cac - nasa · investigate optimum way of ... change in the phase relation between the input pilot...
TRANSCRIPT
MCE-72-198
INVESTIGATE OPTIMUM WAY OFADDING WIDEBAND CAPABILITY AND
RECOMMEND A DESIGN FOR MODIFICATIONOF ONE GOVERNMENT FURNISHEDAM BASEBAND DEMULTIPLEXER
STUDY REPORT(FIN L)
CONTRACT NAS8-29039
CAC
PHASE I
PREPARED FOR
National Aeronautics and Space AdministrationGeorge C. Marshall Space Flight Center
Huntsvllle, Alabama 35812
Martin Marietta CorporationDenver DivisionP. 0. Box 179
Denver, Colorado 80201
https://ntrs.nasa.gov/search.jsp?R=19720024522 2018-08-27T06:47:20+00:00Z
STUDY REPORT
(FINAL)
CONTRACT NAS8-29039
Phase I
Approved by:
John GoodwinProgram Manager
Prepared for
National Aeronautics and Space AdministrationGeorge C. Marshall Space Flight Center
Huntsville, Alabama 35812
Martin Mareitta CorporationDenver DivisionP. 0. Box 179
Denver, Colorado 80201
FOREWORD
This report is submitted in response to NASA/MSECContract NAS8-29039 in compliance with paragraph III.B .of Exhibit "A".
ill
TABLE OF CONTENTS
PAGE NO.
FOREWORD i
TABLE OF CONTENTS „ ±±±
I. INTRODUCTION " . " . . . 1
II. PHASE I STUDY 1
A. General 1
B. Phase Locked Loops 2
1. Pilot Phase Locked Loop 2
2. Carrier Phase Locked Loops ........ 4
3. Wideband Performance Limitations g
C. Phase Matching and Filter Attenuation forChannels Located at 40 kHz through 56 kHz . . 7
1. Channel Operation 7
2. Wideband Performance Limitations 7
D. Phase Matching for Channels Located at 60 kHzthrough 176 kHz 8
E. Phase Matching and Filter Attenuation forChannels Located at 4 kHz through 36 kHz ... 8
1. Channel Operation 8
2. Wideband Performance Limitations ..... 12
F. Investigation of Wideband Amplifiers 13
G. Wideband Phase Shifter 13
H. Tape Recorder Wow and Flutter 15
1. General ....„ 15
2. Wow and Flutter Test 16
3o Phase Locked Loop Deviation Requirements . 17
4. Out-of-Band Coherently Added Pilot .... 18
5. Wow and Flutter Removal Circuit 18
6. System Output When the Phase Locked LoopsDo Not Follow the Flutter ......... 20
iv
PAGE. NO.
III. CONCLUSIONS 23
A. General 23
B. Modification 1 - Adcom Demultiplexer WidebandConversion Using No Pilot Filtering 23
C. Modification 2 - Adcom Demultiplexer WidebandConversion Using Pilot Filtering 25
D. Alternative 1 - Two Channel Breadboard WithDouble Conversion o ....... 26
E. Alternative 2 - Two Channel Breadboard WithSingle Conversion o 28
IV. RECOMMENDATIONS 29
V. REFERENCES . 31
FIGURES'
1 Open Loop Gain, 64 kHz Pilot PLL 32
2 Pilot PLL Interference Spectrum, Reference at -6 dB 33
3 Pilot PLL Interference Spectrum, Reference at -12 dB 34
4 Open Loop Gain, 84 kHz Carrier PLL . . . „ 35
5 In-Line Channel Filter, 16 kHz Channel 36
6 MJ6 Balanced Modulator, Output Spectrum 37
7 Flutter Spectrum, Track 8, Beginning of Tape .... 38
8 Flutter Spectrum, Track 8, End of Tape 39
9 Flutter Spectrum, Track 13 40
10 Flutter Spectrum, Track 8, «r=3 41
11 Flutter Spectrum, Direct Recording ......... 42
12 Open Loop Gain PLL Requirements 43
13 Wow and Flutter Removal System . 44
1.
I. INTRODUCTION
This report discusses the Martin Marietta effort to analyze
and test the Teledyne/Adcom Model G-146 Demultiplexer to determine
the feasibility and optimum method(s) for modifying the unit for
broadband operation. The desired bandwidths under consideration
included 2, 4 and 8 kHz for double sideband (DSB) and quadrature double
sideband (QDSB), and 4, 8 and 16 kHz for single sideband (SSB).
Results from a previous study made by Martin Marietta Corporation during
the performance of contract NAS8-25987 are directly applicable to this
contract and reference will be made to theory and results as detailed
in Martin Marietta Corporation Report No. MCR-72-81, Linear Modula-
tor, Final Report, Contract NAS8-25987, Modification 4, April 1972;
referred to as Reference 1.
II. PHASE I STUDY
A. General
Currently the demultiplexer unit provides 1 kHz bandwidth
operation in the DSB and QDSB modes, and 10 Hz to 2000 Hz operation
in the SSB mode. The channel locations are from 4 kHz to 176 kHz
with 4 kHz spacing, and a maximum of 24 channels can be utilized simul-
taneously. The pilot tone may be located at 4, 8, 16, 32, 64 or 128
kHz, and the reference channel may be at any odd multiple of 4 kHz.
Two different demodulation techniques are employed in the demultiplexer.
The lower channels, 4 kHz through 36 kHz, are up converted with a 160 kHz
carrier, filtered and then down converted to baseband. Channels located
at 40 kHz through 176 kHz are demodulated directly. The demodulation
methods chosen require in-line filtering in all data channels, and
filter matching is employed between the pilot channel and each indivi-
dual data channel to reduce the effect of frequency modulation (FM)
of the input signals caused by tape recorder wow and flutter.
2.
B. Phase Locked Loops (PLL)
The phase locked loop designs were investigated through the-
oretical analysis and on-line testing to determine tracking character-
istics and the amount of FM sidebands in the carrier outputs caused
by the harmonics and sum frequencies produced internally in the
PLL's.
1. Pilot Phase Locked Loop (64 kHz)
The pilot loop performance is independent of the choice of
the six available frequency location and the loop bandwidth is
selectable in 5 steps between 100 Hz and 2 kHz. The 64 kHz location
and 2 kHz bandwidth were selected for evaluation. The loop employs
an "EXCLUSIVE OR" type phase detector with a +15 volt output driver.
The output duty cycle varies from 50 to 100% for a [0°, 0°] to
[0 , 90 ] change in the phase relation between the input pilot and
the voltage controlled oscillator (VCO) output. The average dc
voltage which is the loop control voltage, therefore, varies from 0 too
15 Vdc for a 90 variation in detector input phase. The phase detector
gain is:
K , = 15 volts/degree = 0.167 = -15 dBPd 90
The gain of the VCO referrred to 64 kHz was measured to be:
K = 1 kHz/volt = 3.6 • 10 degree/sec/volt = 111 dB
With the Bandwidth Selector Switch in the 2 kHz position and the
Pilot Frequency Selector Switch in the 64 kHz position, the loop
amplifier/filter is as shown below.
-MS
dc Reference
3.
The amplifier gain is:
A = 14.1 = 23 dB
The filter provides a lag-lead attenuation of 46.5 dB with corner
frequencies at 1.45 Hz -and 308 Hz; and a continuous roll off with ._
corner frequency at 800 Hz.
The dc loop gain is:
AK K , = 8, 5 • 105 = 119 dBpd
A plot of the PLL open loop gain is shown in Figure 1. The unity
gain crossover is at 800 Hz, indicating a loop lock-in range in
excess of 800 Hz as shown by Garnder (Reference 2). The lock-in
or capture range was measured to be approximately 2 kHz.
The phase error in the loop is a function of the open loop
frequency difference between the input pilot and the VCO frequency,
^W (i.e., input signal deviation) as follows:9 = AW
AKK ,pd
The derivation of this formula is detailed in chapter VIII of Reference 1.
Assuming a maximum input signal frequency difference caused by
tape recorder wow and flutter of +17» of the 64 kHz pilot (see paragraph
II.H.3., discussion on wow and flutter requirements), the loop static
phase error will be:
0 = 2ff • 640 • 360 degrees = +1.7°875~- 105
1.7 will be the static phase error for 1% input signal frequency
deviations at rates below 1.45 Hz. Above this frequency the loop
filter reduces the loop gain and the phase error will increase.
4.
The interference rejection performance of the loop will now be
considered. Assuming the system reference tone at 68 kHz to be -6 dB
(or lower) with respect to the pilot, the hard limiter on the PLL
input will, by the capture phenomenon, improve the signal-to-noise
ratio by an additional 6 dB (Reference 3). The phase detector beat
frequency at 4 kHz will, therefore, be -12 dB with respect to full
scale control signal, or 15/4 volts. At 4 kHz the amplifier/filter
gain is 1/65 so that the VCO input signal at 4 kHz is approximately
57 mV. This interference signal will frequency modulate the VCO
causing output sidebands at 64 -4 kHz. The sideband amplitudes are
governed by the first order Bessel function J (m) with the modulation
index as the variable. The VCO deviation caused by 57 mV is 57 Hz and
the modulation index is:
= °'01425
This modulation index corresponds to first sideband amplitudes at -42 dB.
For input 68 kHz reference at -12 dB with respect to the pilot level, the
theoretical output first sideband level would be -48 dB. Tests show the
sidebands to be at -46 dB (Figure 2) and -52 dB (Figure 3) respectively.
The 4 dB discrepancy may come from the fact that the input signal passes
through the 10.5 kHz bandwidth pilot filter giving the reference tone an
attenuation of 3-4 dB.
2. Carrier Phase Locked Loop (84 kHz)
The 84 kHz carrier PLL was selected as representative of
all the carrier loops. It differs from the pilot loop in the following
areas :
*NOTE: To obtain this 6 dB clipping improvement, the pilot must be
greater than the algebraic sum of ALL other signals in the
composite signal.
5.
(i) The phase detector is of the EXCLUSIVE OR type; however,
the VCO frequency is the 7th harmonic of the input frequency (12 kHz)
so that the phase detector gain is reduced by a factor of 14. Also,
the output is limited to a +2.5 V swing rather than the +15V swing
in the pilot PLL. Phase detector gain is:_3
K , - 2.5 volts/degree = 2 •- 10 = -54 dB .P d 9 0 - 1 4
(ii) The VCO is of the multivibrator type and the measured
gain referred to 84 kHz is:
K = 37.8 kHz/volt =1.36 • 10 degree/sec/volt = 143 dB
(iii) The amplifier/filter lag-lead corner frequencies are at
1.59 Hz and 338 Hz respectively with the continuous lag corner frequency
at 3.38 kHz. The amplifier gain measured in the 100 kHz PLL was 3.
Assuming equal loop gain in the 100 and 84 kHz loops the amplifier
gain must be adjusted up in the 84 kHz loop by a factor of 1.19 dueo
to different VCO gains and by a factor of 1.4 due to the fact that the
100 kHz loop is locked to the 5th rather than the 7th harmonic of the
input frequency. The amplifier gain is therefore assumed to be:
A = 5 - 14 dB
The total dc loop gain is AKK , = 1.35 • 10 = 102 dB. Thepd
loop crossover frequency is 180 Hz as shown in Figure 4, and the static
phase error for a +1% (or 84 kHz) input frequency deviation is:
9 = 27r -840 • 360 = +14°T735 - HP
Comparing the carrier loop with the pilot loop, it is seen that
the loop gain differs by approximately 20 dB. With tape recorder
flutter of up to +1% frequency deviation the 14° carrier PLL static
phase error will be intolerable for adequate system performance. In
this report it will therefore be assumed that if system conversion
to wideband operation is contemplated the carrier PLL's will be brought
in line with the pilot PLL performance of a +2° maximum static phase
6.
error for a +1% input frequency deviation. This can be done by in-
creasing the carrier PLL amplifier gain by a factor of 10.
3. Wideband Performance Limitations
The pilot loop and the carrier loop are in series with
respect to phase error on individual channels, and the total phase
error is, therefore, +4° for +1% input frequency deviations at.a. rate
equal to the input signal modulation or flutter frequency. The
effect of this phase error is as follows:
(i) In DSB a 0.3% amplitude modulation (amplitude loss)
on the channel output (see chapter III of Reference 1)
(ii) In QDSB a 0.3% amplitude modulation on the wanted
signal as in DSB; and quadrature feedthrough of up to 7% of full
scale. The quadrature feedthrough represents only about -23 dB
quadrature suppression. (See Chapter III of Reference 1)
(iii) In SSB the 4 carrier phase error will appear as
phase delay of the audio output with the equivalent time delay varying
from 0.14 millisecond at 80 Hz to 0.7 microsecond at 16 kHz. (See
Chapter V of Reference 1). Now let us consider spurious sidebands on
the PLL outputs. In paragraph II. Al it was seen that the reference,
if at a level of -6 to -12dB below the pilot, may be located at 68 kHz
i.e., only 4 kHz from the pilot. Also analysis and measurements
for the sample carrier PLL at 84 kHz showed all spurious responses
down 50 dB. However, the PLL roll off in this system starts at approxi-
mately 1.5 Hz so that for flutter frequencies in excess of 1.5 Hz the
static phase error will be degraded even further than the above
performance predictions indicates. For high quality wideband perform-
ance including 50 dB quadrature suppression in the QDSB mode an increase
in loop gain and change in loop roll offs are required as indicated
in Chapters III and VIII of Reference 1, with modifications caused by
relaxed frequency deviation requirements as outlined in paragraph II H.3
of this report.
7.
C. Phase Matching and Filter Attenuation for Channels Located
at 40 kHz through 56 kHz
1. Channel Operation
The block diagram for the demodulation of these channels
is shown below:
CompositeSignal4-176 kHz
Bandpass
Filter, BP1
Balanced
ModulatorOutput
Demodulation Carrier
Since a balanced modulator is employed with square wave carrier input
(a modification to sin-wave carriers would include rebuilding the
whole system synthesizer which would be prohibitively expensive),
the third harmonic of the carrier will demodulate an unwanted channel
at three times the carrier frequency unless the input bandpass filter
is present. The attenuation requirement for the filter is, therefore,
-40 dB at three times the channel frequency in question. The flatness
of the filter must be 0.1 dBpp. over the channel bandwidth for adequate
DSB and SSB operation, and 0.02 dBpp. for 50 dB quadrature suppression
in the QDSB mode. (See Chapters III and V of Reference 1). Also
in the filter passband the phase slope of the filter must be linear
to better than +0.18° for 50 dB quadrature suppression. Due to the
tape recorder wow and flutter there must be a matched filter in the
pilot/carrier line with phase matching to better than +0.18 over the
frequency deviation range of the input signal.
2. Wideband Performance Limitations
For 16 kHz SSB operation with a choice of upper or lower
sideband, the passband will be 32 kHz, with attenuation characteris-
tics equivalent to a 0.01 dB 5 pole Chebychev low pass filter, and
an equivalent filter must be placed in the pilot line. With 8 kHz
DSB/QDSB operation in some other channel the 0.18 linearity require-
ment must be placed on the pilot filter, and therefore, also on the
8.
SSB channel filter, with the added requirement that also the DSB/QDSB
filter must be 32 kHz wide rather than the 16 kHz required by consid-
ering the DSB channel alone.
With a number of wideband channels in different frequency
locations a separate filter development must be accomplished for
each wideband channel, each filter matching the pilot filter pha_se
slope. It should be noted here that the above filter requirements
are compatible with proven performance of DSB and QDSB performance
already obtained under contract NA.S8-25987; however, in that contract
double conversion demodulation was found to be the optimum demodula-
tion technqiue for all channels, the conversion being performed
in such a way that the filter center frequency is common in all channels.
This greatly reduces the cost of the system filters.
. D. Phase Matching for Channels Located at 60 kHz through 176 kHz
These channels employ direct demodulation similar to channels
at 40 through 56 kHz. Since the third harmonic of the carriers will
be outside the spectrum of the composite signal, no in-line
filter would be needed. However, since the pilot line contains a
filter, a matching filter with requirements per paragraph II.C.
must be inserted in the data line on all channels 60 kHz through
176 kHz also.
E, Phase Matching and Filter Attenuation for ChannelsLocated at 4 kHz through 36 kHz
1. Channel Operation
Since square wave carriers are used for demodulation,
the odd harmonics of the carrier will be present and will demodulate
harmonic channels as described in Paragraph II.C. In-line filters are,
therefore, required. For channels 4 through 36 kHz an additional
demodulation problem arises. For these lowest carrier frequency
with in-line filters a "broadband" demodulation problem occurs as
described in Chapter III-F of Reference 1. To eliminate this problem
double conversion demodulation is required as shown below:
t-176 kHz
all channels
9.
Lowpass
Filter LPl
Common for
NP
Balanced
Modulator
1
v>
Bandpass
Filter BPl**
Balanced
Modulator
I k
160 kHz 160 .kHz-n'4 kHz
Neglecting the effect of the input low pass filter the frequency spec-
trum at the output of the first balanced modulator will be as follows:
Hie
./fcr '74
ZL IZIL_/to
•>-ffete
• Output Spectrum of #1 Balanced Modulator
Spectrum A is the fundamental spectrum, and B is caused by the
third harmonic of the 160 kHz translation frequency at -10 dB.
The wanted channels 4 through 36 kHz has been translated to 156 through
124 kHz respectively, and this will be the appropriate second balanced
modulator carrier frequencies for down conversion to baseband.
In the second modulator the third harmonics (at -10 dB) will
be in the range 372-468 kHz and the fifth harmonics (at -14 dB) will be
in the range 620-770 kHz. For 50 dB suppression of unwanted signals
(caused by demodulation of spectrum B above) the in-line filter must
provide 30 dB attenuation at 372-468 kHz and 24 dB at 620 kHz
10.
(and beyond) since spectrum B is already down 10 dB.
As an example of demodulation of an unwanted channel let us con-
sider the 4 kHz channel. The down conversion carrier will be 156 kHz
with third harmonic at 468 kHz. The 468 kHz will demodulate the
12 kHz channel unless the in-line filter provides 30 dB attenuation
of the 468 kHz (12 kHz channel) in spectrum B. This 30 dB attenu-
ation is required for frequencies above 372 kHz and may be provided
by the in-line bandpass filter, BP1.
Tests were performed on the demultiplexer to verify the required
attenuation in the present narrow band unit. Figure 5 shows the
attenuation characteristics of the in-line bandpass filter at 144 kHz
for the up-converted 16 kHz channel. It is evident that the high-Q
two pole filter provides adequate interference rejection, however,
the filter flatness in the passband (+1 kHz for DSB and QDSB) shows
almost 0.5 dB or 57« difference in gain between the upper and lower
sidebands. In DSB this would cause an error in output amplitude of
approximately 2.570. In QDSB the effect will be a "best" quadrature
suppression of 32 dB (if all other factors affecting the quadrature
suppression are perfect). This mistuning of the filter could be
corrected; however, an extremely accurate and complicated tuning
procedure (of all channels) is required, and upon completion of the
tuning, component changes in excess of 0.17o cannot be tolerated.
The above considerations are with regard to the amplitude flatness of
the filter. In QDSB the linearity of .the phase slope in the filter
is of importance. In our opinion, the best phase linearity with this
high-Q filter will be in the order of 0.5-1 degree over a +1 kHz
range with a resulting best quadrature suppression of 42-36 dB.
11.
To verify the need for the input line low pass filter tests were
run to determine the amount of second harmonic spectrum present on
the output of the #1 balanced modulator. This second harmonic spectrum
is caused by nonlinearities in the modulators internal multiplication
process.
Figure 6 shows the output spectrum of the first multiplier in
the up conversion method, the input signals being the 160 kHz conver-
sion carrier and a single 16 kHz tone representing dc modulation of
the 16 kHz channel. Previous discussions showed the requirement for the
suppression of the harmonic spectrum caused by odd harmonics of the
conversion carrier. Figure 6 also shows that the low pass input filter
IS REQUIRED to suppress the higher channels, notably 164kHz through
176, since the multiplier produces a carrier second harmonic spectrum
at -38 dB. This spectrum will fold back and make the 164kHz - 176 kHz
channels overlap the wanted channels 16 through 4 kHz respectively
which results in the output of these three lower channels containing
signals from the higher channels at -38 dB (without the input low
pass filter).
Attenuation requirements for the input low pass filter is there-
fore:
(i) -12 dB minimum from 164 kHz to 176 kHz due to the second
harmonic foldover effect caused by the inband translation frequency
(160 kHz)
(ii) Approximately -15 to -20 dB to make sure feedthrough
of the composite signal in the 124 to 156 kHz range is down at
least 50 dB at the output of the first balanced modulator.
It should be noted that for the range 3 kHz through 37 kHz the
0.02 dB flatness and 0.18 linearity requirement per paragraph
II. C2 also applies to this input low pass filter.
12.
2. Wideband Performance Limitations
With wideband on one of the lower channels we may have
either 8 kHz DSB on the 12 kHz channel, or 16 kHz SSB on the 20 kHz
channel with the required guardband in both cases occupying the re-
maining channel spacing up to and including the 36 kHz channel. The
output spectrum from the first balanced modulator will be as shown
below:
1 — 1I - .'14
34, 4
Y///,\ \
I loo
*>o 3fe
i — 1 ' — -'• -! i i1 J ! _J L»J
M 33 t i ! „Wo Mi
J
The combination BP/LP filter must provide 20 dB attenua-
tion between 124 and 156 kHz and 30 dB attenuation beyond 420 kHz,
and must be flat to within 0.02 dB over the passband. Let us consider
the effect of using the present approach with a wider BP filter in the
channel line than in the pilot line, and keeping the input low pass
filter with no modifications. Assuming that the new band pass filter
has the required 0.18 linearity on the phase slope, let us consider
the effect of the phase slope of the current 5 pole Butterworth in-
put filter. The phase delay of the filter is shown below:
13.
The signal being considered is 8 kHz QDSB centered at 16 kHz. The
difference in phase angle between the upper and lower sidebands
with respect to the carrier phase is approximately. From Chapter IIIB
of Reference 1 we get the resulting "best" quadrature suppression in
the QDSB mode:
-1Quadrature suppression = sin [ 3" ] = 32 dB(fl
In the DSB mode the sideband phase error effect causes a loss
in output signal amplitude of 1 - cos 1.5° A 0.1%. The SSB
effect will be a 3° discrepancy (for 8 kHz SSB) in phase correlation
if other SSB channels on higher carrier frequencies operates on
linear phase delay versus audio frequency.
F. Investigation of Wideband Amplifiers
The operational amplifiers used in the current demultiplexer
design are of the yi/A709 type. This amplifiers will not provide
adequate system operation in the 8/16 kHz wideband modes.
Investigation of the manufactures specifications on the Harris
Semiconductor HA2600 series amplifier show that they will provide
stable operation at unity gain and will have sufficiently low phase
shift to provide satisfactory operation in a 10 pole all pass Poly-
>phase network for SSB generation of a 16 kHz SSB channel. These
amplifiers will also operate satisfactory as input buffer amplifiers
and have the same pin configuration as the Fairchild it A709 being
used in the present system, therefore, providing a satisfactory
replacement for the proposed wideband operation. The Teledyne
Philbrick 1321 operational amplifier will also perform satisfactorily
as a substitute.
G. Wideband Phase Shifter
The amount of sideband suppression that is achieved for the
phase shift method SSB operation is determined by the accuracy with
which quadrature is achieved for the input data channel frequencies ando o
for the demodulation carrier. A quadrature accuracy of 90 +0.18 is
necessary at all data frequencies to obtain sideband suppression
14.
of 50 dB. For 16 kHz SSB wideband operation the Martin Marietta
Corporation proposes a variation of the current polyphase network design
as shown below.
Cos Wmt
Network
A
Network
B
9)
Cos (Wmt + 9 + 90 )
Networks A and B are Chebychev equal ripple type approximations of the
type developed by D. K. Weaver (Reference 4). The principle difference
with respect to the current design is the implementation of the ten
pole-zero pairs as shown below.
j - SCR,I -f SCR,
Ten pairs are required for 50 dB sideband suppression, and the pole-zero
positions for each of the ten pairs were determined through the use of a
computer program based on D. K. Weaver's design procedure. This has been
performed on a parallel NASA/MSFC contract NAS8-25987, Modification 5. .
15.
Board layout has been accomplished on this contract, and these same boards
will be used as substitute boards in the comtemplated wideband modification
of the wideband demultiplexer.
Experience on the alignment of the polyphase networks in the FDM
modulator system (Contract NA.S8-24682 and Contract NA.S8-25987, Modifi-
cation 5) showed that it is difficult to calculate the exact values to
within +1% because of the stray effects of shielding-and stray coupling
on the PC board. Since component values must be adjusted to within about
.05% for proper sideband suppression it will be necessary to use trim-pots
for proper alignment of the polyphase and modulator networks. However, low
temperature coefficient resistors will be used which will provide a
more temperature stable polyphase network.
The alignment frequencies for each individual pole-zero pair ampli-
fier for 16 kHz SSB operation with lower band cut-off frequency at
80 Hz is AS follows:
NETWORK A NETWORK B
27.3 Hz 94.45 Hz234.7 Hz 411.1 Hz1808.0 Hz 1583.0 Hz3113.0 Hz 6236.0 Hz13546.0 Hz 46369.0 Hz
These are the frequencies for which the individual pole-zero pairs have
exactly 90 phase shift.
H. Tape Recorder Wow and Flutter
1. General
The current NA.SA/MSFC specification of frequency modulation
(FM) caused by tape recorder wow and flutter is +170 frequency deviation at
rates up to 500 Hz. This requires ultra-high-gain/wideband phase locked
loops in DSB/QDSB/SSB receiver stations. The feasibility and performance
of such loops have been proven by the Martin Marietta Corporation on
contract NAS8-25987 (Chapter VIII,Reference 1). However, with this
requirement it is extremely difficult to extract the pilot from the com-
posite signal.
16.
Telemetry Standards do not state maximum allowable wideband
tape recorder flutter limits other than steady state condition of +0.27»
per IRIG Document 106-69 paragraph 5.6.2.2.6. This refers to maximum
cummulative wow and flutter within a 10 kHz bandwidth. This should,
therefore, .be considered the maximum low frequency peak frequency deviation
input to the baseband demultiplexer. To determine the spectrum of the wow
and flutter, tests were run on MSFC's MINCOM Model 30 Recorder/Reproducer.
2. Wow and Flutter Test
The tape recorder tests were performed on MSFC's MINCOM Model
30 recorder, which has specifications of maximum cumulative peak ac flutter
of 0.125% at 0.1 Hz to 10 kHz bandwidth, and a dc speed tolerance of 0.1%.
This gives maximum low frequency wow and flutter of 0.225%.
The flutter tests where performed in both the direct record and
in the indirect record modes using a 900 kHz FM subcarrier, with the tape
speed at 120 inches per second. The test method was as follows: a
108 kHz signal from a crystal oscillator was recorded, played back and FM
demodulated. The audio output was analyzed with a spectrum analyzer with
2 Hz bandwidth. 1 kHz FM submodulation of the 108 kHz signal with a known
deviation was performed to calibrate the output data (the tape recorder
flutter spectrum) in percent FM deviation. The flutter spectrum of the
tape recorder is shown in Figures 7 through 11. The tests were repeated
for an outside and the middle track and also for signals recorded at the
beginning and at the end of the tape reel to check for variations in
flutter caused by tape-scew or mechanical loading on the recorder motor
compensation system. The results differed insignificantly. From the
graphs it is seen that the maximum single frequency deviation is approx-
imately 0.0157, at approximately 100 Hz. The spurious amplitudes in the
300 Hz to 800 Hz range have deviations below 0.005%. The cumulative
flutter between 300 Hz and 800 Hz amounts to approximately 0.02%, and
approximately 0.1057o between 2 Hz and 300 Hz.
17.
3. Phase Locked Loop Deviation Requirements
From paragraph II H2 and referring to Chapter VIII of Reference 1
the tracking requirements of the system PLL's are as follows. Assuming
system channel occupancy of dc to 200 kHz and a maximum static phase error
of 0.18 at the highest channel and a maximum input frequency deviation of
0.225%, the dc loop gain AKK .must-be 135 dB. At 2 Hz the high sidepd _
cumulative jitter is .125%, and the loop gain is 130 dB. At 300 Hz the high
side cumulative jitter is 0.02%, and the loop gain may be reduced to
114 dB. Assuming, figures 7 through 11, that the cumulative jitter in
the 2 Hz to 300 Hz range is equal above and below 100 Hz, the jitter to
be tracked by the loop beyond 100 Hz is approximately 0.067o with loop
gain of 124 dB. Figure 12 is a plot of the required loop gain AKK ,pd
as a function of input deviation rates. The dc gain is increased by
10 dB for improved dc loop performance, with a 6 dB/octave roll of be-
tween 0.33 Hz and 1 Hz, and the high end beyond 300 Hz is assumed to
continue at 6 dB/octave or more (compatible with loop stability). The
loop unity gain crossover will be at 5 kHz or below, depending on the
roll-off beyond 300 Hz.
Using the dc gain figures from the loop design example, Chapter
VIII of Reference 1, and the attenuation of Figure 12; the attenuation
of the reference signal, if located 4 kHz apart from the pilot and at a
level of -12 dB, will be such that the 4 kHz interference sidebands on the
PLL output are at -16 dB with respect to the fundamental (this is only
true if the loop crossover is below 4 kHz).
From this it is concluded that with the above interference side-
band a phase locked loop as described above cannot be used to extract the
pilot from the composite signal, and at the same time track the flutter
adequately. Two solutions to this problem are presented below.
18.
4. Out of Band Coherently Added Pilot
During on-line reception of the composite signal the wow and
flutter IS NOT PRESENT, and a narrow band phase locked loop may be used to
recover the pilot directly. The output of this loop is the new pilot,
the frequency of which is well above the highest frequency in the composite
signal, say 250 kHz. This signal is phase locked to the original pilot
and is recorded on the tape after being summed with the composite signal.
During playback the pilot phase locked loop at 250 kHz will have the gain
(and tracking characteristics) per Figure 12; however, the nearest inter-
ference signal will be at 200 kHz causing a loop beat frequency at 50 kHz. The
attenuation in the loop can be designed to give £ 50 dB suppression of the
unwanted +50 kHz sidebands.
5 . FM Wow and Flutter Removal Circuit
The circuit shown in Figure 13 overcomes the flutter problem by
electronically removing the FM caused by the tape recorder wow and
flutter before the composite signal is demodulated. The following dis-
cussion describes the system operation.
When the composite signal is recorded a reference tone is summed
with the composite signal and is also recorded on the tape. The reference
tone (250 kHz) is far enough removed from the composite signal (DC to 200 kHz)
that it can be easily separated (filtered) from the composite signal.
This summed signal FM modulates a carrier frequency which is then recorded
on the tape. When the tape is played back the composite signal has FM
modulation on it due to the tape recorder wow and flutter. The 250 kHz
reference tone after being demodulated by the FM receiver discriminator,
is compared to a reference crystal controlled VCO in a phase detector. The
phase detector output signal is due to the frequency modulation on the 250 kHz
reference caused by the tape recorder wow and flutter. This signal is then
filtered by LPFl in loop 1 to remove the interference caused by the demod-
ulated DC to 200 kHz composite signal. The loop 1 filtered phase detector
output is used to control the FM receiver local oscillator in such a way
as to reduce or remove the wow and flutter FM from the hetrodyned IF.
19.
Therefore, the wow and flutter FM is removed from all of the demod-
ulated signals producing a clean FDM composite signal at the receiver
output.
The phase detector output is also filtered by LPF2 in loop 2 which
has a very low frequency cut-off and controls the nominal frequency of
-the 250 kHz VCXO. This compensates for the tape recorder being played
back at a slightly different speed or small changes in the VCXO
frequency which can occur between the time the signal is recorded and
when it is played back for demodulation.
The degree to which the wow and flutter FM is removed is a
function of the gain of loop 1 and can easily reduce the FM by orders of
magnitude.
It should be noted, however, that long term changes in speed of the
tape recorder will cause long term changes in the frequencies of the recorded
signals. These long term or very low rates of frequency changes cannot
be corrected for by this system but can be compensated for by controlling
the tape recorder motor speed.
With the removal of the FM as described above, the pilot phase locked
loop does no longer have the gain requirements per Figure 13. The dc gain
(145 dB) remains unchanged; however, the loop roll off may now begin at
a very low frequency, and the loop attenuation of the 4 kHz beat caused
by the FDM system ambiguity/reference tone can be made so as to provide
the required 50 dB suppression of the interference signals.
It should also be noted that with the use of the FM removal circuit
it may be possible to have sin-wave rather than square wave output from
the system phase locked loops. This would facilitate direct demodulation
of all channels using true multipliers. Such an FDM receiver system would
be less complex and cheaper than a system using the double conversion
method as described and recommended in Reference 1.
20.
6. System Output When the Phase Locked Loops Do Not Follow
the Flutter
Assuming the carrier phase locked loop in a given channel does
not follow the flutter of the tape, let us consider the effect on the
demodulated output„
Assume the signal recorded on the tape initially to have been
F-, = A cos w-i t, representing dc modulation in the w^-channel. During
playback frequency modulation is present and the signal is transformed
to: (see Reference 5)
F(FM) = A JQ (m) cos w-^ t
+ A Ji (nOfcosCv^-K/^t - cos (w1-wm)t]
+ A J2 (m) [cos (wj w t + cos (wi-2wm)t] +
where
m = modulation index
w = modulation frequency
J = Bessel function of first kind and nth order
Let us first consider direct modulation with no in-line filters, the
demodulation being performed in a true multiplier, the demodulation
carrier being F£ = B cos w^t. Neglecting all sum frequency term (fil-
tered by the channel output filter) we have the multiplier output as:
F(FM) F£ = AB J0(m) cos o
+ AB J2(m) 2 cos 2 wmt
+ AB J4(m) 2 cos 4 wmt
21.
The wanted output signal (line 1) is reduced by a factor J0(m), and
spurious signal will appear as sidebands spaced at even numbers of
the modulation frequency. The amplitude of these sidebands are de-
termined by the modulation index of the frequency modulation inter-
ference. From tables of Bessel functions of first order the limit-
ing condition for 50 dB spurious suppression is found to be approx-
imately 0.1 J2(0.1) = 0.00124 . The loss in the wanted signal for
m = 0.1 is 0.257o. For a given percentage of frequency modulation,
the amount of spurious signals will be worst in the higher channels
since the modulation index is proportional to the actual FM devia-
tion.
Example: In the 176 kHz channel the maximum deviation for
m = 0.1 and wm = 100 Hz will be 10 Hz or 0.0057%, while the corres-
ponding maximum deviation for the 4 kHz channel would be 10 Hz or
0.257o. For 10 Hz modulation frequency the maximum deviations will
be 0.000577= and 0.0257= for the 176 kHz and 4 kHz channels respect-
ively.
From the above it is concluded that in a demodulator with no
in-line filtering and assuming the carrier phase lock loop is NOT
FOLLOWING the tape recorder flutter, the maximum allowable modula-
tion index in the input signal is approximately 0.1.
If in-line filters are used the formula for the FM input signal
will be modified as follows:
F(FM) = A J (m) cos w,t
+ A J1(m)|fcos (w, + w ) t + 9n~?-cos [~(w, - w )t - 90)"| +l I ! - 1 m 1 J L l m i - * J
22.
where 9 and 9 are determined by the phase slope in the filter.
Similarly to the QDSB analysis in Chapter III of Reference 1 it can
be found that the cancellation of the J,(m) - term is dependent
upon the linearity of the phase slope in the filter. For
9 - Q < 0.18 , the J, (ro) term will always be suppressed by at
least 50 dB.
23.
III. CONCLUSIONS
A. General
As pointed out in paragraph II.B., C, D and E, the Adcom Model
G-146 FDM Demultiplexer has not been designed in such a way as to be
easily converted to a high performance wideband system. However, to prove
the feasibility of such conversion two approaches are presented. The
pilot and the reference signals will be maintained in present locations;
however, if the demultiplexer is to be used with a combination of both
narrow and wideband channels, either two pilots must be used, or the
filtering on all channels must be wideband. Two alternative approaches
to prove the feasibility of wideband demodulation are also presented.
These approaches include breadboarding efforts rather than rework of the
Adcom demultiplexer. The contemplated output signals of the FDM as modi-
fied, by the Wideband Modification Contract, NAS8-25987 Mod. 5, are shown
below.
PIUTT
I V
Y"M . -Lii$ 3fe m SZ to 4.4 fct 7z £4 tot '76
It is assumed that these signals will be the input signals to
the demultiplexer with the addition of a quadrature DSB signal at 44 kHz.
B. Modification 1 - Adcom Demultiplexer Wideband Conversion Using
No Pilot Filtering
This modification includes the demodulation of the above signals
by shorting the demultiplexer's in-line filters on channels 44 and 104 kHz
and also the 64 kHz pilot filter and the 68 kHz reference filter. No
other channels can be demodulated during the demodulation of the two wide
24.
band channels; however, other channels may be demodulated by reinsert-
ing the pilot and reference filters (by a simple switch) and rerunning
the tape. Some higher channels may be demodulated simultaneously with
the wideband channels by shorting their in-line filters. . These channels
are 148 kHz to 176 kHz. 4-28 kHz cannot be demodulated on the same tape
run as wideband channels since filtering is always required for the up
converted channels.
Channel modifications will include:
(i) New 16 kHz polyphase network, these networks have been developed
by the Martin Marietta Corporation on Contract NAS8-25987.
(ii) Two 8 kHz lowpass output filters with 50 dB attenuation at
16 kHz and 0.1 dBpp ripple for the 44 kHz channel.
(iii) 16 kHz low pass output filter with 50 dB attenuation at
32 kHz and 0.1 dB ripple for the 104 kHz channel.
The demultiplexer block diagram is as shown below for the two
wideband channels.
From FDMModulator"
>Mor 7
104 /(fr IT--
i n/. >Q
AGC
3° kH- ~
1° ,.„
—^
t
^
)
. . . N/
Vk— — ?
1
1
J
x
^
y
X
V
P™ ™~
X
L>cV
£v
— 44°/C
_^
44°»
""" '
LP, 8
)° kHz
LP, 8
to
SL kHz
Poly-phase
kH-
kH;
z
s
^
>
[•
V
—"-> 0° CHANNEL
Note: 132 +16 kHmust not bused.
~~~ 90° CHANNEL
. , LP, 16 VK7.
25.
It should be noted that this system deviates from the design
principles of the Adcom demultiplexer in the respect that channel-to-
pilot filter matching is not employed. This modification is useful
to prove wideband operation in general.
The performance of this system will be sub-optimum because of
the high PLL static phase errors as pointed out in paragraphs II-B,
C, and D.
C. Modification 2 - Adcom Demultiplexer Wideband Conversion
Using Pilot Filtering
This modification includes the modification 1 changes with
bandpass filters in the signal lines as shown below.
' J JV A /-ip 1 . v1 1 >i
V
f
10
10
BP1
BP2
4/0"
4/0°
\r
r v
BP3 .
X 'V
^
' 44 /4o
'
X
t_X
t
'
A-X
— 44
•v>
V
7
ZkPP
LP
° kHz
LP
0°_ kHz
oly-
r U CHANNEL
>• 90° CHANNEL
f & r "*- lo kHz' SSB
• ^ Pilot
BPl must be flat to within .02 dB over 32 kHz to provide operation for
both LSB and USB with 50 dB sideband suppression. The pilot filter BP3
must be matched to BP2 for FM cancellation. This means a 32 kHz wide pilot
26.
with minimal interference rejection. BPl although only requiring 16 kHz
bandwidth to pass the DSB signal must also be 32 kHz wide. For the DSB
signal BPl must provide 40 dB attenuation at 3 x 44 kHz because of carrier
third harmonic demodulation of the 132 kHz channel if this channel is
used. To provide this attenuation and the 32 kHz flatness, a minimum of
5 poles and 5 zeros are required in the filters. The alternative to sharp
cut-off filters is the deletion of the channels at 132 _16 kHz. In
QDSB 0.18° linearity and matching is required for 50 dB suppression. A
performance trade-off must be made between phase locked loop static phase
error (paragraph II.B) and the filter requirements. Suboptimum perfor-
mance will result with the present PLL's. All other channels in use must
have the matching 32 kHz in the data line, or two pilots may be employed
with wideband filtering/matching of the one pilot and the wideband channels,
and narrowband filtering/matching of the second pilot and the narrowband
channels. With two pilots the time correlation between wide and narrowband
channels will be degraded.
D. Alternative 1 - Two Channel Breadboard with Double Conversion
This alternative can be used if the modifications of the Adcom
Demultiplexer are deemed unsatisfactory. The alternative suggests demodu-
lation of the FDM modulator signal by a separate two channel breadboard.
For DSB and QDSB this has already proved feasible for 16 kHz bandwidth on
linear modulator contract NAS8-25987. The breadboard will not include
AGC and Ambiguity Resolver since these features have already been proven
by the Adcom Demultiplexer design. The breadboard will, however, prove
feasibility of up to 16 kHz bandwidth in SSB. The suggested breadboard
is shown by the following diagram.
27.
BLOCK DIAGRAM:
0 CHANNEL
90 CHANNEL
SSB
NOTE: Reference tone
will not be
included.
The performance of the breadboard will be state-of-the-art for wideband
operation with 50 dB quadrature suppression.
The breadboard will utilize modified printed circuit boards from the
FDM modulator rework (Contract NAS8-25987, Modification 5) for SSB
demodulation; and the 2-channel wideband DSB/QDSB breadboard developed under
Contract NAS8-25987, Modifications 1 through-4. The DSB/QDSB is built to
accept the tape recorder wow and flutter and the only modifications re-
quired is to change the carrier frequency from 64 kHz to 44 kHz, and
modify to accept 64 kHz pilot rather than 128 kHz.
28,
New designs, if required, include the output filters only.
The DSB/QDSB breadboard has presently 16 kHz rather than 8 kHz
output filters. A 16 kHz filter will be required on the SSB
output.
The SSB demodulation will be single conversion, with double
conversion as a more expensive option. This option should be
exercised if SSB carriers are to be used in the lower third of
the composite signal, since square wave carriers are used in
this alternative.
E. Alternative 2 - Two Channel Breadboard Using Single
Conversion
As recommended in the FDM system study, contract
NAS8-25987, the double conversion of all channels is required
in the demultiplexer when operating the demultiplexer off a tape
recorder. If, however, an FM removal system as described in
paragraph II.H.5 is used, it will be possible to provide sin-
waves inside the-carrier phase locked loops. Single conversion
using true multipliers will, therefore, be possible and there
is no longer a need for in-line filtering (except for output
filters). This alternative is contingent upon the development
of the FM removal system, and proposes the use of single con-
version SSB, DSB and QDSB. The block diagram of the breadboard
is shown below.
/ 0i f)/, /r\ VKw
_ .CompositeSignal *
104 / y(J KHZ—
/• /, / C\ Ij-TJ n» «* 1f\H / JJ KHZ
/
\r
\/
\' /
\
X
t
X
tx
tX
\/
^7
\
\
i
Network
8 kHz LP
kHz , LP
^ 1 f. l-U™v v ID KHZ,* "7 LPT
.. ^ n PTTAMNTTT
^~ 90 CHANNEL
44/90 kHz f
29.
The breadboard uses true multipliers, and a new design is required for the
synthesizer phase locked loops to provide sin-wave outputs.
IV. RECOMMENDATIONS
The study to evaluate a wideband conversion of the Adcom
Demultiplexer shows that it is possible to convert the unit as described
in paragraph III.B and III.C. However, such a conversion will have the
following disadvantages:
(i) The system will not operate at both narrowband and wideband
simultaneously without using two or more pilot frequencies.
(ii) The system will not provide optimum performance without re-
design of all phase locked loops.
(iii) The matching of pilot and channel filters, with the channel
filters all being at different frequencies, is extremely
difficult and costly.
In addition to the above disadvantages, the system will not prove any new
circuitry usable in a state-of-the-art FDM demultiplexing system.
The basic problems in the design of an FDM demultiplexer results
from the tape recorder wow and flutter. Two approaches are available to
overcome these problems. A separate coherent pilot well outside the com-
posite frequency spectrum can be recorded simultaneously with the composite
signal as described in paragraph II.He4, with the optimum demodulation
method being double conversion as recommended in Reference 1. To verify*v
this approach the following steps should be taken:
(i) Develop circuitry necessary to add a new pilot on the tape
as described in paragraph II.H.4.
(ii) Develop a demodulator breadboard as described under Alternative 1,
paragraph III.D.
(iii) Test the breadboard system.
The second approach requires the development of the FDM removal
system as described in paragraph II.H.5. If this system proves feasible,
the optimum demodulation system will be single conversion, this system being
less complex and cheaper than the double conversion system. To verify this
approach, the following steps should be taken:
30.
(i) Develop the FM removal system as described in paragraph II.H.5,
(ii) Develop a demodulator breadboard as described in paragraph
III.E.
The QDSB system is the optimum high accuracy FDM data system as
concluded in Reference 1, and the results of a successful completion of
either of the two above approaches will prove the feasibility of the wide-
band QDSB system with better than 50 dB suppression.
It is, therefore, our recommendation that either or both of the
above approaches be performed rather than' reworking the Adcom demultiplexer
unit.
31.
V. REFERENCES
1. Martin Marietta Corporation Report No. MCR-72-81, "Linear
Modulator, Final Report, Contract NAS8-25987, Modification 4
April 1972. . - . . . . .. .
2. F. M Gardner, "Phaselock Techniques", Wiley and Sons, Inc.,
1966.
3. T. Kaiser et. al., "Multiple Access to a Communication
Satellite with a Hard Limiting Receiver", Institute of
Defense Analysis, Report No. R-108, January 1965.
4. D. K. Weaver, "Design of RC Wideband 90 Degree Phase
Difference Network", Proceeding of the IRE, April 1954,
PP 671-676.
5. B. P. Lahti, "Communication Systems", Wiley & Sons, Inc.,
1968
Figure 1.
sss ^ 33 F=
64 kHz Pilot PLL OutputSpectrum in 2 kHz BW mode
System Input: )64 kHz at 0 dB68 kHz at -6 dB
Pilot level: 2 vpp
Figure 2.
$4 -s§p.--;- . - hm -r mEHffiSlfc. ,J'.::..iJi;aiM Ltjte
64 kHz Pilot PLL OutpuSpectrum in 2 kHz BW mode
Pilot Level: 2 vpp
System Input: 64 kHz at 0 dB68 kHz at -12. dB
Figure 4.
• ••BBBBB«B •••••••••• ••••• BBBB•BBBBBBBBB BBBBBBBBBB BBBBB BBBI••BBBIBBBBBBBBBBBBBHBBBBBBflBI•BBBB BBBBBBBBBBBBBBI BBBBB BBBIBBBBB BBBBB BBBBB BBBBB BBBBBBBBI•BBBBBBBBBBBBBBBBBBHBBBBBBBBI•••BBaBBaMBBBBBBBBBIBBBBBBBBI•BBBBBBBBB ••BBBBBBBiBBBBBBBBB• •BBaaaan •••••••••• BBBBBBBBIBBBBBBBBBB BBBBBBBBBB BBBBBBBBI•••BBCaaB* BBBBBBBBBB BBBBBBBBM
BBBBBBBBBB BBBBB BBBBB BBBBBBBBB•BBBB BBBBB BBBBB BBBBB BBBBHBBBIBBBBB BBBBB BBBBB BBBBB BBBBB BBBBm m m m m m m m m m BBMBBBBBBB BHBBBBBBI• BBBBBBBBBMBBBBBBMBBBBBBBBBI••••••BBBBRBBBBBBBBBBBBBBBBBI•••••••BBHBBBBBBBBBBBBBBBBBBI•••BBaflBBMBBHBBBBBBBBBBBBBBBa
BBBBB BBBBB BBIIBB BBBBB BBBBB BBBB•••BBBBBBHBBBBBBflBBBBBBBBBBBI• B BBB BBBBBMMBaBBBBB BBBBBBBBB
khf iaaBa^aM BBBBB BBBBBBBBBBBBBI• • Bh3*BhM0BB«H BBBBB BBBBB BBBR• BBBJHbWBMBHBBBBVB BBBHBBBBIm m m m m m w ^ n m m m m m m m m m m ••aaaaaafl
!:::::: s :::'•:;.;; «;%•::::• BBBBftB •!•*•> •••• •ISBB^'VBBl• •BBBBB B* jf mamm •••••IJLVBI••BBBBB ''.mam mmmm mmmmmaS.mm•BUBBBr BBBBB BBBB •llBBBBBkA,
mm* ;•• BBBB BBBB •BBBBUBBBBB' amo m m m m mmmm IBBBBBBBM
BBBBBB BBBB BBBB IBBBBBBBI•••••• BBBB BBBB IBBBBBBBIm m m m m m m m m m m m m m m m m m m m m m m•••••• •••• •••• ••••••••••BBBRB •••• mmmm •BBBBMBBM
»•••«» •••• BBBB •••••••••mmm mm* •••• BBBB •BBBBBBBI•••••• •••• •••• •BBBBBBBfl
• • • ••• ••• m m m m m m m amm •'••• •• •'••• ••••
M *• m m m m mm mmmm m m m m* mm •••• •• a mmm "mrn'mm• m m m m m m m m m'mmm m m m m• »• * ••• •• m m m m m m m m
•;«:«»::» &»»•:»••»
ijjtmjjffl S^ALE: 10 dB/l.b cm flf||H||||i||||||||||||||||||||||W::::::::; 15.7 kHz/cm \ ~ \ \ \ \ \ \ \ \ : \ \ \ \ \ \ \ - : : : l \ \ \ \ \ \ : : : ::::! : : : E ; \ l \ \ \ \ \ \ \ \ \ \ : ] : l \ l \ : l \ : \ : \ \ \ \ \ : \ \ ::::'-:::: :::i;:::: ::::::
i i l l i ; ; i : : : : : : i ; i : : ; E : : I I E : : : -0- : : I E : ::::::::::!::: i::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
TJI i i i i i^ ' i35*9 IK iliMlLLuiiy I I I rmTTTl H I 1 frr-Trm/ir • • " ^ " ^ n l l l l l H I I I I I I H m I I H I I I I I I I I I I I M
i::::::::::::::::::::::::: 1Q :::::::: ::j: : 148.8 :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
• • - - : : - ; : : • - • : : • • • : : : : : • • - - - TITl 1 1 TmTrlW : :: :: . - - . . . . . . . - . . . - - - . ::::
2 ' ''; '-I
; E ; ; ^ ; E ; E ; ; ; ; E E ; ; ; ; E ; ; ; E ; E ; ; ; ; ; E ; ; ; : E E ; ; ; ; E E ; ; J E E ; ; E E E E ; E E ^ ^ ^1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II 1 1 111 [ 1 ttl n 1 1 1 IF1 I I F I 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 III 1 1 11 F N J I I 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 II! 1 11 1 1 ' 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M )•
, . . . j -.. . . . - j - --- -I i • --- - - - I1 -•
m m m m m m m m m m m m m m m m m m m m m mmmmmmmm,mmat{»^f :••"""" •••••••••• •••••••••••••••••••• nmunL<«BnHnmunnuHnnnwuBUun>un«nB»«HH»nMmml
• •••BBBBBBH •••&••••&• ••••BBBBHMBBMBan^BBBBaBBBBBBBaaaHBBBBHMBBBMHBBBHBa BBBBB BBBBfl •••••BHBBBBVHBBBC^HBBBBHVBBiBBB BKBBBHBBBB BBBBBBBBBB B3MSBBBBBM BBBHB0BBHBBBBBI1 BBBBB BBBBB BBBBBBBBBMBBBBBBBBBB BBBBBBBBBB BBBBB BBBBB BBBBBBBBBN BBBBBBBBBB HBBBB BBBBB BBBBBBBBBB BBBBBBB^^x* IBBBBBBBBB BBBBBBBBBB BBBBBBBBBB ililBBBBBBBB BBBBBgBBBB BBBBI1 BBBBB BBBBB BBBBB BBBBB BBBBBBBBBB BBBBBBBBB BBBBB BBBBB BBBBBBBBIBIBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB&Bh'^aBBBBBBMB BBBBB BBBBBBBBBBBBMB*BBBBBBBBBIBBBBBiBBBBBBIlBIIBBBBBBBMBB BBBBBBBBBB BBBBBBBBBBBBnB^BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBflBBBBBBBBBBBBBBBBBBBBBBI^IMli^^ BBBBBBBBBBBB BBBBBBBBBBBPBBBBBBBBBBBBI BBBBBgBBBB BBBBIIBBBBaBBMBB BBBBBBBBBB BBBBBBBBBBBBBj::BBBBBflBHBBBBBBBBBBBBBBBBBBBBBBBBBBBB •••••••BBB ••••BBBBBB BBBBBBBTlBM •BB^TBBBBB BBBBBBBBBB BBBBEBBMBB BBBBBBBBnil BBBBB»BBBBBBBBI5 BaaHB BBBBB •BBBBBBBBB BBBBa *aBBB BBI/il ilBBBaB •BBaBBBBBB BBBBBVaBBaBBBBBBBBBBBBaBBBBaaB aBBHBBBI^MHBB^5:±S*''M* MBPBB^:?BBB BBIBBB BBBBB BBBBBBBBBB •••BBBaBBU BBBBflllBBBB BBBBI::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ! ;!:::s:::::::::::::::::::::::::::::::::::::::::::BQ H::::::i;m m m m m m m m m m m mmmmmmmmmmmmmmmmmiwmjmff 'fMMMMMMimmmmmmmmmmmmmmmmmmmH^-^j^-^sx^BBBBHBBBBBB BBBBB BBBBBBBBBflBBIBrABBBlBBBBBBBBI BBBBBBBBBB BBMHMBMBfla BBBBB •BBBBBBKrftfBflBBBBBBBBBBlBBraBBBBBBBBBBBBBBBBBlBBBBBB BBBBB •K^T«BBBMB^^=r? 1BBBBBBBBB BBBBB BBBBB BBBBI
• BBaBBBBBBBBBBBBBMBBBBBBBBMF^BffBIKA*it*jfBfJaBBBBBBBBBB ••••BBBBBB •••••BBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBM BBBMBBBBBBBBBBB BBBBI •••£.'£••••••••••••••••••••*•••••••••BBBBBBBBBMBBHBBBBalBBaaBBBBB(/IB
flB
afBlBlBlBf>lBl|lBB'HaB BBBBB ••••BBBBHB BBBBB BBBBB BBBBBBBBBB BBBBB BBBBB •••••••••• •••••BBBBB BBBBMBBBBBBBBBf^Z!*BB BBBBB •••BRBBBBBgBBBB BBBBIB BBBBB BBBBB BBBBB ••BBCBBBBB;'.'BI^^^^^^^^^Ht BBBBB BBBBB BBBBBBBBBB BBBBB BBBBB BBBBBBBBBB BBBBB BBBBB BBBBBBBBBBIBBBB BBBBB BBBBBBBBBB •BBBBBK*.'£. '7 JMIBBBBBBBI BBBBB iBBBB BBBBIB BBBBB BBBBB BBBBBBBBBIBBBBrVBlR'IBBVlPBBBWl BBBBB MBBBB BBBBB BBBBBBBBBB BBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBBIBBBB BBBBB BBJBBBBBBBB BBBBB BBHBa£?^«BBBBBB BBBBB B°BBBB BBBBIIBBBBaBBBBBBBBBBBBBBNBBBBJJBBBaBaBBBBBBBBBBBBBBBBBBflBBBBBBBBHMflBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBeBBBBIBBBBBBBBBBBBBBflBBBBBBBBBBBBBBBBB^.'T.^BBBBBBBBIIBBBBBBBBI• BBBBB BBBBB BBBBB aBBaa BBB/BBBBBB BBBBBBBKBBBBBBBBBBBBBHBB BBBHBBBBBBHBBB •BBBBBBBBB BBBBB BBBBBVUBBBBBBBM BBBBBBBBBB BBBBBBBBBBBBBBBBBBBB •••BBM^^'MBBMBBSBBBBBaBBI
:::::::::::::::::::%-::::::: B| |:::::::::::::::::::: ^:::::::::::::::::: ::::::K!!!!!!!!!!!R;:::::::::::::::::::H:»:::::::::::::::»:»• BBBBHBBBBB BBBBB r,*BBI BBBBBBBBBB BBBBB BBBBI BBBBB ••«•!] BBBBB BBBBB BBBBB BBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBBIBBBBBBBBBBBBBIBBBBB BBBBBBBBBB BBBBBBBBBBBBBBIBBBBBBBBll• BBBBBBBBBB BBBjBBijBBB* BBBBB BBBBI BBBBBBBBKI BBBBB MBBBBBBBBBBBBBKIBBBB BBBBI BBBBB BBBBB BBBBB BBBBB BBBBBBBBBBMBBBI BBBBB BBBBB BBBBI BBBBBBBBBB BBBBB HBBBB BBBBB BBBBB BBBll• BBBBB BBBBBBBBBr.lBBaBBBBBaBVBaaiaBBBBBBailBBEaBBBBBBBBBBBBBBiaBBBaaBBBB BBBBB BBBBB BBBfla BBBBB BBBBBBBBBB BBBBB BBB«B BBBBB BBBBB MBBBBBBBBB BBBBBBBBBBBBHMB BBBBB BBBll• BBBBBaBBBBBBBBivaBBBBBBBBBBflBaaaBBKIBaiiBaBBBHBBBBVB BBBBB ••••• BBVBBBBBBB BMBBBBBVHH BflBBB BBBBB BBBBBBBBBB •••BBaBBMBBBBBBBBBBB •BBMBBBaBa BBBHBBMBBH BBBBBBBBBHBaBllI BBBBIBBBBIBBBF/flBBBBIBBBBBia^BBflJBBBBBBBBBBBBBl1*BBB*B""l •••*•••••• BBBBBBBBBB BBBBBBBBBB BBBBBIBBBB BBBBBBBBBBBBBBBBBBBBBBBBBBBBaBBBBBBBBBBB BBBBBBBBBBBBBBBBBBBI BBBllIBBBBBBaBBBBBfl/iaBBBBIBBBBBI^^^^^^^^^^HlBBBBBBBBflBBBBBBBBBBBBBBBBBaBBBBBBflBBBBBBBBBBBBBBBMBBBBBBBBBBBBBBIBBSBBBBBBBBBBBBBBBBBBBBBIBBBBBBBBBBBBBBSBBBBBBBllmmm a mmmmmHmmmr l fmmmmmmmmmmmi m Mm^^ m ^^^Mimmmmmmmmmmmmmmmmmmmmmmmm •BBaiBBBBBBBBBBBBBBBBBBBRMBBBRBBaflBiBBBMBiBBBBaBBiBlIaBiBaaiBBBBBMiiflBBMBBiBBaaiSiBBiiiBafl• Ba^aBBDMaBBa^HBBBaaBBBBBBI^^^'T^lf'-^^H-.VKBaBBBBBBaBBBaaBBRBBaaBaBBBB BBBBB BBBBH aBBaaaBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBHBBBBaaBBBB BBBBI• ••*«• •!••(••••'••• •BBBa'BBBBBI^^^^^BAB^^^^BlBBBBBMB«BBBBBBBHaaaaBBHBBBBHBBaBBBBBBBaBHBaBaBHHBBaBHBBBBBBBBIBBBBaBHBBHBHaBBBHB BBBBB BBMHaBHBaBBBBaBBHBHB ••••••&&••
BBBB r/'BBfl BBBBB BBBBBI^^^^^^^^^^^HlBBBB BBBBBBBBBB BBBBI BBBBB BBBBI BBBBBBBBBB BBBBB BBBBI •BBBIBBBBBBBBBIBBBB1B BBBBB BBBBI BBBBBBBBBB BBBBB BBBBB BBBBBBBBBB BBBBMBBBB r/«BBBIBBBB BBBBB •HHHHBBfl99BPiVBBBBllBBBBHBBBCIBBBBBBBBI BBBBB BBBBB BBBBB BBBBI BBBBB BBBBB BBBBI BBBBBIBBBBBBBMB BBBBB BBBBI IBBBB BBBBB BBBBB BBBBB BBBBBBBBBB BBBBB
•BBBHBIINBI^aBIIIMBBIBBaVIIBaaillHMBBBBBBRBaHaMHiaBBBBBaBaaBIBBBBBflaBanaBBIBBBBBaBBBBBaBBaaaBaiiaaBIIBBBBiaBBBaBHMIiaBBIIBBilBBBBIBBBflBBBBBMIBBBiiBBaiiaBBIm m m m m m m m m m m f t m m m m ••••• BBBBB BBBBB BBBBBBBBBBBBMBBBBBBB BBBBBBBBBB BBBBB BBBBB BBBBBBBBBB BBBBBBMBBB »•••&•••• HBBBBBBBB BBBBBBBBBB BBBBBBHHBB BBMBBBMBBB BBBBBBBBBB BBBBRI •BBBBBBBBflBBBBIBBBBIBBBBBIBBBI BBBBBBBBBB IBBBHBBBBI BBBBBBBBBI BBBBBBBBBBBBBBBBBBBI BBBBBBBBBB HBBBBBBBBB IBBBBBBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBBBBBBBBBBBI BBBBIIIBBBBBBBB^HBBBBIBBBHBBBBBBBBBIBBBBBBBBBIBBBBHBBBBBBBBBBBBBBIBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBMBBBIBBBBBIBBBBBBBBB BBBBBBBBBB IBBBIBBMBBBBBBBBBBBBBBBBBBBBBBBBBMI•BBBBBBBBB/IIBBBBBBBBR BBBBB BBBBB BBBBB BBBBI BBBBB IBBBI •••••••••• BBBBIBBBBB HBBBBBBBBB BBBBBBBBBB BBBBI BBBBBIBBBB BBBBB BBBBBBBBBB IBBBB BBBBB BBBBIIBBBB BBBBB BBBBI IBBBIB BBBBI BBMV/CBBBMBBBBBIBBBBa BBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBM BBBBBBBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBB •BBBBBBBBIBBBBBBBBBB BBBBBBBBBB BBBBB BBBBIBBBBB BBBBB BBBBI•'•BBBB m m m t ' a m m m m m m m m m m ' m m m m m m m m m m ' m m m m m m m m m m ' m m m m m m m m m m ' m m m m m m m m m m ' m m m m m BBBBB •••••BBBBH •••••••••• BBBBBBBBMB ••••••BBB* BBBBBBBBBB BBHBBBBBBBBBBMB BBBBB BBBBBBBBBB'BBVBB•HBBBBBBB^IB BBBBB BBBB1 BBBBIBBBBB IBBBB BBBBB BBBBB IBBBRIB BBB BBBBBIBBBB BBBBI BBBBB BBBBI BBBBB BBBBH BBBBI BBBBBlBBBB BBBBB BBBBBBBBBB BBBBI BBBBB BBBBB BBBBB BBBBBflBBBB IBBBB• RBBBBRaM'tMIBBIBRBBBIBBBBBBaHBRRUBMIBBBIIHBBNIIIBIBIBBBflBBBBHBaBBBliRBIBBBBBBBBBIKMBBBBBBBBBIIBBBBBBIHBBBIIflBBRIBBBBBBflBIRBBBRBBBBHBBBBMIBaBBBBBBBBBIBMMBBBI• •BBBBBBt|«HBBBBBIBBBIIBBBBBBBBI BBBBIBBBBB IBBBBBBBBM BBBBBBB.BBHBBBBBBBBBB BBBBBBBBBIBBBBBBBBBB BOBBBBBBBB IBBBBBBBBI BBBBBBBBBIBBBBBBBBBB BBBBBBBBBBBBBBBBBBBB BBBBI<BBKBflBBlT/BM,MBBBBBBBBB ••BBBBBBB*BBBBBBBBBBBBBBMBBB«MHBBBBBMBBHBBBBBBBBBBBBBBBBBBB* BBBBBBBBBB BBBBBBBBBBIBBBBBBBBABBBBBBBBBBBBBBBBBBIBB BBBBBBBBBB BBBBBBBBBB BBBBI
MMBBiiBBBivn m m m m m m m m m m m m m m m i m m m m m m BBBBB m m m m m BBBBB BBBBIIBBBBBBBBB BBBBB BBBBBBBBBB BBBBB BBBBB BBBBHMBBBH BBBBB •••BBBBBB* BBBBB BBBMIB BBBBB BBMHBBBBMB BBBBB BBBBBBBBBB BBBBB(«»••• m m n m m m m m m m m m m m m BBBBB aBBBa •••••••••••••••••BBB ••BaBBaaaa •••••••••• •••BBB»BMa«aBBBBa»a •••••••«!••••••••»•••••••• ••••••••••••••••••••••••»•••••••_•<•• ••••B• TiBBBB ••'«•• laBBBBBBBRBIBBB BBBBB BBBBBIIBBBBBBBBBIBBBI BBBBBBBBBB •••••••••• BBBBBBBBBI IBBBBBBBB* BBBBB ••••••••••••••B BBBBBBBBBBIBBBBBBBBB BBBB afafafafBfflffflffflffflffjffBlBBBBBBBBl• fc'v«BBBBCJ*B««BBBB«BBB**BBBBIBBBBBBBBIBBBB»aBBB»BBBBHBBBBBBBBBI BBBBB BBBBB BBBBBBBBBI BBBMBBBBBB IBBBBBBBBB IBBBBBBBBI BBBBB BBBBI BBBBBBBBBB BBBB ^^^^^^^^^^^^^^^H• •^BBB •!]•••••••• ••••• BBBBBBBBBIBBBBB BBBBB BBBBB BBBBI •BBBBBBBBM BBBBB BBBBB BBBBB BBHBB ••••• BBBBB •••BBBBBBB •BBBBBBBBB BBBBH ••••& ••••BBBMBBaBBBBTflBBBBBBBBBBBBBBBBBBBBBBBB• •^••••'TBEB ••••••••••••••• BBBBB BBBBB •••••'•BBBB BBBBM •••••••••& BBBBBBBBBB BBBBBBBBBB ••••BBBBBB'BBBBBBBBaBBBBBBBBBBBBBBBBBBBBB BBBBB BBBBB BBBB ^^K^V7B^B«BBBBVTBBBBlIBB^a«Bfl«B«BaBBBBBBBaBBBBBBBBBBBBBBBMBBBIHBBBBBBBBIIBBBBBBBBBBBBBBBBBBaiBBBBBBBBBIBBBBBBBBIBBBBIBBBBIIBBBBBBBBlBBBaBBBBBBBBBBBBBBBBBBB*^Bff^jVQ!C;V3!T*TB|• •B\tBffr/JB«B BBBBBBBBBI •BBBBBBIBVBBBBBBBBBB BBBBBBBBBI BBBBBBBBBB BBBBBBBBBB IBBBIBBBBRBBBBBBBBBMBBBBBBBBBBBBBBBBKBBBBBBBBBBBBBBBBBIflBBBBBBBB^BlXlflLBUBi^BiZlIlU:m.V<l »?)••• BBBBBB BBBBI aBBBBBBBBB IBBBBBBBBI •••BB BBBBI IBBBBBflaBI BBBBB BBBBB BBBBB BBBBI •BBBBBBBBB BBBBBBBBBB •••••••••• BBBBBBBBBB BBBBH BBBBB mmmmfmmmmmmmmm^mmmmmmmmmmi^mmmmml
BBaVM/BBBaBBBBBBBBBBBBBBaBBBBBaBBBB BBBBB •••BB BBBBI •BBBBBflBBB BBBBBBBBBB BBBBB BBBBB BBBBB BBBBB BBBBB BBBBB •BBBBBBBBM BBBBBBBBBB BBBBB BBBBB •••• •"VH^lr^BBBBf!^BHBBHfV
m m m u w U m a m m m m m m m m m m m m m m m m m m m m m m m BBBBBBBBBB BBBBB BBBBB •••••BBB»BaBBBBaBBBBB««BBBBBBa ••••BBBBBB BBBBBBBBBB BBBBBBBBBB BBBBBBBBBB BBBBB BBBBB BBBB *am*ijjlL*I^f<l^LlJBT-i"IBBBBX<BBBB BBBBBBBBBI IMBBBBBBBB BBMBBBBBBBIBBBBBBBBI •••••BBBBB BBBBBBBBBI IBNBBBBBBIBBBBBBBBBB BBBBBBBBBB BBBBBBBBBK BBBBBBBBBB BBBBBBBBBB BBBB •••••••••l kBi• •BBB(»BBBBBBBBaflBBBR BBBBB BBBBB BBBBB*BBB« BBBBB BBBBI BBBBBBBBBB BBBBB UBBBI •BBBBBBBBI IBBBB BBBBB IBBBBBBBBB IIBBBBBBBB IBBBBBBBMB BBBBBBBBBB BBBB 'aBBBBBBBBBBBBBBBBBBBBBBBBlIBBBBffBBBflBBBBBBBBBBI BBBBBBBBBB IBBBBBBBBI BBBBBBaBBIBBBBBBBBBB CBBBBBBBBB BBBBBBBBBI BBBBB BBBBI BBBBBBBBBI •IBBBBBBBl •••BBBBBHB ••BBBBBBBB ••BBHlBBBBBBBBBBBBBBBBBBBBBBBBlm-mmmmilmmm*m m m m m m m m m m m 9 m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m BBBBBBHBBBBS«BBBBBBBIBBBBBBBBB BBBBBBBBBB ••BBBBBBB* •BBBBBBBBBHSaaBeaBBBBaBB^BBBBBBBBBBBBBBiBBBBBBBBBl•.'••••U ••»•«••••••••••'••••• BB»BBBBBBB •••••'••••• ••••••••••••••••*•••••••• H««B*B-«BMM •••>•••••••••••••••«•••••••••*'•••••••••• ••BBBBBBB* aaaMB ••••• BaBBB BBBBB •••••BBBHBBttHBBBflIBIIB BBBBR BBBBBBBBBB BBBBBBBSBB BBBBBBBBBI BBMBBBBBBI BBBBBBBBBB BBBBBBBBBM IBBBBBBBBI BBBBBBBBBI BBBBBBBBB* BBBBB BBBMB BBBBB BBBBB BBBBB BBBBB BBBBBBBBBB BBBBIBlB_B_BBBBaaBflBBBBaBIHaBBiaaBBB BBBBB BBBBB BBBBB BBBBB •BBBIBBMBBBBBBB BBBBB BBBBB •••BBBBBBI •BBBBBBBaB BBBBBBBBBB BIBBB BBBBH BBBBBBBBBI BBBBB BBBBB BBBBBBBBBB IBIBBBBBBB BBBBI^^••^^^••aBBCBBBflVBaBBBBBaBflaBBBMBBBBBBBaBBBBfllBBBBflBBBBaBBBBBBBBBBIBBBBBBBBBBBBBBBBBBB BBBBBBBBBI •BBBBBBBBN BBBBBBCBBB BBBBBBBBBB BBBBBBBBBI IBBBBBBBBB BBBBI^B7^^HRBflB«BBaiaBBBflBBBBBBBBBBBBBBBBBBBBBBBBIBBBBBBBBBBBBBBBBBBaBBBBMBBBBIBBHBBBBaBBBBBBBBBBBBBBBBBBBBBBBHBBBBHBflBaBBflBBBflBBB«BBBBBBBBBBBBBBBBBBBBB^•B»J»lB_BRBB|lBaBaaBBIBBBBaBB*'BaBBBBBBBBBIMBBBaaaBBaBHBBBaaBBBBBBaB BBBBB B«B"« ••••••••••••••• ••••••••BB ••••BBBBa* ••••••BBBB BBBBB BBBBB BBBBaBBBBB BBBBBBMBBB BBBBBI^^SM^^HRBBBBBBBIBaBBBBBBBB BBBBBBBBBB BBBBBBBBBI BBBBBBBBBB BBBBBBBBBB BBBBBBBBBI BBBBBBBBBB BBBBBBBBBI ••BBBBBBBBBBBBBBBBMH BBBBIBBBBB BBBBBBBBBB BBnBBBBBBBBBBBI^••^^^•iBBBBBBBIBBBBB BBBBB BBBBBBBBB« BBBBBBBBflfl ••••BBBBBB BBBBBBBBBB ••••••BBBB IBBBBBBBBB BBBBBBBBBB •••BBBBBBH BBBBBIBBBB IBBBBBBBBB BBBBBBBBBI IBBBBBBaBB BBBBIBBBBBBBrMBBBBBBl1 ••• BBBB1 BBBBB BBBBH BBBBBBBBBB BBBBB BBBBB BBBBBBBBBB BBBBB BBBBB BBBBBBBBBI BBBBB BBBBa BBBBBBBBBI IIBBBflBBBH BBBBB BBBBB IBBBBBBBBB BBBBB BBBBI BBBBB BBBBB BBBBB----•• •••••.•••BBaBBBiaa BBB • • BBB BBBBB •••••BBBBBBaBAlBBBBBBBBBB BBBBB BBBBB aBMBBBBBBU BBBBB BBBBBaBBflBBBBBaBIBBBBBBBIIBBHBBBBBB1BaB«MBBBBBBBBBBflBBBBaBBBBBB' lBBB BBBBB• ••••• •BB>aBBBBVMBB«BaBBBaBaB*MJ1BBBBBBBBBBBBBBBBaBBBBBaaflaBBBBBBaBBBaBBMMBBBBBBB'flaBBBBBBBB<BBB«iBBB«Ba'«BaiMaBBBB«aH«BBBBBB1BBBBBflBBBBB<BBBBBBHBBBBBBBMB«BBBaBBBa••BBBB •BBB,aBBBBBBBBIBBBBBBBBaBBBKBBBBBBBBBBBBi«BBBI BBBBBBBBBB BBBaBBBBBB BBaBBBBBIMHBBBBBBBBI MBBBBBBBBR IBBBBBBBB* BBBBBBBBBB BBBBBBBBBR BBBBBBBBBI IBBBBBBBBBBBBBI• IBBBI BBB •,••&& • BBB Bt BBBBB BBBr.aBaBBBBHBVBBBBBBBiBBB BBBBB •BBBBBaBBBBflBBBBBBBBBKBfllflBBBB BBBBB BflBBBBBBBIBBBBBBBBBP BBBBBBBBBBBBBBBBBBBB BBBBBBBBBI BBBB.BBBBBB BBBBB• ••BBB BBBOl««B«BBBBBI«BBBflBB«'JBl1>S»<lflMVIBBBBBBMaBBBB •BBB«»BB»B BBBBBBBBBB BBBBB BBBBI BMBBBBBBBB BBBHB BBBBB •BBBBBIBB* BBBBI BBBBB BBB BBBBBBBBBBBB BBBBB BBBBBBBBBB BBBBB• ••aaa ••••,•••&« •••••••••a a Mb BBB«r.«BBBBBBBaaBBBBBMHMBBBM«aaBiBBBBaBaaaaHiBBa*aBBai BBBBB aaBBH ••aaaaBBBH •••aaaaBB* BBBBB BBBMBBBBBBBBBBB BBBBB BBBBB BBBBB BBBBB BBBBBm m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m ' m » m m m » m m m m m m m m m m m m m m m m m m m m m m •••••BBB •••••aaaaBBBRaBaa a BBBB •BBHBBBBBBBBHBBBBBBBBBHBB BBBBB BBBBB BBBHB Ba»BaBBBBHHBBB«B* BBB BBBBBm m m m a m •••• •••••••••••••••••••••r. •••••••••••••••••••••••••••• •••••••••.HvBBBBBiiiBBaBBinaSiiHBiRiBRKinaBBBiiBEaBBaiiaSBBMaiiiBMMBiBBBiBBBBBi ••BBBBMBBBBBBBIB B B D B B BBBB.iaBBKIIBailBBBB«BBBBBaBBBB*BBBBBBBBBIIBBBBBaBaRBBBBBBBB1flBBBIBBBBB«aBaRBaBBB BBBBB •BBBBBBBBB •OBBBBBBB* BBBBIBBBBB •BBBB«BBBI •••BBBBBBB BBBBBBBBBB BBBBB»,•»••• ••••,•••••••• •,••••••••••••••••••••••••••••••.••• BBBBBBB •••••. ••••••BBBBBBBBI »•••••• •• BBBBBBBBBB IBBBB BBBB* BBBBB BBBBB BBBBB BBBBB •BBBBBBBBI IBBBa BBBBJB &••••m m m m m m •••• •••••••« • ••••••••••BB* •••••••••••••••••••••••••••••••••••• «•'•'••'• •••••'&••*•'••••• m m m m m m m m m m ' m m m m m m m m m n •••BBBBBBB BMBBBBBBBB •••BBBBaaakBaaaaa'aBB BBBBB• ••••• m m m t m m m m m m m m •••BBBMBBBBBPBBBBBBBBIBBBaBMBBBBBBBBKBBB>BBBBBBBBBlt»-*^»««IBaiKB««»BBBBiiBBBBBIIBaBlilBBBBBBBSBBBBBBBBBBIBBBBBBBBaBBBBB>BBBBBBBBBBBBBflBBBai• «•••• •••••••••••• •.•••••••BBBBBBBBBBBB>.BBBBBBBBBBBBHHBRBHBIIBBBBBBBBaM^:^^'.W«**B •••!!•••••• •BBBBlflBBBIBBBBBBBBBBBBBBBBaBBBBBaBBBaaB BBBBBBBBBB BBBBBBBBBBBBBBI•,••••• ••••,•••••••• I •••••••••••••••••&•••••••••••••••••••&••••••• ••IBB ••••*V*VaVBB«lt *••••• ••BBHaBBBMBHBBBBBBBVBBaflBBBBBBIBBBaaaBBBiaBBBBBBBXBBBBH BBBBB BBBBIm m m m m r . m m m m ' m m m m m m m m u m m m m m m m m m m •••••••••• m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m - m m m m m m m m m m m m m m m m m m m m BBBBBBBBBI •BBB**BBBB BBBBIm t m m m m m m a m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m mmm m ••••••••••••••••••••••BBBB ••••••»•» •••••••,&•• BaBa»»»»»«'»»»"»»»B«*»»aB»»nBa»B««"BaBB«B •••••••••»•••••••••• ••••••>•••••••••;•••••••• •••BBB ••••••••••••••• BBBBBBBBBB BBBB* •BBBI.««BHBBBBaB BBBBBBBBBI IBBBIBBBBB aBBBBBBBBa •BBBBBBBBBBBBBBBBBBB BBBBBBBBBI BBBBBBBBBB BBBBBBMBBB BBBBB
•'••""••••••'"•••••ll 1 ••••••••••'•"••••!••••"••••"••• "*BB•I* » mm m'm m •»•:•_••• •'< • « • ••••••••••:»••••••••*_••••••«•*• ••••••••••.••••• •••••BBBBM B B B B B B B B M * ••••••••••••••• ••••••••a« ••»••••••• ••••••••&•'•••••••••• •••••••••• •••••
!•••••••••••••••••••••••• !••••••••••••• •••••••••••••••••••••BBBP ••»••••••• ••••••••••••••••••••••••»••••» IUBBIBBBBI
I IHI i
illliillliiy
::::::::::::::: : : °:::::::::::::::::::::::::::::: :::::::::::::::::::: i:::::::::::::::: :::::::::::::::::::::::BBBHBBSSBHBBS B • BBBI BBBBBBBMBIIBBBBBBBMBBBBBBB HBB BBBBBBBBBB BBBBBBM BB BBBBBBBBBB BBBBB BBBBBBBBBBBBBB IBBBBBBBB::::::::;:::::: : i :::::::::::::::::::::::::::::: :::::::::::::::::::: ::::::::::::::::: :::::::::::::::::::::::
SHOH i 1 1!"•I""!!!!!!!!! SS*S!SSSS!SSSii!SSSSS!SSuSS!SSSS!!SSSS!SSSS!SSS!iS I!!!!!!!!!!!!!!!! S!SSSSSSSSSSS!SSS!!SSSS.....[filiilliiifll Elll[|j[jl!!!lj!|j|jjj!! jlilfiii
••:::::::;i:::::::::::::;::EE-:::::::E:::::::: ::::::•:!:::••:::•••:::••::::•:: :»:i:i:i:::::ii :::::::i:::::::::iBBBBBBBBBBBBBBBBBBBBBBBBIBBMBUBRBBBBBBBBBBBBBi BBBBBBBB IBBBBBBBMBBBBBBBBBBB BBBB BBBBB BBBBI BBBBBB BBBBBBBB IBBJBBBBBBC
!
i I L I I I I I I UNI i i i i i i i imt im
Channel inbut signal harmonics 1st 0 dB2nd -54 dB3rd -60 dB
This implies that the 176-168 kHzchannels will show up when demod-ulating channels 12-4 kHz respect-ively at -38 dB (if input LP isnot present).
Output MJ6 Multiplier (1st up-conversion stage)with 16 kHz channel signal input to Jpw passfilter (Pig. 5.) converter. i
Figure 6,
ir:r!rci ii-jfc'iiiVittnrf^- W
Figure 11.
44
Figure 13