has PMin and PMax values. For example, 17.4A. If the main signal has a frequency of F,
if the only error is the accuracy discussed the harmonics have frequencies of nF, where
previously (e.g., 0.50 dB), a level set of, say, n is an integer. For example, the second har-
−10 dBm will produce an actual output monic is 2F, and the third harmonic is 3F. In
power level of −9.5 to −10.5 dBm. many cases, the 3F harmonic is stronger than


~ 30 dB
~ -45 dB
~ -65 dB



nF/2 F FOther nF



-3000 -300 +300 +3000

Fig. 17.4 Output signal noise: (A) spectrum output; (B) residual FM.

the 2F harmonic, although in general the nal frequency, nF/2 represents the subhar-
higher-number harmonics are weaker than monics. Typically, unless something inter-
lower-number harmonics. feres with the output signal, subharmonics
Sometimes, there are subharmonics as are not as prominent. One thing that makes
well. These are integer quotients of the subharmonics prominent, however, is the
main signal. Again, if the F is the main sig- use of frequency multiplier or divider
Receiver Tests and Measurements 245

stages (which is the case in many modern signal generators. Figure 17.5 shows a simple
generators). analog architecture that once was common
Finally, miscellaneous spurious signals on even high-grade instruments and still is
(spurs) are found on some generators. These common on service-grade instruments.
might be from power supply ripple modulat- The signal is generated in an L-C con-
ing the output signal, parasitic oscillations, trolled variable frequency oscillator. The VFO
digital noise from counter or phase-locked typically has a band switch for selecting dif-
loop circuits, or other sources. ferent frequency ranges. A calibrated tuning
Harmonics and spurs usually are mea- dial gives the user an approximate idea of the
sured in terms of decibels below the carrier output frequency. However, because of drift
(dBc), where the carrier is the amplitude of and the mechanical aspects of calibrating the
the main output signal. In general, the fewer dial, these dials are not terribly accurate.
are the unwanted components, the better the Some instruments have an output am-
signal source. plifier, although for many decades even qual-
Phase noise warrants some special con- ity signal generators lacked power amplifiers.
sideration. It usually is measured in terms of The output of the VFO was fed directly to the
dBc/Hz; that is, decibels below the carrier output level control.
per hertz of bandwidth. This noise is concen- Service-grade generators of this archi-
trated around the main signal frequency and tecture usually have a crude form of output
normally graphed on a log-log scale to per- level control. Higher-quality instruments, on
mit compressing both close-in and further- the other hand, have some variant of the out-
out noise components on one graph. put circuit shown in Figure 17.5. High-level
output sometimes is provided to permit the
user to route the signal to a frequency
ARCHITECTURES counter so that an accurate determination of
frequency can be achieved.
Despite many different configurations, with Two attenuators are in the output level-
different “block diagram” representations, setting circuit. A coarse attenuator is used to set
only a few architectures are used in designing a relative output meter to some calibrated







Fig. 17.5 Simple analog signal generator.










Fig. 17.6 Frequency synthesizer.

point. In most cases, the meter would be cali- rate, stable fixed-frequency source such as a
brated with 0 in the center of the analog scale. crystal oscillator. The frequency of the refer-
The coarse attenuator is adjusted to center the ence section must be precisely adjustable
meter pointer over the 0 point in the center of over a small range so that it can be compared
the meter. When this is done, the settings of the to a higher-order standard, such as a cesium
fine output attenuator are valid. beam oscillator or WWVB comparator re-
A synthesizer architecture is shown in ceiver, for purposes of calibration.
Figure 17.6. This type of signal generator is Because it controls the frequency syn-
more modern and can produce very accurate, thesizer, the stability of the reference section
high-quality signals. This signal source has determines the overall stability of the signal
three main sections: reference section, fre- generator. The stability of ordinary crystal os-
quency synthesizer, and output section. Figure cillators is reasonably good for many pur-
17.6 breaks down each section into further poses but not for use in the reference section
components for easy analysis. of a signal source. For that purpose, either
temperature-compensated crystal oscillators
(TCXO) or oven-controlled crystal oscillators
Reference Section
(OCXO) are used. The TCXO typically ex-
The reference section is at the very core of hibits crystal aging of better than ±2 ppm/yr
the signal generation process. It is an accu- and temperature aging of ±1 ppm/yr. The
Receiver Tests and Measurements 247

OCXO is capable of 0.1 ppm/yr for aging Suppose, for example, a signal source
and 0.01 ppm/yr for temperature. has a reference of 5 MHz and is divided by
The crystal oscillator usually is oper- 20 to produce a 250 kHz reference. If the
ated at a frequency such as 5 MHz, but frequency synthesizer divide-by-N stage is
lower frequencies often are needed. To gen- set for, say, N = 511, then the VCO output
erate these lower frequencies, a divide-by-N frequency will be 0.25 MHz × 511 = 127.75
digital counter is provided. This circuit divides MHz. Band switching, operating frequency,
the output frequency by some integer, N, to and frequency resolution are controlled by
produce a much lower reference frequency. manipulating the reference frequency and
Many signal generators provide ref. out VCO divide-by-N settings. In some cases,
and ext. ref. in capability (these connections the frequency is entered by keypad, and
often located on the rear panel of signal gen- this tells the signal generator the values to
erators, so may be overlooked). The ref. out set. Alternatively, “tunable” signal genera-
connector provides the reference signal to tors may have a digital encoder shaft con-
other instruments or is used for calibration nected to a front panel control.
purposes. The ext. ref. in allows the use of an The noise component of the output sig-
external reference source in place of the inter- nal is composed of thermal noise and phase
nal reference section. This sometimes is done noise. Of these two, the phase noise tends
to lock two signal generators that must work to dominate the performance of the signal
together or to substitute a much more accu- source. Noise from both the reference oscilla-
rate reference such as a cesium beam clock. tor and VCO phase contributes to the overall
output phase noise. There is a 20 log (N)
degradation of the phase noise performance of
Frequency Synthesizer
the signal source because of the divide-by-N
The actual signal is produced in the frequency nature of the PLL. Fortunately, the bandwidth
synthesizer section, which is generated by a of the PLL tends to limit the contribution of the
voltage controlled oscillator whose output is VCO phase noise to the overall phase noise
compared to the reference signal. Voltage performance.
variable capacitance diode (varactors) can be
used for the VCO, as can surface acoustical
Output Section
wave oscillators (which are used at higher fre-
quencies and in the microwave bands). The output section performs three basic
The frequency of the VCO is set by a functions: It boosts power output to a speci-
DC control voltage applied to its tuning input fied maximum level, it provides precision
line. This control voltage is generated by in- control over the actual output level, and it
tegrating the output of a phase detector or keeps the output level constant as frequency
phase comparator that receives the reference is changed.
frequency and a divide-by-N version of the The power amplifier is a wideband am-
VCO frequency as inputs. When the two fre- plifier that produces an output level of some
quencies are equal, the output of the phase value in excess of the required maximum
detector is 0, so the VCO tuning voltage is at output level (e.g., +13 dBm). A calibrated
some quiescent value. precision attenuator then can be used to set
But, if the frequency of the VCO drifts, the actual output level to any lower value re-
the phase detector output becomes nonzero. quired (e.g., −136 dBm to +13 dBm).
The integrator output ramps up in the direc- The accuracy of the output power set-
tion that will cancel the frequency drift of the ting depends on keeping constant the RF
VCO. The VCO frequency is continuously power applied to the attenuator input, even
held in check by corrections from the inte- though oscillators (including VCOs) tend to
grated output of the phase detector. This exhibit output signal amplitude changes as
type of circuit is called a phase-locked loop. frequency is changed. In older, manual signal

sources, the RF voltmeter at the input of the diagram of the output section of a signal gen-
attenuator had a 0 center and the output erator that offers amplitude modulation. The
coarse attenuator level could be set manually depth of modulation for most signal genera-
to 0, making the calibration of the fine atten- tors is adjustable over at least 0–30%, while
uator meaningful. Modern signal sources, some offer 100% AM.
however, use an automatic level control cir- The amplitude modulator is called a
cuit to accomplish this job. burst modulator in some signal generators
The automatic level control modulator and placed between the ALC modulator and
essentially is an amplitude modulator con- the output power amplifier stage. The input
trolled by a DC voltage developed by rectify- signal used to modulate the RF carrier can be
ing and filtering a sample of the RF output derived from an external source or provided
level. The ALC driver compares the actual by an internal function generator that pro-
output level with a preset value and adjusts duces at least sine waves. The function gen-
the control signal to the ALC modulator in a erator also might produce square waves,
direction that cancels the error. sawtooth waves, triangle waves, or pulses,
depending on the nature of the signal gener-
ator. When the CW mode is desired, the
MODULATORS modulating source is turned off.
Note that two modulators are in Figure
A signal generator is said to differ from a sig- 17.7: ALC modulator and burst modulator.
nal source in that it will provide modulation The ALC modulator is designed to keep the
of the CW signal. Although I believe this is a output RF level constant, while the burst
distinction without a practical difference, it is modulator wants to vary the output RF level
nonetheless a commonly held usage. in accordance with the applied audio signal.
These two circuits, therefore, are in conflict
with each other. The ALC hold-off circuit is
Amplitude Modulation used to moderate this conflict.
Amplitude modulation (AM) conveys intelli-
gence over a radio carrier by varying the Frequency and Phase Modulation
amplitude in accordance with the applied
audio. The sine wave RF carrier is described There are two basic forms of angular modu-
by a(t) = Ac sin (2πfct), where a(t) is the in- lation: frequency modulation (FM) and phase
stantaneous amplitude, Ac is the peak ampli- modulation (PM). In FM, the carrier fre-
fier, fc is the carrier frequency, and t is time. quency is varied in accordance with the
When another signal is used to vary the am- modulating frequency; while in PM, the car-
plitude of the carrier, amplitude modulation rier frequency remains constant but its phase
results: is varied. Both FM and PM are considered es-
sentially the same. Consider the case of FM.
s(t ) = Ac sin (2πfc t ) × [1 + µ sin (2πfmt )] (17.5) When no modulation is present, the RF car-
rier (fc) remains constant. But, as a sine wave
where audio signal is applied, the carrier frequency
µ is the depth of modulation; will deviate from fc to a lower frequency (F1)
fm is the modulating frequency; on one peak of the modulating frequency
(fm), and to a higher frequency (F2) on the
and all other terms are as previously defined. alternate peak of fm. The deviation is the dif-
You will see the terms linear AM and ference between the carrier and either ex-
logarithmic AM. The former is used when the treme; that is, +Fd = F2 – fc and –Fd = fc – F1.
depth of modulation is expressed as a per- In most cases, signal generators use symmet-
centage and the latter, when it is expressed rical sine wave modulation, so +Fd = –Fd and
in decibels (dB). Figure 17.7 shows a block is called simply Fd.
Receiver Tests and Measurements 249









Fig. 17.7 Output section of a frequency synthesizer.

The peak voltage of the carrier remains sometimes is generated by using a reactance
constant, defined by modulator to vary the phase of the oscillator
V = V (t ) sin [2πfc + βm (t )] (17.6)

In FM, the depth of modulation is ex-

pressed as the modulation index (β), which SWEEP GENERATORS
is defined as the ratio of the deviation to the
modulating frequency, or Fd/fm. The FM Sweep generators (sweepers) are used to
process produces a large number of side- produce a signal that changes frequency over
bands, and at certain values of β, the carrier a specified range. Although the sweeper
will go to 0. The sidebands are described by nominally resembles an FM generator, there
mathematical entities called Bessell functions. are differences. For one thing, the sweep
In phase modulation, the modulation index range usually is quite a bit larger than the de-
(β) is related to the variation in the phase; viation of an FM signal generator. Another
that is, β = ∆φpeak. thing is that the sweeper tends to change fre-
Straight frequency modulation is cre- quency from one limit to the other, then
ated by varying the frequency of the oscilla- snaps back to the first limit.
tor in the frequency synthesizer (see Figure There are three basic forms of sweep:
17.8). Phase modulation, on the other hand, linear (“ramp”) sweep, stepped sweep, and









Fig. 17.8 Frequency modulation and phase modulation in a frequency synthesizer.

list sweep. All these forms use a voltage- converter produces an output voltage pro-
controlled oscillator frequency synthesizer; portional to the binary number applied to its
the difference is in the waveform used for inputs. The maximum output voltage is set
sweeping the output signal. by an internal or external DC reference volt-
Figure 17.9A shows linear sawtooth age and the applied binary number. In this
waveform sweep. At time T1 the ramp ap- case, the DAC input is driven by a binary
plied to the VCO is at 0 and begins ramping counter, which in turn is incremented by a
up. The frequency begins to move upward digital clock (square wave generator).
from 0 V in a linear manner until time T2, at The maximum number of steps that can
which point it snaps back to 0 V. be accommodated using any given binary
Stepped sweep (Figure 17.9B) uses a counter is 2N, where N is the bit length of the
series of discrete voltage steps to change the counter. For the 16-bit counter shown in
frequency of the VCO. This method does not Figure 17.10, therefore, a total of 216 = 65,536
produce an output on every frequency in the different levels (including 0) can be created.
output voltage range from T1 to T2 but only The maximum output voltage is less than the
at specific values determined by the steps. reference voltage by the 1 lsb (least signifi-
The steps are produced in a circuit such as cant bit) voltage, 1/2N. If a 16-bit counter is
shown in Figure 17.10. A digital-to-analog connected to a DAC with a 10.00 V reference
Receiver Tests and Measurements 251

A T1 T2


Fig. 17.9
Waveforms used for
sweeping the output sig- T
T0 T1 T2
nal: (A) sawtoothed;
(B) stepped. B TIME


Fig. 17.10
DAC-binary counter cir- DC DIGITAL-TO-ANALOG
cuit for generating a
stepped waveform.

source, then the 1 lsb step is 0.00015 V and share a common output connector to the an-
the maximum output voltage is 10.00 – tenna. If the signal generator is connected to
0.00015 V = 9.99985 V. the antenna connector to test the receiver, and
Many signal generators are used to test ra- someone accidentally keys the transmitter,
dio communications transceivers. Transceivers then serious damage can result to the signal
contain both transmitter and receiver, which generator. High-quality instruments provide




Fig. 17.11 Open-loop sweeper.




Fig. 17.12 Closed-loop sweeper.

architecture with output reverse power protec- is used to drive the divide-by-N counter in the
tion. This circuitry prevents damage to the sig- synthesizer PLL feedback loop.
nal source if a high RF power level is applied
to its output connector.
List sweepers also have a DAC to con-
trol the VCO but drive it with a series of
values stored in a digital memory similar to
those used in computers. This permits arbi-
Improving Spectral Purity
trary waveforms to be created.
Figure 17.11 shows one way sweepers Certain high-quality measurements are very
are built: open loop. The frequency synthe- sensitive to extraneous signals coming out of
sizer loop is broken and a linear sawtooth, a signal generator. Many signal generators
stepped waveform, or list-generated wave- put out harmonics that are –30 dB down
form is applied to its DC control voltage input. from the main signal (carrier), while other
A closed-loop approach is shown in Figure signals may be either higher or lower than
17.12. The frequency synthesizer loop is not this level. The way to get rid of these extra-
open; rather, a binary counter or list memory neous signals is to place a frequency selec-
Receiver Tests and Measurements 253

tive filter between the output of the signal toriously unpredictable when you terminate
generator and the device under test. them in an impedance other than the design
Figure 17.13 shows the use of a low-pass impedance. Understand the passband re-
filter to eliminate the harmonics and any spurs sponse and the insertion loss of the filter be-
above the main signal. Select a filter with a –3 fore using it.
dB point somewhere between the main signal
and the first extra signal, and an attenuation
Improving Mismatch Loss
slope steep enough to reduce the “bad” signals
as much as possible. If there are any subhar- Mismatch error occurs because the load
monics (or spurs lower than the main signal), and the signal generator are not impedance
then either use a bandpass filter or add a high- matched. In any electronic circuit, the maxi-
pass filter with a –3 dB cutoff between the mum power transfer occurs when the imped-
main signal and the subharmonic. ances are matched. An inherent mismatch
Take care, however. Real filters lack the problem may be in either the signal genera-
nice flat response seen in some textbooks. tor or the load and almost certainly in the ca-
They have passband ripple and some odd re- bles or other devices connected in line with
sponses out of band. Also, L-C filters are no- the signal generator.




Fig. 17.13 Use of a low-pass filter to eliminate harmonics and spurs.




Fig. 17.14 Use of an attenuator to improve range: (A) DUT connected in the normal way; (B) solution to
the problem.

For example, assume we have a signal generator output and the DUT. You have to
generator with a VSWR of 1.9:1, and a device adjust the signal generator output level control
under test with a VSWR of 1.6:1 connected in 10 dB higher than normal to compensate for
the normal way (Figure 17.14A). The mis- the extra attenuation. This works because
match loss can be found once we know the fixed resistive attenuators tend to be designed
reflection coefficients. with very low reflection coefficients. Suppose
The source coefficient is we have the same components in Figure
17.14B as in Figure 17.14A but add an attenua-
SWR1 − 1 1.9 − 1 0.9 tor with ρA = 0.31. The mismatch loss becomes
ρS = = = = 0.31
SWR1 + 1 1.9 + 1 2.5
ML = 20 log [1 + (ρS ρD (ρA)2)]
The device under test is = 20 log [1 + ((0.31)(0.23)(0.31)2)]
SWR1 − 1 1.6 − 1 0.6 = 20 log [1 + 0.0069]
ρD = = = = 0.23
SWR1 + 1 1.6 + 1 2.6
= 20 log [1.0069]
The mismatch loss is = (20)(0.003) = 0.06 dB
ML = 20 log [1 + (ρS ρD)]
= 20 log [1 + (0.31)(0.23)] = 20 log [1.07] Improving Third-Order Intercept
= (20)(0.03) = 0.6 dB
One of the most important specifications for
Figure 17.14B shows the way of dealing an amplifier or radio receiver is the third-
with this problem. Insert a 10 dB fixed attenu- order intercept point (TOIP). This specifica-
ator in line with the line between the signal tion tells us something about the device’s
Receiver Tests and Measurements 255

dynamic performance, especially in the pres- ence level established equal to the minimum
ence of multiple input signals. If you listen to discernible signal or, in some procedures, an
any shortwave receiver, AM or FM BCB re- S1 signal level. The receiver is tuned to the
ceiver, or any scanner receiver, you will real- third-order product frequencies and the atten-
ize that most areas of the country are polluted uator decreased (raising the signal level) until
with too many radio signals. When multiple the same reference level is produced. The
strong signals are received at the same time, TOIP can be calculated from these data points.
receiver (or amplifier) nonlinearity occurs But, look what happens if the signal
and heterodyne products are created. generator is working properly (Figure 17.16).
If F1 and F2 are two input signals (one The spectrum of Figure 17.16A is what we
of which might be the desired signal), these hope to see. Frequencies F1 and F2 stand up
“intermodulation” products will have fre- smartly above the noise level (which ideally
quencies equal to mF1 ± nF2, where m and n is quite low). Figure 17.16B shows the same
are integers. The third-order harmonics are spectrum with the third-order IM products.
those in which m = 2 and n = 1, or m = 1 and When you see this at the output of an ampli-
n = 2 (i.e., m + n = 3), and these are the most fier or receiver being tested, you might as-
difficult to handle. The worst case is usually sume that the IM products are generated in
the 2F1 – F2 and 2F2 – F1 third-order prod- the DUT—but not always. Sometimes, the
ucts because they fall close to F1 and F2 and signal from one generator gets into the out-
may be within the device passband. A prob- put stages of the other generator and causes
lem with these products is that they increase an IM response due solely to the test setup.
at a rate three times the number of decibels as A couple things can be done to prevent
the fundamental signal. If F1 or F2 goes up 1 the problem. First, if the combiner used to
dB, then the third-order products go up 3 dB. merge the signals into one line is a resistive
Figure 17.15 shows the basic setup for star circuit, only 6 dB of isolation lies be-
measuring the third-order intercept point (also tween the ports. Using a hybrid combiner
certain other parameters). The two signal gen- with more port-to-port isolation helps tre-
erators produce F1 and F2. They are set to mendously, because it reduces the signal
identical output levels, usually quite high, such reaching the other generator’s output stages.
as −10 or −20 dBm. Initially, the receiver is Another fix is to insert 10 or 20 dB fixed
tuned to one frequency (e.g., F1) and a refer- attenuators in each signal line (Figure 17.17).




Fig. 17.15 Combining two signal generators for measuring the third-order intercept point and other

F1 F2
Extending Upper Output Range
Signal generator output controls are calibrated
in terms of output voltage, usually microvolts
(µV) or millivolts (mV), or the power level
(e.g., dBm, decibels relative to 1 mW in 50 Ω).
A typical generator produces output levels up
A to some value from about 0 dBm or perhaps
+20 dBm (or some value in between). What
do you do if the signal generator maximum is,
say, +10 dBm and you need a signal level of
+30 dBm (1 W)? Or, what if you have a signal
generator like mine that is wonderful at lower
levels but falls apart at higher levels?
The solution is simple and obvious:
Amplify the output. But take care. Figure 17.18
shows an external amplifier used to boost the
output level of the signal generator. Because
F1 F2
all amplifiers can become nonlinear and pro-
duce a bit of harmonic distortion in their own
right, a low-pass filter is inserted in the path
between the amplifier output and the DUT.
F1 F2 Also make sure that the selected ampli-
B fier can do the job. Make sure the TOIP speci-
fication of the amplifier is sufficiently high that
the signal generator cannot overdrive it. The
IM IM maximum input drive level (usually specified
in dBm) and the output power level (also ex-
pressed in dBm or possibly watts) must be suf-
ficient to handle the job. Otherwise, adding
IM IM the amplifier might add problems as well.

Fig. 17.16 Signal generator spectrums: (A) ideal; Reducing the Output Level
(B) with third-order IM products.
You might want to reduce the signal genera-
tor output level. One reason for doing this is
that you need a very small signal at the DUT
These attenuators provide additional isola- but a higher signal to act as a reference or be
tion between the signal generators. Of fed to a frequency counter.
course, you have to adjust the output levels Sometimes this is done when using an
of the signal generators to overcome the ex- old analog signal generator that has an inac-
tra loss. curate analog frequency dial but a well-
Take care, however. Be certain that the calibrated output attenuator. To get a higher
signal generator output can be cranked to a level for the counter while providing a low-
higher level without producing spurious out- level signal to the amplifier or receiver being
put signals, harmonics, and other extraneous tested, use a setup like that in Figure 17.19.
signals. One of my signal generators works In other cases, you simply need a lower sig-
well from 0 to 90% full output; but at output nal level than the generator can provide.
levels >90%, the spectrum blossoms with un- The attenuator should be a calibrated
wanted signals. type. You can obtain a continuously variable
Receiver Tests and Measurements 257






Fig. 17.17 Solution to spectrum with third-order IM products.




Fig. 17.18 Improving the output level of a signal generator with an external amplifier.






Fig. 17.19 Using an economy signal generator to get a high level for the counter while providing a low-
level signal to the DUT.

calibrated attenuator, but these are costly (al- Practical

though some tend to come on the surplus
market). A lower-cost alternative is to use a
precision step attenuator. These devices have
switch selectable attenuation levels in vari-
ous steps. The total attenuation is the sum of
all the individual attenuations. You can build 10
step attenuators, but for precision work, you
are well advised to buy one (the resistors for Minimum
precision attenuation levels are really odd Signal (MDS)
values, although they can be approximated).
3 dB


Radio receiver specifications can be verified Fig. 17.20 SNR for two levels: 3 dB and 10 dB.
using test equipment and a few simple pro-
cedures. Such tests are made to evaluate re-
ceivers, troubleshoot problems, and verify widths. The value of the noise at any given in-
performance. A number of receiver parame- stant depends on the bandwidth of the chan-
ters are important, but perhaps the most nel. For most receivers, this means the IF
commonly discussed one is sensitivity. We selectivity bandwidth, although in some cases
now look at how these tests are done. the audio bandwidth is less than the IF, so that
number dominates.
Figure 17.20 depicts two different defi-
nitions of SNR. Basically, you cannot hear
Receiver sensitivity is a measure of how well signals down in the noise. The minimum dis-
the receiver picks up very weak signals. As cernible signal is operationally defined as the
with most engineering measurements, the signal level that is the same as the noise floor
notion of sensitivity is an operational defini- or the signal level 3 dB above the receiver’s
tion. In other words, standard procedures noise floor. But that sensitivity is not very
yield coherent results by which different re- useful for most applications. Perhaps there
ceivers (or the same receiver before and after are some people who can discern a signal
repairs) can be compared. only 3 dB above the noise floor, but most
Sensitivity basically is a problem of sig- people require a higher SNR to notice a sig-
nal-to-noise ratio or, more properly, the sig- nal. Although some definitions use 6, 12, or
nal-plus-noise-to-noise ratio [(S + N)/N ]. Every 20 dB, the standard for practical sensitivity is
receiver or amplifier has a basic noise level, the signal level that produces a 10 dB SNR.
consisting of the noise produced external to This definition is found on most CW, AM,
the receiver and noise produced inside the re- and SSB receivers.
ceiver. Even a receiver with its antenna input
terminated in a shielded matching resistor, SIGNAL GENERATOR
rather than an antenna or signal generator, The signal generator selected to make sen-
shows a certain amount of thermal noise. sitivity measurements must have very high
One important consideration when mak- isolation figures. Most service-grade signal
ing sensitivity measurements (or comparing generators are useful for troubleshooting but
receiver sensitivity specifications) is band- unsatisfactory for sensitivity measurement.
width. Thermal and other forms of noise are The reason is that signal escapes around the
Gaussian distributed over all possible band- cabinet flanges and control bushings. If you
Receiver Tests and Measurements 259

have a sensitive receiver or spectrum ana- of such equipment are listed on the World
lyzer, then you can detect this signal. Wide Web. If you do not need the latest dig-
Want to give it a try? Connect a shielded itally synthesized signal generators, you can
dummy load to the output of the signal gen- find good signal generators at low cost.
erator and turn the signal generator’s output
to 0. Connect a whip or wire antenna to the
Test Setup
receiver’s antenna input and tune the receiver
across the signal generator frequency with Figure 17.21 shows the test setup for most
the RF gain cranked all the way up. receiver sensitivity measurements. The atten-
The signal generator also needs cali- uator is optional and may not be needed if
brated output control. The correct calibrations the signal generator is adequately equipped
are either in dBm (power decibels relative to 1 with a good-quality calibrated output attenu-
mW in a 50 Ω load) or microvolts (µV). Some ator. When measuring an AM receiver, set
signal generators have an output meter that the signal generator modulation for 30%
can set relative output but become “cali- depth and 1000 Hz.
brated” if a calibrated step attenuator is con- The receiver output level is measured
nected between the output of the signal using an audio AC voltmeter. Ideally, the in-
generator and the receiver under test. You can strument should be calibrated in decibels as
find the exact level, if you can measure the well as volts and have RMS reading capability.
high level output of the signal generator. The receiver must be correctly set up or
Laboratory-grade signal generators may the measurement will be in error. In most test
be beyond the means of many people, but setups, the receiver’s RF and AF gain controls
there is a relatively vigorous market in used are turned to maximum and the squelch is
or surplus equipment. A number of sources turned off. Further, the automatic gain con-






Fig. 17.21 Test setup for most receiver sensitivity measurements.


trol must be either turned off or, in some FULL-POWER SENSITIVITY

models, clamped to a DC level according to Some older receivers use the full-power sen-
the manufacturer’s directions. sitivity figure. This is the signal level that will
produce the full rated audio output power.
MINIMUM DISCERNIBLE SIGNAL SENSITIVITY Set the signal generator for 1 kHz modulation
To measure the minimum discernible signal with 30% depth. Tune the radio and signal
(MDS), we need to find the signal level in generator to the same frequency and crank
dBm or µV that is 3 dB above the receiver up the output until the audio output power is
noise floor, using the following steps: at the full-rated power level (e.g., 400 mW, 1
W). The signal level that produces this condi-
1. Connect the equipment as in Figure
tion is the “sensitivity” of the receiver.
17.21 and set the receiver and signal
generator to the same frequency.
10 DB (S + N)/N TEST
2. Turn the signal generator output all the The 10 dB test method is the same as the 3
way down to 0. dB MDS method, except that the signal gen-
erator level is increased until the output is 10
3. Crank the RF gain and AF controls all the
dB above the noise floor.
way up (you may want to set the audio
An alternate method sometimes is used
output control to a convenient level if
on AM receivers:
you have no dummy speaker load).
1. Set up the signal generator and receiver
4. To measure the MDS, first measure the
as discussed previously.
rms value of the noise (“hiss”) output
on the AC voltmeter, then increase the 2. Set the output of the receiver to produce
signal generator output level until the at least 0.5 W audio output, or if the
receiver output level increases 3 dB. rated output power is lower than 1 W,
set it for at least 50 mW audio output
You can determine the numerical value
of the receiver noise floor by the same ap-
proach. Measure the output noise level, then 3. Turn off the modulation. If the audio
find the MDS by this procedure. The receiver output drops at least 10 dB, then the
noise floor level will be the same as the sig- signal generator setting is at the 10 dB
nal generator output level (less any attenua- S + N/N level. If the level drops less
tion in the line). than 10-dB, then readjust the signal
generator output level upward a small
STANDARD OUTPUT CONDITIONS amount and try again.
A sensitivity specification used for consumer
radio receivers uses a standard output ap- ON-SITE EFFECTIVE SENSITIVITY TEST
proach. A typical receiver sensitivity spec This test is done on-site only when the re-
might read “xx µV for 400 mW in an 8 Ω load ceiver is installed. It is intended to get some
when modulated 30% by 1 kHz.” The same idea of how well the receiver performs in its
equipment setup (Figure 17.21) can be used actual installed environment (Figure 17.22
for this measurement. A power of 400 mW shows the test setup):
(0.4 W) in an 8 Ω load is the same as 1.789 V
1. Measure the 10 dB S + N/N sensitivity
rms, which can be read on the AC voltmeter.
as discussed previously (see Figure
Use an 8 Ω noninductive resistor for the load
17.21 for setup) and write down the
rather than the loudspeaker—the sound lev-
els are pretty annoying. Adjust the signal
generator output level for an rms output volt- 2. Connect the hybrid combiner, two-
age of 1.789 V and read the output level from position coaxial switch, antenna, and
the signal generator controls. dummy load into the circuit.
Receiver Tests and Measurements 261




Fig. 17.22 Setup for on-site test of receiver sensitivity.

3. Set the switch to the dummy load and tions receivers as well. Today, the 12-dB
measure the sensitivity. It will be consid- SINAD method is preferred.
erably worse than the 10 dB sensitivity.
4. Set the coaxial switch to the antenna
The 20-dB quieting method relies on the ca-
and again measure the 10 dB sensitivity.
pacity of the FM detector to suppress noise
It should be lower still.
once the limiting signal level is reached. The
The effective sensitivity is SNR10dB − well-known capability of FM to eliminate noise
(SNRLoad − SNRANT), where SNRANT is the SNR is because most noise amplitude modulates the
when the antenna is connected. The figure carrier. If the amplitude can be clamped below
SNRLoad − SNRANT is the degradation factor. For the level where the noise is effective, then the
example, suppose the 10 dB SNR is −122 frequency variations can be detected to recover
dBm, the SNR when the load is connected is the audio. This effect is called quieting, the re-
−77 dBm, and when the antenna is connected duction of noise as the signal level increases.
it is −70 dBm. The effective on-site SNR is To measure the 20 dB quieting sensitivity,
SNREFF = SNR10 dB − (SNRLoad − SNRANT) 1. Connect the receiver and signal gener-
ator as in Figure 17.21. Keep the signal
= −122 dBm − [(−77 dBm) − (−70 dBm)]
generator output at 0. The modulation
= −122 dBm − [−7 dBm] (deviation) should be set to whatever
is appropriate for the class of receiver
= −115 dBm being measured.
The effective sensitivity is valid for only 2. Turn the RF gain all the way up. Set
the given site and conditions present when the audio output to produce a conve-
the test is performed. If the site or the noise nient reading in the high end of the
generators and other signals present change, AC voltmeter scale.
then the test must be repeated.
3. Measure the output noise level and
write it down.
FM Receiver Sensitivity
4. Turn up the signal generator output
There are two basic methods for measuring level until the reading on the AC volt-
the sensitivity of FM receivers: 20 dB quieting meter drops 20 dB. The signal genera-
and 12 dB SINAD. The 20-dB quieting method tor output level that accomplishes this
typically is used on FM broadcast band re- is the 20 dB quieting sensitivity (typi-
ceivers and once was popular for communica- cally less than 1 µV).

SINAD SENSITIVITY of noise and distortion is 25% the signal volt-

The sensitivity of FM receivers is often ex- age. As signal levels get higher, the SINAD
pressed in terms of SINAD (signal-noise dis- and 10 dB SNR values tend to converge.
tortion). This approach to the signal-to-noise Figure 17.23 shows a typical test setup
ratio recognizes that the problem of detec- for the SINAD measurement. The output AC
tion depends on not simply signal and noise voltmeter is augmented by a total harmonic
level but also distortion. The SINAD method distortion analyzer, both of which measure the
is described by equation 17.7: output signal across the audio load (speaker,
load resistor, etc.). To test SINAD sensitivity,
Signal + Noise + Distortion
SINAD = (17.7)
Noise + Distortion 1. Set the signal generator frequency and
receiver frequency to the same value.
In terms of decibels, the following equation
2. Set standard conditions: modulating fre-
is used:
quency 1 kHz sine wave, deviation set
SINAD (dB) = to 60% the peak deviation used for that
V Signal + V Noise + VDistortion  (17.8) service. For example, for an FM BCB re-
20 log   ceiver, deviation is ±75 kHz, so set the
 V Noise + VDistortion  signal generator deviation to 0.6 × ±75
where kHz = ±45 kHz. For a communications
receiver designed for ±5 kHz deviation,
SINAD(dB) is the SINAD sensitivity ex- set deviation to 0.6 × ±5 kHz = ±3 kHz.
pressed in decibels (dB);
VSignal is the output voltage due to the 3. Adjust the receiver audio output to ap-
signal; proximately 50% the receiver’s rated
VNoise is the output voltage due to noise; audio output.
VDistortion is the output voltage due to 4. Adjust the signal generator output until
distortion. the input signal is high enough to pro-
The standard 12 dB SINAD sensitivity duce 25% distortion. This is the 12 dB
corresponds to a 4:1 SNR, in which the sum SINAD sensitivity.





Fig. 17.23 Setup for on-site test of SINAD sensitivity.

Receiver Tests and Measurements 263

Special SINAD sensitivity meters are bandwidths specify the noise floor and sensi-
available that combine the THD analyzer and tivity in the narrowest available bandwidth.
audio voltmeter functions in one instrument. However, that may not be useful if the normal
mode for that receiver requires a wider band-
width. For example, an H.F. communications
RECEIVER NOISE MEASUREMENT receiver may have filters for AM mode (BW = 6
kHz), single-sideband (BW = 2.8 kHz), and
Radio reception is a problem of the signal-to- CW (BW = 500 Hz). If the mode required for
noise ratio. For this reason, it sometimes is your application is SSB, then do not rely on
necessary to measure noise levels and the sensitivity and noise floor measurements made
noise performance of receivers. on the 500 Hz bandwidth for the CW mode.
The noise floor of a receiver is the
same as the minimum detectable signal and
1 dB Compression Point
measured in the same way. The MDS is de-
fined as the signal level that causes an out- The 1 dB compression point is the input sig-
put 3 dB above the noise floor. Hence, when nal level at which the receiver gain drops 1
you measure the MDS you also measure the dB (Figure 17.24). The gain of the system is
noise floor. depicted by the Pout/Pin line. This charac-
Setting standard conditions is necessary teristic is linear up to a point; that is, a 1 dB
to take this measurement properly. For increase in input signal level causes a propor-
example, one caution is to use a signal level tionally scaled output level change. At some
at least 10 dB above the expected sensitivity point, however, the receiver is saturated and
of the receiver, then work down to find the 3 cannot accommodate any further input signal.
dB increase level. The operational definition of where this oc-
A standard condition is to ensure that the curs is the 1 dB compression point.
noise floor measurements are made in a stan- To measure this point, use the standard
dard bandwidth. Often, receivers with multiple test setup of Figure 17.21. Bring the input

1 dB
Output Level (Pout )

Fig. 17.24
The 1 dB compression point. Input Level (Pin )

power level applied to the receiver antenna INTERFERING

terminals up from some low value in 1 dB SIGNAL
steps until the receiver output level drops 1
dB. The input level at which this occurs is
the 1 dB compression point. DESIRED

Dynamic Range DF
Dynamic range represents the total range of
input signal levels that can be accommo-
dated. The classical dynamic range measure
is the blocking dynamic range, but the third-
order IMD dynamic range is used as well.


F1 F2
The blocking dynamic range (BDR) mea-
sures the difference between the receiver Fig. 17.25 The blocking dynamic range.
noise floor and the level of an off-channel
undesired signal to reduce the sensitivity to
on-channel signals by a specified amount the desired signal, while F2 is the interfering
(Figure 17.25). In other words, it is a measure signal.
of the range from MDS to a specified desen- The amplitude of F1 should be set such
sitization of the receiver. that it is above the receiver MDS by at least
Connect the test setup of Figure 17.26. the mount of the required minimum SNR
Two signal generators are coupled through (e.g., 10 dB) but below the point where it be-
an isolating hybrid to a step attenuator then gins to increase IMD products. Using a higher
to the receiver. An AC audio voltmeter is level minimizes the noise error contribution.
used to measure the output level of the re- The exact level depends somewhat on the
ceiver. Figure 17.25 shows the signal situa- test procedure. If the receiver’s automatic
tion for the receiver input. Frequency F1 is gain control can be disabled, then set the sig-





Fig. 17.26 Setup for testing the blocking dynamic range.

Receiver Tests and Measurements 265

nal level about 10 dB below the 1 dB com- difference between the MDS and the level de-
pression point. If the AGC cannot be dis- termined by this procedure is the TOIMDDR.
abled, then the signal level should be lower,
say, about 20 dB above the noise floor.
A Cautionary Note
Frequencies F1 and F2 must have some
specified spacing (∆F). In most cases, an HF In both this measurement and the IMD mea-
receiver will call for a 20 kHz spacing, while surement (discussed next), it is important to
a VHF/UHF receiver will call for a 20 kHz or ensure that only the receiver IMD products
100 kHz spacing. Some special purpose mi- are measured. Anytime a receiver or amplifier
crowave receivers sometimes have consider- is overdriven into a nonlinear operating re-
ably larger spacing. Whatever the case, the gion, the possibility exists for the creation of
same spacing should be used on different re- unwanted distortion products. Even a single
ceivers when comparisons are made, be- signal can produce first-order effects. If, for
cause it seriously affects the results. example, frequency F is applied to the input
At the start of the measurement, the F1 in sufficient power to overload the front end,
is turned on, F2 is turned off, and the re- then harmonics at 2F, 3F, and so forth might
ceiver is adjusted for maximum output of F1. be generated. Similarly with normal mF1 ±
Turn on F2 at some level, such as −100 dBm. nF2 IMD products higher than first order.
Increase the level of F2 in 1 dB steps until Several mechanisms exist and must be
the output level of the receiver drops 1 dB. addressed. First, the hybrid coupler could be
The level at which this occurs represents the nonlinear. This can occur when broadband
blocking dynamic range. ferrite or powdered iron-core transformers are
If this test is performed on automatic test used in the hybrid coupler. The coupler must
equipment, then the output levels are auto- produce very low levels of IMD or these will
matically stepped. Also, in some cases, the fre- affect the results of the test. Typically, a third-
quencies also are stepped up and down the order intercept point (TOIP) of >50 dBm is re-
band, although maintaining the frequency quired of the coupler. If the hybrid coupler
separation ∆F. Adequate settling time between being used has a lower TOIP or if the receiver
measurements must be programmed into the has a particularly high TOIP (>45 dBm), then
protocol. The signal generator setting time use no test configuration that places attenua-
rarely dominates this case, but be wary of the tion between the receiver input and the out-
AGC settling time. AGC circuits have a time put of the coupler. That configuration forces
constant, and it might be a relatively long pe- the coupler to operate at too high a level.
riod (e.g., 5 sec). Understand the AGC time Or, the signal from one signal generator
constant before programming a frequency or could enter the output circuits of the other.
signal level stepping protocol. Most amplifiers produce IMD products when
signals arrive from the outside. This problem is
THIRD-ORDER IMD DYNAMIC RANGE reduced to negligible levels if the signal gener-
A test setup similar to that for BDR is used ator outputs have a high degree of inherent
for the third-order IMD dynamic range isolation, if attenuation is used between the
(TOIMDDR). Set the signal generators for a coupler and the signal generators, or if the in-
convenient output level such as −20 dBm and put-to-input port isolation of the hybrid is
the frequencies in band with a specified ∆F high. Try to get as much isolation as possible
spacing. Set the attenuator for a high degree between signal sources and between the
of attenuation (low signal level). Increase the sources and the receiver (>90 dB). Also pru-
signal by decreasing attenuation in 1 dB steps dent is using a signal generator with a high-
until a third-order response appears in the level output (approximately +15 to +22 dBm).
output. The actual signal level applied to the Another problem with coupling between
receiver, as usual, can be calculated from the signal generators is that many high-quality sig-
signal generator and attenuator settings. The nal generators employ a feedback-controlled

automatic level control that samples and rec- Second-order IM products (F1 ± F2) tend
tifies the RF output signal and then feeds to fall outside the band, so they are filtered
back the derived DC control voltage to an out easily. Third-order IM products (2F1 ± F2
amplitude modulator. When external signals and 2F2 ± F1) usually are outside the band.
arrive, they can affect the ALC in two ways. Several methods are used to measure the
First, the extra signal level may influence IMD performance of a receiver. Figure 17.28
feedback levels. Second and more likely, the shows one standard setup. Two signal gener-
IMD products may get into the feedback sys- ators are used to provide the two different sig-
tem, causing beat notes that modulate the nals required for the IMD test. Each signal
regular output signal. generator is equipped with an adjustable at-
tenuator, which may or may not be external to
the generator. In some cases, both internal
Tidbit: How do you tell if the IMD is due to and external attenuators may be used.
the receiver or the test fixture? In most cases, Optional bandpass filters sometimes are
reducing the signal levels will give you the used to clean up the signal generator output
answer. If the ratio of desired to IMD re- spectrum. These filters are used to suppress
sponses changes when the input signal level
to the receiver is changed, then it is a reason- harmonics of the output frequency. If the sig-
able assumption that the IMD is generated in- nal generator has sufficiently low harmonic
side the receiver. If the ratio does not change, output, then these filters can be eliminated.
then it is probably due to the test fixture. Keep in mind that some filters use ferrite or
powdered iron cores, so they may saturate
and cause IMD products of their own. The
two signals are combined in a two-port hy-
Intermodulation Distortion brid. Following the hybrid is another attenu-
Intermodulation distortion occurs when two ator. This attenuator supplies signal to the
frequencies, F1 and F2, combine to produce receiver input.
heterodyne products that were not in the The output signal is monitored by any
original set of signals. Figure 17.27 shows of several means. Some procedures use the
this behavior. When F1 and F2 are suffi- audio AC output level, as measured by an AC
ciently strong, the receiver becomes nonlin- voltmeter. In others, the spectrum of the au-
ear so mixing will occur. When this happens dio output signal is measured using a spec-
IM products rise up out of the noise. trum analyzer. An alternative might be to use
a frequency selective voltmeter (wavemeter).
The last method is out of favor because spec-
F1 F2 trum analyzer prices have fallen significantly.
Some people will use the receiver’s S meter
(if it has one) to make this measurement. Still
others couple the IF signal to an RF/IF spec-
trum analyzer, which may show more infor-
mation but requires entry inside the receiver.
The other methods treat the receiver as a
“black box,” requiring no modification of, or
IM IM entry into, the receiver. The IMD test is best
PRODUCT PRODUCT run in one of the linear reception modes
(SSB or CW), but that is not always possible
(e.g., when the receiver is FM or AM only).


The audio output level is monitored on an
Fig. 17.27 IMD products (close-in frequencies). audio spectrum analyzer (or measured on a
Receiver Tests and Measurements 267









Fig. 17.28 A standard setup for a complete receiver test.

wavemeter). The signal levels are turned up ceiver noise, so it makes the IMD look
until the IMD product being investigated worse. Also, AGC action could interfere with
rises from the noise level. this test.
The spectrum analyzer method can be
particularly useful for measuring products S METER METHOD
below the noise floor of the receiver. Recall In this method, the level of the IMD product
that the noise floor is proportional to band- is noted on the receiver’s S meter. A refer-
width. Typical bandwidths vary from 500 Hz ence signal is then provided that matches the
for CW receivers to 200 kHz for FM broad- S meter reading. This yields the level of the
cast receivers (and more for microwave radar IMD product. Problems with this method in-
receivers). If the bandwidth filter on the au- clude the tendency of some receivers to
dio spectrum analyzer is set to some narrow compress gain when the signal level gets to a
value, such as 5 or 10 Hz, then the noise is level above S9 or S9 + 10 dB. However, this
much lower, so low-level IMD problems is a better method than some for measuring
show up better. the IMD performance of receivers with very
high IMD performance.
The SNR method of measuring IMD uses ei- STANDARD METHOD
ther an audio signal-to-noise ratio meter or a The usual method for measuring the IMD
SINAD meter. The audio output is set to pro- performance is to set the signal generators to
duce a 1 kHz signal for this method. Care some convenient high-level output (e.g., −20
must be exercised to prevent excess noise dBm). Select test frequencies (F1 and F2) and
contribution from the signal generator output calculate the third-order products (2F1 + F2,
noise. This noise is indistinguishable from re- 2F1 − F2, 2F2 + F1, and 2F2 − F1).

Set the receiver to a channel frequency, 1. The minimum discernible signal at

F1. If possible, turn off the AGC or clamp it to 162.55 MHz requires an attenuator set-
a low value (highest receiver gain), if possible. ting of –89 dB, so PIM = (−10 dBM – 89
If the receiver uses a front-end RF or IF atten- dB) = −99 dBm.
uator, then set that to 0 dB. If an RF preampli-
2. The response at 162.53 MHz required
fier is used, turn if off. Adjust both the receiver
19 dB of attenuation for the third-order
tuning and F1 to the same frequency and
response to equal the MDS level, so PA
maximize the receiver output. Set the second
= (−10 dBm – 19 dB) = −29 dBm.
signal generator to F2, a specified spacing
(e.g., 20 kHz) away from F1. Set both signal Because this is a third-order response, N = 3 so
generators to a convenient output level, such
as –10 dBm. Set the in-line attenuator to the [3 × ( −29 dBm)] − ( −99 dBm)
highest setting (most attenuation). 3−1
Once the setup is complete, turn off the −87 dBm + 99 dBm
signal generators and measure the receiver =
output noise level on an AC audio voltmeter. +12 dBm
Turn on signal generator F1 and decrease the = = +6 dBm
attenuator setting in 1 dB steps until the output
noise level of the receiver increases 3 dB. This Once the PA and PIM points are found, any IP
is the minimum discernible signal reference can be calculated using equation 17.9.
level. Return the attenuator settings to maxi-
mum. Record this signal level as PIM = −10
dBm – (Attenuator setting). Selectivity Testing
Tune the receiver to either of the close-in Selectivity is a measure of the receiver’s abil-
third-order product frequencies (2F1 − F2 or ity to reject off-channel signals. It is mea-
2F2 − F1), while leaving the signal generators sured in terms of bandpass, so it has the
at F1 and F2 (both −10 dBm output). Reduce units of frequency. The operational specifica-
the attenuator setting until the receiver output tion of selectivity is the bandpass (Figure
response at this frequency increases 3 dB (the 17.29A) between the two points (F1 and F2)
same as the reference MDS). Record this level where the frequency response drops −3 dB
as PA = −10 dBm – (Attenuator setting). from its midband point (Fo).


IPN = (17.9)
N −1 It is possible to use a standard CW generator
to take this measurement, although it is te-
where dious and may not yield the best informa-
tion. Set up a signal generator and output
IPN is the intermodulation product of indicator as per Figure 17.21. Center the sig-
order N; nal generator frequency inside the receiver
N is the order of the intermodulation passband and adjust the receiver output for a
product; convenient level indicated on the AC volt-
PA and PIM are signal power levels in meter. Adjust the signal generator below Fo
dBm to a point where the output drops the speci-
fied amount. This frequency is F1. Repeat the
EXAMPLE procedure above Fo until F2 is determined.
A 162.55 MHz receiver is tested using fre- The receiver bandwidth is F2 − F1.
quencies F1 = 162.55 MHz and F2 = 162.57
MHz. The close-in third-order products are SWEEP METHOD
2F1 − F2 = 162.53 MHz and 2F2 − F2 = The CW method produces a raw indication but
162.59 MHz. lacks certain information, at least in its simplest
Receiver Tests and Measurements 269

form. We are not simply interested in the F2 − sweep method produces an oscilloscope trace
F1 value but also in the shape of the passband. that can be photographed.
Ideally, a receiver passband is flat inside the Figure 17.29B shows the test equip-
passband, rolling off gently on the upper and ment setup for a simple sweep test. A sweep
lower ends. This may not be the actual case. generator will sweep through a specified
By using the sweep method, we can deter- band of frequencies. In the simplest form,
mine the shape of the passband as well as its it consists of a voltage-controlled oscillator
bandwidth. Surprises can be lurking here, so with its tuning voltage in the form of a saw-
the information is needed. To get shape data tooth waveform (modern sweepers are a bit
by the CW method requires collecting a lot of more sophisticated but the idea is the same).
data points and manually graphing them. The The same sawtooth can be used to control

F1 FO F2








Fig. 17.29 Selectivity testing: (A) frequency response; (B) use of a marker generator.

the horizontal deflection of an oscilloscope. the critical squelch level just barely quiets the
The sweep signal is applied to the receiver’s receiver when no signal is being received.
antenna input, while the receiver output is Indeed, when a particularly high level noise
applied to the vertical deflection input of the burst is heard, it might degrade the squelch
oscilloscope. The result is a trace represent- enough to pass through momentarily. Tight
ing amplitude vs. frequency. squelch requires a much larger signal to
A marker generator can be used to indi- cause break through; here, the squelch con-
cate specific frequencies within the passband trol is set to maximum.
(typically, F1, F2, and Fo). In some cases, dis-
crete frequencies are used, while in others the CRITICAL SQUELCH
harmonics of a standard frequency may be Connect a signal generator to the antenna in-
used. For example, a 1 kHz marker generator put of the receiver. Adjust the signal genera-
that is sufficiently rich in harmonics will pro- tor frequency on channel.
vide markers every 1 kHz throughout the pass- For AM receivers, set the signal genera-
band. However, harmonic markers are limited tor modulation depth to 30% with a modulat-
in that their upper harmonics may not be ing frequency of 1000 Hz. For FM receivers,
strong enough to produce the desired display. use a deviation approximately 60% the nor-
The sweep rate setting deserves some mal deviation for that receiver, with a 1000
comment. The sweep rate is the repetition Hz modulating frequency. Then:
rate at which the signal passes through the
1. Set the signal generator output level to
swept band. This attribute is controlled by
the lowest possible level so that no sig-
the frequency of the sawtooth generator. If
nal is heard in the receiver output.
the sawtooth rate is too slow, then the dis-
play will flicker and be hard to read. If it is 2. Set the squelch control to the point
too high, then there may be ringing in the re- where the output noise just disappears
ceiver’s IF filters, causing a distorted reading. when no signal is present.
Flicker fusion tends to occur in the 8–10 Hz
3. Bring the signal generator’s output
range for most human operators, so the min-
level up very slowly until the squelch
imum sweep frequency will have to be at
breaks. Record the output level as the
least these values. When you get above 40
critical squelch point (dBm or µV).
Hz or so, the danger of causing a ringing re-
sponse in high-Q IF filters becomes more
likely. Therefore, select a sweep frequency
The tight squelch test is identical to the criti-
between roughly 8 and 40 Hz for manually
cal squelch test but with the squelch control
operated measurement systems.
on the receiver set to the maximum (most
squelched) position. The signal level re-
Squelch Tests quired to break tight squelch will be much
Communications receivers often are equipped
with a squelch circuit. These circuits turn off
the receiver’s audio output when no signal is
The squelch range lies between critical
being received. Noise occupies the bandwidth
squelch and tight squelch. It may be ex-
when no signal is received, and it is uncom-
pressed in signal level units (dBm or µV) or
fortable to hear. When the squelch circuit de-
in dB. For example, the following signal lev-
tects noise, rather than signal, it turns off the
els lead to the squelch range listed here:
receiver output. The operator sets the squelch
from the front panel. Critical Squelch: 0.4 µV (−115 dBm)
The two levels of squelch are critical Tight Squelch: 58 µV (−71 dBm)
squelch and tight squelch. A control set to Squelch Range: 57.6 µV (44 dBm)