THE POSSIBLE TRANSMISSION BAND IN THE CASE OF LONG-DISTANCE TROPOSPHERIC PROPAGATION
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP81-01043R002900040003-9
Release Decision:
RIPPUB
Original Classification:
K
Document Page Count:
65
Document Creation Date:
December 27, 2016
Document Release Date:
February 4, 2014
Sequence Number:
3
Case Number:
Publication Date:
November 14, 1958
Content Type:
REPORT
File:
Attachment | Size |
---|---|
CIA-RDP81-01043R002900040003-9.pdf | 7.22 MB |
Body:
ii
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Area and subject category
Russia (in Russian)
2. Title of Project in English
Radio Engineering
RADIO ENGINEERING
(RAD I OTEKHN I KA)
vol. 11, No. 9
September 1956
Pages 1-80
3. Author
?????????????1?11.
4. Security Classification
Unclassified
5. Foreign Language Title of source material Radiotekhnika
6. Date and data on publication Vol. 11) No. 9, September 1956
.1(7. sookrrts- to+,
PREPARED BY
LIAISON OFFICE
TECHNICAL INFORMATION CENTER ? STAT
MCLTD
WRIGHT-PATTERSON AIR FORCE BASE. OHIO -
Declassified in Part - Sanitized Copy Approved for Release_@ 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
__TABLE OF CONTENTS
Page
GJ 1. 'hie-Possi - - -
- Possible Transmission Case of Long-Distance
"i Tropospheric Propagation by V.N.Troitskiy -1
2. A Method of Calculating Propagation Constants in Waveguides
-3 with Non-Ideally Conducting Wallaby L.N.Loshakav 8
121
; The Effect of Precipitation on the Electrical Properties
74 1 of Wire Antenna Array Surfaces by. V.K.Paramonov
: 4. Receiving Antennas and Industrial Radio Interference by
V.V.Roditi and M.S.Gartsenshteyn '
1811
...-
_. 5. Mutual Correlation of Fluctuation Noises at the Output of
This translation was prepared under the auspices of 20. Frequency Discriminators by M.V.Ma.ksimov
. -
the Liaison Office, Technical Information Center, Wright- 22_1 6. Study of Shock Excitation and Forced Quenching of Oscillations
of Quartz by T.N.Yastrebtseva and I.G.Akopyan
24 -1
Patterson AFB, Ohio. The fact of translation does not
- 7. Feedback in Transistor Circuits by Ya.K.Trokhimenko
guarantee editorial accuracy, nor does it indicate USAF
__A 8. Diminishing the Pulse-Front Distortions in Video Amplifiers
Using Junction Transistors by T.M.Agakhanyan
approval or disapproval of the material translated. 1
- 9. The Principal Properties and Characteristics of the Synchronous
Filter by N.K.Ignatlyev
39_J
10. The Use of Harmonics in Calculating Envelopes by A.A.Kulikovski.y
3
11. All-Union Scientific Session Devoted to Radio Day (continued)
3C
-J
-- Section of Receiving and Transmitting Equipment 108
40__
-- Television System 111
-- Radio Broadcasting and Acoustics Section 114
44_1
Section for Semiconductor Devices 116
1 46=j 1
i _J Radio Engineering Section . 1
117
48_1
-J New Books 119
50?
Electronics Section
27
38
53
63
74
STAT
82
102
107 I
?STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
? THE POSSIBLE TRANSMISSION_BAND_IN.THE CASE OF LONG-DISTANCE
TROPOSPHERIC ;PROPAGATION
by
V.N.Troitskiy
1
10
1.1-1
18
20_
24-1
26_1
28-1
Determination of the band width of the communications spectrum that can be
30_
1
The question of the distortion of transmission in the case of long-
distance tropospheric propagation is considered under the assumption
that the atmosphere is anisotropic;and that the horizontal inhomo-
geneities of the dielectric constant are smaller than the vertical
inhomogeneities. Formulas are given for determining the band of
1
the spectrum of communications that can be transmitted without dis-
tortion. The influence of antenna directivity on the possible trans-
mission band is analyzed.
?
transmitted without distortion in the case of long distance tropospheric propagation
? 32_
'is a problem of great practical importance for practice.
34_
The fundamental cause of such distortions is the formation by the field at the
1
place of reception of a large number of waves with unequal propagation times, owing
31.1-1
:to the different length of the path of each such wave. If we bear in mind that the
40._J
- troposphere is in fact anisotropic, and that the horizontal inhomogeneities of the
42_1 1
_dielectric constant are smaller than the Vertical inhomogeneities, then, as has been(
441
Shown in my previous paper (Bib1.1), the field of a tropospheric wave may repre-.
46-J
_isented as the superposition of the waves separately reflected from inhomogeneities of
1 1
__different dimensions. In this case there will be a continuous spectrum of these in-
t
- homogeneities, as defined by the KoImogorov-Cloukhov law.
52-1 !
4
54:1---- It has been shown (Bib1.1) that under! this assumption the field strength of a I
__tropospheric wave is defined by the equation i
56 i
1.
- -
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
mi
a
0
',6?6
?
0 0 1 00 8 cp
E2 C p4 I 3 e-4'Psinv'S e- xdx,
0 0
6_.where_C tconst, p 1 1 = dimension of_the_inhomogeneityx.X -the w.avelengt111
a
o
the rays and the earth (Fig.1),-x= height of point of reflection above
the limit of direct visibility on the side of the receiving antenna and
C--11 = distance between stations, p = a = geocentric angle between
10_,:tact between
lq?jthe point at
points
of con-
on that of the transmitting antenna.
15 1
_j
18_J
1
29:1 If originally the trailing front of the wave had an ideal form (Fig.2a), then
this front will become blurred (Fig.2a) at the place of reception, owing to the ar-
__
25__Irival of a number of reflected waves with different lags. The value of the ampli-
The time lag of the individual rays will be
tc= 2x sin ch
C
fc__tudet of the received signal at any point T of the frailing front will be
30_1
co co
321= C p4/ 3 e--UpsinT.4
x,
_d
V._ where
36d
401
4211
44
46_1
CT
' 2 sin Tv
! -
Integrating, by analogy to our earlier procedure (Bib1.1), we get
n
shiF-50
= quantity characterizing the intensity of
1
the front of the received
by the equation:-i
8p
x
s dx,
4::--where Eo = field in free
the inhomogeneity.
-1
Thus, by eq.(4),
5. 'be of the form defined
space,
and Fo
[1+
U
R sin2 ffo T
CT
(2)
(3)
(4)
2
signal of the tropospheric wave will
STAT
STAT
6
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
4
? .6.-11 ?
3.
where E
- 4*
7r.
10--
12__acts
it?._- that a band filter with a'finite pass band also
16
18_-
20___
24--]
t .
1 Z, --
1 CT )\ r
. - - - MO 70
1
77 ?
111
?
?
?
(5)
= amplitude of tropospheric wave under stationary conditions.
Consequently the medium in which long-distance tropospheric propagation occurs
1
on the signal in about the same way as a certain band filter. It is well known
26:d
23,
__This blurring is usually characterized by the
30__
_.the output signal from. 0.1 to 0.9 of its steady-state
Fig.1
causes blurring of the wave front.
b)
Fig.2
time of variation of the amplitude of
32-1
?ideal filter with
34_1
(Bib1.2) that the
__- filter
33--
40_1
49_
In our case
a 4.
46__
43_
value. In the case of an
a rectangular frequency characteristic, it is commonly known
lag time T is inversely proportional to the band width of the
0.86
-
the lag lag tine may be easily determined from the equation
Max "12,5 IZ sin To
(7)
?rag
(6)
?c
On determining the value of Af from eq.(6), and substitution
50__
_max from eq.(7), we get the "pass band of the medium,' in the case of tropospheric
_
of the value of
c2-1
?
--propagation:
SA
56-I
= 0,393 _ s_cn
R. 1.
(8).
yap
J
S.
?
It is easy to determine the angle Tolfrom geometrical considerations,. if we
know the distance and height of the antennas:
4
6
e???????
_where
10
12
^ ?
20
92_1
24:j
25
70
--not
32_
- greater
sin To =
act.= equivalent radius of the earth,
1
1
q
R1+R2
(9)
R2 = distance from horizon to transmitting and receiving antennas
1
On substituting the value of sin To from eq.(9) in eq.(8), we obtain:
respectively (Fig.1).
Aimax
ca2
= 1,37 .
R2(1 - q)2
(10)
We have until now been considering the distortion of a signal for the case of
very strongly directional antennas, when the angle of directivity is markedly
- lected.
36_J
a
2,9-J 1 1
ceiving
1
42_1
_J
45,
1
,40-1
ev ?1
50-1
52,73
than the angle
so that the directivity of the antennas could be neg-
Let us now consider the general case of directional antennas, assuming that
nd f2(a2) are the characteristics of directivity of the transmitting and re- ?
antenna (al and a2 are angles measured from the principal direction).
1
Then eq.(2) takes the form:
Co oo
I
-E2 = C 1.p4r- --1 j4cP Sin To
cll j
1
By geometrical reasoning we get
$142
e? R x dx.
2x
212 (12)
-
For an approximate evaluation of the influence of the antennas on the pass
4
STAT
_
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
?
l_band, we shall approximate the directional characteristics of the antenna in the
2-4 i
.... form of the right triangles of width 40, where go g....angle of_directiyity of_antennas
4....) !
In that case the inteeal of eq.(11) will be of the form
i
_ - - -
8? co
C f p41? Fe ?4 '7' sin 7? di f e R x dx ,
-p-.2
10-1
12
where
14--
15-1
10_
20_
22_
24_
Under such conditions the band Af varies with time. The value of these varia?
the actual spectrum of inhomogeneitiea.at_the ndii
Integrationis
As a result we get
1
Pw=0,191/7.:17f1,13 11
25_ sing-fs
dh CT
R sin) f
28_
The form of the wave front is defined by the.equation:
41 42 43 44 0,5
Fig.3
of time differs from the Kolmogorov spectrum. If there are gaps in
then selective fading is possible, leading to additional distortion.
2 3 4 5
Consequently,
32--
- Troitskiy,V.N. ? This Journal.
- 2. Siforov,V,I. 40_4
? Radio Receiving
46
48_
52
5.et
Paper received by editors 2 April 1956
BIBLIOgRAPHY
11 (5) (1956)
i ?
Equipment, Chapter 16. 1954
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
0--
ihere
2-4
ea = pass.band_in the .case of directional antennas,
a =t pass band in the case of nondirectional antennas.
6 1,4
Under such conditions the band pf varies with time. The value of these varia-
2:1
;tions.depends on how strongly the actual spectrum of inhomogeneitiea.atthe_indiyi-,
Q1
r .
.50
?,?...
....1 2 sin is
Q ( sina?14 ) %=2'5 1( 1 + ? a? ?) { 10-2[( 1+
.10 --1_1 11 3
(18)
P
?
-i- 2 sin ?70) ? 1] + 1 ] " ? i}.=.1.1
_J .
14_
Thefactor Qtc2 sin To ) defines the influence of antenna directivity.
I
16
ao
18:: For ? 1,
sin yo
20 Q V:--) ? 1.
ma gpo
22__ ao
Figure 3 shows the relation between the value of Q and the parameter sin To ?
24-1 (4o
Under the condition that 1, we may obtain from eq.(18):
-- sin
26-1 co i
1
Q5 sin's,
287 V. ? ( 19 )
__ I
Consequently, in this case 1
Rs (1? q)
As will be seen from eqs.(10) and (17), the pass band of the medium depends
sharply on the distance. In the case of actual traffic lines 200-300 km long, this
-
4/--band is rather wide. Thus a communication can be transmitted with a rather wide
spectrum by tropospheric propagation without appreciable distortions.
46__
It must, however, be noted that all these conclusions are true only for average
43--lconditions, since the assumption that the .spectrum of inhomogeneities is of Kolmo-
__
so.:gorov character is evidently true only on the average. At individual instants of
time 1
the spectrum of inhomogeneities differs significantly from the Kolmogorov
54.J
_,spectrum.
56.?J
4
2 --
41 42 43 44 0; 47 I 2 3 4 5 7 V
sky:
26_1
Fig.3
23:1
32__dua1 instants of time differs from the Kolmogorov spectrum. If there are gaps in
32_.the spectrum, then selective fading is possible, leading to additional distortion.
34__ Paper received by editors 2 April 1956
BIBLIOgRAPHY
1. 3E--
Troitskiy,V.N. - This Journal. 11 (5) (1956)
40
2. Siforov,V.I. - Radio Receiving Equipment, Chapter 16. 1954
42
44
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
a
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
0 W..?????????????????????1
2-4
1
A METHOD OF CALCULATING PROPAGATION CONSTANTS IN WAVEGUIDES
4' 1
:IF WITH NON-IDEALLY.CONDUCTINGWALLS
6
15..) I
?1;
by
0
.-- 1
? __ L.N.Loshakov,
i
Full Member of the Society
1
???????I
16 -
18_
? 20
The author describes an approximate-method of calculating propagation
constants in waveguides with non-ideally conducting walls, based on
the application of the conjugate lemma, gives a confirmation of the
1
correctness of the method, and considers the results of its application.
24 1. It is well known that the rigorous, theory of the propagation of waves in
non-ideally conducting walls is based on the solution of the Maxwell
1
application of the Shchukin-Leontovich boundary condition, which
1
__As true for sufficiently high conductivity of the material of the wall. Such theory,
30_1 1
however, cannot be extended to the case of a waveguide of rectangular cross section.
32_4
? __1The method of approximate calculation of damping in waveguides, which is termed the
34_1 ?
_iskin-effect method (Bib1.1), and is less rigorous, but more general, has therefore
36_1
-lenjoyed widespread use in engineering practice. This method yields the damping con-.
38-1 1
-Jstant for operation of a waveguide at frequencies higher than the critical, but does=
_.waveguides with
26-1
equations and the
28
40_1
--snot permit calculation of that constant for the critical frequency, nor at all oi [
22_
-- the propagation constant in a waveguide.
44-
In an earlier paper (Bib1.2) I have pointed out the possibility of approximate
46-1
--;calculation of the propagation constant for a waveguide by the aid of a method base
48 '
50 on the conjugate lemma. The proposed method possesses considerable simplicity and
-1
-: generality, and involves no restrictions on the working frequencies. .
52.:1
In the belief that this method might findpractical application, and in view
5A
- of the absence of any discussion of the possibilities of that method in my earlier
56
8
_ _ .
: - a-- -al --
_paper (Bib1.2), I present below accuracy of the method
2-4
?and give certain results obtained byits_use.
4_I
We note that the results given below may also be found by means of the Umov-
.
Toynting complex theorem; and that method of calculation has been discussed in a
"lemma method" described below is more con-
10._Ivenient.
1 19
2. It follows from my earlier paper (Bib1.2) that by applying the conjugate
0
:11Ork by G.V.Kisuntko (Bib1.3); but the
16 -
'lemma (the Lorentz lemma, written on the replacement of one of two independent
;fields by a complex conjugate field) to the element
1
we may obtain the starting relation:1
f(f-Eiri;1] [E1,74])71,ds = f
The following notation was used in writing this relation:
of length dz
VrTfir.litd/.
17,3! = electric and magnetic fields in the line,
9C_1
'ideal conductivity of the walls (the field sought), E11, Tin = electric and magnetic
20_J
fields in the same line with ideal conductivity of the walls (the auxiliary field),
.s = cross section of the line, L = contour of the walls, -a; = unit vector of longi-
of the transmission
(1)
taking account of the non-
3 c O. If
16 0, then, expressing cosw 1x in terms of t.,xponential functions, the inner inte -
18 gral may represented in the form:
20
22
24
26
28
30
32
34
36
3
40
42_
co
2 == 1 e-22ix?iv1-41 [e(--i-
pNox+
j dx.
2
Dividing the entire domain of integration into two, we get
S-1-y
-- 2 eY-x) [e(-"-lw)x e(-P-lah)x dx
2
0
CO
4_ I $
e-22[X^(TfA fe(?P?lotaX e(-p-kmx dx.
2
?
Taking account of the expression for 12, after performing the integration of
q.(19) and the corresponding transformations of the relation so obtained, we find
44
46
48_
__where
50_
52?
. 54
A
(tov,o (e?th [A COs B sin ov]
- e2
(21?P)(2a + .
[(2a? [3)2-1-.4] [(22 -011 + .31
2. cp + -o[2o +i)' + A*2 -F (2,i+ A )1 E42t+,4- PE- f8 4
up-Fir +(2Jii A.)1f0 Tr+A oJEC Pr+.2 4 (4a2?P) +441
4aul1i(3 4-7)'? 2Arrol-1.21[40+ [31 +
.)21 [01-01+1 'MK 2- ?
?????
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
( 20 )
L
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
,
2. values
Knowing
-I-44(01(54-7) [2 -4-7)2-1-1Ste24-(2usi-Eilta)21
2.4010+02) +1
A? - 01,
of I1 for 't < 0.
it is easy to find K(T)
For the case
T >
K ec): 4 - 5 (i.e., if the separation between the resonant fre-
42__
44?quencies of the filters 11 and .T.2 exceeds the pass band of one of them by a factor
-Jof not less than four or five).
46--
48 As a rule, in a considerable part of the radio devices met in practice, there
50 (is always such a separation of the resonant frequencies of the separation filters.
5
54
In conclusion I may state that the parameters of the resonant system of a radio
eceiver likewise affect the function of mutual correlation of the noises at the -
56-1Output of the separating filters.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
???
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
0
2
41-J
6 11. Bunimovich,V.I. - Fluctuation Processes in Radio Receiving Equipment. Publishing
-4-1
8-1 House."Sovetskoye Radio", 1951
-12 Karnovskiy,K.I. - "Contribution to the Question of Energetic Summation."
10 ?
?
BIBLIOGRAPHY
12--
14:]3?
16--
18__
20__
24-1
26_
28_
30-
32__
34__
36_
3
40
42
44
46
48
. 50-1
52--
? 5
Collection of Papers. Institute of Motion Picture Engineers, No.II,1954
Karnovskiy,M.I. - Contribution to the Question of the Interference of Complex
Signals. Trudy komm.po akust, No. (1955)
Paper received by Editors 19 January 1956
?
52
1
0
-
2-
STUDY OF SHOCK EXCITATION AND FORCED QUENCHING OF OSCILLATIONS OF QUARTZ
11 Y
T.N.Yastrebtseva and I.G.Akopyan
A method of shock excitation of the, oscillations of quartz by the aid
of a balanced circuit is considere1 . Several different methods of
forced quenching of the oscillations so obtained were studied. A cir-
cuit is given which allows rapid excitation and quenching of the.
i
oscillations of quartz, and is suitable for practical use.
1. Introduction
Modern pulse engineering makes widespread use of devices creating an electrical
26d:time scale.
The accuracy of time measurement here is determined by the stability of
98__the marker generator.
30_
It is well known that quartz resonators have highly stable oscillatory proper-
39 ties. When a quartz resonator is used to 'measure the duration of single, nonperiodi
-_]
341and unsynchronized processes it is necessary to have the oscillations of the quartz
,6crystal occur with constant initial phase and amplitude at an arbitrary instant of
0
38 time coinciding with the beginning of the process under investigation. It is obvi-
40:-jou3ly possible to satisfy this requirement by using shock excitation of quartz, fol-
42 lowed by its damping to complete rest, by the arrival of the following exciting
44_2pulse.
46 Papers (Bib1.1,2) have been devoted to the production of marker pulses by the
48 aid of the quartz resonator, but the circuits proposed in them have a number of
50__shortcomings.
52-- It is therefore of interest to take a further study of various methods of ex-
--
? 54_citing_and_quenching_the oscillations _of a quartz.resonator,_whichmay__be_:utilized_tO
56_build_a_reliably operating pulse generator of high-precisionmarkw_isignals.
53
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
The present paper presents a theoretf-Cal and experimental investigation o
excitation of quartz oscillations, and considers several methods for their forced
The coefficients A, B, G,0 , T1 pl are expressed in terms of the constants of
the system, and, as shown by research, A m 1, B m -a, pl< 0, G> 0, 0 < 0.
.The first two terms in eq.(1) give an increasing exponent, the third gives
10-_I i
_Jdamped harmonic oscillations.
b) Excitation of the quartz by a rectangular pulse of amplitude Eo and dura-
tion ?
In this case we have, for t < T , u (t) Eoh (t).
For t
2. Shock Excitation of a Quartz Resonator
262 Let us investigate the relation between the oscillating part, u (t) .and the
28?parameter .r ?
30__
32__
24--
26__
To study the shock excitation of quartz, the balanced circuit shown in Fig.1
was assembled. The exciting signal,
through the capacitance Co, equal to
applied to the input of a paraphase amplifier,
the capacitance of the crystal-holder, is fed
to the quartz crystal Ky. At the instant of shock, the quartz may be considered in
first approximation to be of capacitance Ce,. At the output (points alb) the excitin
signal, when the circuit is in exact balance, may be considered to be practically
absent; at the same time a signal taken from the crystal appears between these
points. Thus only the oscillations of the crystal will act on the input of the am-
plifier, connected to the points alb.
n calculating the amplitude and phase of
the oscillations being excited, the ordinary
equivalent circuit of the quartz is used. The
equivalent circuit of Fig.11 used in the calm-
36
lation, is presented in Figs.2. Here r =
2RaRi IR
40 Ra t Ri K Ra +Ri ; LI CI R., Co = equiva-
lent parameters of the crystal under study.
44
The computation was performed by the operational method. Two cases were in-
vestigated:
a) Excitation of the quartz by a voltage drop.
50
Denoting the voltage on the plates of the quartz by u (t), we find the transfer
59
function h (t) for an input voltage of the; form E =a0(t)? where a0(t) is a unit
function
Thus the amplitude of the useful signal depends on and may vary from 0 at T
21t
kT to 2E0G for T (k )T, where T =
2, 3 4000
To compare the theoretical results with the experi-
RI mental data, the formulas obtained were calculated for a
quartz generator, type K-11 (cut X + 50), of frequency
Fig.2
100 KC. The calculations showed that the coefficient G
7.65 X 10r4, i.e., that the amplitude of the useful signal is about
litude of the exciting pulse. The results of the calculation agreed with
Fig .1
a) Input
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
0
_../the experimental data within the limits oi accuracy of measurement.
2-- 1
The relation between the amplitude of the_ oseillations_and_the_d_unation_of_the__
exciting pulse that had been found from. -thecalculation was also checked experi-
_Jmentally. Figure 3 gives the results of the measurements.
? Oscillograms of the oscillations of Ile quartz crystal were taken at various
8--
durations of the exciting pulse (Fig.4). lAs will be seen from them, at T ????
?-?????
12-- 2
= 5 microsec (Fig.4b), the oscillations ex-
cited are close to sinusoidal and have a.
considerable amplitude. At T =
= 10 microsec, (Fig.40, when according to
the calculations no oscillations should be
excited, they still do exist, but have a
very small amplitude and a complex form,
differing in frequency from the fundamen-
16--
18--
20--
22 2-11,
4
26?
_
?
S.
S.
S.
S.
S.
S.S.
S.
r
al 42 44 45 qd 17t8 4R /RF
Fig.3
28-- a) Experimental; b) Theoretical
30__
_oft (for instance, Fig.4a shows an oscillogram at T " 0.3 microsec) the form of the
32__
osciUations excited is also complex, indicating the presence of oscillations of
--1
tal oscillations. At intermediate values
34__
__various frequencies.
36:]
38?
These data tend to support the view that it is desirable to use particularly
__monofrequent quartz crystals for producing time marks.
40_
42_
44_
46_
48_,
a)
b)
Fig.4
50_ Studies were made on quartz plates vibrating in flexure. The plates, figured
52-lor vibration in thickness, gave poor results.
54..)3._Quenching.:the_Osci3lation3 of a Quartz Resonator
5611 Thelorced_quenching of a quartz resonator was studied- by various methods__
' I
56
0a) Quenching the Oscillations of the Quartz Crystal by Shunting of an
2_4 Active Resistance
41
The simplest method of quenching the loscillations in ordinary resonant circuits
A
lis by shunting the circuit with an active resistance. .A check of this method for
10.11
12
!present paper the influence of the value of the shunting resistance on the dapping
16
quartz is of interest. In an earlier reference (Bib1.3) the impossibility of effec-
tive quenching of the oscillations by this method has been demonstrated. In the
18_
of the quartz crystal is calculated and eyperimentally checked. Figure 5 gives an
equivalent circuit for a quartz crystal shunted by an active resistance. The equa-
tion of this system is of the form:
dnr, + + d14 " LC0
d211( ,
az
I
R 2
26 Vhere: 1.= =--_
s 'cc,? LC ?
2Ed
. The character of the damping of the quartz crystal may be established by study-
30i 1
ing the roots of the characteristic equation. It will be easily seen that at r = 0
32
. 2
the characteristic equation takes the form p2 + a P 4. cd o
\ di, -
+ 21) ?'
(4)
i c
34
I ; f...i
li A
U II ?7 --? C-L and the damping decrement of the system equals the natural
36 .4.- !
I i damping decrement of the free quartz crystal. The influ-
!
ence of the intermediate values of p was investigated by
!
numerical solution of the characteristic equation.
i
The calculations showed the damping to be maximum at a definite value of r
1
40 Fig.5
42
44R
about 1
_,
1 but to increase by only one order of magnitude over the value for th
co c
A64 00 1
_tree damping of the crystal. Figure 6 gives a comparison of the calculation with
48__1 I
--the experimental values (e = damping factor of the system).
50:1 !
i
52--b) Damping of Oscillations in a Circuit with Negative Feedback
? 54-1
Figure 7 shows the simplest circuit for imposing a negative feedback on the
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04 ? CIA-RDP81-01043R002900040003 9
?
STAT
STAT
0
2
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
4_J
quartz crystal. Its equivalent circuit i4 analogous to the circuit given in Fig.2.
-0
,4170
.??
???
b)
300
200
?e
eee
WC
.'
J
. t , .
t ?
The_equation_of_the_systemds
. of the form:
d311 R . 14K
dt ?
L r Co
+d2
d[
1
/2 4_ _L
LC a L Co
,
no .11.
d i L +
14_ 0 5 AO 15 20 25 30 35 40 45 (1)'
_ ___
_I_ - 1 -
BD]
4 Lr Co dl
16-- . Fig.6
1 . 1 + K : A
18__ a) Experimental; b) Theoretical; c) Natura1. .
-t- ? ir ---- v .
LO Co -
20_3 damping; d) r, kiloohms
where K - overall amplification Lac-
.(5)
22.._
_tor of the system.
24--
If we put r* = 1 K
then we-arrive at the same eq.(4), which describes the.
+
34'
36
behavior of a quartz crystal shunted by the resistance r*.
Thus the variation of K is equivalent to the
variation of r. The case r- 0 corresponds to the
case K -00. For K = 0, r* = r. Consequently such a
method cannot provide a greater damping than that
given by shunting with an active resistance of op-
3 .
i
Fig.? timum value.
4(L'
,
a) Quartz crystal; ,
42_ c) Quenching the Oscillations of a Quartz Crystal in
b) Amplifier a Balanced Circuit
44
The circuit shown in Fig.1 was now studied. To
46..J
?effect
--effect the quenching, the oscillations of 'the crystal from the points (a, b) must be
---,red through an amplifier and phase converter back to the input of the circuit. Fig-
50-d I
--ure 8 gives an equivalent circuit for this connection.
52--
The equation of the system
the form:,
is of K
--
? 54._
u
2
=0,
L 2Lce
56-1
? _
0 i
-1R y
1
2-7where a = Y wo =
yCo
4
For:I(=4 -c? ?2 ? 4 a the coefficient of the last term of eq. (6) vanishes.
6
8 - n9
2-11
1 0 ???????????
12_
22_
Fig.8
,17
Fig.9
a) Theoretical; b) Experimental
24__For such a value of K, the roots of the characteristic equation are of the form:
261 ?
20_
30_
=
a
2 ?
ia y 11 (COO-r 2 , I
- + aT ) ?
k 2 ) L(4
32_ As may be seen from a calculation, the damping is characterized by two summands:
.34:Ja and y. In practice, a = 102, Y = 105 - 107, i.e., the damping has been increased
36__by several orders Of magnitude over the natural damping.
38-- The relation (3 (K) was studied for a specific quartz resonator by numerical so-
--
40__lution of the characteristic equation. The same relation was also studied experi-
_J
42_6n-tally. The calculated damping, and that experimentally-found (Fig.9) are over
44__three orders of magnitude greater than the natural damping of the quartz crystal,
467Jwhich permits the successful practical use of such a system of forced damping. Thus,
42__Ifor instance, an amplification factor of 200 in the feedback circuit assures a de-
50--crease in amplitude by a factor of 2.8 in only eight periods.
52--d) Shock Quenching of a Quartz Resonator 1
? 54:1
It was found in studying shock quenching that by varying the duration of the
_59
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
? ?
? ????'-,4 ?????????????a,??????.???? ??-??*???????
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
0 ,
Jexciting pulse the amplitude of the extit d oscillations could be varied. If a rec-
2-1 tangplar pulg!!.1 however, is caused to act n_a_quar_t_Z_C_Mtal_alrea_dy excitedl_theT4,
41byvarying the duration of the pulse, its amplitude, and its delay with respect to
mo.w rri..,:,4??.* ? -4 ? ? ?? J ,
the exciting pulse, the amplitude of the oscillations can be considerably reduced
(aImost complete quenching).
28--
,
30__
32__
34_
36
38-,
The oscillograms of Fig.10 show shock quenching for the case when the exciting
50?and quenching pulses are Of the same duration(T = 5 microsec) while the delay
52ime of the second pulse with respect to the first is varied (Fig.10a : tdel =
5-microseck-Fig.10b- tdel' 10 microsec; Fig.10t-t- tdel -=-15 microsec).
dvitable-to-use this method of quenching in-cases when the-amplitude -of--
1
60
0
--.
the oscillations must be rapidly reduced. romplete quenching is difficult, owing
the non-monofrequency of quartz crystals that has already been mentioned.
-1
As a result of our investigations, a ibalanced circuit of excitation and quench-
6 jiliCarthe-oscilliticiff.1-Of The complete circuit diI
? 8--
10?
gram is given in Fig.11. The operation o1 the auxiliary equipment, permitting
quenching of the oscillations to be accomplished at any desired instant after 'their
12_ excitation, is not considered, in viev of lits
14
Thus the equipment presented allows the oscillations of a quartz crystal to be
16--Iquenched after a few periods; the instant of turning on the feedback and its dura-
tion being easily changed.
241
26-1
crystal has been performed. A complete solution has been obtained for the oscilla-
itions excited for the cases of action of aisignal of stepped or rectangular shape on
30 Ithe quartz crystal has been found.
The optimum pulse duration at which the excited oscillations are at maximum am-
litude and have the smallest number of harmonics has been established.
Conclusions
A theoretical
and
experimental investigation of the shock excitation of a quart
. 32
34_
36
Four methods of forced quenching of quartz crystal oscillations have been stu-
died: shunting the crystal with an active resistance, quenching ii a circuit with
40:1
negative feedback, quenching by mans of a balanced circuit, and shock quenching.
4fl
I It has been found that the damping ofithe quartz crystal can be increased by
44-1
--only one order of magnitude when the first two of these methods are used.
46
Calculation .and experimental study of thequenching of quartz crystal oscilla-
48=1.1tions by means of the balanced circuit proposed in the present paper have shown that
5Q1
'the damping of the quartz crystal oscillations may be increased 104 times by this
_inethod. Further than that, the quality factor of the crystal maybe artificially re
54..]
duced to any required value by permanent feedback connection.
61
Declassified in in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04 ? CIA-RDP81-01043R002900040003 q
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
-
0
A balanced circuit ha i been developed for exciting and quenching quartz-crystal
4
^
Icacillations. The circuit is suitaillp..10
6J
84Physics, Moscow State University.
_JMigulin for suggesting the theme of
10--
12-
14
16
practigal_uae.
This Work has been done at the Department of Oscillation Theory, Faculty of
18
tion.
The autrors express their thanks to Professor V.V.
this research and for his help during its execu-
Paper received by Editors 9 April 1955; resubmitted after revision 6 January 1956
-BIBLIOGRAPHY
t.
20:1 ? blynall,D.J. - Jour.Inst.Elec.Engrs. 931 -A (7), (1946),PP.1207-1214
92-12. natrebtseva,T.N. and Galkin10.P. 7 This Journal. (7), (1955)Ipp.69-73
24-- Proc.Inst.Radio Engrs. 29(4), (1941),PP.195-199
26
28
30
32
34_
40
42
44
46
48__
50_
52--
54
56d
0
2
4 ? ?..44.. ?
6-1
12 ?
24?
YEAMACK IN TRANSISTOR CIRCUITS
by
Ya .K.Trolthimenko
Full Member of
the Society
This paper gives a technique of analysis of the principal types of
feedback in transistor circuits.
1. Introduction
The role of feedback in transistor circuits is even more important than in cir-
cuits using vacuum tubes. There are two reasons for this.
1. The transistor has a deeper internal feedback than the vacuum tube.
2. In modern transistors the instability of the equivalent parameters is con-
6d
23siderably greater than in vacuum tubes.
The technique of designing radio is controlled primarily by the properties of
::the active elements. The transistor, as an element of the circuit, differs from the
34--vacuum tube in the value and character of its feedback and in its low input resist -
36-Hance.
tures
4 Sof the internal
42__paifiers may be
It is the last-named fact 'that is primarily responsible for the specific fea-
1
of the analysis of circuits using transistors. Thus, depending on the ratio
resistance R of the signal source and the input resistance
arbitrarily divided into three classes:
44_1
45-
1. Voltage amplifiers, for which the following inequality holds:
>>
48
50-
2. Current amplifiers, for which the reverse inequality holds, i.e.:
52,-
mcrit, the circuit will operate under oscillatory conditions:
Sc = SCO
1
t
- ? a
1 i a.e rn Ram ? Jr + 69 %?
e sin (2- b-1), ' (17)
b
_ _ .
b 1
1= a rctg ? a (;2 -11-- ma, ); b
(42-1- ' 2m
In the oscillatory regime, the build-up time
r--
fn = e? 8,
a
44
ma2
(18)
(19)
is a coefficient indicating how many times tn is smaller than it is in the
case m 1. The build-up time is shorter, the greater the allowable value of the
0:7
5 _
52?
0,7
0,6
0,5
' 1.0 1,2
0 48
Fig.3
in
;first overshoot:
! - 4
a
4 (20)
Tr(aat -- 1)21- b2.e
Here
and
78
tnt?
bin
a.? arctg 1 ma
for ma bin
b 17 ?1 . (22)
_ .
Figures 3 and 4 give the values of n9' and A related to m and ae for aa .. 0.05
(Tifi-e-n ad varies from 0 to 0.7, aand A vary only slightly).
When reCe >Tal i.e. in circuits using high-frequency transistors (P 1-I, junc-
tion transistors), the effect of the capacitance of the emitter junction, Ce, cannot
be neglected. This capacitance, too,
leads to an increase in the flow of
charge carriers into the base region,
when the voltage drops are steep, and,
consequently, also leads to shortening
of the build-up time of the pulse fronts
It is precisely for this reason that the
transfer characteristic of a stage with
grounded emitter, using a junction tran
sistor, is of oscillatory character at
AZ
10
8
6
4
2
ar2
aet7
-.42
?.Add
1,4 1 -0 2,0. 2,2 g4 2,6 at
Fig.4
32_
low values of R .
34
?
The analysis of the circuit of Fig.1:with high-frequency triodes is somewhat
36
cumbersome in the general form. But in practice, in such circuits, the value of
3
R0 C, is found to be close to that of reCe' In this case, i.e., when
1
42_ - RoCo? rt Ce > Tit
44-
- the mutual conductance Sc may be expressed by eq.(8), and all the quantities neces-
46__
- sary for the calculation (t n, A, etc.) may be determined by the formulas given
48_1 -- above, if we consider
40
. (23)
50-
52?
? 54
A Ro re
? = 1 ? ae=-- 1 l-
ase=
R9 r br.b
- ?
(24)
and
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
79
STAT
STAT
?
?
0
? t- t. .2. It. ',wt.:ft...POW. taw?. . ? ? ?
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
21
. (25)
4-
6
.3. Circuit with Combination Current Distr bution Circuit and Feedback
???????I
10?,
12_
14_
16
18_
20--
22__
24
26
30
The source of the input pulses for the intermediate stages of the amplifier is
the stages preceding them, which may in practice be represented as current genera--
tors. To diminishlthe distortion of the pulse fronts in such
circuits, the use of what are termed current-distributing
circuits is recommended (Bib1.3). The simplest of such cir
Fig.5
cults consists of a resistance and an inductance connected in
series with it. A
disadvantage of such circuits is that an
appreciable shortening of the build-up time can be obtained
only at high values of the inductance.
In circuits with a combination of a current-distributing circuit and feedback,
induction coils are in practice no longer needed. In this case the current-
distributing circuit consists of a resistance connected in parallel with the input
32 of the stage (Fig.5).
36
3
40
The current amplification factor for such a circuit is as follows:
re, a A!
di;n rb Z1-f- Za (P Ts 711- 1) -- a 0 -11- /10 -II- ZezJ
(26)
42__ The structures of the expressions for K and Sc are entirely the same, and
i
44 therefore the formulas for tn, A, etc. are also the same, if Rn is substituted for
46 RG in them. For an assigned amplificati* the build-up time of the circuit of
48 Fig.5 is the shorter, the greater the values of RH and Ro. In actual circuits, the
50 resistor RH is connected to the collector circuit of the preceding triode. An in-
59 crease in this resistance, or in Ro, makes it necessary to increase the voltage of
54-ithe_supp1y_source.__In practice. it is sufficient.to_confine ourselves to the values_
567.1RH ...5_2=3_kiloohms,..since_on_further increase of this resistance ..the shortening_of
80
0
tn is very slight.
2--
4
6
The circuit with complex feedback is-merelk a singe special case, shio-iiing how
4. Conclusion
V.?????:??????%1
8_
effectively the distortions of the pulse fronts may be diminished by redistributing
10--
the carriers in the base region. The redistribution of the carriers may effected
12_
by the most varied methods: use of feedback, employment of current distributing cir
14:1
cults, etc. The feedback circuits show the greatest promise, since they not only
16--
shorten the build-up time but also lead to qualitative and quantitative improvements
18_
20_
22__
24--
26 _
28_
34_
36_
38
1 of many other properties of the amplifier
This method of diminishing the distortions of the fronts is applicable to pulse
i
equipment designed either for amplificatiOn or for shaping pulses with steep drops.
I
Paper received by Editors 27 April 1956
BIBLIOGRAPHY
1. Adirovich,E.I. and Kolotilova,V.G. - Zhur.eksp.i teor.fiz. 29,No.6(12), (1955),
p.770
2. Chaw,W.F. and Suran,J.J. - Proc.Inst.Radio Engrs. 41(9),(1953),P.1125
3. Shea,R.F. - Principles of Transistor Circuits. New York 1953
40__J 4. Schaffner,J.S. and Suran,J.J. - J.APpl.Phys. 24(11), (1953),P.1355
48_
50-
52-
54_,
?
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
STAT
STAT
1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
-
2.1
THE PRINCIPAL PROPERTIES AND CHARACTERISTICS OF THE SYNCHRONOUS FILTER
4__
. 6
8?
.
N .1C.Igpat I yev
10-- The author discusses the action of an electric filter designed to sepa-
rate or absorb oscillations of.arbltrary form repeated at an assigned
14_
16--
18__
22__
24 --
26-31
28:
.1
34_
36_
38?
_
40_
42_
frequency. The basic parameters of, the filter are established.
Relations for the frequency, phase and transfer characteristics
of the filter are derived. The concept of the filtration factor with
respect to the noise voltage is introduced, and the relations of this
factor to the parameters of the filter are established.
The possibility of using a long open line as an accumulator for
the filter is shown.
1. Introduction
In modern radio engineering it is rather often necessary to separate a complex
oscillation into two components, one of which is an arbitrary function of time, per-
iodically repeating at the assigned frequency fo, while the other is either a non-
periodic or a periodic function of time, repeating at a frequency other than fo. The
first component may be termed synchronous, and the second, asynchronous, with re-
spect to the assigned repetition frequency ft:). In some cases it is necessary to
44 isolate the synchronous component and eliminate the asynchronous (for instance, in
46_controll1ng interference in radar work); in other cases, on the other hand, the
1
48?asynchronous must be isolated and the synchronous eliminated (for instance, in the
-j
1 50 Iselection of moving radar targets).
1'
A number of devices to solve such a problem have already been constructed,
1
54._testedy-and-described in-the literature (Bib1.1, 2,-3, -4, 5). -Accordinuto_the_purn
- ealizi ? it, and the methods-of-describing the principal-prop--
56-71.
62
:e
W.: 7, 2;73
0
? ???? ...It.% et, 1????1
__Jerties, these devices have been given the 'post varied names. In some cases they are
2-4
leaned canceling devices, in others, synchronous accumulators, in others, comb fil-
? 4_1 1
,ters or synchronous filters, in still others, integrating devices, and so on.
6:1 _
. 8:] Wi-ih-5.11-COriiideFEhe action of all-these devices (f-e-g-gfaiesof s the form in
I
which they are realized, or of their purpose) as special cases of the realization of
___Ia particular type of filter, hereafter termed synchronous filter. This
10:1
I
12 --Ivoted to the elucidation of the principal properties and characteristics of such
14:1
generalized filter.
paper is de-
2611
3 2_1
.34
36_1
3 C-
40__
cumulators may be
42-1
'Thus, for example, the accumulators that have been practically realized and de-
44:1
'scribed in the literature, include capacitative storage devices with cathode-ray
The synchronous filter is
a natural generalization of the resonant filter. In
contrast to the usual resonant filter,
which is able to separate (or absorb) only
Ud a) sinusoidal oscillations of assigned fre-
quency, the synchronous filter can sepa-
rate (or absorb) oscillations of arbitrary
form, repeated at- the assigned frequency
1'0, to which it is "tuned".
The accumulator which must "remember"
the form of the oscillations fed to it, is
a principal and inseparable part of the
synchronous filter. The action of the ac-
based on the utilization of the most varied physical phenomena.
./.1'. : ... ....
/14( \\
II
li %
t
It
'I'
CCCCCCC,."".0
Fig.1
a) Differentiating
grating output
output;
b)
b) Inte-
46_1
'switching (Bib1.1,2), with vacuum-tube switching (Bib1.4),
48-1 __switching (Bib1.5), electrolytic storage devices (Bib1.3),
501
__istoring information in the form of waves propagated along a line, and others. Ac-
52-1
-icording to the form of the accumulator, the circuit and design of the synchronous
? 54 1
and with mechanical
wave storage devices,
be radicall modified, but its basic properties as a filter must never-
83
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04 ? CIA-RDP81-01043R002900040003 q
STAT
STAT
Declassified in Part - Sanitized Cop Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
7
?
0
theless be .maintained unchanged. ,
As the starting stem ,_serving to elucidate the principle of operation of the
4__
. synchronous filter, and to obtain the fundamental relations, an idealized system of
6
10-
12-
14
16-
18
20_
22_
a c o f --rer with -a CapaCitifOrgr011irge de ce F3. .1 1.).'"r-al be used ilMildr
paper. .It is considered as an equivalent
form.
circuit for a synchronous filter of any
2. The Synchronous Filter with Capacitative Accumulator
The commutator K, rotating at the angular velocity wo = 2xfo, periodically
passes a number of contacts, uniformly distributed along an arc of the circumference,
Idealizing the operation of the commutator, we shall consider that, in passing from
contact to contact, it does not connect them with each other, and at the same time-
?
24 --
_Jdoes not remain separated from one of them. A capacitor of capacitance C is con-
?"61
_Inected to each contact; the leads from the opposite plates of the capacitors are con-
nected together. The system consisting of the commutator and the set of capacitors
30_
32_
34::
. 36_
3
40
42_
with a sufficient degree of accuracy, in the form of a sequence of rectangular
44?
pulses, each of duration At.
46_
If
If the value of the time constant RC is taken sufficiently great by comparison
is called an accumulator. The resistance R is connected in series with the commuta-
tor.
The input voltage u is applied to the storage tube through the resistor R. It
is assumed that the time At, expended on a single switching of each capacitor of
the storage -tube, is sufficiently short by comparison with the time during which u
can vary appreciably in value; in other words, the input voltage may be represented,1
48._
with the time At, then a large number of revolutions of the commutator will be re-
'sol
__quired for each of the capacitances of the accumulator to become charged to a vol-
52--
__tage close to the input voltage. Under such conditions, the component of the input
, 54
_Jvoltage that is asynchronous with respect to the frequency dt-rOiiiion-fO of the ?1
56 1 -
84
1
4
4
0
_Jcommutator can produce but slight fluctuat-ons of the voltage on the capacitors .of
2-A
!the accumulator. At the same time the component of the int_p_jt_L_roltage that is syn-
4__
with respect to the frequency fo Lill fully charge those capacitors after
4._.1
61 1
the appropriate number of revolutions of t: e commutator (i.e., each of the ciipacitors
:1
8
will be charged up to that instantaneous value of the synchronous component of the
10:1 1
input voltage which is taken on by that voltage at the instant of commutation of, the
12
__given capacitor). As a result of this, the voltage u.
14_1 1
__cummlator, andtwill approach the synchrondus component
16 ?
18_
20_
will be separated on the ac-
of the input voltage, which,
1
as it were, it "remembers", while on the resistor the voltage ud will be separated,
1
and will approach the asynchronous component of the input voltage. Accordingly (by
1
analogy to an elementary filter, consisting of resistances and capacitances con-
22-4 1
_.-nected in series), this 6ystem of a synchronous filter has two outputs: an inte-
,
__grating output and a differentiating output.
24_1
26
The smaller the value of the coefficient
28_
30_ (1)
, 32_
whichweshalltermthedampingcoefficient,theclosertLidllbe to the synchrom-
1
34_
__ous component, and ud to the asynchronous component, of the input voltage. For
36_
__practical purposes the cases when b ?c. 1 are of the greatest interest.
at
b..
RC'
38?
The commutation frequency fo and the damping coefficient b are the basic par-
40:] t
__ameters of the synchronous filter, completely characterizing its action on the form
42_
_of the input voltage. 1
44_..] 1
,
-- In view of the fact that during the process of commutation each commutated ca-
46__ 1
--pacitance is able to vary somewhat, the form of the voltage taken from the accumula-
48d1
tor will prove to be "sawtooth". In analyzing the processes taking place in this
? 50__i .
?synchronous filter system, we shall not take this "sawtooth" character into account,
52-1 1
and shall take, as the true value of the voltage at each commutated contact, the
. 54
becomes the end of its commutation.
$5
STAT
STAT'
Declassified in Part - Sanitized Cop Approved for Release ? 50-Yr 2014/02/04 ? CIA RDP81 01043R002900040001
L
-,
_13. Derivation of the Starting Formula
t: 2-1
5
3
4_ We shall commence our analysis of th operation of the filter by deriving the
i 6__3_tarting_tor_mulazelat_ingthev_oltage_at_ii.t_soutpatt_o_the__v_oltage at its input.
,
8__For this purpose we introduce the following additional notation: t = current time;
. ,
i
t 10 _-n = number of complete revolutions of the commutator from time t = 0; x = coordinate
e
c
12 of relative displacement of commutator (in fractions of a complete revolution),
1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
, liameasured from the position occupied by the commutator at time t = 0; To = period of
i
,
t
, 16__commutation:
. --. ---I
Tof?
(2)
18__
I 1
,, ? , ? 1
20__ -
92__ According to the meaning .of the notation__ just introduced, the current time may
z 24--be expressed in the form
4__ ' This permits analysis of the process rf variation of the voltage at any eapaci-
tor_(f or_any_x)__as_a_function-of_t he. number-of--revolutions_of...the_commutator_m_Know:_-
__
8?ing the function gi(n), the voltage on the capacitor for any n may be found as the
10R sum of voltages, each of which is the result of the action of one of the successive
i2?pulses of the input voltage, i.e.,
14_ _
uiRn + x) T 0] :----. u[x T01(1 ? e?b ) e-nb-Fti[(1 +x) To] X
514
22__
24_
26__
28_-
where k = a whole number. 1
The expression so obtained may be rewritten in the form
28_-
x-0
This formula, relating the form of the voltage at the integrating output to the
form of the voltage at the input,
will serve us as our starting point
36
for the solution of all problems,
3
both of steady and transient char-
.
acter.
42__
To find the form of the vol-
a
tage at the differentiating output
the obvious equation maybe used:
Let a single voltage pulse, of duration At, be fed to the filter input from some
39__source of DIF, and let it fall on one of the capacitors of the accumulator. This
34__capacitor will be charged to the voltage 1 - e-b, and then, at each successive coma,-
-- I
36 mutation, it will be discharged by a factor of e-b. As a result, a sequence of
38 pulses will appear at the integrating output of the filter,- and the value of these
I] 40 pulses will vary by the law
44_-
46_.
The function gi(n), represented in Filg.2, may be considered as the reaction of
48?the integrating filter to the unit pulse function (i.e., as its transfer character-
__
50 --istic).
52-- Using eq. (3), the input voltage, which is assigned in the form of the function
54--of -time -uctly-may-be represented in the form of .a .function of- the number_o1.revolu-
--1
g i(n). (1 ? e?b ) e?nb. (4)
ud (0 = u (4 ?us (0, (6)
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/02/04: CIA-RDP81-01043R002900040003-9
?????
??. ??
?
0
2
:1 forms of input voltage which, on the one hand, are of the greatest practical inter-
est, and, on the other, most completely disclose the characteristics and properties
A
.of the filter. Such forms of input voltaic: include: a periodic voltage of arbitrary
4
-6 f
form, a harmonic voltage, and "noise" voltage.
? 8.-
10_h. Action of Periodic Voltage
12-
Beginning at time t = 0, let the periodic voltage u(t) act on the filter input.
14_-
__Let that voltage have a repetition periodiT, differing little from the commutation
16--
period To, i.e. let the following inequality hold:
18__
1
20_ AnC7; (7)
22_ where
. I
2.4- AT. T. - T? (8)
? 26--
If the form of such voltage is obser4ed on an oscillograph with sweep beginning
28--
3019n the left part of the screen and having la repetition period that is .a multiple of
_ To, then for A T> 0, the oscillogram will ibe seen as moving from right to left,
. 32
while forAT < 0, on the contrary, it will be seen as moving from left to right.
34
Using eq.(8), we may write
? 36
3 u(t? KT)) = K AT ---- KT)
40 or, taking account of the fact that T = period of repetition of the function u(t),
49_while k is an integer, we may. write:
-
I ,
44_
1
(9)
46_
48_ Thus, according to eq.(5), the expression for the voltage at the integrating
50-output of the synchronous filter in this ease may represented in the form
5 1
al(t)=-(1-e-a)E-uo-g A e'rb. (10)
* 54
x-o
56 J If the inequality eq.(7)_holds, then variations of the quantity k assuming
0
2--
only integer values, will lead to relativ6ly small variations of the function u(t -
I
- kAT), and this latter function ma_y be considered, with a sufficient degree of ac
4_
__curacy, to be a continuous function of the variable k. Similarly, assuming the fol-
lowing inequalat-Tto had:
8?
.
10--
--
1
12-
.....4the function e-kb may likewise be considered to be a continuous function of the var-
14-4
=liable k. Here, as the limiting value for k, we may use, instead of n (which assumes
b