Figure 6-6 How frequency multipliers increase carrier frequency and deviation.

Frequency multipliers

Crystal
6.5 MHz 156 MHz
4.8-MHz deviation
4 3 2
4  3  2  24

crystal frequency can be “pulled” slightly from its natural resonant frequency. By mak-
ing the series capacitance a varactor diode, frequency modulation of the crystal oscil-
lator can be achieved. The modulating signal is applied to the varactor diode D1, which
changes the oscillator frequency.
It is important to note that only a very small frequency deviation is possible with
frequency-modulated crystal oscillators. Rarely can the frequency of a crystal oscillator be
changed more than several hundred hertz from the nominal crystal value. The resulting
deviation may be less than the total deviation desired. For example, to achieve a total fre-
quency shift of 75 kHz, which is necessary in commercial FM broadcasting, other techniques
must be used. In NBFM communication systems, the narrower deviations are acceptable.
Although it is possible to achieve a deviation of only several hundred cycles from
the crystal oscillator frequency, the total deviation can be increased by using frequency
multiplier circuits after the carrier oscillator. A frequency multiplier circuit is one whose Frequency multiplier circuit
output frequency is some integer multiple of the input frequency. A frequency multiplier
that multiplies a frequency by 2 is called a doubler, a frequency multiplier circuit that Doubler
multiplies an input frequency by 3 is called a tripler, and so on. Frequency multipliers Tripler
can also be cascaded.
When the FM signal is applied to a frequency multiplier, both the carrier frequency
of operation and the amount of deviation are increased. Typical frequency multipliers
can increase the carrier oscillator frequency by 24 to 32 times. Fig. 6-6 shows how
frequency multipliers increase carrier frequency and deviation. The desired output fre-
quency from the FM transmitter in the figure is 156 MHz, and the desired maximum
frequency deviation is 5 kHz. The carrier is generated by a 6.5-MHz crystal oscillator,
which is followed by frequency multiplier circuits that increase the frequency by a fac-
tor of 24 (6.5 MHz 3 24 5 156 MHz). Frequency modulation of the crystal oscillator
by the varactor produces a maximum deviation of only 200 Hz. When multiplied by a
factor of 24 in the frequency multiplier circuits, this deviation is increased to
200 3 24 5 4800 Hz, or 4.8 kHz, which is close to the desired deviation. Frequency
multiplier circuits are discussed in greater detail in Chap. 8.

Voltage-Controlled Oscillators
Oscillators whose frequencies are controlled by an external input voltage are generally
referred to as voltage-controlled oscillators (VCOs). Voltage-controlled crystal oscillators Voltage-controlled oscillator (VCO)
are generally referred to as VXOs. Although some VCOs are used primarily in FM, they Voltage-controlled crystal
are also used in other applications where voltage-to-frequency conversion is required. As oscillator (VXO)
you will see, their most common application is in phase-locked loops, discussed later in
this chapter.
Although VCOs for VHF, UHF, and microwaves are still implemented with discrete
components, more and more they are being integrated on a single chip of silicon
along with other transmitter or receiver circuits. An example of such a VCO is shown
in Fig. 6-7. This circuit uses silicon-germanium (SiGe) bipolar transistor to achieve an
operating frequency centered near 10 GHz. The oscillator uses cross-coupled transistors
Q1 and Q2 in a multivibrator or flip-flop type of design. The signal is a sine wave whose
frequency is set by the collector inductances and varactor capacitances. The modulating

FM Circuits 177
 Figure 6-7 A 10-GHz SiGe integrated VCO.

DC control or VDD
modulating
voltage

L L

D1 D2

Q3 Q4

Out Out
Q2
Q1

Current source
V

Typical
current
source

voltage, usually a binary signal to produce FSK, is applied to the junction of D1 and D2.
Two complementary outputs are available from the emitter followers Q3 and Q4. In this
circuit, the inductors are actually tiny spirals of aluminum (or copper) inside the chip,
with inductance in the 500- to 900-pH range. The varactors are reverse-biased diodes
that function as variable capacitors. The tuning range is from 9.953 to 10.66 GHz.
A CMOS version of the VCO is shown in Fig. 6-8. This circuit also uses a cross-
coupled LC resonant circuit design and operates in the 2.4- to 2.5-GHz range. Variations
of it are used in Bluetooth transceivers and wireless LAN applications. (See Chap. 20.)
There are also many different types of lower-frequency VCOs in common use,
including IC VCOs using RC multivibrator-type oscillators whose frequency can be

178 Chapter 6
Figure 6-8 A CMOS VCO for a 2.4-GHz FSK.
VDD

Current source

Q3 Q4

Out Out

Control or
modulating
voltage D1 D2

Q1 Q2

N-channel
MOSFET

Figure 6-9 Frequency modulation with an IC VCO. (a) Block diagram with an IC VCO.
(b) Basic frequency modulator using the NE566 VCO.

V+

R1 V

6 8 R2 R1
1.5 kV
VC
Current Schmitt Buffer 0.001 mF
sources trigger amplifier 3
Modulation 5 C2
input
6 8
Modulating 4 FM
signal 5 NE566
out
Buffer 4 3
amplifier 7 1
1 R3
7
10 kV
C1
C1

(a) (b)

controlled over a wide range by an ac or dc input voltage. These VCOs typically have
an operating range of less than 1 Hz to approximately 1 MHz. The output is either a
square or a triangular wave rather than a sine wave.
Fig. 6-9(a) is a block diagram of one widely used IC VCO, the popular NE566. NE566 IC VCO
External resistor R1 at pin 6 sets the value of current produced by the internal current

FM Circuits 179
 sources. The current sources linearly charge and discharge external capacitor C1 at pin 7.
An external voltage VC applied at pin 5 is used to vary the amount of current produced
Schmitt trigger circuit by the current sources. The Schmitt trigger circuit is a level detector that controls the
current source by switching between charging and discharging when the capacitor charges
or discharges to a specific voltage level. A linear sawtooth of voltage is developed across
the capacitor by the current source. This is buffered by an amplifier and made available
at pin 4. The Schmitt trigger output is a square wave at the same frequency available at
pin 3. If a sine wave output is desired, the triangular wave is usually filtered with a tuned
circuit resonant to the desired carrier frequency.
A complete frequency modulator circuit using the NE566 is shown in Fig. 6-9(b).
The current sources are biased with a voltage divider made up of R2 and R3. The mod-
ulating signal is applied through C2 to the voltage divider at pin 5. The 0.001-μF capacitor
between pins 5 and 6 is used to prevent unwanted oscillations. The center carrier fre-
quency of the circuit is set by the values of R1 and C1. Carrier frequencies up to 1 MHz
may be used with this IC. If higher frequencies and deviations are necessary, the outputs
can be filtered or used to drive other circuits, such as a frequency multiplier. The mod-
ulating signal can vary the carrier frequency over nearly a 10 :1 range, making very large
deviations possible. The deviation is linear with respect to the input amplitude over the
entire range.

Phase modulator 6-2 Phase Modulators

Most modern FM transmitters use some form of phase modulation to produce indirect
FM. The reason for using PM instead of direct FM is that the carrier oscillator can be
optimized for frequency accuracy and stability. Crystal oscillators or crystal-controlled
frequency synthesizers can be used to set the carrier frequency accurately and maintain
solid stability.
The output of the carrier oscillator is fed to a phase modulator where the phase shift
GOOD TO KNOW is made to vary in accordance with the modulating signal. Since phase variations produce
Simple phase shifters do not frequency variations, indirect FM is the result.
produce a linear response over a Some phase modulators are based upon the phase shift produced by an RC or LC
large range of phase shift. To
tuned circuit. It should be pointed out that simple phase shifters of this type do not
produce linear response over a large range of phase shift. The total allowable phase shift
compensate for this, restrict the
must be restricted to maximize linearity, and multipliers must be used to achieve the
total allowable phase shift to desired deviation. The simplest phase shifters are RC networks like those shown in
maximize linearity. Multipliers Fig. 6-10(a) and (b). Depending on the values of R and C, the output of the phase shifter
must also be used to achieve the can be set to any phase angle between 0 and 90°. In (a), the output leads the input by
desired deviation. some angle between 0 and 90°. For example, when Xc equals R, the phase shift is 45°.
The phase shift is computed by using the formula
XC
ϕ 5 tan21
R
A low-pass RC filter can also be used, as shown in Fig. 6-10(b). Here the output is
taken from across the capacitor, so it lags the input voltage by some angle between 0
and 90°. The phase angle is computed by using the formula
R
ϕ 5 tan21
XC
A simple phase-shift circuit can be used as a phase modulator if the resistance or
capacitance can be made to vary with the modulating signal. One way to do this is to
replace the capacitor shown in the circuit of Fig. 6-10(b) with a varactor. The resulting
phase-shift circuit is shown in Fig. 6-11.
In this circuit, the modulating signal causes the capacitance of the varactor to
change. If the modulating signal amplitude at the output of amplifier A becomes more

180 Chapter 6
Figure 6-10 RC phase-shifter basics.

Vc Xc Vc
  tan1  tan1
Carrier (fc) R VR
Output   0ⵑ90
C
VR (output) lead

Vg R VR

1
Xc 
2␲fcC Vc Vg
(a)

Carrier (fc)
Out put VR
R


Vg C

Vg
Vc lag
R VR (out put)
  tan1  tan1
Xc Vc
  0ⵑ90
(b)

Figure 6-11 A varactor phase modulator.

fc
Carrier  V (bias)

Vg R R1

C1 C2
Output

Dv R2
(C )

fm
A

Inverting amplifier
Modulating signal

positive, it adds to the varactor reverse bias from R1 and R2, causing the capacitance to
decrease. This causes the reactance to increase; thus, the circuit produces less phase
shift and less deviation. A more negative modulating signal from A subtracts from the
reverse bias on the varactor diode, increasing the capacitance or decreasing the capac-
itive reactance. This increases the amount of phase shift and the deviation.

FM Circuits 181
 With this arrangement, there is an inverse relationship between the modulating signal
polarity and the direction of the frequency deviation. This is the opposite of the desired
variation. To correct this condition, an inverting amplifier A can be inserted between the
modulating signal source and the input to the modulator. Then when the modulating signal
goes positive, the inverter output and modulator input go negative and the deviation
increases. In Fig. 6-11 C1 and C2 are dc blocking capacitors and have very low reactance
at the carrier frequency. The phase shift produced is lagging, and as in any phase modulator,
the output amplitude and phase vary with a change in the modulating signal amplitude.

Example 6-2
A transmitter must operate at a frequency of 168.96 MHz with a deviation of 65 kHz.
It uses three frequency multipliers––a doubler, a tripler, and a quadrupler. Phase modula-
tion is used. Calculate (a) the frequency of the carrier crystal oscillator and (b) the phase
shift ¢ϕ required to produce the necessary deviation at a 2.8-kHz modulation frequency.
a. The frequency multiplier produces a total multiplication of 2 3 3 3 4 5 24.
The crystal oscillator frequency is multiplied by 24 to obtain the final
output frequency of 168.96 MHz. Thus, the crystal oscillator frequency is
168.96
f0 5 5 7.04 MHz
24
b. The frequency multipliers multiply the deviation by the same factor. To
achieve a deviation of 65 kHz, the phase modulator must produce a
deviation of fd 5 5 kHz/24 5 6208.33 Hz. The deviation is computed
with fd 5 ¢ϕ fm; fm 5 2.8 kHz.
fd 208.33
¢ϕ 5 5 5 60.0744 rad
fm 2800
Converting to degrees gives
0.0744(57.3°) 5 64.263°
The total phase shift is
64.263° 5 2 3 4.263° 5 8.526°

A simple formula for determining the amount of frequency deviation fd represented

by a specific phase angle is
fd 5 ¢ϕ fm
where ¢ϕ 5 change in phase angle, rad
fm 5 modulating signal frequency
Assume that the lowest modulating frequency for the circuit with the shift of 0.75 rad
is 300 Hz. The deviation is fd 5 0.75(300) 5 225 Hz or 6112.5 Hz. Since this
is PM, the actual deviation is also proportional to the frequency of the modulating
signal. With the same maximum deviation of 0.75 rad, if the modulating frequency is
3 kHz (3000 Hz), the deviation is fd 5 0.75(3000) 5 2250 Hz or 61125 Hz.
To eliminate this effect and to generate real FM, the audio input frequency must be
applied to a low-pass filter to roll off the signal amplitude at the higher frequencies in
a low-pass filter.

182 Chapter 6
Example 6-3
For the transmitter in Example 6-2, a phase shifter like that in Fig. 6-10 is used, where
C is a varactor and R 5 1 kV. Assume that the total phase-shift range is centered on 45°.
Calculate the two capacitance values required to achieve the total deviation.
The phase range is centered on 45°, or 45° 6 4.263° 5 40.737°and 49.263°. The
total phase range is 49.263 2 40.737 5 8.526°. If ϕ 5 tan21 (R/XC ), then
tan ϕ 5 R/XC.
R 1000
XC 5 5 5 1161 V
tan ϕ tan 40.737
1 1
C5 5 5 19.48 pF
2πf XC 6.28 3 7.04 3 106 3 1161
R 1000
XC 5 5 5 861 V
tan ϕ tan 49.263
1 1
C5 5 5 26.26 pF
2πf XC 6.28 3 7.04 3 106 3 861
To achieve the desired deviation, the voice signal must bias the varactor to vary over
the 19.48- to 26.26-pF range.

6-3 Frequency Demodulators Frequency demodulator

Any circuit that will convert a frequency variation in the carrier back to a proportional
voltage variation can be used to demodulate or detect FM signals. Circuits used to
recover the original modulating signal from an FM transmission are called demodulators,
detectors, or discriminators.

Slope Detectors
The simplest frequency demodulator, the slope detector, makes use of a tuned circuit and Slope detector
a diode detector to convert frequency variations to voltage variations. The basic circuit
is shown in Fig. 6-12(a). This has the same configuration as the basic AM diode detector
described in Chap. 4, although it is tuned differently.
The FM signal is applied to transformer T1 made up of L1 and L2. Together L2 and
C1 form a series resonant circuit. Remember that the signal voltage induced into L2 appears
in series with L2 and C1 and the output voltage is taken from across C1. The response
curve of this tunedCcircuit
1 is shown in Fig. 6-12(b). Note that at the resonant frequency fr
the voltage across peaks. At lower or higher frequencies, the voltage falls off.
To use the circuit to detect or recover FM, the circuit is tuned so that the center or
carrier frequency of the FM signals is approximately centered on the leading edge of the
response curve, as shown in Fig. 6-12(b). As the carrier frequency varies above and below
its center frequency, the tuned circuit responds as shown in the figure. If the frequency
goes lower than the carrier frequency, the output voltage across C1 decreases. If the
frequency goes higher, the output across C1 goes higher. Thus, the ac voltage across C1
is proportional to the frequency of the FM signal. The voltage across C1 is rectified into
dc pulses that appear across the load R1. These are filtered into a varying dc signal that
is an exact reproduction of the original modulating signal.
The main difficulty with slope detectors lies in tuning them so that the FM signal is cor-
rectly centered on the leading edge of the tuned circuit. In addition, the tuned circuit does not
have a perfectly linear response. It is approximately linear over a narrow range, as Fig. 6-12(b)
shows, but for wide deviations, amplitude distortion occurs because of the nonlinearity.

FM Circuits 183
 Figure 6-12 Slope detector operation.

D1
T1
Recovered modulating signal
L1 L2 C1
R1 C2

(a)

fr
High frequency
produces higher fc
voltage
Voltage across C1

Lower frequency
produces lower
voltage

(b)

The slope detector is never used in practice, but it does show the principle of FM
demodulation, i.e., converting a frequency variation to a voltage variation. Numerous prac-
tical designs based upon these principles have been developed. These include the Foster-
Seeley discriminator and the ratio detector, neither of which is used in modern equipment.

Pulse-Averaging Discriminators
Pulse-averaging discriminator A simplified block diagram of a pulse-averaging discriminator is illustrated in
Fig. 6-13. The FM signal is applied to a zero-crossing detector or a clipper-limiter that
generates a binary voltage-level change each time the FM signal varies from minus to
plus or from plus to minus. The result is a rectangular wave containing all the frequency
variations of the original signal but without amplitude variations. The FM square wave
is then applied to a one-shot (monostable) multivibrator that generates a fixed-amplitude,
fixed-width dc pulse on the leading edge of each FM cycle. The duration of the one
shot is set so it is less than one-half the period of the highest frequency expected during
maximum deviation. The one-shot output pulses are then fed to a simple RC low-pass
filter that averages the dc pulses to recover the original modulating signal.

Figure 6-13 Pulse-averaging discriminator.

Zero- Low-pass
A crossing B C filter D Recovered
FM One-shot
detector modulating
input multivibrator
or clipper/ R signal
limiter  C
Load


184 Chapter 6
Figure 6-14 (a) FM input. (b) Output of zero-crossing detector. (c) Output of one shot.
(d) Output of discriminator (original modulating signal).

(a)

(b)

(c)

(d )

The waveforms for the pulse-averaging discriminator are illustrated in Fig. 6-14. At
low frequencies, the one-shot pulses are widely spaced; at higher frequencies, they occur
very close together. When these pulses are applied to the averaging filter, a dc output volt-
age is developed, the amplitude of which is directly proportional to the frequency deviation.
When a one-shot pulse occurs, the capacitor in the filter charges to the amplitude
of the pulse. When the pulse turns off, the capacitor discharges into the load. If the RC
time constant is high, the charge on the capacitor does not decrease much. When the
time interval between pulses is long, however, the capacitor loses some of its charge into
the load, so the average dc output is low. When the pulses occur rapidly, the capacitor
has little time between pulses to discharge; the average voltage across it therefore remains
higher. As the figure shows, the filter output voltage varies in amplitude with the fre-
quency deviation. The original modulating signal is developed across the filter output.
The filter components are carefully selected to minimize the ripple caused by the charg-
ing and discharging of the capacitor while at the same time providing the necessary
high-frequency response for the original modulating signal.
Some pulse-averaging discriminators generate a pulse every half-cycle or at every
zero crossing instead of every one cycle of the input. With a greater number of pulses
to average, the output signal is easier to filter and contains less ripple.
The pulse-averaging discriminator is a very high-quality frequency demodulator. In
the past, its use was limited to expensive telemetry and industrial control applications.
Today, with the availability of low-cost ICs, the pulse-averaging discriminator is easily
implemented and is used in many electronic products.

Quadrature Detectors
The quadrature detector uses a phase-shift circuit to produce a phase shift of 90° at the Quadrature detector
unmodulated carrier frequency. The most commonly used phase-shift arrangement is
shown in Fig. 6-15. The frequency-modulated signal is applied through a very small
capacitor (C1 ) to the parallel-tuned circuit, which is adjusted to resonate at the center
carrier frequency. At resonance, the tuned circuit appears as a high value of pure resis-
tance. The small capacitor has a very high reactance compared to the tuned circuit
impedance. Thus, the output across the tuned circuit at the carrier frequency is very close

FM Circuits 185