TPS 7.2.1.

Frequency Multiplication

by K. Breidenbach

Revised and updated edition

October 1998
“The sensitive electronics of the equipment contained in the present experiment litera-
ture can be impaired due to the discharge of static electricity. Consequently, electro-
static build up should be avoided (particularly by utilizing appropriate rooms) or
eliminated by discharging (e.g. at the panel frames or similar).”
 TPS 7.2.1.1 Table of Contents

Measurement Procedures
1 Practical Information on the Experiments ............................................................... 5
1.1 The Oscilloscope ....................................................................................................... 5
1.2 The Spectrum Analyzer ............................................................................................ 7
1.3 Measurement Assembly ........................................................................................... 9

TPS 7.2.1.1 Frequency Multiplication

1 An Analog Method of Frequency Multiplication ................................................. 13
2 Frequency Synthesis, PLL ...................................................................................... 19

Solutions .................................................................................................................. 25

Appendix ................................................................................................................. 37

3
 TPS 7.2.1.1 Measurement Procedures 1.1

1 Practical Information on the Experiments

In Sections 1.1 - 1.3 some fundamental questions

are dealt with so that the experiments in the fol-
lowing chapters can be conducted more easily.
These chapters may be omitted by students whose
knowledge is sufficiently advanced.

1.1 The Oscilloscope

The oscilloscope is one of the most important
measuring instruments used in electrical engi-
neering and electronics. It is used to graphically
display voltages u(t) with respect to time and can
thus be regarded as a 2-dimensional voltmeter. Fig 1.1-1: The function of the oscilloscope
A : Amplitude window
The signal is represented in the form of cartesian T : Time window
coordinates with the abscissa (x-axis) usually rep-
resenting the time scale and the y-axis the voltage
scale, e.g. msec./Div. and V/Div., see Fig. 1.1-1.
There are, of course, other voltage measuring in- a voltage which is strictly linear with respect to
struments such as the moving coil meter and the time. This voltage operates as the horizontal de-
digital voltmeter and one can ask why, then, the flector in a cathode-ray tube. The saw-tooth gen-
oscilloscope should be so indispensable for volt- erator is a component of the time base. The
age measurement. The answer lies in the fact that vertical deflection is controlled by the measured
normal pointer meters or digital meters are only signal itself. This results in a stationary represen-
suitable for measuring voltages that have a de- tation of the voltage signal on the screen when the
fined voltage – time curve. They can normally trigger is functioning correctly. It is interesting to
only be used to measure DC or harmonic AC volt- note that the screen displays only a small part of
ages. Each time a non-sinusoidal voltage is mea- the signal curve. When the time base is set at e.g.
sured with such an instrument, an incorrect T = 1 µs/Div. and the screen has a grid of ten divi-
reading results. However, for known signal curves sions, then a “picture window” of duration
the incorrect reading can be converted using a T = 10 · 1 ms = 10 ms is formed on the screen. In
correction factor. Naturally, for this the corre- real-time operation, the oscilloscope has, there-
sponding form factor of the signal must be fore, a slow-motion function. Processes which are
known. If the signal curve is unknown, then a much too fast to be observed by the human eye
voltage measurement can only be carried out us- can be made visible.
ing an oscilloscope. Here, we must distinguish
between two cases: 2. The voltage signal to be measured is either not
periodic, or has a very long period.
1. The voltage signal to be measured is non-sinu- The oscilloscope must then be operated in storage
soidal, but periodic with a “high” frequency. mode.
The oscilloscope is operated here in repeating Note:
real-time mode. This is the mode which is most It is not practical, within the scope of this book, to
often used. The principle is simple and will be give more than general guidelines about the use of
dealt with here only briefly, since it is assumed the oscilloscope. Please take definite setting in-
that the student is already familiar with it. A saw- structions from the manufacturer’s operating in-
tooth generator is started each time the periodic structions. With reference to the experiments
signal to be measured exceeds a level (trigger described in this book, it is recommended to use a
level), which can be set. This generator produces low-cost storage oscilloscope with 2 channels and

5
 TPS 7.2.1.1 Measurement Procedures 1.1

a bandwidth of 20 MHz. All the experiments have examples. The experiments can, however, to a
been carried out using such a device (LD Cat. No. great extent, be carried out using only a purely
575 29) and produced the results given here in the real-time oscilloscope.

Fig. 1.2-1: Construction of a spectrum analyzer

A: Signal path (amplitude section) B: Oscillator section
 Input amplifier/attenuator V1 $ VCO
Mixer % Saw-tooth generator
! Band filter C: Display unit
" Rectifier
# Output amplifier V2

6
 TPS 7.2.1.1 Measurement Procedures 1.2

1.2 The Spectrum Analyzer

The Spectrum Analyzer spondingly shifted along with it. The principle of
Signals can be equally validly described in terms the operation of the spectrum analyzer is shown in
of their spectrum as well as in terms of their Fig. 1.2-2. Its assembly is shown in Fig. 1.2-1.
behaviour with respect to time. Spectrum or fre-
quency analyzers are used in measurement tech- The Time Rule of Electrical Telecommunications
nology for the recording of signal spectra. These Technology
analyzers operate either in digital mode using When using analyzers according to the sweep fil-
mathematical algorithms such as e.g. the fast Fou- ter principle, one of the most fundamental laws of
rier transformation (FFT) or in analog as filter electrical engineering must be taken into consid-
banks or sweep analyzers. The latter principle is eration. According to the time rule of telecommu-
realised in the spectrum analyzer, training panel nications technology, the transient response of a
726 94 and should, thus, be investigated more low-pass filter system to a voltage jump becomes
closely. all the longer, the smaller its bandwidth b is. This
The analyzer is made up of a signal path (ampli- is also true for band pass filters. If pulses of very
tude component) A, an oscillator component B short duration are to be transmitted via a low-pass
and an indicating unit C (see Fig. 1.2-1). The har- filter (LP) then the output signal of the LP needs
monic signal supplied by the VCO is fed into the more time to reach the pulse amplitude, the
mixer together with the input signal. Depending smaller the bandwidth b of the filter is. The time
on the spectral quality of the input signal and the law of electrical telecommunications technology
oscillator frequency, an IF signal which lies ex- is similar to the uncertainty principle in atomic
actly in the pass range of the band filter appears at physics. It states that it is impossible to decrease
the mixer output . The AC voltage signal at the the duration of a signal as well as the bandwidth
output of the band filter is produced for various of the transmission channel for that signal at the
spectral components in the input signal at corre- same time. When tuning the VCO, the mixer out-
spondingly different frequencies of the VCO. If put signal is all the longer in the pass range of the
the VCO frequency is linearly dependent on its following band pass filter (BP), the slower the
control voltage, then it can be used for the X de-
flection of a cathode ray tube (or an XY-recorder).
In this way, the X axis also receives a linear fre-
quency division. The desired frequency-depen-
dent amplitude representation of the input signal
is displayed on the monitor when the VCO varies,
if, after rectification and corresponding amplifica-
tion, the output voltage of the band filter is con-
nected to the Y deflection of the beam tube (or
XY-recorder). Therefore, the spectrum analyzer
represents a process of the well-known super-
heterodyne principle used in radio technology,
whereby the band filter can be regarded as a spec-
tral window. The position of this window within
the frequency range is determined by the VCO
frequency. The width of the window is deter- Fig. 1.2-2: The principle of the spectrum analyzer
mined by the selected bandwidth of the band pass A: Amplitude transfer function of the bandpass
filter. (spectral window) with bandwidth b.
The centre frequency of the BP shifts
When the VCO sweeps the set frequency span, the with V = SPAN/T.
center position of the spectral window is corre- B: Arbitrary signal spectrum

7
 TPS 7.2.1.1 Measurement Procedures 1.2

Fig. 1.2-3 The time rule of electrical communications technology.

A: Input signal Pulse with duration T and amplitude a
B: Transmission channels with LP character fg1 < fg2
C: Output signals

change in frequency of the VCO is. The change in Where

frequency based on the time unit depends on: fmax : maximum frequency
1. The period T of the saw-tooth generator fmin : minimum frequency
(SCAN TIME) b : bandwidth of the filter
2. The absolute frequency range (SPAN) T : saw-tooth period SCAN-TIME
through which the VCO passes. The difference fmax – fmin is called frequency
SPAN.
If, for example, a spectrum analysis has to be car- Examples:
ried out in a wide frequency range and, in addition 1. SPAN = 200 kHz
to this, a very short saw-tooth period is selected, T = 1/25 s
then a very large change in frequency per time b = 10,000 Hz.
unit is the result. The mixer output signal passes Compare the adjustment possibilities of the spec-
the center frequency of the BP correspondingly trum analyzer 726 94. Using a real time oscillo-
quickly. According to the time rule, the bandwidth scope you can display the spectrum with
of the BP must now be selected “sufficiently” b = 10,000 Hz and SPAN = 200 kHz!
large, if the BP is to build up to the input ampli- The spectral resolution is:
tude, see Fig. 1.2-3. The spectral resolution capa- R = 20,
bility of the analyzer drops with a large bandwidth in other words: 20 spectral lines could theoreti-
of the BP. For this reason, the application of the cally be distinguished.
spectrum analyzer requires a constant compro- 2. SPAN = 1,000 Hz
mise between the spectral RESOLUTION and T = 20 s
fault-free amplitude reproduction. For the rela- ⇒ b = 33 Hz
tionship between SCAN-TIME T, bandwidth b ⇒ R = 30.
and the frequency window SPAN the following For a display on the oscilloscope we require a
approximation is valid: storage oscilloscope.
20
b= ( f max − f min ) Eq. 1.2-1
T

8
 TPS 7.2.1.1 Measurement Procedures 1.3

1.3 Measurement Assembly

Preliminary note – Using the frequency counter and the oscillo-

Here, an introductory investigation of the spec- scope, set a square-wave signal with an am-
trum and the time curve of a symmetrical square- plitude AR = 5 V and a frequency fR = 2 kHz
wave is made. This is done in order to practice the on the function generator. The TTL A input
handling of the equipment of the complete mea- of the frequency counter must remain con-
surement set-up which is used in the following nected to the analyzer by means of bridging
experiments. By complete measurement set-up plugs. Testing the signal frequency fR takes
we always mean the equipment which is required place via connecting leads, and, after switch-
to permit the investigation of signals in the time over of the counter, via the analog input.
and spectral domains. This means that such a 1. Draw the true-to-scale curve with respect to
work place always consists of a spectrum ana- time of the square-wave signal.
lyzer, oscilloscope and counter. The measurement 2. Record the spectrum of the square wave sig-
assembly is also a part of the experiment set-up. nal in the frequency range of approx.
The analyzer and the counter are connected to- 1.5 kHz...20 kHz. The experiment is divided
gether as a single unit. Therefore, it initially mea- into Points 2.1, 2.2 and 2.3. However, first
sures the VCO frequency of the analyzer (toggle read the following notes on using the spec-
switch in position TTL, rotary switch to freq. A). trum analyzer. The terms used are the same
If frequency measurements are carried out at other as in Section 1.2.
measurement points, the bridging plug connection
between counter and analyzer is to be removed if Operating instructions for the spectrum
necessary, or the analog input of the counter is to analyzer
be used after actuation of the toggle switch. A. Setting the signal path
It is advantageous to set the gains V1, V2 as high
Equipment: as possible, as this increases the sensitivity of the
1 Function generator signal path. However, avoid overdriving your sys-
0...200 kHz/230 V 726 96 tem (indicated when the LED OVER lights up).
1 Power supply unit ±15 V, 3 A 726 861
1 Spectrum analyzer 726 94 B. Setting the oscillator section
1 Frequency counter 0...10 MHz 726 99 Important operating instructions for the frequency
1 Multimeter M3E 531 57 tuning can be derived from the time law of elec-
1 Oscilloscope HM 205/2 575 29 trical communications engineering.
2 Probes 10:1 (switchable) 575 231 The VCO of the spectrum analyzer is controlled
1 XY Recorder optional 575 662 by the saw-tooth generator. The amplitude and
1 Set of bridging plugs 501 511 scan time of the latter are set using the control el-
1 Set of connecting leads 500 414 ements SCAN-MODE (fmin, fmax) and SCAN-
TIME. The choice of the correct run time must
Objectives: be based on the time law and can, therefore, not
The student should: be fixed. It is dependent on both the chosen
– attain practice in handling the equipment. bandwidth b/Hz and on the frequency window
– recognize the effect of the time law of elec- SPAN = fmax – fmin. In your own experiments use
trical communications technology on prac- either Eq. 1.2-1 or trial and error. The LD experi-
tical measurement. ment instructions always offer corresponding set-
– recognize the structure of line spectra. ting notes.

Experiment procedure Setting the frequency window (SPAN)

– Assemble the experiment according to plug- First, the lower frequency limit is set. This is car-
in diagram MP-1. ried out in SCAN-MODE fmin with the corre-

9
 TPS 7.2.1.1 Measurement Procedures 1.3

sponding control. Then, the upper frequency limit set parameters must be changed, this is only pos-
is set in SCAN-MODE fmax. The frequencies can sible in SCAN-MODE STOP. In this mode, the
be read from the connected counter. Here, locking of the RESET function is cancelled. In ad-
SPAN = fmax – fmin holds true. dition, the analyzer can be operated manually in
this setting by means of a toggle switch. The setting
Starting the analyzer UP of the toggle switch makes the VCO run in the
In SCAN-MODE RUN, the VCO will now sweep direction fmax, while the button setting DOWN re-
the set frequency range once. An LED indicates duces the frequency. When fmax has been reached,
when the upper frequency limit has been reached. the analyzer carries out an automatic zero balance
The VCO stops at fmax, so that an inadvertent over- (Auto-Zero). During the Auto-Zero phase, the in-
writing of the spectrum is avoided when the XY put signal is switched off internally.
recorder is used. For the same reason, a premature
reset of the VCO to fmin during the sweep cycle is Note:
also not possible. Reset in RUN-mode is, therefore, Since the XY recorder cannot be used for a sweep
only possible after fmax has been reached. However, period T = 1/25 s (see C), the set frequency span
should one desire to interrupt the spectrum record- is here repeatedly swept in RUN-MODE. This
ing during the sweep cycle, which can last up to makes it possible to use the spectrum analyzer
160 s, because e.g. one notices at an early stage that also as a sweep generator.

Plug-in diagram MP-1

10
 TPS 7.2.1.1 Measurement Procedures 1.3

C. Connection of the display unit The X+ input of the recorder is connected to

External measurement instruments are used as the X socket of the analyzer, while the Y+
display units. input is connected to the analyzer’s output.
1. Analog voltmeter The recorder inputs X– and Y– are con-
2. XY recorder nected to each other and to ground. Both
3. Storage oscilloscope axes of the recorder must be calibrated. The
Special features of the different display units for X-axis is set to fmax, and the Y-axis is
the experiment procedures are closely dealt with matched to the greatest spectral amplitude.
in more detail under Points 2.1, 2.2 and 2.3. You may need to practice this procedure
first. It is sensible to calibrate the recorder
2.1 Application of the spectrum analyzer as a so that the spectrum you are looking for is
frequency selective voltmeter (advantageous reproduced in a size which conforms to your
for quantative measurements). recorder paper.
Connect an analog voltmeter, 10 V DC, to A. Setting the signal path
the analyzer output (see plug-in diagram V1/2 = 1
MP-1). In automatic operation, switching over the gain V2
A. Setting the signal path during the scan process does not improve the mea-
V1 =1 surement results.
V2 = 2, 5, 10 b = 500 Hz.
V2 is matched to the magnitude of the re- B. Setting the oscillator section
spective spectral line and selected as large Frequency range fr : 20 kHz
as possible. SPAN = fmax – fmin : 20 kHz...500 Hz approx.
b = 500 Hz. SCAN-TIME T : 160 s
B. Setting the oscillator section The analyzer cycle is started when the unit is
Frequency range fr : 20 kHz switched to SCAN-MODE RUN. Record the
SPAN = fmax – fmin : 20 kHz...500 Hz approx. spectrum in scan mode RUN. Do not forget to
SCAN-TIME T : 20 s lower the pen (Pen Down). Unlike manual opera-
– Now, record the spectrum of the square- tion, the scan procedure is not stopped, but rather
wave signal by starting the VCO in SCAN carried out to the end in RUN mode.
MODE RUN. In the range of spectral en-
ergy, the output signal shows a short dip. 2.3 Using the storage oscilloscope as a display
Stop the VCO in SCAN-MODE STOP and unit.
control it UP/DOWN around the spectral We are using plug-in diagram MP-1 again
line in manual operation using the here.
pushbutton UP/DOWN. In manual opera- Set the analyzer as when using the XY-re-
tion, the VCO runs with a fixed scan time of corder. Set the scan time to, e.g.
T = 640 s when the pushbutton is actuated. T = 20 s, T = 40 s.
Stop the scanning process when you find a The easiest way to proceed is to record the
spectral line. You then have time to read the spectra with the storage oscilloscope in
spectral amplitude S(n) from the voltmeter. ROLL mode. This eliminates the difficulties
Enter these amplitude values S(n), the sweep of triggering. The storage time base is set so
indices n and the corresponding frequencies that the period is greater than or equal to the
f in Table 1. Also note down there the SCAN TIME set at the analyzer. The
analyzer settings. This should always be analyzer output is connected to the Y-input
done, as it allows you to check your meas- of the storage oscilloscope. Try to optimize
urement results by calculation at a later date. your use of the screen area for displaying
Draw a graph of the spectrum. the spectra by selecting a suitable Y-gain.
When the spectrum you are looking for is
2.2 Automatic operation of the spectrum displayed completely on the screen, you can
analyzer with the XY-recorder cancel the ROLL function by pressing the
Connect the spectrum analyzer as shown in button marked SINGLE. The contents of the
plug-in diagram MP-1. screen are then “frozen”. If the spectrum is

11
 TPS 7.2.1.1 Measurement Procedures 1.3

to be stored long-term, it is recommended

that the memory contents be saved with the Signal: Analyzer:
HOLD function. Draw the spectrum in a AR = V V1 =
diagram. f0 = Hz b = Hz
fr = kHz
3. Check the time law of electrical communi- T = s
cations technology. Use either an XY-re- SPAN :
corder or a storage oscilloscope as a display
unit. Repeat Experiment 2 using the follow- Measurement Theory
ing settings: f S(n) SR(n) SR(n)
n V2
b = 100 Hz / 10 Hz kHz V V V
T =  5 s, 40 s, 160 s
What do you observe?

4. Describe the envelope curve of the spectrum

of a symmetrical square-wave signal? Draw
the curve you expect as a dotted line in dia-
gram 1. What general statement can you
make about the spectrum?

12
 TPS 7.2.1.1 Frequency Multiplication 1

1 An Analog Method of Frequency Multiplication

Equipment: The 4-quadrant characteristics field is the joint

1 Power supply ±15V, 3 A 726 86 representation of 4 characteristics:
1 Frequency counter 0 - 10 MHz 726 99 1st Quadrant (I) : Output characteristics
1 Function generator field
0 ... 200 kHz/230 V 726 96 2nd Quadrant (II) : Current transfer
1 Spectrum analyzer 726 94 characteristic
1 Frequency multiplier 736 25 3rd Quadrant (III) : Input characteristic
1 Oscilloscope 531 29 4th Quadrant (IV) : Voltage reaction
2 Experimentierkabel 100 cm 500 444 (not drawn)
1 Analog-Multimeter C. A. 406 531 16 The operating point (OP) of the transistor is deter-
2 Probes, switchable 10:1 575 231 mined using the base potential dividers R1 and R2.
2 Sets of bridging plugs 501 511 The controlling signal u1(t) is superpositioned on
1 Frame the bias voltage. Consequently, the resulting tran-
sistor base-emitter voltage UBE also takes on the
Objectives curve form u1(t) with respect to time. From the
1. Understanding systems with non-linear input characteristic, we can see that the corre-
characteristics sponding base current IB has a curve which devi-
2. Linear and non-linear distortions ates from u1(t). The distortions occurring in the
3. Principle of analog frequency multiplication base current are dependent on the position of the
operating point on the input characteristic and on
Theoretical background the level of control caused by the input signal. As
There is generally a complicated relation between a current controlled component, the bipolar tran-
the input and the output signal for a non-linear sistor converts base current variations into corre-
quadrupole. This can be described by a curved sponding variations of the collector current with a
characteristic. An example of a non-linear qua- current gain factor β. The variations in the curve
drupole is a transistor at a high control level. The of the collector current produce voltage variations
response of the transistor becomes clear through across the load resistor RL which are decoupled
consideration of the 4-quadrant characteristics via capacitor C2. There is, therefore, an amplified
field (Fig. 1-1). This is, for example, based on an voltage u2(t) across C2, whose curve with respect
amplifier in a common emitter according to to time deviates from the controlling signal u1(t).
Fig. 1-2. If the controlling signal is harmonic and periodic
with period T0, the output signal has the same pe-
riod of oscillation, but is non-harmonic (dis-
torted). The output signal can, therefore, be

Fig. 1-1:  The 4-quadrant characteristics field Fig. 1-2:  Transistor amplifier in common emitter

13
 TPS 7.2.1.1 Frequency Multiplication 1

The power series contains an infinite number

of components of increasing powers. As long as
u1/uT ≤ 1, the following approximation holds:
I B ( UB + UB )

U  u1 
= I B 0 exp.  B  1 +  Eq. 1-3
 UT  uT 
  

In this case, the transistor operates as a linear

amplifier.
Fig. 1-3: The input characteristic For larger input voltages u1, ever more compo-
nents of higher orders in the series expansion
must be taken into consideration. Even the quad-
ratic component has a frequency multiplying or
subjected to Fourier analysis. Spectral component modulating effect. (See advance question 6). In
frequencies which are an integral multiple of the general, if a characteristic of the nth degree is con-
fundamental frequency f0 = 1/T0 are expected in trolled with a harmonic signal of frequency f0, fre-
the spectrum of the distorted output signal. The quency multiplication up to the nth multiple of the
structure of the output spectrum is largely depen- input frequency f0 is obtained in addition to sum
dent on the choice of OP, on the input characteris- and difference frequencies.
tic and the degree of control. An exponential input
characteristic form is assumed for a quantitative Note: Up until now, our discussion has been
investigation. The input characteristic is repre- limited to the exponential characteristic
sented in Fig. 1-3. Let us assume that it has an ex- of the base-emitter junction in the for-
ponential curve form given by Eq. 1-1: ward direction. The transition from the
forward range to the blocking range caus-
 U +u  es an additional strong non-linearity in
I B = I B0 exp.  B 1
 Eq. 1-1
 uT  the characteristic; we will see in the fol-
  lowing experiment that this to a large ex-
IB : Base current tent accounts for the harmonic compo-
IB0 : Bias current at OP nent.
UB : Base bias voltage
u1 : Input voltage The Frequency Multiplier
uT : Temperature voltage The operating point (OP) of the transistor is fixed
The characteristic described by Eq. 1-1 can be ex- in the range 0 ≤ UB ≤ 0.7 V by the base bias volt-
panded around the OP using a Taylor series: age. The angle of current flow of the emitter cur-
rent is controlled by the choice of the operating
I B ( UB + U B ) point. In some cases the transistor is only con-
trolled briefly by the input signal if the base bias
U  u  current is small. The input characteristic sup-
1
= I B 0 exp.  B  exp.  
 UT   uT  presses the negative input voltages completely,
   
and distorts positive input voltages to a greater or
U   lesser extent. In each case, spectral components of
u1 u2 u3
= I B0 exp.  B  1 + + 1 2 + 13 + ... a higher order occur in the emitter current (collec-
 uT  uT 2uT 6uT  tor current). The resulting harmonic oscillations
  
are selected according to frequency using an anti-
Eq. 1-2 resonant circuit.

14
 TPS 7.2.1.1 Frequency Multiplication 1

When the frequency multiplier is in operation, For the arrangement of the voltages in the fre-
the oscillating circuit is fixed to the frequency quency multiplier, see Fig. 1-4. Here:
f0 = 100 kHz. The selection of the desired har- U1 : Input voltage
monics is carried out by changing the frequency UE : Voltage across the emitter resistor
f1 of the input signal. UC : Voltage between collector and ground
Example: U2 : Output voltage
Harmonics with frequencies: Due to unavoidable characteristics dispersion,
fn  =  nf1 n  =  1, 2, 3.. amplitude measurements using different devices
can be expected for f1 = 10 kHz when a suitable may vary. This is unimportant and does not affect
choice of the OP and the amplitude of the input the student’s ability to understand the experiment.
signal is made.
Thus, the oscillating circuit tuned to
f10 = 10 f1 = 100 kHz selects the 10th harmonic.

Fig. 1-4:  Frequency multiplier

15
 TPS 7.2.1.1 Frequency Multiplication 1

Plug-in diagram 1-1

Advance Questions Experiment procedure

– Assemble the experiment according to plug-
1. What properties does the output signal of a
in diagram 1-1.
non-linear system have when a harmonic,
1. Investigate the output signal U2 (transformer
periodic signal (e.g. sinusoidal) is fed into
output) for the occurrence of harmonics in
its input?
relation to the OP and the control signal U1.
2. How are non-linear distortions character- 1.1 Input voltage U1 sinusoidal, 100 kHz, (50-100)
ised? mVPP, operating point set to right limit.
3. Distinguish between linear and non-linear - Display the output signal U2 (10:1 probe
distortions. must be used) and the input signal U1 on the
4. Draw, in the 4-quadrant characteristics field oscilloscope. Oscilloscope settings:
of a transistor in common emitter, the signal TB: 20 ms/Div., Y1, Y2: matched to the cor-
curve for a high sinusoidal control level. responding signal size.
What happens to the output signal? – Adjust the frequency of the resonance cir-
5. What is the total harmonic distortion of a cuit until U2 reaches a maximum. For fine
transistor current dependent on? tuning: U1 and U2 must have a phase-shift of
6. The transfer characteristic of a pn-FET has a 180°.
quadratic curve for UGS ≥ UP. – Now, change the amplitude of the input sig-
nal U1 so that the output signal U2 is almost
  
2  10 VPP. (This typically requires a voltage
U U1 ≈ 60 mVPP.)
ID = I 0  1 −  GS  
  U P 
  Draw the voltages U1 and U2 for the fre-
  quency multiplier, which has been tuned to
Ignoring the fixed component I0, we have: the input signal in the described manner.
2 Why does a phase shift of 180° occur when
U  the frequency adjustment is exact? How
ID ≈ − I 0  GS  large is the resonance impedance of the os-
 UP 
  cillating circuit when the emitter resistance
RE is 330 Ω?
What happens when UGS is the sum of two har- Starting with this pre-setting, now change
monic signals s1(t) and s2(t)?) only the frequency f1 of the input signal U1.
s1 ( t ) = A1 cos( 2 π f1 t ) Set the following frequencies one after the
other:
s 2 ( t ) = A2 cos( 2 π f 2 t )

16
 TPS 7.2.1.1 Frequency Multiplication 1

f1 = 100.0 kHz – Vary the frequency of the resonance circuit

f1 = 50.0 kHz until U2 reaches a maximum.
f1 = 33.3 kHz – Now alter the amplitude of the input signal
f1 = 25.0 kHz U1 so that the output signal U2 is 10 Vpp (you
f1 = 20.0 kHz will usually need a voltage of
f1 = 16.6 kHz U1 ≈ 1200 mVpp).
f1 = 14.3 kHz Determine the amplitude and frequency of
f1 = 12.5 kHz the output signal U2 as a function of the fre-
f1 = 11.1 kHz quency of the input signal U1. Once again
f1 = 10.0  kHz carry out your measurements for the same
Measure the amplitude and frequency of the frequency settings of the input signal listed
output signal U2 and enter the values into in 1.1:
Table 1. Use Table 3.
Draw the curve of the output signal for
Note: The measurements become much f1 = 20 kHz and for f1 = 16.6 kHz.
easier to make when, after f1 has been set,
the resonance circuit is carefully reset. 2. Connect the output of the frequency multi-
plier to the input of the spectrum analyzer,
1.2 Sinusoidal input voltage, i.e. connect the spectrum analyzer in parallel
100 kHz, (500...800)  mVPP, operating point to the oscilloscope. Recalibrate the spectrum
setting control to center position analyzer as described under Point 1.3. To
– Again display signals U1 and U2 on the os- obtain U2 = 10 Vpp you will need an input
cilloscope. signal U1 ≈ 1700 mVpp.
– Adjust the frequency of the resonance cir- Set the frequency of the input signal to
cuit until U2 reaches a maximum. f1 = 20 kHz.
– Now change the amplitude of the input sig- Record the following spectra in a purely
nal U1 so that the output signal U2 is 10 Vpp qualitative manner (in this case, the storage
(you will usually need a voltage of oscilloscope or the XY-recorder is used as
U1 ≈ 600 mVpp). the display unit, cf. Chapter 1.3)
Once again, determine the amplitude and 1. The spectrum SU2 of the output voltage U2
frequency of the output signal U2 as a func- 2. The spectrum SU1 of the input signal U1.
tion of the frequency of the input signal U1. The spectrum analyzer must be connected in
Carry out your measurements for the same parallel with the input of the frequency mul-
frequency settings of the input signal listed tiplier for 2.
in 1.1: Analyzer settings:
Use Table 2. V1 = 10
Draw the curve of the output signal for V2 = 1
f1 = 33.3 kHz and for f1 = 25 kHz. Draw the b = 500 Hz
envelope curves in dotted lines. fr = 200 kHz
1.3 Sinusoidal input voltage, 100 kHz, SPAN : 5 kHz...110 kHz
(1000...1500)mVpp, T = 160 s
operating point setting control to full left de- Draw the spectra.
flection. 3. Interpret your results
– Display the signals U1 and U2 on the oscillo- Note:
scope again. The tables can be found in the appendix.

17
 TPS 7.2.1.1 Frequency Multiplication 1

18
 TPS 7.2.1.1 Frequency Multiplication 2

2 Frequency Synthesis, PLL

Equipment: – Speed control

1 Power supply ±15 V, 3 A 726 86 – Tracking filters
1 Frequency counter 0...10 MHz 726 99 We can see from this list that the PLL is particu-
1 PLL frequency multiplier, larly “at home” in telecommunications technol-
series 2 or better (see below) 736 26 ogy. There it is used, e.g. for picture
1 Oscilloscope 531 29 syncronization in TV sets. If this circuit should
2 Experimentierkabel 500 411 fail, the TV world would degenerate into chaos.
2 Probes, 10:1, switchable 575 231 Headless figures would run along the top of their
1 Analog multimeter, e.g. 531 57 TV screen, their legs jerking helplessly; mean-
2 Sets of bridging plugs 501 511 while, the missing heads, separated from their tor-
1 Panel frame sos by a big black bar, would populate the bottom
of the screen. This well-known effect occurs
Objectives: whenever there is a synchronization fault in the
The function and design of the phase-locked loop vertical deflection. To ensure perfect reproduction
(PLL) of the picture, the vertical deflection in the trans-
Design of a frequency synthesizer mitter and the receiver must coincide completely
Note: with respect to frequency and phase.
It is only possible to feed the bipolar signals of A PLL is used to achieve the synchonization. It is
the function generator to the phase detector with a control circuit which has the task of matching
series 2 or better; cf. Point 1 of the experiment the phase and frequency of an oscillator to a refer-
procedure. ence oscillation. In steady-state operation, the os-
cillator of the PLL will oscillate with the same
1. PLL applications frequency as the reference signal. Depending on
Frequency multiplication can be carried out using the type of PLL, the phase shift will either be 0°
analog and digital methods. The analog method or 90°. In either case, the phase shift will not
utilizes the generation of harmonics from the con- change in time.
trol of a non-linear characteristic and subsequent
frequency selection (see Chapter 1). The result is 2. Design and function of the PLL
a harmonic output signal in the desired frequency As the PLL is a control circuit, it can be explained
band which is relatively clean with respect to the using block circuit diagrams such as are common
spectra. This method is used, for example, for fre- in control technology. The voltage-controlled os-
quency conversion in transmitters. cillator (VCO) is the controlled loop, the phase
The digital method of frequency multiplication detector (PD) compares set and actual values, the
and frequency stabilization is referred to as fre- loop filter is the controller and actuator.
quency synthesis. In this case, the output is gener-
ally a quartz-stabilized square-wave signal with a Steady-state operation of a PLL
frequency that is a whole or even rational multiple Let us assume that an input signal s1(t) with the
of the fundamental frequency. Means of selection frequency f1 (in terms of control technology, this
such as those used in analog transmission meth- is the reference variable W) is applied to the phase
ods are not required. detector. The VCO will probably not be so
In this chapter we will carry out frequency synthe- friendly as to oscillate at the same frequency right
sis using a phase-locked loop (PLL). Its function away. Its frequency f2 will thus initially be differ-
is to be examined in detail in the experiments. In ent from f1. Thus an AC voltage is present at the
addition to frequency synthesis, PLL circuits are output of the phase detector with a frequency
also used in many technical areas. Some examples equal to f2 – f1. This AC voltage is then led to the
are: input of the VCO via the loop filter. The VCO
– Demodulator circuits responds to an AC voltage at its input with a cor-
– Synchronisation circuits responding frequency change in its output signal.

19
 TPS 7.2.1.1 Frequency Multiplication 2

Fig. 2-1: Principle of the PLL

Loop filter
Phase detector
!VCO

Thus, the VCO is frequency-modulated by the AC cally, i.e. uncontrolled. The mean phase position,
voltage. The changing frequency of the VCO is however, remains constant. Noise superposi-
immediately recognized by the phase detector PD. tioned on the wanted signal thus does not influ-
If we are lucky, the VCO generates the “right” ence the PLL in the first approximation.
frequency f1 for a brief moment. This causes a DC
voltage to appear at the PD output. This too is led Dynamic behavior of the PLL
to the VCO via the loop filter, where it causes the Up to now we have examined the PLL in steady-
VCO to generate a fixed oscillation of the input state operation only. The reference signal has the
signal with the frequency f1. The VCO thus “locks constant frequency f1, and the PLL was fortu-
into” the frequency of the input signal. The con- nately able to lock into this frequency. In this
trol-circuit design of the PLL obviously ensures case, its task was to control minor disturbances. In
that each deviation of the VCO frequency from examining the dynamic behavior, we are inter-
the input signal frequency results in an automatic ested in the ability of the PLL to lock into the ref-
resetting of the VCO. The controller adjusts the erence signal at all, which, under certain
VCO until the VCO frequency again matches the circumstances, can differ greatly in terms of fre-
input frequency. The controlled loop exhibits in- quency when the unit is switched on. It is obvious
tegral behavior. For this reason even the smallest that it cannot accomplish this for every frequency.
of frequency differences leads in time to consider- The frequency range in which the frequency f1 of
able phase shifts which control the VCO. The the input signal (reference signal) must lie if the
phase shift Φ causes a voltage UΦ after the PD. PLL is to lock into it is called the capture range
This is led to the input of the loop filter. The phase of the PLL. The size of the capture range depends
shift Φ thus directly influences the DC control on the time constant TI and the gain Ap of the loop
voltage Uf of the VCO. There is normally a linear filter (controller). If the PLL has already locked
relationship between Φ and the output voltage of in, the VCO frequency f2 will follow the
the PD: frequency f1 of the input signal across a particular
UΦ = kΦΦ Eq. 2-1 range. The frequency range in which f1 may
change without the PLL reaching its tracking
Noise behavior of the PLL limit is called the hold range. The hold range is
In its locked state, the phase detector determines larger than, or at least as large as, the capture
the phase difference Φ between the VCO and the range. It depends on the loop gain and the control
reference signal. If the reference signal contains characteristic of the VCO.
noise, the phase difference will change stochasti-

20
 TPS 7.2.1.1 Frequency Multiplication 2

3. Components of the PLL

The function blocks of the PLL shown in Fig. 2-1
can be assembled with discrete parts or realized as
ICs. In the PLL-ICs commonly in use today
(e.g. CD4046, NE560), the loop filter is realized
in the form of a passive RC lowpass and must be
provided externally, while the other two compo-
nents, the VCO and the PD, are integrated on a
single common chip.
The VCO
At low frequencies, harmonic oscillators based on Fig. 2-2:   Realizing a loop filter as a passive
two cascaded integrators and one inverter are PI controller
used. Triangular-square-wave generators or simi-
lar are equally possible. For higher frequencies,
emitter-coupled multivibrators or LC-oscillators
are preferred. The important thing to know about
a VCO is the control characteristic. This describes
the relationship between the input voltage Uf and dividers R1 and R2. R1 and C are mainly respon-
the output frequency f2 at the VCO. For a linear sible for the time constant TI of the I component.
control characteristic, we obtain a relationship For R2 ≤ R1 the following is approximately true:
between f2 and Uf according to Eq. 2-2, where: R2
k : VCO constant (unit: Hz/V) AP = TI ≈ 2 π f R1 C Eq. 2-3
R1 + R2
f0 : VCO zero-current frequency with no
control voltage present If the P component is too large, the tendency of
f2  =   kUf + f0 Eq. 2-2 the PLL to oscillate increases. Too large a time
constant makes the loop slow. The art of con-
The phase detector structing an otherwise fully integrated PLL is in-
The behavior of the PLL is significantly deter- deed limited to the correct dimensioning of the
mined by the phase detector. If the phase detector loop filter in many cases.
is realized as a circuit in analog technology, e.g.
as a multiplier functioning as a synchronous recti- 4. Frequency synthesis using the PLL
fier, then the entire PLL is considered analog. frequency multiplier
Digital PLLs have, for example, sample and hold If a frequency divider with the divisor N2 is placed
elements, EXOR gates or flip-flop circuits as between the VCO output and the PD, the VCO
phase detectors. The simplest form of digital PD locks into the frequency f2 = N2 f1. If an additional
is certainly the EXOR gate. It is comparable to an voltage divider with a divisor of N1 is connected
overdriven multiplier circuit. between the input of the PLL and the second input
of the PD, the input frequency f1 is reduced by the
The loop filter factor of this divisor. According to Fig. 2-3, the
The type of loop filter used depends on the PD. If, following is true for the relationship between the
for example, it takes the form of a sample-and- input frequency f1 and the VCO frequency f2:
hold element, the output voltage changes in
jumps. For this reason, the loop filter will be de- f1 f2
=
signed as a PI controller with no D-component. A N1 N2
common controller circuit with PI characteristic is Eq. 2-4
shown in Fig. 2-2. This is a passive lowpass cir- N2
f2 = f
cuit. The P component Ap is formed by the voltage N1 1

21
 TPS 7.2.1.1 Frequency Multiplication 2

Fig. 2-3: Principle of frequency synthesis

Thus, using this relationship, any quartz-stabi- – Connect the VCO output directly to the
lized, rational frequency ratios can be derived “lower” input of the PD using a bridging
from a reference oscillation with the frequency f1. plug.
Display the signal at the VCO output on
The PLL frequency multiplier your oscilloscope. Use the corresponding
The training panel contains the following mod- control to adjust the loop filter so that the
ules: PLL locks in.
1. Clock oscillator, square-wave output signal, The PLL is locked in when the VCO out-
τ = 50%, f1 = 100 kHz, unipolar 0/+10 V. put signal has a symmetrical duty cycle
2. Phase detector constructed as sample-and- and has approximately stationary slopes.
hold element. – Measure the VCO frequency using the
3. Loop filter with PI characteristic (lowpass). counter (toggle switch in position: ANA-
The P component is adjustable. LOG). Connect the multimeter to the VCO
4. VCO input and set the range to 3 V DC. Note
5. Adjustable frequency divider with each value for the VCO frequency f2 and
N2 = 2...10 control voltage Uf in Table 1. You must
vary the frequency f1 of the input signal in
Advance questions the range from 20 kHz...200 kHz in steps
1. Draw the function circuit diagram of a PLL. of approx. 20 kHz..
What functions do the individual elements Note: The equation f1 = f2 is valid only when the
have? VCO is constantly locked in. Sketch the
2. Name some possible ways of realizing a graph of the control characteristic, i.e. f1
VCO and a phase detector. versus Uf.
3. Describe the lock-in process. What is meant What happens when you disconnect the
by capture range and hold range? function generator?
4. How must the PLL circuit be expanded in
order to be able to use it for frequency syn- 2. Investigate the tracking behavior of the
thesis? PLL. You can vary the input frequency,
5. Where else are PLL circuits used? among other means, by switching the fre-
quency decade on the function generator.
Experiment procedure Observe whether the locked-in PLL loses its
Set up the experiment according to plug-in dia- synchronization; in other words, draw your
gram 2-1. conclusions regarding the hold range.
1. Record the control characteristic of the VCO Test whether the PLL is able to lock into the
– Apply an external square-wave signal to respective input frequency. In doing so, you
input 1 of the PD from the function genera- are investigating the capture range of the
tor with f1 = 100 kHz and 20 Vpp. PLL. To do this, it is necessary to briefly

22
 TPS 7.2.1.1 Frequency Multiplication 2

break the connection between the function ponent. Once again, display the signal
generator and the PD. Vary the input fre- curves at the VCO input and output on your
quency as much as is necessary or possible. oscilloscope, but this time with a greater
Note: time resolution (TB: 1 ms/Div.). Draw the
You may need to adjust the loop. signals in a diagram. Compare your results
3. Replace the function generator with the here with those obtained in Point 4 and in-
clock generator of the training panel. Using terpret them.
the oscilloscope, display the signals at the 6. Assemble a frequency synthesizer. To do
two inputs of the PD: Y1,2: 0.5 V/Div., this, connect the adjustable frequency di-
TB: 2 ms/Div. Adjust the PLL again until it vider in the feedback line from the VCO
locks in. Measure the frequency f1 of the output to the PD input. Set the divider ratio
quartz oscillator. Draw the input signal and from N = 2 to N = 10 in consecutive order.
the VCO output signal (the two oscilloscope Display the signal at the VCO output on the
signals) in a diagram. What do you observe? oscilloscope and, if necessary, adjust the
4. Use your oscilloscope to display the curves loop filter so that the PLL remains locked-
of the signals at the VCO input and VCO in. Measure the frequency of the oscillator
output. Draw these signals in a diagram. using the counter (gate time: 1 s). Measure
5. Investigate the PLL in its unlocked state. To the frequency of the VCO output signal with
do this, start from the settings given in Point the counter for all nine divider settings using
3 and move the adjuster control of the loop the same time resolution of the counter.
filter very slightly to the left. Take another Record your results in Table 2. Display the
look at Fig. 2-2. The adjuster control of the signals at the inputs of the PD for all divider
loop filter acts on resistor R2. When the con- settings, i.e. the reference oscillation and the
trol is at its right limit, then R2 = min. Ac- output signal of the frequency divider.
cording to Eq. 2-3, the gain of the P What do you observe?
component is also a minimum. Correspond-
ingly, when the control is at its left limit, Note
R2 = max., i.e., maximum gain of the P com- The tables can be found in the appendix.

Plug-in diagram 2

23
 TPS 7.2.1.1 Frequency Multiplication 2

24
 TPS 7.2.1.1 Solutions MP

Measurement set-up
Results
1. 2.1 Table 1:
Spectrum of a symmetrical square-wave signal
Signal: Analyzer:
AR = 5 V V1 = 1
fP = 2,0 kHz b = 500 Hz
fr = 20 kHz
T = 20 s

Measurement Theory
f S(n) SR(n) SR(n)
Curve of a square-wave signal with respect to time n V2
Parameters: AR = 5.0 V kHz V V V
fR = 2.0 kHz 1 2 6.6 1 6.6 6.37

3 6 4.2 2 2.1 2.12

Note: 5 10 6.6 5 1.32 1.27

The set square-wave amplitude is proportional to
the measurement result. As the amplitude cannot 7 14 4.7 5 0.94 0.91
be determined more precisely than ±5% with the
9 18 3.6 5 0.72 0.71
oscilloscope, we must accept deviations of this
magnitude.
The amplitude SR(1) of the fundamental oscilla-
tion is greater than the square-wave amplitude mation of the spectrum of a symmetrical square-
AR by a factor of 4/π = 1.27. The law for the for- wave signal is:
4 AR
SR ( n ) = n = 1, 2 , 3,...
π ( 2 n − 1)

Spectrum of a square-wave signal

Signal : AR = 5.0 V Analyzer: V1 = 1 V2 = 2, 5, 10
fR = 2.0 kHz b = 500 Hz T = 20 s
fr = 20 kHz Span : 500 Hz...20 kHz
Display in manual operation

25
 TPS 7.2.1.1 Solutions MP

2.2

Spectrum of a square-wave signal

Signal: AR = 5.0 V Analyzer: V1 = 1 V2 = 1
fR = 2.0 kHz b = 500 Hz T = 160 s
fr = 20 Hz Span : 500 Hz...20 kHz
Display with XY-recorder

2.3 cal communications technology. The filters

of the analyzer no longer respond.
If mechanical measurement instruments,
which are subject to inertia, are used as dis-
play instruments, e.g. a multimeter or an
XY-recorder, the low-pass behavior of the
whole system is increased still further. The
needle or pen shows practically no deflec-
tion. It is thus a good idea when using an
XY recorder to select the scan time one step
larger than is required by the time law for
Spectrum of a square-wave signal the analyzer filters.
Signal : AR = 5.0 V 4. The envelope in the spectrum of the sym-
fR = 2.0 kHz
Analyzer : V1 = 1 V2 = 1 metrical square-wave signal has the hyper-
b = 500 Hz T = 40 s bolic curve 1/f.
fr = 20 kHz Span : 500 Hz...20 k Hz
Representation on storage oscilloscope
Further general comments on the spectrum of the
symmetrical square-wave signal:
3. The reduction of the bandwidth b narrows – The spectrum has a line structure.
the spectral window. The manual frequency – The spectral lines occur at odd multiples of
setting thus becomes more and more diffi- the fundamental frequency fR (3 fR, 5 fR,
cult although the spectral lines become more 7 fR,...)
pronounced. Even when a storage oscillo- – The amplitudes SR(n) are inversely propor-
scope is used as an (almost) inertialess dis- tional to the odd multiples of the funda-
play instrument, the spectrum cannot be mental frequency (1/3, 1/5, 1/7,...).
reproduced at full amplitude. The reason for – The amplitudes SR(n) are proportional to
this is to be found in the time law of electri- the amplitude AR of the square-wave signal.

26
 TPS 7.2.1.1 Solutions 1

An Analog Method of Frequency Multiplication

Answers
1. The output signal will be periodic but not har- 6. U GS = s1 + s 2
monic. As the distortions occurring here do = A1 cos( 2 π f1t ) + A2 cos( 2 π f 2 t )
not depend on the frequency of the input sig-
2
nals, they are termed non-linear distortions. A U 
non-linearly distorted output signal always ID ≈ − I 0  GS 
 UP 
contains new spectral components which  
were not present in the original signal. I0
2. The degree of non-linear distortion is de- = − ( s1 + s 2 ) 2
scribed by the distortion factor k. U P2
3. Linear distortions are frequency dependent.
( s1 + s 2 ) 2 = ( s12 + 2 s1 s 2 + s 22 )
We distinguish between attenuation distor-
tion and delay distortion. Attenuation distor- = A12 cos 2 ( 2 π f1 t ) +
tion is described using the amplitude
2 A1 A2 cos( 2 π f1 t )cos( 2 π f 2 t ) +
frequency response A(f) of the system under
study. Delay distortion is described by the A22 cos 2 ( 2 π f 2 t ) )
frequency-dependent phase difference. Non-
linear distortions depend on the level of the
input signal but not on its frequency.
With: cos x cos y = 1/2 [cos(x + y) + cos
(x–y )] we obtain:

We can recognize the following components

in the drain current:

I0  A12 A12
ID ≈ −  2 + 2 cos( 2 π 2 f1 t ) +
U P2 

A1 A2 cos[ 2 π ( f1 + f 2 ) t ] +
A1 A2 cos[ 2 π ( f1 − f 2 ) t ] +

A22 A22 
+ cos( 2 π 2 f 2 t ) 
4. Due to the bent input characteristic curve, 2 2 
distortions in the base current occur for high
control voltage at the base. With corre- 1. DC voltages proportional to
sponding gain, these cause a distorted col-
lector current. The collector current flows A12 A22
,
across the load resistor RL; thus, the ampli- 2 2
fied non-linearly distorted output voltage 2. Spectral components with twice the signal
can be taken from this resistor. frequencies 2 f1 and 2 f2:
5. The distortions are dependent on the curve
of the input characteristic, the degree of con- A12 A22
trol and the selection of the operating point cos( 2 π 2 f1 t ),cos( 2 π 2 f 2 t )
2 2
(OP). For example, if the operating point is
3. Spectral components with the sum and dif-
fixed at 0 V, the transistor blocks com-
ference frequencies f1 + f2, f1 – f2:
pletely for negative input signals! If we ad-
ditionally select a high control level, i.e. A1A2 cos[2π(f1 + f2)t], A1A2 cos[2π(f1 – f2)t]
input voltage, many higher-order terms of The pn-FET can be used for frequency dou-
the power series according to Eq. 1-2 must bling because of 2, and for modulation be-
be taken into consideration. This in turn re- cause of 3.
sults in stronger distortions.

27
 TPS 7.2.1.1 Solutions 1

Experiment results
1. Table 1:

Y1 : ~ 10 V Y2 : ~ 60 mV
U1 = (50...100) mVss, SINE
OP : right limit
f1 U1 f2 U2
kHz mVss kHz Vss

99.91 50 99.9 100

49.9 50 99.7 0.75

32.9 50 __ 0.1
25.0 50 __ __
Adjustment of the frequency multiplier

1.1 The frequency multiplier is a frequency se- 1.2

lective amplifier in common emitter. Let us
designate the impedance of the frequency-
selection element of the circuit, the
antiresonant circuit, as ZC. Let RE be a pure
ohmic emitter resistance, through which
negative feedback is applied to the circuit.
The frequency-dependent voltage gain is
then calculated as:
Z U
VU ( f ) = − C = − 2
RE U1
Output voltage U2
In the case of resonance, ZC is also a purely Input voltage U1 ≈ 600 mVpp, f1 = 33.3 kHz
OP center position
ohmic resistance, i.e., there is no additional
phase shift.
Resonance resistance:
Z CRES = VURES RE

10 . 0
VURES = = 200 s.Table 1.
0 . 05

Z CRES = 200 ⋅ 330 Ω ≈ 67 kΩ

The output transformer is not taken into con-
sideration here!
Output voltage U2
Note: Input voltage U1 ≈ 600 mVpp, f1 = 25.0 kHz
Output voltages below 300 mVpp are no longer OP center position
recognized by the analog input of the counter.

28
 TPS 7.2.1.1 Solutions 1

Table 2:

U1 = (500...800)mVss, SINE
OP : center position
f1 U1 f2 U2
kHz mVss kHz Vss

100.5 600 100.5 10

51.2 600 102.5 7.2

33.3 600 100.2 4.5

Output voltage U2
24.5 600 98.1 2.4 Input voltage U1 ≈ 1200 mVpp, f1 = 16.6 kHz
OP left limit
20.1 600 100.5 0.9

16.6 600 99.6 0.3 Table 3:

14.3 600 __ __ U1  =  (1000...1500)mVss, SINE

OP : left limit
Note: f1 U1 f2 U2
The strength of the output signal depends on the kHz mVss kHz Vss
characteristic of the particular transistor you are
using, the choice of the operating point and the 97.7 1200 97.7 10.0
control level. As all quantities are subjected to
50.3 1200 100.6 7.0
scatter, it is unlikely that you will obtain the
same numerical values as those here. The sample 33.6 1200 100.8 5.2
solutions thus are only intended as reference val-
ues. 25.2 1200 100.9 3.0

1.3 20.0 1200 100.3 2.0

16.6 1200 99.6 1.4
14.3 1200 100.6 1.0
12.4 1200 99.2 0.5
11.1 1200 (88.8) 0.3
10.0 1200 (10.0) 0.15

Output voltage U2
Input voltage U1 ≈ 1200 mVpp, f1 = 20.0 kHz
Note:
OP left limit When measuring with f1 = 11.1 kHz and
f1 = 10.0 kHz, the counter only responds to the
signal components with amplitudes greater
than the sensitivity limit of the analog input.
For f1 = 11.1 kHz, it has apparently “over-
looked” one oscillation: instead of
9 · 11.1 kHz = 99.9 kHz, it has only counted
8 · 11.1 kHz = 88.8 kHz. For f1 = 10.0 kHz, it
has only counted 1 oscillation out of 10.

29
 TPS 7.2.1.1 Solutions 1

Output voltage U2 subjected to load by the spectrum Output signal U2 not subjected to load by the spectrum
analyzer. analyzer.
Input signal with f1 = 20.2 kHz, the fundamental period is Same input signal as on the left.
the “high” peak. Output signal frequency f2 = 101.2 kHz

The value of the input resistance Ri of the spec- circuit affects the spectrum of the collector volt-
trum analyzer is 10 kΩ. The output transformer of age UC. Thus the use of the spectrum analyzer
the frequency multiplier converts RE with the leads to two significant error sources:
square of its transformation ratio n2 into the 1. When the analyzer is connected to the out-
antiresonant circuit. The additional damping due put of the frequency multiplier, the input
to the spectrum analyzer causes a reduction in the spectrum before the oscillating circuit (at
quality of the circuit. This lowers the selectivity the collector) changes.
of the oscillating circuit. Under certain circum- 2. The bandwidth of the oscillating circuit be-
stances, the output signal U2 can now contain in- comes wider, i.e. it permits more spectral
terfering spectral components which would have components to reach the output.
been filtered out if the selectivity had been In spite of this, the spectra allow us to draw the
greater. At the same time, the damped oscillating following qualitative conclusions:

Output spectrum SU2 and input spectrum SU1

V1 = 10 fr = 200 kHz
V2 = 1 SPAN : 5 kHz...110 kHz
b = 500 Hz T = 160 s

30
 TPS 7.2.1.1 Solutions 1

– When a non-linear characteristic is control- lective enough. For this reason, the output
led, harmonics are formed from a harmonic signal U2 contains more than just the desired
input signal U1 (the single line at f = f1 in the spectral component, and the resulting curve
spectrum SU1) at f = nf1, n = 1,2,3,... of the output signal with respect to time is
– The desired harmonics can be selected distorted.
through the use of a filter with a high-qual- – The oscillation with the fundamental fre-
ity selection device. quency f1 is the highest peak in each case. It
3. is followed by n–1 oscillations with decreas-
– The frequency of the output signal f2 is al- ing amplitude.
ways 100 kHz in spite of excitation by a low Example:
frequency input signal. f1 = 20 kHz
– At a high control level and in the absence of f2 = 100 kHz
a base bias voltage, many harmonics are n = 100/20 = 5
formed. At low frequencies of the input sig- In this case, there are four oscillations of de-
nal U1 these harmonics are so close together creasing amplitude.
that the antiresonant circuit is no longer se-

31
 TPS 7.2.1.1 Solutions 1

32
 Solutions
TPS 7.2.1.1 2
Frequency Multiplication

Frequency Synthesis, PLL

Answers
1 The PLL can be described using the terminol- modulates the VCO. An oscillation with the
ogy of control technology. The VCO is then frequency f1 of the input signal briefly appears
the controlled system, whose frequency and at its output. In this way the AC voltage com-
phase are to be fed to an external reference ponent is eliminated from the control voltage
variable. The phase detector is used to recog- of the VCO. The VCO continues oscillating at
nize the phase difference, which is the control the desired constant frequency f1. The capture
difference; it compares the set and actual val- range is the frequency range within which the
ues. The loop filter serves as the controller. frequency f1 of the input signal must remain if
The loop filter generates the control voltage Uf the PLL is to lock into it.
required for the VCO as a function of the out- Once the PLL is locked in, then it will follow
put signal of the PD. an input signal with its VCO frequency in a
2. Methods of realizing the VCO: particular frequency range without losing its
LC-oscillators synchronization. This frequency range is
called the hold range.
Phase-shifting oscillators
Multivibrators 4. The frequency synthesizer consists of a PLL
Methods of realizing the PD: which contains a frequency divider in the
Multiplier feedback line from the VCO output to the PD.
Flipflop If the divider is N, then the VCO must gener-
EXOR ate N-times the output frequency in order to
Sample and hold element lock in. If the divider can be set or pro-
grammed, any multiple of a quartz-controlled
3. Before the reference signal with the frequency original frequency can be generated from it
f1 is switched on, the VCO oscillates to a with the same frequency stability.
center frequency, for example, f0. The fre-
quency difference f0 - f1 results in an AC volt- 5. PLLs are used in:
age at the output of the PD. This is modified Demodulators
by the loop filter and fed to the input of the Speed controls
VCO. This control voltage initially frequency- Tracking filters

The functional circuit diagram of the PLL

 Loop filter
 Phase detector
! VCO

33
 Solutions
TPS 7.2.1.1 2
Frequency Multiplication

Experiment results
1. If the connection from the PD to the function 2. The capture and hold ranges are approxi-
generator is interrupted, the PLL can no mately equal. We can determine only the
longer lock in. The output signal of the VCO lower cut-off frequency, as the function gen-
causes a slightly superpositioned AC voltage erator with fmax = 200 kHz is not sufficient to
at the loop filter output. This in turn causes a determine the upper cut-off frequency limit.
continuing frequency modulation of the VCO. The PLL can be pulled from f1 ≈ 2 kHz to at
least f1 ≥ 200 kHz without losing its “hold”. It
Table 1:
can also synchronize to an external frequency
The control charac- within the same frequency range.
teristic of the VCO
3. The signals are in phase. The oscillator fre-
f2 Uf quency is:
kHz V fosc = f1 = 99.990 kHz
99.1 2.18
80.4 2.12
60.2 2.06
0
40.3 1.96
19.0 1.87
120.7 2.25 0
139.8 2.31
Reference signal and VCO output signal
160.5 2.37
4. The control voltage Uf is essentially a DC
180.8 2.43
voltage of approx. 2.2 V.
200.3 2.50

Control voltage 7f and VCO output signal in locked-in

state

The control characteristic of the VCO

34
 Solutions
TPS 7.2.1.1 2
Frequency Multiplication

5. The DC control voltage now has a clear AC Observation:

voltage superpositioned on it. The VCO signal The PLL is stable; in other words, the output sig-
shows FM (apparent in the multiple slopes). nal of the divider and the reference signal are in
The PLL as a unit has become unstable, i.e., it phase at all divider settings, i.e., Φ = 0°.
oscillates. The PLL was stable in Point 4. The column f2/N shows that the output signal is
always an integral multiple of the input frequency
f1. With the aid of a synthesizer is possible to gen-
erate output signals having stable frequencies
0 which are integral multiples of f1 from a quartz-
stabilized original frequency.

Control voltage 7f and VCO output signal, not in locked-in

state.

6. Table 2:

Frequency synthesis
Frequency of clock oscillator
f1 : 99.990 kHz
Gate time of counter: 1 s
f2 f2
N
kHz N
2 199.980 99.99
3 299.970 99.99
4 399.960 99.99
5 499.950 99.99
6 599.940 99.99
7 699.930 99.99
8 799.919 99.99
9 899.910 99.99
10 999.899 99.99

35
 Solutions
TPS 7.2.1.1 2
Frequency Multiplication

36
 TPS 7.2.1.1 Appendix 1

Worksheet Table 2:
The 4-quadrant characteristics field
U1  =   mVss, SINE
AP :
f1 U1 f2 U2
kHz mVss kHz Vss

Table 1: Table 3:

U1  =   mVss, SINE U1  =   mVss, SINE

AP : AP :
f1 U1 f2 U2 f1 U1 f2 U2
kHz mVss kHz Vss kHz mVss kHz Vss

37
 TPS 7.2.1.1 Appendix 1

To chapter 2:

38
 TPS 7.2.1.1 Appendix 1

Table 1: Table 2:

The control charac- Frequency synthesis

teristic of the VCO
Frequency of clock oscillator
f2 Uf f1  :  99.990 kHz
kHz V Gate time of counter: 1 s
f2 f2
N
kHz N

39
 TPS 7.2.1.1 Key Words
Appendix 1

Key Words Page

4-Quadrant characteristics field ................... Frequency Multiplication 1 ................................... 13

Angle of current flow ................................... Frequency Multiplication 1 ................................... 14
Attenuation distortion .................................. Frequency Multiplication 1 (Solutons) ................. 27
Capture range ............................................... Frequency Multiplication 2 ................................... 20
Delay distortion ............................................ Frequency Multiplication 1 (Solutions) ............... 27
Distortion factor ........................................... Frequency Multiplication 1 ................................... 16
Filter bank .................................................... Measurement Procedures ........................................ 7
Form factor ................................................... Measurement Procedures ........................................ 5
Frequency conversion .................................. Frequency Multiplication 2 ................................... 19
Frequency multiplication ............................. Frequency Multiplication 2 ................................... 19
Frequency stabilization ................................ Frequency Multiplication 2 ................................... 19
Frequency synthesis ..................................... Frequency Multiplication 2 ................................... 19, 21
Hold range .................................................... Frequency Multiplication2 .................................... 20
Input characteristic ....................................... Frequency Multiplication 1 ................................... 13
Linear distortion ........................................... Frequency Multiplication 1 (Solutions) ............... 27
Loop filter..................................................... Frequency Multiplication 2 ................................... 20
Non-linear distortion .................................... Frequency Multiplication 1 (Solutions) ............... 27
Output characteristics field .......................... Frequency Multiplication 1 ................................... 13
Phase detector............................................... Frequency Multiplication 2 ................................... 19
Phase difference ........................................... Frequency Multiplication 1 (Solutions) ............... 27
Phase-locked loop (PLL) ............................. Frequency Multiplication 2 ................................... 19
Resolution ..................................................... Measurement Procedures ........................................ 8
Spectrum analyzer ........................................ Measurement Procedures ........................................ 8
Sweep analyzer ............................................. Measurement Procedures ........................................ 7
The time rule of electrical
telecommunications technology ................... Measurement Procedures ........................................ 7
Time base ..................................................... Measurement Procedures ........................................ 5
Transfer characteristic .................................. Frequency Multiplication 1 ................................... 13
VCO .............................................................. Frequency Multiplication 2 ................................. 19
Voltage reaction ........................................... Frequency Multiplication 1 ................................... 13

40