B 417326

ARTES
LIBRARY
1837
VERITAS
SCIENTIA
OF THE
UNIVERSITY OF MICHIGAN
PLURIBUS UNUM
SI QUAERIS PENINSULAM AMOENAM
CIRCUMSPICE
DEPARTMENT
OF
ENGINEERING

ENGINEERING
LIBRARY
TK
5535
C91
DRC
201

SYNCHRONOUS AND OTHER MULTIPLE
1
TELEGRAPHS
SOME METHODS OF OBTAINING INDEPENDENT TELE-
GRAPH CIRCUITS ON A SINGLE WIRE BOTH
WITH AND WITHOUT SYNCHRONISM
BY
ALBERT CUSHING CREHORE, PH.D.
Member of the American Institute of Electrical Engineers
Author of "Alternating Currents" (with F. Bedell)
WITH ILLUSTRATIONS AND WORKING DIAGRAMS
NEW YORK
MCGRAW PUBLISHING COMPANY
1905
MICH UNIV
012
LIBRARY

TIRRVKA
ор
ПIСН ЛИГА
Copyright, 1905
BY
ALBERT C. CREHORE
ROBERT DRUMMOND, PRINTER, NEW YORK

PREFACE.
THE subjects included in this volume admit of a natural
division into three Parts. The first Part is devoted to the
general subject of methods of obtaining independent telegraph
circuits by the use of direct and alternating currents on the
same wire. This is the first technical description of these
methods which has been published, and I have some satis-
faction in making them public at the present time because
the principles have been tested for a number of years under
rather difficult service conditions.
The arrangements of circuits throughout the book are
with almost no exceptions not merely theoretical results, but
have been successfully operated' either in the laboratory or
under service conditions. The fact that these results, worked
out in detail in the laboratory on artificial wires made to
imitate real wires, have afterwards been applied to real wires:
with almost identical results serves to prove that good results.
in studying new systems of telegraphy may be obtained in the
laboratory. The convenience of this mode of investigation is:
much greater than is attainable under the most favorable con-
ditions with actual wires.
The nature of these subjects is such that investigation
must almost of necessity be experimental rather than largely
theoretical; for the number of conditions which must be ful-
filled is very great, and to overlook one of them might be fatal
to purely theoretical results. There is also a personal ele-
ment which it is not possible to put into an equation.
The name of this work is perhaps only justified by reason of
its brevity. It is evident that only certain phases of these
extended subjects can be included in a volume of this size.
In particular, the part relating to the operation of synchronous
145222
iii

iv
PREFACE.
motors has so many omissions as to demand some apology.
It is intended mainly to present the more important aspects
of the problem.
The subject concerned in the first Part is of considerable
importance, as the methods employed furnish the means of
obtaining double the telegraphic service from a single Morse
wire, while retaining nearly all the advantages of single Morse
circuits. The fact that single Morse circuits are now so exten-
sively used shows more clearly than anything else that no
substitute has yet been found which fulfils all the desired con-
ditions as satisfactorily; and the methods here described per-
haps add a smaller number of limiting conditions to those of
single Morse wires than any other methods.
These methods are not related to the so-called sine-wave
systems where an alternating current is automatically opera-
ted upon at or near the zero points for purposes of telegraphy,
accounts of which have been published elsewhere.
The second and third Parts relate to methods of obtaining
telegraph circuits by means of the synchronous rotation of
two bodies at distant points, the second Part concerning the
means of obtaining the synchronous rotation, and the third
Part the means by which the rotation may be utilized for
obtaining independent telegraph circuits. The subject has
not been treated from an historical standpoint, as it is an old
and consequently a large one, many investigators having
worked on the problems made possible of solution by syn-
chronism. The object is rather to assist the reader to obtain
a clear conception of the subject from a practical and experi-
mental point of view. Some of the fundamental principles
upon which such systems depend are not mentioned and natu-
rally are not explained in the available works on this subject.
To avoid conveying an incorrect impression concerning
this part of the work it should be stated that the experience
which I have had with synchronous telegraphs has been obtained
in the laboratory; but all of the results have been obtained
by the use of actual instruments on artificial wires, and I antici-
pate with entire confidence that these results will be verified
in all essential particulars when real wires are used, as was
found to be the case with the systems described in Part I.
99
(C
In the second Part the use of special wires for the sole
purpose of synchronizing is advocated, and one method of

PREFACE.
V
operating such motors is given. Many different plans for
obtaining synchronism have been proposed, but nearly all of
them are designed to operate telegraphs on the same wire at
the same time.
When a whole wire is to be used for nothing else, the prob-
lem presented resembles in some respects the long-distance
transmission of power. It is not power, however, which is
needed at the distant points as much as it is synchronism,
and the problem is really the transmission of time-signals.
The necessary power may be supplied by local sources at each
place, the motor merely being regulated or governed by the
transmitted time-signals. If it were as easy as not to trans-
mit sufficient power to operate all the receiving motors, this
would be the simplest way to solve the problem; but it is
not easy to do so with the conditions imposed by the nature
of the case, as the distances concerned in telegraphy are very
great for power transmission.
Power transmission to these distances requires electro-
motive forces higher than would be permitted on a telegraph
pole-line as it is used to-day, where linemen are required to
handle the wires. The inductive disturbances caused by a
comparatively large transmitted current would interfere with
the operation of other telegraph wires on the same pole-line.
It seems probable, therefore, that the plan to be used will
employ some form of generator of alternating or reversed
currents, distributing them from a central point in waves
having a constant periodicity, which will be received by relays
that control some form of synchronous motor driven by power
from a local source.
The operation of synchronous motors on single Morse wires
and on duplex wires described in Part II depends upon the
principles explained in Part I, and these methods of operating
motors are new as far as I am aware.
The measurements in synchronous telegraph circuits taken
under working conditions given for convenience in the table
will, I believe, be of interest, and will assist in making this
part of the subject more clear. The nature of the phenomena
in this class of telegraphs makes it difficult to operate long
wires in opposite directions at the same time unless some
special provision is made. There are three methods described
for doing this, one of which, making use of artificial wires and

vi
PREFACE.
a single main-line relay, is well known; but the other two
methods are new, in one of which the line is artificially lengthened
and adjusted to the speed of the trailers, and in the other two
independently adjustable sections of the receiving sunflower
are provided, the latter arrangement being adapted to any
wire. In these two methods the ordinary artificial lines required
for a duplex balance are dispensed with; but only half as many
circuits can be obtained as when the artificial lines are used.
The third Part is concluded with an account of some prac-
tical methods of operating Morse systems by the synchronous
method, showing how circuits operated over the same trunk
lines may be distributed to different points.
The operation of Morse circuits is not, I believe, the most
important application of synchronous telegraphs, and it was
intended to give some account of printing telegraphs both
with and without synchronism in this volume; but it seems
best not to delay the publication of this part of the work until
those subjects are completed.
A. C. CREHORE.
LABORATORY, YONKERS, N. Y.

CONTENTS.
PART I.
METHODS OF OBTAINING INDEPENDENT TELEGRAPH
CIRCUITS BY THE USE OF DIRECT AND ALTERNATING
CURRENTS ON THE SAME WIRE.
..
INTRODUCTION..
CHAPTER I. THE RECEIVING APPARATUS IN THE DUPLEX-DIPLEX
SYSTEM.
Principles of the duplex-diplex. Use of condenser and induc-
tance to separate the direct from the alternating current.
Resistance shunt. Balancing condenser
Balancing condenser by alternating-current
relay or by an inductance-coil. Objections to very high and
very low frequencies. Frequency used. Value of condenser
employed. Division of currents in different branches of
receiving-circuits. Value of balancing-coil for condenser branch.
Construction of compensator. Reception of signals on sounder.
CHAPTER II. THE TRANSMITTING APPARATUS IN THE DUPLEX-
DIPLEX SYSTEM.
Stations having direct-current transmitters only. Effect
of condensers on direct-current transmission. Difference
between closed-circuit and open-circuit working. Effect of
leakage with direct current. Reason for placing battery at both
ends of the wire. Reduction of alternating current along the
wire. Working alternating current on the closed-circuit plan.
Alternating current on the open-circuit plan, with direct cur-
rent on either the open-circuit or the closed-circuit system.
Working alternating-current side on lines having considerable
leakage.
CHAPTER III. APPLICATIONS OF THE DUPLEX-DIPLEX SYSTEM....
Working installation on actual railroad line. Repeater
on alternating-current side. Eight-pole alternating-current
magneto-generator. Application to lines now worked with
PAGE
3
5
16
27
vii

viii
CONTENTS.
polar duplex.
Advantages of duplex-diplex. Limiting dis-
tance of working. Operating diplex only. Use of load-coils
in wire.
CHAPTER IV. THE ALTERNATING- AND DIRECT-CURRENT QUADRU-
PLEX...
CHAPTER V. LOW-FREQUENCY DUPLEX-DIPLEX.
INTRODUCTION..
PART II.
METHODS OF OBTAINING SYNCHRONISM AT DISTANT
POINTS.
CHAPTER VI. THE OPERATION OF SYNCHRONOUS MOTORS ON
WIRES WHICH ARE USED FOR NO OTHER PURPOSE.
INTRODUCTION..
Method of controlling motors. Cities connected by a system
of synchronous motors operated from a central point. Effect
of heavy leakage on a line devoted to synchronizing. Control
of synchronous motors by relay contacts. Special form of
two-pole synchronous motor adapted to synchronous telegraphs.
Damping out oscillations by use of mercury cup.
CHAPTER VII. THE OPERATION OF SYNCHRONOUS MOTORS ON
WIRES WHICH ARE ALSO USED FOR TELEGRAPHS.
....
Obtaining synchronism on single Morse wire. Use of duplex-
diplex principle to obtain synchronizing circuit. High-fre-
quency synchronous motor adapted to synchronous tele-
graphs-general appearance and construction in detail. Ob-
taining synchronism on duplex Morse wire. Synchronism on
synchronous multiplex wire.
PART III.
PAGE
SYNCHRONOUS TELEGRAPHS.
CHAPTER VIII. MEASUREMENTS IN SYNCHRONOUS TELEGRAPH
CIRCUITS.
The instruments required. Method of measuring current.
The measurements, as given in table, and discussion of the
readings.
CHAPTER IX. EXPLANATION OF THE MEASUREMENTS.
Theoretical calculation of current at receiving end of line.
Single current waves as transmitted and received in synchronous
....
37
40
45
47
56
69
71
85

CONTENTS.
ix
system, at different frequencies. Difference between real and
artificial line.
CHAPTER X. DIFFERENT CONDITIONS OF INSTRUMENTS AND LINE
LEAKAGE..
...
Effects of opening circuits leading from segments. Escape
of line charge from both ends of line. Influence of receiver-
coils upon received current. Lines with leakage. Effect
of leakage on synchronous telegraphs.
WORKING IN OPPOSITE DIRECTIONS ON THE SAME
CHAPTER XI.
WIRE...
Effect of attempting to work backward on wire without
special arrangement. Method of working in both directions
in which line is adjusted to speed of trailer. Method of employ-
ing two independently adjustable sunflowers. Use of grounded
segments.
....
INDEX...
..
CHAPTER XII. MORSE CIRCUITS: GENERAL CONSIDERATIONS AND
PARTICULAR SYSTEMS..
Comparison of synchronous with other telegraphs. Use of
one wire for all messages in one direction and another for those
in opposite direction. Use of double currents, as compared
with single currents. Minimum time necessary for trans-
mission of a signal. Speed of fastest Morse operators. Opera-
tion of sounder on local circuit of polar relay. Method of dis-
tribution of circuits. Use of artificial lines.
....
PAGE
92
97
105
117

2

PART I.
METHODS OF OBTAINING INDEPENDENT TELE-
GRAPH CIRCUITS BY THE USE OF DIRECT
AND ALTERNATING CURRENTS ON THE
SAME WIRE.

7
4

INTRODUCTION.
THE system of telegraphy in most general use in America
is the single Morse system, which is essentially that employed
since the earliest days of telegraphy. There have been many
improvements in the telegraphic art which increase the message-
carrying capacity of a wire, such as the well-known duplex
and quadruplex circuits, and automatic Morse and printing
telegraphs; but notwithstanding the single Morse continues
to be used, and is apparently the only system which com-
pletely satisfies some requirements.
Some of the advantages of the Morse system are the ease
with which it may be handled and understood, the simplicity
of the instruments, and the ability to introduce a station at
any point of a wire so as to include few or many offices. Of
the other systems mentioned the simplest is perhaps the duplex
Morse, and this is considerably more complex, involving the
use of artificial wires which must be adjusted to suit weather
conditions; and it does not permit stations to be readily inserted
in the line. In many cases the advantages gained by it do not
compensate for the disadvantages, and at best only two mes-
sages can be transmitted at once in opposite directions, but
not in the same direction.
A system which permits the transmission of two messages
over the wire in either the same or in opposite directions, and
at the same time requires no artificial wires and permits of
stations being inserted in a manner very similar to the single
Morse system, evidently has advantages over the common
duplex. A wire equipped with such a system may be used
to supply two independent circuits, in one of which, on one
side of the wire, two operators may communicate without
3

4
MULTIPLE TELEGRAPHS.
interruption from other operators, two of whom may at the
same time be communicating from different points on the
other side of the same wire.
The different effects which different kinds of instruments:
have upon the flow of alternating and direct currents furnish
a basis for obtaining independent circuits, because these instru-
ments may be so disposed that the alternating current, although
passing over the wire together with the direct current, will at
the stations become differentiated from it, each current going
in one of two parallel paths so as not to disturb the instru-
ments operated by the other. This principle may be applied
to a single Morse wire, using the instruments now employed
for the direct-current circuit, and adding others to obtain an
alternating-current circuit, the resulting arrangement being
known as a Duplex-Diplex system; or it may be applied to a
common duplex circuit, the resulting arrangement giving a
Quadruplex system.

CHAPTER I.
THE RECEIVING APPARATUS IN THE DUPLEX-DIPLEX
SYSTEM.
THE ordinary Morse telegraph is operated by the direct cur-
rent, that is to say, the current in the circuit continues to
remain of the same strength as long as the operator's key is
closed. The additional circuit in the duplex-diplex system
is created on the same wire by means of an alternating
current, by which it may be understood that the current
is reversed many times per second as long as the operator's
key remains closed. Both the alternating and the direct
current flow over the same wire, being superposed, one upon
the other, in the wire itself; but use is made of the differing
physical characteristics of these currents to separate them
at the stations, each flowing in an independent path parallel
to that of the other. With a direct current the principal
effects of inductance and capacity are noticed only when the
current is made, broken, or changed in any other way; but
with an alternating current the inductance and capacity have
a continuous effect upon the current. Advantage is taken of
this fundamental difference to separate the direct current from
the alternating current as completely as possible at each station.
An obvious way to separate such currents is to connect
in parallel, as in Fig. 1, a condenser, C, and an inductance-
coil, L, dividing the line wire into two paths at A, and uniting
them at B. If both a direct and an alternating electromotive
force are continuously applied to the line wire, it is evident
that all of the direct current must pass through the inductance-
coil, and most of the alternating current may pass through
the condenser, the relative proportion of alternating current
through each path depending upon the values of the coil and
condenser.
In general this principle of separation of the currents is
5

6
MULTIPLE TELEGRAPHS.
made use of, but such an arrangement alone would not afford
a means of obtaining a practical system, and this is largely
due to the fact that the direct current cannot remain steady,
but, in order to transmit Morse signals, must be changed with
each stroke of the key. This causes the direct current to
partake more or less of the properties of the alternating cur-
A
~~
Goog
C
음
​L
B
FIG. 1.-General principle employed in separating the direct from the
alternating current.
rent, at each change of current which the operator makes;
and thus causes an effect of brief duration to be manifested
on the alternating-current side by the working of the direct-
current side, making itself evident by kicks or brief disturb-
ances in the alternating-current receiver. It is this inter-
ference between the two sides of the system which must be
overcome before a practical duplex-diplex system is obtained.
In Fig. 1, as a receiver for the direct current, an ordinary
standard Morse relay might be inserted with the inductance-
coil, or the relay may form this inductance-coil itself; and a
receiver for the alternating current, an ordinary polarized
relay,
might be inserted in the condenser branch. When the
direct current is flowing the Morse relay will operate, and the
polar relay will receive no current. If an alternating current
is received in addition to the direct current, almost all of it
will pass through the condenser and the polarized relay, and
cause the armature to vibrate in unison with the current.
The small proportion of alternating current passing through
the Morse relay is not sufficient to affect its armature, either
against the retractile spring when no current is flowing, or
against the magnetic effect of the direct current when it is flow-
ing. It would thus be possible to adjust the instruments to
receive independent messages, one by the alternating current
and the other by the direct current; but at each station it is
not only necessary to receive but also to transmit messages.

DUPLEX-DIPLEX SYSTEM: RECEIVER.
7
Let us assume that the direct-current side is operated
on the closed-circuit Morse system. Ordinarily it is merely
necessary to insert a key in series with the Morse relay in order
to transmit a message, and it might occur to one to insert a
key in series with the relay in the branch L, in Fig. 1. This
would enable the operator to transmit his message, but at the
same time the operation of his key would cause violent kicks
upon the polar relay in branch C, which would entirely pre-
vent the reception of a message on the alternating-current side,
and is not, therefore, permissible. It has been found better
never to disturb the arrangement of the receivers on a wire
in order to insert a transmitter; that is to say, the transmitters
for both the alternating- and the direct-current sides should be
inserted in the wire outside of the receiver set.
Before considering the arrangement of transmitters a
description of the receiver set, with some of the values which
are found suitable for practical use, will be given. It will
be observed that the circuit in Fig. 1, ALBC, forms a closed
circuit having resistance, inductance, and capacity, which might,
with proper values, form a resonant circuit. To have this
circuit resonant would be extremely undesirable for the pur-
pose, for it would magnify the alternating current received by
the Morse relay and would tend to cause mutual disturbances
between the two sides of the system.
The values chosen in practice do not make this circuit res-
onant at the frequency used, and as a still further precau-
tion against interference between the two sides a non-inductive
resistance is connected between the points A and B, as in Fig. 2,
which has the effect of separating the two sides of the system
more completely, by providing an additional path for the dis-
charge of either the condenser, C, or coil, L, each of which
would otherwise have to discharge through the other, or through
the line.
On the other hand, it is desirable to make the branch from
A to B, containing the condenser and the polar relay, resonant
in itself; that is, to so proportion the condenser, C, and the
inductance-coil, which is in this case the polarized relay, PR,
that they balance each other at the particular frequency of
the alternating current used. The relation which must exist
in order to balance the induction and capacity is expressed
by the formula

8
MULTIPLE TELEGRAPHS.
I
Cw
A
where w=27 times the frequency; from which it is evident
that a large variety of values can be found satisfying the con-
dition, and it is experience only which finally determines the
best practical values.
Of the four elements to be considered, that is, resistance,
capacity, inductance, and frequency, perhaps frequency is the
$for
Lw=0,
C
R
B
PR
op
FIG. 2.-Standard form of receiving-circuits.
most important, for the separation of the two currents is effected
solely by their differing physical characteristics, and a constant
current might be considered as an alternating current with a
frequency of zero. An alternating current with a frequency of
one would evidently not differ sufficiently from one with a fre-
quency of zero to form the basis for separating the currents.
In
practice it is found that a frequency of 50 or 60 cycles per second
is more difficult to use than higher frequencies. A very high
frequency would afford still more complete and easy separation.
There are, however, objections which arise with the higher
frequencies which do not exist, to so great an extent, with
the lower frequencies. One of these is that the static capacity
of a telegraph line has the effect of continuously reducing the
strength of an alternating current the greater the distance
from the alternating-current generator, in an analogous manner
to that which would be the case with a direct current, if there
were a uniformly distributed leakage along the wire; and this
reduction in current strength with distance is greater the
higher the frequency. This reduces the length of line which

DUPLEX-DIPLEX SYSTEM: RECEIVER.
9
can be operated very materially, because a certain current
strength at the receiving station is necessary, and to obtain
it on a long wire the transmitting current and electromotive
force become too large. Too high an electromotive force
causes sparking, and too large a current causes interference
on the direct-current side. Another objection to a very high
frequency is that the mass of the tongue of a polar relay is
unable to follow the vibrations if it is of an ordinary construc-
tion, and it seems very desirable to employ the polar relay
to operate a sounder on a local circuit.
The value of the frequency adopted is about 150 cycles
per second, which seems to avoid the objections both to very
low and very high frequencies. With 150 cycles the value
of w in the above formula is 942.5, and with this numerical
1.13X10-6.
This permits of a variation between the values of L and C,
but this again is involved with the resistance of the circuit,
for no coil can be made having zero resistance. We have not
succeeded in making the apparent resistance (not impedance)
of the coil for alternating currents even very nearly approxi-
mate the ohmic resistance, as measured for direct currents.
This is due to energy losses of hysteresis and eddy currents
in the coil and core, which add a considerable proportion of
the apparent resistance to any such coil as it seems practicable
to construct.
value the formula may be expressed LC:
=
I
(942.5)2
If it is assumed that a condenser of, say, one microfarad
is to be employed, the value of inductance required to balance
106
it at 150 cycles is L = (942.5)2=1.13 henrys. It is more diffi-
cult to construct a coil of 1.13 henrys so as to keep its real
and apparent resistance low than it is to construct a coil with
a smaller value of L. This points to the desirability of using
a larger condenser.
On the other hand, if the condenser is given too great a
value there are other objections, since this condenser, as will
be shown later, in some cases bridges a Morse key, and in this
situation it has the effect of interfering to some extent with
the direct-current signals, rendering them less sharp and clear.
The largest condenser value in which this interference with
the Morse signals cannot be detected is that most advantageous

IO
MULTIPLE TELEGRAPHS.
to use, and in practice it is found that three microfarads fulfils
these requirements.
Assuming, therefore, three microfarads as the standard
condenser to be adopted for the system, the balancing value
of inductance, at 150 cycles, is .375 henrys. Much depends
upon the construction of this coil. Winding coils to produce
this value of inductance has been tried with large wire having
low resistance, and with small wire having higher resistance,
with a large amount of iron and with a small amount of iron,
and with both open and closed magnetic circuit; and it does
not reduce the apparent resistance of the coil to alternating
currents to use very large wire, in order to secure small ohmic
resistance.
Taking the three-microfarad condenser and coil together,
about the lowest impedance which we have obtained for the
combination is from 110 to 125 ohms, total. Without the
coil the condenser impedance alone is
I
106
Co 3X942.5
=
=
354 ohms.
The presence of the coil, therefore, makes the impedance of
the condenser branch 354-125=229 ohms less than it would
be without the coil; and when there are many stations on
the wire this reduction per station aggregates a considerable
amount. The actual ohmic resistance of these coils is 50 ohms.
The advantage of balancing inductance and capacity in
the condenser branch will be apparent if the currents and
potentials in the three branches of Fig. 2 are calculated. Assume
that there is an alternating fall of potential of four volts between
the points A and B. Taking a case where the condenser is
balanced with inductance first, the polar diagram Fig. 3 shows
the currents in the various branches. It should be stated
that the impedance of a standard Morse 150-ohm relay with
its working adjustment is about 3,000 ohms, for a current
of 150 cycles. Also that the value of the resistance branch, R,
which it has been found practicable to use with the 150-0hm
relay is 400 ohms, non-inductive resistance.
Letting the line OE in the diagram represent the pressure
of four volts between the points A and B, the current in the
resistance, R, must be 4/400 ampere, or 10 milliamperes, and

DUPLEX-DIPLEX SYSTEM: RECEIVER.
II
AD
is represented by the line OR, in phase with OE. The cur-
rent, IL, in the Morse relay, is 4/3000 ampere, or 1 milli-
amperes, and is represented in the diagram
by the line OL, lagging approximately 90°
behind the line OE. The current, Ic, in
the condenser branch is represented by OC,
and, the capacity being balanced by induc-
tion, is in phase with OE, and has a value
of 4/125 ampere, or 32 milliamperes. The
sum of the three currents, OL, OR, and
OC, is OD, and has a value approximately
the same as the arithmetical sum of OR
and OC, namely, 42 milliamperes. The ratio
of the alternating current in the various
branches to that in the main line is about
3 per cent. in the Morse relay, 24 per cent.
in the non-inductive resistance, and 76
per cent. in the condenser branch.
C
E
R
OL
OE=4
IL-OL=0.0013
IR OR=0.010
Ic=OC=0.032
1=OD=0.042
FIG. 3.-Vector dia-
gram for circuits
of Fig. 2 when the
relay coil, PR,
balances the con-
denser C.
Contrast this with the result obtained
by removing the balancing-coil, which is
represented in Fig. 4, where the letters have
the same meaning as previously. The differ-
ence is that the condenser current, OC, in-
stead of being in phase with, is now 90° in
advance of the pressure, OE, and if OE were
taken as four volts now, the condenser cur-
rent would be much smaller than in the last
example. If OE is taken 11.3 volts now, the
condenser current, which means the current through the alter-
nating-current receiving-relay, is the same as in the preceding
example, namely, 32 milliamperes. The current, OL, in the
Morse relay is about 3.8 milliamperes, and the current, OR,
in the resistance, R, is 28.3 milliamperes, while the line cur-
rent, OD, is about 40 milliamperes. In this example the ratio
of the alternating current in the various branches to that in
the main line is about 9 per cent. in the Morse relay, about
70 per cent. in the non-inductive resistance, and 80 per cent.
in the condenser branch. (It should be noted that the line
current, due to phase differences, is less than the sum of the
currents in the three branches.)
In comparing the case in which the condenser is balanced

12
MULTIPLE TELEGRAPHS.
with that in which it is unbalanced, it is seen that the fall of
potential between the points A and B is about 2.8 times as
much in the latter case as in the former; while the current
in the condenser branch, or the receiver, has the same value,
and the line current is approximately the same.
This means,
where there are a large number of stations on the wire, that
D
及
​Cen
OE=11.3
IL =OL=0.0038
IR =OR=0.0283
lc =0C=0.032
I
AE
- AR
O
74
1=OD=0.040
FIG. 4.-Vector diagram for circuits of Fig. 2 when the relay, PR, in
removed.
a high voltage would be required to drive the same current
in the latter case as would be obtained by a comparatively
low voltage in the former case. This amounts to saying that
the impedance of each station to alternating current is much
larger in the second case.
Another improvement in using the first arrangement is
that the alternating current passing through the Morse relay
is only about one-third as great as in the latter case, and this
makes considerable difference in the ability to use the direct-
current apparatus.
The non-inductive resistance, R, provided for reducing

DUPLEX-DIPLEX SYSTEM: RECEIVER.
13
interference between the two sides of the system, also has
the effect of shunting some of the main-line currents, both
direct and alternating, past the station without passing through
the instruments. In the case of the direct current, with a
150-ohm relay shunted by 400 ohms the current in the relay
is about 73 per cent. of the main-line current. If the current
in the relay is to be 30, the main-line current must be about 41.
When there are many direct-current stations on the wire,
the total reduction in the line resistance from this cause is
considerable. Suppose there are 30 stations in all. Without
the shunts the resistance per station is 150, and the total 4,500
ohms, due to instruments alone. With the shunts, each sta-
tion being 150 in parallel with 400, that is, 109 ohms, the total
is 3,270 ohms, 1,230 less than without the shunts. It thus
appears that to obtain the same current in the relays the bat-
tory does not have to be proportionately increased.
Moreover, it is not necessary to have the same current
in the relay as before to obtain just as good working; for it
is not the smallness of current which makes Morse working
on the closed-circuit system difficult provided the current is
not reduced unduly, since a Morse relay can be adjusted to
work well on a very small current, but it is the variation due
to line leakage that causes the trouble. These variations are
always less the smaller the total resistance of the line, and
the fact that the total line resistance is materially reduced in
the case cited more than compensates for a slight loss of cur-
rent in the relay, since in any case there is plenty of margin
of current for operating the relay.
The resistance shunt also has a compensating effect on the
alternating-current side. It is obvious that there is loss of
current at the receiving end of the line in the alternating-cur-
rent instrument due to this shunt, but at all stations nearer
the generator end of the line there is more current than is
required for operating the instruments, and it is an advantage
to shunt some of the current past the station. At the same
time the station impedance is thereby reduced, and the total
current received at the end of the line increased, which tends
to compensate for the loss due to the shunt.
some
It is not necessary at any given point of the wire to have
both alternating- and direct-current receivers. In
instances it may be desirable to have one, or the other, or

14
MULTIPLE TELEGRAPHS.
both. No matter which receiver is to be omitted from the
diagram of Fig. 2, the circuits are always connected as in that
diagram, but instead of the relays coils may be inserted which
are equivalent to the relays.
At stations where the alternating-current receiver is to
be omitted, in place of the polar relay is substituted an induc-
tive coil. This coil is put into the same box with the non-
inductive resistance R, having a common terminal, B, Fig. 2,
which is brought to the middle binding-post of the com-
pensator-box, Fig. 5. The left binding-post on this box is
i t t
FIG. 5.-Compensator.
the other terminal, A, of the non-inductive resistance; and
the right binding-post is the other terminal, C, of the induc-
tive coil. The base of the compensator-box measures 41 by 51
inches.
On a long telegraph wire the alternating current will per-
haps vary from 100 milliamperes at the generator end to 20
or 30 at the receiver end, and if there are a large number of
these compensator-coils distributed along the wire, it is evi-
dent that some of them will carry large currents and others
small currents. It is important that the inductive values of
the coils should not change with or be dependent upon the
current, as they are there for the purpose of balancing a fixed
condenser. If they are constructed with closed iron circuits,
it is found that the coefficient of induction varies to a large
extent with the magnitude of the current. and while balancing
the condensers at one value of the current they are far from
balancing them at another value of the current. This difficulty
is almost entirely overcome by leaving an air-gap of the proper
magnitude in the iron circuit of the inductive-coil, and the
compensators are constructed in this manner.


DUPLEX-DIPLEX SYSTEM: RECEIVER.
15
The plan adopted for the main-line receiving circuits has
now been described. In operating the sounders on local cir-
cuits the usual arrangement for the Morse relay on the direct-
current side is maintained, and in installing the duplex-
diplex apparatus on any Morse line which is now in operation
it is not necessary to change any of the instruments used in
the direct-current working.
On the alternating-current side, however, it is desirable
to use a repeating-sounder between the polar relay and the
reading-sounder, as is done in the common quadruplex system.
The polar relay is adjusted so that a bias holds the tongue
against its contact, and immediately upon receiving a signal
the relay armature is set into rapid and strong vibration,
which continues until the signal ceases. The repeating-sounder
is opened immediately upon the beginning of the vibration,
and this closes the reading-sounder, so that the signals on the
alternating-current side come in firm and clear on the sounder,
and are read exactly as in ordinary Morse working.

CHAPTER II.
THE TRANSMITTING APPARATUS IN THE DUPLEX-
DIPLEX SYSTEM.
IT has not been shown in the preceding chapter on the
receiving circuits how either a direct-current or an alternating-
current transmitter may be connected to the line so as to
cause no interference with the other side of the system. With
respect to the transmitter circuits, it should be noted that
the closed-circuit system or the open-circuit system may be
employed for either the direct-current side or the alternating-
current side, independently of that used on the other side.
Since the closed-circuit Morse system is used almost exclu-
sively in this country, the circuits as employed in closed-cir-
cuit working on both sides are considered first.
If it is desired to introduce a station in a wire in one cir-
cuit only, for example, the direct-current side, without having
any transmitter or receiver in the alternating-current side,
then, since there is no receiver present at the station on the
alternating-current side to be interfered with, the ordinary
closed-circuit Morse key, K, may be introduced in the circuit,
as shown in Fig. 6. This arrangement becomes identical with
the receiver circuits in Fig. 2 if the key is closed, and opening
the key has the effect of stopping all the direct current in the
wire. When the key is closed the condenser, C, is discharged
through the Morse relay, L, and resistance, R, in parallel; and
when it is opened the condenser receives the full pressure of
the direct-current batteries, since it then forms the only obstacle
to the direct current. The condenser should therefore be able
to withstand high pressures without breaking down.
The discharge of this condenser through the Morse relay, L,
is beneficial rather than otherwise, for the moment of closing
the key, when the discharge takes place, is just the moment
when the relay should respond to the signal. When the key
16

DUPLEX-DIPLEX SYSTEM: TRANSMITTER.
17
is opened all current is suddenly removed from the relay except
for its own discharge through the resistance, R, which in prac-
tice has no perceptible effect. The home
The home relay, therefore,
A
K
Sc
C
AL
R
mi
D
m
Tex
B
FIG. 6. Standard arrangement for inserting a direct-current station
only in a wire.
responds rather more sharply and decisively with the con-
denser than without it.
The presence of the condenser, C, tends to prolong the cur-
rent very slightly in all the distant relays upon the line, when
working the key, K, for it is seen that when the key is opened
it is desired to stop the current immediately, but the con-
denser, previously discharged, must then receive a new charge
from current which continues to flow over the wire, through
the distant relays. The amount of the effect, however, with
a condenser of three microfarads, is so small that practice
has shown that t may be entirely neglected.
It should be stated that this effect is not even theoretically
present in an open-circuit system; but seems to be present
only with the closed-circuit plan.
Fig. 7 is drawn as an alternative method which may be
employed to insert a direct current station only on the closed-
circuit plan. It is essentially the same as Fig. 6 and Fig. 2
when the keys are closed. The difference is that it requires
more apparatus, there being two condensers, C and C', and
two balancing inductance-coils, D and D', instead of one of
each. The receiving circuits between the points E and B are
left intact, as in Fig. 2, and for this reason it would be possible
to substitute for the coil D a polar relay and receive the mes-
sages on both sides of the wire, though being able to transmit
but one message.

18
MULTIPLE TELEGRAPHS.
It will presently appear that a transmitter for the alter-
nating-current side may at any time be inserted in the wire
outside of the points A and B without disturbing the wiring
of the circuits of Fig. 7, and thus convert it into a complete
of
üm
R
A
150w
LI
K
D'
FIG. 7.-Alternative method of inserting a direct-current station only in
a wire.
+ fo
100 Volts
station, having receivers and transmitters in each of the two
circuits.
Before describing a method of inserting a transmitter on the
alternating-current side, the difference between closed-circuit.
and open-circuit working will be considered, for this difference
is very much more marked with the alternating current than
with the direct current. The matter will be more clearly under-
stood if the effect of leakage on a wire with direct currents.
is first considered.
K₁
Referring to Fig. 8, let EABE represent a telegraph wire
having a resistance, including instruments, of 2,000 ohms, and
E
850W
C
C
E
D'
1000w
850w
B
ma
K2
150w
Hom
L2
B.
E
FIG. 8.-Illustrative of line leakage in ordinary closed-circuit Morse
working.
battery of 100 volts at the A end only. If there is no leakage,
the current which flows upon closing both keys, Ki and K2,
is 100/2000 ampere, or 50 milliamperes, and this current is
the same in both relays.

DUPLEX-DIPLEX SYSTEM: TRANSMITTER.
19
Let there be a leak of, say, 1,000 ohms resistance connected
from line to earth at its middle point. Then upon closing
both keys the current in Li, near the battery, becomes 66
milliamperes, and that in relay L2, at the distant end of the
line, is only 333 milliamperes; the current flowing to ground
through the leak being also 333. If the distant key, K2, is
opened, the current in L2 falls from 333 to zero; but that in
Li falls from 663 to 50 milliamperes. On the other hand if the
key K1, nearest the battery, is the only one to be opened, the
current falls from 333 to o in the distant relay, L2, and from
66 to o in the home relay, L1.
That is to say, the key K1, nearest the battery, will always
operate all the relays for any working adjustment; but the
relay L1, nearest the battery, must be carefully adjusted to
receive signals from the distant key, K2, which is sending a
message toward the battery, because the current is only changed
from 663 to 50 in relay L1 by the operation of key K2.
The instruments can be adjusted more easily if the battery
is divided into two sections of 50 volts each, one connected
at each end of the wire, which is the common practice, and
has become so for these reasons. Assuming this to be done
in the wire illustrated in Fig. 8, the currents which flow when
both keys are closed are 50 milliamperes in each relay and no
current through the leak, CE. If the key K2 only is opened,
the current in L2 falls from 50 to 0; but in L1 from 50 to 25.
The change of the working current in this relay is now 50 per
cent., whereas in the previous example it was only 25 per
cent.; and a working adjustment is much more easily obtained.
Although the above example assumes the leakage to be
all at one point of the wire, which is seldom the case in prac-
tice, yet it illustrates the effect of leakage as well as if it were
uniformly distributed. If it were uniformly distributed, the
current on the wire would gradually diminish from a large
value near the battery end to a small value at the distant
end, passing through all the intermediate values. In this
condition, for the reasons stated, it is difficult to send a message
toward the source of electrical energy (the battery) by open-
ing and closing the wire at the distant end.
If for the direct electromotive force we suppose that an
alternating electromotive force of 150 cycles is substituted
in Fig. 8, and that there is no leakage, the alternating cur-

20
MULTIPLE TELEGRAPHS,
rent will decrease continuously along the wire from the gen-
erator in a manner analogous to the decrease of the direct
current, with leakage, as far as its effect upon the receiving
instruments is concerned. The cause of this decrease of the
alternating current is the static capacity of the wire.
One difference between the two cases is that there is little
or no energy loss with the alternating current, while there is
an energy loss with leakage. The amount of energy, how-
ever, is of little consequence, since the total energy used on a
telegraph wire is small in any case.
The greater the static capacity of the wire per mile the
more rapidly the current will fall. In underground wires the
capacity per mile is so much greater than in aerial wires that
the working distance is reduced if underground wires are used.
In practice the reduction in the current along the wire is
so rapid that it is only practicable to send messages toward
the source of alternating electromotive force for very short
distances; and since it is not practicable to use two generators
(for they must be synchronized), one at each end of the wire,
so as to obtain the advantage cited above, the application of
the alternating current on the closed-circuit plan is much
more limited than with the open-circuit plan. Generally, a
generator is located at each station on the alternating-current
circuit; but there is one important case where a closed circuit
may be used.
Figs. 9 and 10 illustrate two methods of inserting a trans-
mitter for the alternating-current side on the closed-circuit
A
K
2G
PR
R
imm
B
[5]
FIG. 9.-Arrangement for inserting an alternating-current station only in
a wire.
plan, which are the counterparts of Figs. 6 and 7 for direct
currents. The circuits of Fig. 9 become the same as those
of Fig. 2 when the key is closed, except that the inductance L',

DUPLEX-DIPLEX SYSTEM: TRANSMITTER.
21
Fig. 9, should be larger than L of Fig. 2 is required to be, for
when the key is opened the inductance receives considerable
pressure from the alternating electromotive force, as it then
serves to stop the flow of alternating current, although some
small amount will pass through it. It is not quite as effectual
in stopping the alternating current as the condenser is with
the direct current in Fig. 6.
In opening the key K the direct current is not interrupted,
but flows, as before, through the coil L'. The ohmic resistance
of this coil may be such that the larger part of the direct cur-
rent flows through it at all times, whether the key is open
or closed. When the key is closed, however, the coil, being
shunted by a non-inductive resistance and the polar relay,
loses its reactive property as regards its effect on the direct-
current side.
The full reactive effect is inserted in the direct-current
side when the key is opened. Since but one key on the alter-
nating-current side is opened at a time, only one such coil is
introduced, at some one point of the whole line, this point
being the station which happens to be transmitting the alter-
nating-current message. The value of the inductance of one
coil is not great enough to have a detrimental effect in the
direct-current working.
In Fig. 10 an alternative plan of introducing an alternating-
current station is shown. This arrangement requires one more
A
L
K
E
R
Amm
C
PR
Sop
B
FIG. 10.-Alternative method of inserting an alternating-current station
only in a wire.
coil, L, than is used in Fig. 9, and it retains the receiving cir-
cuits intact between the points E and B, as in Fig. 2, and makes
it possible to have both alternating- and direct-current receivers,
if desired. With this arrangement a direct-current trans-
mitter may easily be introduced in the wire outside the points

22
MULTIPLE TELEGRAPHS.
A and B, thus completing a station with receivers and trans-
mitters on both sides.
If the apparatus between the points A and E of Fig. 7
is inserted in the line beyond A or B or at E, this gives a com-
plete working station, which is illustrated in Fig. 11, where KA
Af
R
im
în
icto
PR
FIG. 11. Complete station with both direct- and alternating-current
transmitters and receivers, and with no generator at the station.
KA
F
Ko
D
G
B
is the key for transmitting on the alternating-current side,
and Kp the key for transmitting on the direct-current side.
When both of these keys are not being used, and are closed,
the points A and G are short-circuited, and the only apparatus
in the line is the receiving-set between G and B.
It should be understood that the operation of the direct-
current key, KD, will send messages in both directions along
the line, and will operate all of the direct-current receivers
in the usual manner, a battery being connected to ground
at each end of the line.
The alternating-current generator, on the other hand, is
situated at one end only of the wire, say on the A side of the
station of Fig. 11, and is connected in series with the battery
furnishing the direct current. The armature of an alternating-
current generator itself may be connected in the wire, or, where
more than one wire is to be supplied with alternating current,
the generator is preferably connected to the primary of a trans-
former, and some secondary coil, giving the desired voltage
for the wire in question, is inserted in the line.
As explained above, the operation of the key KA will not
work the alternating-current receivers which are on the A side
of the line, toward the generator, but will operate all receivers
on the B side of the station. The effect of this is that the
operator nearest the end of the line away from the alternating-
current generator can send no messages on the alternating-
current side.

DUPLEX-DIPLEX SYSTEM: TRANSMITTER.
23
The system in this condition really amounts to a duplex
only, which permits of a number of way stations; for the
direct-current side permits the operator to send a message
toward the A end of the line at the same time that another
operator is receiving one on the alternating-current side. The
utility of this is somewhat questionable, however, for on a
way station line of this character operators at two stations
in communication with each other want to speak back and
forth as though they had an independent wire; but they evi-
dently could not do so unless they controlled both sides.
This closed-circuit system is illustrated because it has an
application where two single Morse wires are available; for
one wire may be equipped with alternating apparatus to send
messages from A to B, the generator being at A, and the other
wire with apparatus to send messages from B to A, the gen-
erator being at B. It is evident that with two single Morse
wires another complete working circuit is obtained which is
capable of sending messages in one direction or the other,
or in both directions at the same time if desired, the additional
service furnished being equivalent to an independent wire with
many way stations, worked duplex and requiring but two
generators.
OPEN-CIRCUIT SYSTEMS.
The diagram Fig. 12 shows a means of working both the
direct- and the alternating-current sides on the open-circuit
D' C'
KD
F
ww
R
G
큰​이어
​C
PR
Ľ
KA
B
Lamm | | x
FIG. 12.-Complete
station with both direct- and alternating-current
transmitters and receivers, and with both a direct- and an alter-
nating-current generator.
system. The receiving circuits between the points F and G
remain the same as in Fig. 2, but the alternating-current key,
KA, is arranged to introduce the alternating-current gen-
erator Y only when it is closed, and to remove the generator

24
MULTIPLE TELEGRAPHS.
when it is open. There is a continuous path through the
station at all times for the direct current. When the key is
open the points G and B are connected together, thus short-
circuiting the coil L', and when closed the path is through
the generator Y. When the key is in transit, however, from
back to front stop, there is a path through the coil L', and
the direct-current circuit is never interrupted by the operation.
of this alternating-current key.
It is undesirable, however, to compel the direct current
to pass through the coil L', even for a brief instant, and this
has led to the use of a continuity-preserving transmitter, which
never. compels the direct current to pass through the coil L'.
This is illustrated in Fig. 13, which is a preferred arrangement,
TA representing a continuity transmitter.
It is now evident that the coil L' may be opened and removed
without disturbing the operation of the system. The reason
why it is shown in the diagram is that in practice this coil is
provided to afford a protection against the circuits being acci-
dentally interrupted, if an operator should chance to open
the circuit in adjusting his transmitter contacts. This coil
receives the full potential of the generator Y, but its induct-
ance is such that only a very small current passes through it.
In a similar manner a direct-current transmitter, KD,
Fig. 12, is introduced in the line between the points A and F,
but a better plan is to use the continuity-preserving trans-
mitter, TD, of Fig. 13. The condenser C' and balancing-
A
To
D'
229
-¶¶¶¶
X
F
www
R
m
G
C
PR
G
TA
B
FIG. 13. Similar to Fig. 12 except that continuity-preserving trans-
mitters are employed.
coil D' are merely inserted here as protective devices to pro-
vide against accidental opening of the alternating-current side
when adjusting the contacts of the transmitter TD. Normally
the alternating current never passes through the condenser C',

DUPLEX-DIPLEX SYSTEM: TRANSMITTER.
25
but the latter receives the full pressure of the battery X when
the transmitter is closed.
A
As previously stated, it is not necessary to operate the
direct-current side on the open-circuit plan because the alter-
nating-current side is so operated, and Fig. 14 illustrates a
KD
D
L
L
J
TA
96
m
Sc
R
PR
B
FIG. 14.-Complete station with both direct- and alternating-current
transmitters and receivers, and with an alternating-current generator
but no direct-current generator. Most common arrangement.
complete station with the alternating current operating on
the open-circuit plan and the direct current on the closed-
circuit plan. This disposition of the apparatus is apt to be
more frequently used than any other, because of the fact that
the alternating current must generally be used with the open-
circuit plan, and the common practice in this country is to
use the direct current on the closed-circuit plan.
One important feature of the system is its ability to
provide a through working circuit on the alternating-current
side on a wire having a large number of way stations on the
direct-current side. The operation of the alternating-current
side takes place independently of anything that is happening
on the direct-current side, and has all the advantages of work-
ing on the open-circuit system, which is, in general, preferable
to closed-circuit working, because the working current received
in the relays always varies between some value and zero,
instead of varying between two values neither of which is
zero, as previously explained.
The messages are always received on the alternating-cur-
rent side, as some of the operators have expressed it, as
clear as a bell" and "as though the sender were in the next
room "; and this independently of the weather conditions,
provided that the leakage on the line does not exceed a value.
which will provide the minimum current required to operate
the receivers. In the closed-circuit Morse the operators are
64

26
MULTIPLE TELEGRAPHS.
continually troubled by the fact that a different adjustment
is required when there is much leakage on the wire, in receiving
from stations at different points. This adjustment of the
relay is in practice found to be unnecessary on the alternat-
ing-current side for the reasons stated.
Another important difference between the alternating-
and direct-current sides in respect to the leakage of the wire
should be mentioned. The percentage of decrease in the
received current on the direct-current side is greater than on
the alternating-current side for a given leakage; and the differ-
ence between times of leakage and no leakage is therefore
more noticeable on the direct-current side than on the alter-
nating-current side.

CHAPTER III.
APPLICATIONS OF THE DUPLEX-DIPLEX SYSTEM.
THE duplex-diplex system described in the previous pages
is capable of being used in a manner very similar to the pres-
ent Morse system, and since ordinary sounders are used as
receivers (and not some device which would cause a buzzing
or humming noise in a telephone, as in some systems) repeaters
may evidently be employed in the alternating-current side
similar to those in the direct-current side. In fact, a repeater
may be introduced in the alternating-current side at a point
where there is no repeater in the direct-current side.
Instead of giving separate examples of the many possible
applications of the duplex-diplex system to wires, a system
will be described which is in actual operation on the Penn-
sylvania Railroad, having been installed in July, 1902, since
it illustrates a variety of different stations, including a repeater
station.
Fig. 15 is a map of the region showing the lines of rail-
road along which the wires are strung from Pittsburgh, in the
western part of Pennsylvania, to Toledo, in the northern part
of Ohio, on Lake Erie; a distance of 261 miles, 175 miles being
the distance from Pittsburgh to Mansfield, O., and 86 miles
the distance from Mansfield to Toledo.
There is another wire, running from Cambridge, O., to
Crestline, O., a distance of 118 miles. Upon this Crestline-
Cambridge wire there are about nine stations, indicated by
crosses, and this line is not equipped with alternating-cur-
rent apparatus, but at Mansfield there is a side-line repeater
on the direct-current side, which unites the two wires into
one system, so that an operator at Cambridge transmitting a
message transmits not only to all points on the Cambridge-
Crestline wire, but to all points on the direct-current side of
the Pittsburgh-Toledo wire as well.
27

28
MULTIPLE TELEGRAPHS.
In the Pittsburgh-Toledo wire there are no stations between
Pittsburgh and Mansfield, and 14 stations on the direct-current
side between Mansfield and Toledo, not including the Toledo
and Mansfield stations. The wire is worked on the direct-
current side straight through from Pittsburgh to Toledo, but
at Mansfield there is a repeater on the alternating-current side,
and this example therefore illustrates the case of a through
TOLEDO
CRESTLINE
N
MANSFIELD
50
CAMBRIDGE
PITTSBURGH
MILES
FIG. 15.-Map giving stations on an actual wire, with scale of distances.
wire with no intermediate stations (Pittsburgh-Mansfield sec-
tion) and of a way wire having 16 stations (Mansfield-Toledo
section).
The through wire from Pittsburgh to Mansfield is partly
copper and partly iron, 100 miles of the distance being No. 9
copper and 75 miles being No. 8 iron wire. The wire from
Mansfield to Toledo is No. 8 iron. The resistance of the line
and instruments from Pittsburgh to Mansfield is about 3,200
ohms. The resistance of the line from Mansfield to Toledo
is about 2,500 ohms, exclusive of instruments, but the large
number of instruments on this wire increases its resist-
ance so that it is more than that of the Pittsburgh-Mansfield
section.
The alternating-current side of this system is designed to
furnish a through circuit between the terminal stations, Pitts-

APPLICATIONS OF THE DUPLEX-DIPLEX SYSTEM.
29
-······
||-|-|-·|-|-|-·
Cambridge
Au
X
* of
ww
Pittsburgh
qfe
A
Fat
to
OF
of
عاد
Mansfield
of
671 ป
Le
i
of
podi
Ho
Toledo
Jo
10
LODE
FIG. 16.-Diagram of apparatus used on the wires shown in Fig. 15. The local wires at the repeater station
are omitted; they are shown in Fig. 17. There are on this wire thirteen other
stations similar to X between Toledo and Mansfield.
Crestline
-·|-·|·|·|·|·|-~
BB
→HHHHHHHH

30
MULTIPLE TELEGRAPHS.
burgh and Toledo, and incidentally it furnishes another sta-
tion at the repeater in Mansfield. Most all of the through
traffic between Pittsburgh and Toledo is sent over the alter-
nating-current circuit, the direct-current side being occupied
with business from intermediate stations.
In Fig. 16 a diagram of the apparatus used at Pittsburgh,
Mansfield, Toledo, and intermediate stations is given, showing
the method of operating a repeater on the alternating-cur-
rent side without having one on the direct-current side. A
side-line repeater is provided, however, on the direct-current
side at the same point, Mansfield, to repeat into the Cam-
bridge-Crestline wire.
V
Fig. 17 shows the alternating-current repeater. The alter-
nating current received from Pittsburgh, on the left of the
drawing, after passing through the receiver-set, between the
points A and D, is compelled to pass to ground through the
condenser C₁, on account of the inductance-coil L', which
prevents the current from passing on through the opposite
receiver-set and thereby causing interfering action.
In a
similar way, if alternating current comes from Toledo the
same coil, L', compels it to pass to ground through the con-
denser C4. There is no connection from line to ground except
through condensers, so that the direct current passing over
the wire has no path to ground.
If Pittsburgh transmits alternating current, the polar relay
PR₁ responds, and its local circuit opens the transmitter T2,
which is normally closed, and thus connects the alternating
generator, Y, from ground, E, through the condenser C3 to
the point B, on the wire to Toledo. The inductance-coil L'
prevents this current from going out through the relay PR2
toward Pittsburgh, and thus helps naturally to prevent the
back action of the repeater, which would happen if it operated
the polar relay PR2; for this relay would operate the trans-
mitter T₁, and send an alternating current from the gener-
ator Y through the condenser C2 back through the line toward
Pittsburgh.
As an additional precaution in preventing the relay PR2
from operating, due to the small current which may pass
through the coil L', an auxiliary relay is provided with con-
tact at M2, which is closed so as to short-circuit PR2 when-
ever transmitter T2 operates. A similar contact is provided

APPLICATIONS OF THE DUPLEX-DIPLEX SYSTEM.
31
←
←
To Cambridge
To Pittsburgh
A
-fc
SIG
PR
D
π
క
[q
-·|·|-··
。』
Mansfield
·-·|·|--
F
M
of
of
·-|·|-·
M2
E
m
------
T₂
C4
GDE
PR2
B
To Crestline
To Toledo
T
FIG. 17. Repeater station, with a repeater on the alternating-current side, but no through repeater on the direct-current
side; there being, however, a side-line repeater on this side. The dotted lines indicate local wires.

32
MULTIPLE TELEGRAPHS.
at M1, to short-circuit the relay PR₁ whenever transmitter Tı
opens.
In other respects the repeater operates in a manner similar
to ordinary repeaters, and the action will be understood from
the diagrams without further description. Only two of the
condensers C1, C2, C3, and C4 are connected at the same time
from the line to ground, and these condensers are made smaller
than the normal duplex-diplex condensers so as not to greatly
increase the static capacity of the line. They are one micro-
farad each.
The generator used on this wire is of special construction,
being an eight-pole magneto-generator, and is shown in Fig. 18.
FIG. 18. Alternating-current 150-cycle eight-pole magneto-generator.
It generates 180 volts when running at 2,250 revolutions per
minute, giving 150 cycles per second. The voltage may be
reduced by adding armatures or keepers to the magnets so
as to partially divert some of the field from passing through
the armature.
A permanent magnet field was desired in these machines
because the power where the generators are is supplied by
the constant-potential 60-cycle alternating-current system,
and this method seemed simpler than to obtain a direct cur-
rent for exciting the field of other types of machines. The
magnets on these machines have remained very permanent,
no noticeable drop in the electromotive force generated having
been experienced. The electromotive force employed on the
line is about 160 volts at each station.

APPLICATIONS OF THE DUPLEX-DIPLEX SYSTEM. 33
The generators are constructed for belt-driving, so that
one type of machine will serve with any kind of power avail-
able; but direct-coupled generators and motors are to be
desired, similar to those employed in the case of direct-current
generators for telegraph purposes.
An important application of the duplex-diplex principle
is to lines now worked with the polar duplex. The system is
not capable of being worked to quite as long distances as the
polar duplex, but in cases where it can be used it has several
advantages over the latter. First, it permits of two messages
being transmitted simultaneously from the same end of the
wire, and may therefore be employed to clear up an excess
of business in one direction. Moreover, each receiving opera-
tor can "break" the operator sending his message independ-
ently of the other pair of operators, which is not true with
the polar duplex; this in effect being the same as supplying
each pair of operators with an independent single Morse wire.
This saves time when corrections or any matter which is
not clearly understood has to be retransmitted, for in the
polar duplex if the receiver B fails to understand something
which A has transmitted he can only inform A of that fact
over this wire by interrupting C, who is at the time trans-
mitting a different message to D, at the opposite side of the
wire. Thus all four persons are involved, whereas in the duplex-
diplex the receiving operator who fails to understand something
merely "breaks the sender and informs him, without dis-
turbing the other two operators.
Another conspicuous advantage is that the two independent
circuits on the one wire in the duplex-diplex system may be
treated as if they were two wires, and rented or leased to inde-
pendent parties, each being served with a single Morse wire.
The method of accomplishing this, showing different branches
or legs, distributed to different sets of subscribers, each side
of the wire affording a number of Morse stations, is shown in
Fig. 19. This furnishes the subscriber with nothing but stand-
ard Morse instruments, which are operated on a local circuit.
The two legs, indicated in the diagram by AC and DC, corre-
sponding to the alternating-current and direct-current sides
of the duplex-diplex, unite at a central office, where all of the
apparatus connected to the main trunk wire is situated.
If two wires, each having a ground-return circuit, are used

7
青​利
​A
11
DC
AC
4······~
H
HO
····
۱۰۱۰ ۱۰۰۰
#
-0%
LINE
【三​工
​····
Hil
۱۰۱۰۰۰-
|
| | | | | |
DC
AC
FIG. 19.-Method of distribution, furnishing each of two sets of subscribers with an independent Morse wire service.
34
MULTIPLE TELEGRAPHS.

APPLICATIONS OF THE DUPLEX-DIPLEX SYSTEM. 35
for connecting the central station with either the alternating-
current or the direct-current set of stations, then no automatic
repeater at the central station is necessary; but such a repeater
seems to be necessary if only one wire is employed for each leg.
Either of these methods is applicable to both the direct-current
and the alternating-current sides, but in the illustration the
direct-current side is represented as having an automatic re-
peater at the central station, and the alternating-current side
has two wires, but no automatic repeater. On one of these
wires there are a series of ordinary keys, normally open, con-
nected in parallel. With these keys circuit-closers are not
required, and each key must be left open when the operator
has finished sending. On the other wire there are a series of
standard Morse relays and sounders.
Another advantage gained by the duplex-diplex is that there
is no artificial line, which sometimes requires adjustment with
the polar duplex. Also, stations may be inserted anywhere in
the wire without the necessity of inserting a repeater station,
which is required with the polar duplex.
Since the polar duplex can operate to greater distances, this
alone is a reason why it will not be replaced by the duplex-
diplex in many cases. But it is not to be understood that the
duplex-diplex system can work to only very short distances.
The limit of working is reached when operating the system as
a duplex rather than as a diplex, and is due to the interference
between the two sides of the system, which increases with the
length of the line.
The longer the line the greater the alternating electromotive
force to send sufficient current to the distant end, and this
means that the alternating current at the transmitting end of
the wire increases with the length of the wire and therefore
must finally reach a point where it will interfere with the best
operation of the direct-current relay, which at this station is
being used as a receiver. In general, as near as any limit can
be stated from present experience, it may be said that a line
having a KR of 15,000 or under may be operated duplex-diplex.
If, however, a line is equipped to operate diplex only,
the length of line which can be worked may be considerably
increased, for then the two generators, direct-current and
alternating-current, are at one and the same end of the wire
only, and there is no difficulty in separating the small received

36
MULTIPLE TELEGRAPHS.
currents at the distant end so as to operate their respective
receivers.
If two line wires between the same points were equipped.
with duplex-diplex apparatus they might be operated so as to
obtain advantages over the polar duplex and yet each wire
always be worked diplex, so as to retain the advantage of long-
distance working. Thus, if A, B, C, and D are four operators
in one city, X, corresponding with four other operators, A', B',
C', and D' in another city, Y, all four operators, A, B, C, and
D may simultaneously transmit messages to A', B', C', and D"
from X to Y; and again, the four operators A', B', C', and D,
may all send simultaneously from Y to X. Also, the two
operators, A and B, on one wire may send to A' and B', from
X to Y, at the same time that the operators C' and D' are trans-
mitting two messages to C and D, from Y to X, this last instance.
being equivalent to the use of the wires as two polar-duplex
circuits.
The above examples have assumed that the system has
been operated merely as a diplex, but have even thus shown
advantages over the polar duplex.
The limiting distance of working the alternating-current
side of the system has heretofore been considered with the
understanding that the telegraph wires upon which the system.
is used are those which are commonly found, without being
equipped with inductance-coils. The advantage which it is
stated has been gained by the use of so-called "load-coils"
distributed along the wire (according to certain systems of
which that due to Prof. M. I. Pupin may be cited as an example)
for the purpose of extending the working distance in telephony
should be obtained with much less difficulty if such a system
were applied to a line employing a simple sinusoidal current,
in the duplex-diplex system; for in such a telegraph system
there is only one frequency and wave length to deal with,
whereas in telephony the number of different frequencies is
unlimited. By such means transmission by the duplex-diplex
system should become practicable to much greater distances
than are otherwise possible.

CHAPTER IV.
THE ALTERNATING- AND DIRECT-CURRENT QUADRUPLEX.
THE principle of using an alternating current to obtain an
additional circuit over a wire has been applied to the ordinary
polar duplex system, and as a result, in addition to the direct-
curent duplex, an alternating-current duplex is also obtained
over the same wire, which is independent of the direct-current
duplex. This is known as the quadruplex, and as to the handling
of messages, sending four messages simultaneously, two in each
direction, does not differ from the common quadruplex.
A diagram of the system is shown in Fig. 20. An artificial
line is employed, but instead of connecting the two receivers in
series, as in the common quadruplex, they are connected to the
line and artificial line in circuits leading from two independent
apexes, A and B. The direct-current transmitter, TD, is con-
nected to the apex A through an inductance-coil, L', and the
transmitter TA is connected to the apex B directly, but the
branches leading from this apex to the line and artificial line
each contain a condenser.
When a balance is obtained a current from the transmitter
TD, passing through the coil L' to apex A and thence to line
and artificial line, does not affect the home polar relay, PR₁,
and does not flow around to ground via the condensers and
apex B because of the presence of the condensers. When the
direct current is altered or reversed, however, a brief impulse
is transmitted through the two condensers to ground through
apex B; but this does not affect the relay PR2, because equal
currents pass differentially through the two halves of the
winding.
One purpose of the coil L' is to reduce the suddenness of
this current impulse, but the main purpose is to prevent the
alternating current which is applied to the apex B by the trans-
37

38
MULTIPLE TELEGRAPHS.
mitter TA from returning backward to ground through the coils
of the direct-current relay PR1, coil L', and transmitter TD, and
~1∙Moldo
compel it to pass to the line.
A certain small amount does
pass backward through this
coil, L', but this does not
affect the relay PR1, be-
cause it divides differentially
in the two windings. Thus
it is seen that neither relay,
PR₁ nor PR2, is disturbed
by the transmission of either
the direct or the alternating
current.
Relay PR3, at the other
end of the wire, receives the
direct current, and its tongue
is held firmly against one
side or the other, as there
is always a positive or a
negative direct current on
the wire. The path through
the polar relay PR4 at the
distant end offers a lower
impedance for the alter-
nating current than PR3,
and this relay will respond
alternating currents
to
only.
Line
PR3
PRI
B
A
Artificial Line
O
F
THE
ドロ
​min
mi
Hooldada
FIG. 20.-Diagram of the direct- and alternating-current quadruplex.
One distinguishing differ-
ence between this quad-
ruplex and the common
quadruplex is that the plan
of using a high voltage for
one side of the system and
a low voltage for the other
is not employed, the voltage
on the alternating-current
이이이​이이​이이​이이​아
​side being approximately
the same as that required on the direct-current side; and
the currents, instead of passing through the same receivers,

ALTERNATING- AND DIRECT-CURRENT QUADRUPLEX. 39
and being differentiated because of their differing magnitude
and direction, are here caused to pass through different
channels in which the separate receivers are located, this
being accomplished by the differing characteristics of the two
currents.

CHAPTER V.
LOW-FREQUENCY DUPLEX-DIPLEX.
THE duplex-diplex system has been operated in an experi-
mental way upon loop wires having both terminals in New
York City, the circuits going and returning by different routes,
so that the wire actually surrounded considerable territory.
These experiments were made to ascertain the extent to which
the operation of the duplex-diplex system on wires strung
on the same pole-line carrying telephone circuits would affect
the telephones.
For this purpose an alternating current of 150 cycles was
applied to a certain wire, and the other wires on the pole-line
were examined with a telephone receiver for inductive disturb-
ances. With the first values tried there was considerable inter-
ference with the telephones, and the alternating electromotive
force, and consequently the current, was then gradually re-
duced, leaving the frequency the same, until no audible sound.
could be detected on any of the wires. The value of voltage
thus permitted was, however, so low, something like 10 or 15
volts, that it was not sufficient for operating the duplex-diplex
system.
Experiments were then undertaken to ascertain the effect.
of reducing the frequency. As the frequency was reduced,
the hum on the wire which gave the loudest noise disappeared
when the frequency had fallen to between 50 and 60 cycles,
and at this frequency a sufficient current could be transmitted
without affecting any telephones to enable the duplex-diplex
to be operated.
When attempting, however, to operate the 150-cycle duplex-
diplex apparatus at this low frequency there was serious inter-
ference between the two sides of the system itself, and an attempt
was made to readjust the values of inductance and capacity
so as to be able to work the apparatus. This, however, proved
not to be practicable, and it then appeared desirable to develop
some system utilizing a frequency of 50 cycles
40

LOW-FREQUENCY DUPLEX-DIPLEX.
41
The system which has been devised for operating at a low
frequency differs from the 150-cycle system described chiefly
in the arrangement of the receiving circuits. This is shown
in Fig. 21, which represents receivers for both sides of the
miny
C₁
A
F
C2
E
Ri
心
​R2
[209
K
B
C3
FIG. 21.-Arrangement of receiving circuits which may be employed with
an alternating current of comparatively low frequency.
system. The line wire divides at A into two branches, D and
F, which unite again at B. The branch D contains an induct-
ance, L, and a differentially wound relay, R1, having its two
windings in parallel, the branch D dividing and passing through
the two coils of the relay to unite at the point E.
If it were not for the fact that the condenser C1 is inserted
between the points D and G, in series with one of the windings of
the relay only, that is, if the condenser C₁ were short-circuited,
then any direct current arriving over the line would divide
at D into two equal currents, neutralizing each other and pro-
ducing no magnetism in the relay R1. The condenser C₁
however, interrupts the direct currrent which would otherwise
flow through the half of the winding GE, and compels it all,
except that which passes through r, to pass through the winding
HE, thus operating the relay.
The reason for this arrangement is that alternating current
coming over the line has a very different effect, for the induct-
ance L prevents all but a very little of it from going by the
path ADEB, and compels it to go via the path AFKB, which
has a comparatively small impedance. The small residue of
alternating current which does pass through the coil L divides
almost equally in the two paths DGE and DHE, since the
condenser C₁ has such a value that the impedance of these
two branch circuits is made up almost wholly of the windings

42
MULTIPLE TELEGRAPHS.
GE and HE. That is, the condenser C₁ behaves, in respect to
the alternating current, about the same as the wire DH, and
therefore the alternating current divides equally in the two
windings GE and HE, and thus even the small residue which
passes by the branch D is neutralized by the differential relay.
In the branch FK is inserted, in an analogous manner, a
differential relay, R2, which will respond to alternating currents.
At F it divides into two branches, each passing through a con-
denser, C2, C3, to the relay coils, and uniting again at K and B.
If the coil L' which is connected in one of the branches only
between J and K were removed or short-circuited, then all the
alternating current which passed into the branch F would divide
into two equal parts, which would neutralize each other in the
relay R2.
The presence of the coil L' upsets the balance of this relay
and compels most of the alternating current to pass through the
condenser C3. The relay R2 will thus respond properly to
alternating currents. No direct current can pass through it on
account of the condensers C2 and C3, but the sudden impulses
or kicks when the direct current is made and broken do pass
through these condensers. They divide, however, nearly equally
in the two branches, C2 and C3, because the condensers, for
this kind of current, have the predominating effect, and because
they divide equally the effect of these disturbances is neutralized
in relay R2. This arrangement gives a very perfect way of
receiving both alternating and direct current, which is adapted
for frequencies of 50 or 60 cycles.
A high-resistance shunt, r, is connected around the condenser
C₁ in order to prevent slight oscillations in the closed circuit
DHEG, which are set up through the action of the alternating-
current transmitter. The alternating- and direct-current trans-
mitters are connected in the line in the same manner (Fig. 14)
as in the 150-cycle system, except that the values of the con-
densers and coils correspond to the 60-cycle system.
The apparatus involved in this receiving circuit is more
complicated than that in the 150-cycle system, and the con-
denser capacity required is much greater. The condenser C1
is nine microfarads, and C2 and C3 are each six microfarads,
making a total of twenty-one microfarads, as against three
microfarads with the higher-frequency system.

PART II.
METHODS OF OBTAINING SYNCHRONISM AT
DISTANT POINTS.


INTRODUCTION.
MANY telegraph systems depend for their operation upon
the principle of the synchronous rotation of two bodies at
distant points. The word "synchronism" as here employed
does not mean an approximate synchronism, such, for example,
as is afforded by two well-regulated clocks; but a synchronism
similar to that obtained with gearing, so that the distant
motor makes, say, just 1,000,000 revolutions while the home
motor makes the same number, and not one more nor one
less. This would evidently be an impracticable problem with-
out something to take the place of gearing; and the ether
through which the electromagnetic impulses pass may be
thought of as giving such a connection between the two instru-
ments, which at the same time has sufficient elasticity to permit
of some oscillation or displacement, but never enough to throw
the system out of gear.
In the development of synchronous telegraphs there are two
problems which may be considered independently of each
other: first, the means of obtaining synchronism; and second,
the means of securing a practical telegraph system after the
problem of synchronism has been satisfactorily solved. Although
the practical result depends upon the successful solution of
both of these problems, each may be investigated independently
of the other. Synchronism may be accurately obtained for
experimental purposes by mounting the two revolving bodies
upon a common shaft at one place and operating the telegraph
svstem on a loop wire or an artificial line.
It makes a more systematic arrangement in treating of
synchronous telegraphs to divide the subject into two distinct
parts, and the part relating to the means of obtaining syn-
chronism logically precedes that relating to the means of using
the synchronism for operating telegraphs.
45

46
MULTIPLE TELEGRAPHS.
If synchronous telegraphs ever come into more general use,
it is difficult to say how the problem of obtaining synchronism
will be solved; but it would seem desirable from many points
of view to maintain synchronous rotors in each of the several
important centres which are to be connected by telegraph.
This interconnected system of rotors would be kept running
day and night at all the places, so that it would be easy to
insert any standard set of telegraph instruments, whether a
printing or a Morse set, in the line in a manner similar to that
now practised with the Morse telegraph.
The system of synchronous motors adopted thus becomes
important; for, when once installed, it serves for the operation
of more than one wire, each wire going, if desired, to a different
locality, and being capable of carrying several telegraph mes-
sages at once, either by the Morse or a printing system. Thus
each synchronous rotor is entrusted with as many circuits as, in
the judgment of the manager, it is wise to operate dependent
upon the one synchronous system. The more confidence there
is in the synchronism the greater the number of circuits likely to
be entrusted to it. For these reasons it seems wise to set apart
a special wire, or system of wires, between the various cities, to
be used for nothing else than to maintain synchronism. The
transmitting capacity is considerably increased in this manner,
even though one wire is used for nothing but the synchronism.
In certain cases, however, it will be desired to operate syn-
chronous motors on the same wires used for telegraphs, and
three distinct cases are given showing the operation on a wire
used for single Morse transmission, duplex Morse transmission,
and multiplex Morse or printing telegraphs. In the first two
instances the additional service is obtained on wires now used
for nothing but Morse transmission, without interfering with
the Morse service. In the last instance the synchronism is
obtained on the same wire with the synchronous telegraph
itself; but this is only accomplished by sacrificing some of the
time, which might otherwise be devoted to operating synchro-
nous telegraph circuits, for the motor; and this reduces the
number of circuits that might otherwise be obtained by at
least one.

THE OPERATION OF SYNCHRONOUS
A
LET Some place nearly in the centre of the region which is to
be connected by the synchronous telegraph be selected as a
point from which to send out impulses over the wires radiating
to the outlying districts. A simple way of arranging this is
illustrated in Fig. 22, in which the line MN represents a shaft
www.|·|·|·|·|·|
F
ㄓ​ˊ
MOTORS ON WIRES
WHICH ARE USED FOR NO OTHER PURPOSE.
H
M
N
C
CHAPTER VI.
G
E
|·|·|·|·|·|-
L
Line
PR
E
FIG. 22.-Method of controlling synchronous motors when a wire is
employed wholly for synchronizing.
carrying two trailing brushes, C and D, electrically connected
to each other at J, and insulated from the shaft. The brush D
bears on the continuous ring, H, and makes connection between
the telegraph wire, L, and the trailer, C, which makes contact
with the ring, AB, divided into two equal parts by the open-
ings F and G. To one half of the ring, A, a grounded posi-
tive battery is connected, and to the other half, B, a grounded
negative battery of the same electromotive force.
The shaft,
47

48
MULTIPLE TELEGRAPHS.
MN, is driven by some mechanical power, the most convenient
usually being an electric motor, connected to constant-potential
mains, and geared to the shaft so that the trailer rotates at the
desired speed.
The ring might have been divided into more than two parts,
in which case more alternations of the current would occur per
revolution of the trailer; but if it is desired to cause the trailer
at the receiving end of the wire to revolve in synchronism
that of the transmitter, the same number of waves must be
received per revolution of the trailer as are transmitted per
revolution. If there were eight alternations of the current per
revolution of the trailer, for example, then the receiving-trailer
might run in any one of four positions each differing in phase.
ninety degrees from the succeeding position. The synchronous
telegraph can only be operated, however, when the correct one
of these four positions is found, a process which has been termed
"finding the circuit," and which necessitates the use of some
apparatus not required for any other reason.
With four alternations per revolution there are but two
positions in which the synchronous motor will run; and with
two alternations per revolution there is only one position in
which the motor can run, and this must necessarily be the
correct one. The reasons for the use of two segments in
Fig. 22 are that the trailer will always start in the proper phase
position, and that all switches or apparatus required for finding
the circuit may be dispensed with, thus making the action more
simple to describe, although it is not intended to imply that
only two segments should be employed, especially when a whole
wire is to be devoted to the purpose of maintaining synchronism.
The synchronous motors may be operated, as hereafter explained,
by local power supplied through the contacts of an ordinary
polarized relay, this relay being the only instrument inserted
in the wire at any given station.
.
To show that it is entirely practicable to operate such a
system of motors in all kinds of weather let us consider its
operation more in detail. Suppose, for example, that the
place chosen for locating the transmitter is, as in Fig. 23, at
Buffalo; and suppose the wire leading from the trailer divides
into three separate branches, the first going to Chicago, the
second to Pittsburgh, and the third to Albany, and let the
third branch divide at Albany into two, one going to New

OPERATION OF SYNCHRONOUS MOTORS.
49
York and Philadelphia and the other to Boston. The figure
shows the relative distances of these places, the cities being
indicated by crosses on the wires.
Chicago
If the longest distance from the central station is not more
than 500 miles, then the most distant points may be 1,000 miles
op
Cleveland
Go
Buffalo
L
T
Pittsburgh
100
Miles
200
300
Albany
New York
Philadelphia
Boston
op
FIG. 23.-Map showing a plan for connecting cities by a system of syn-
chronous motors operated from a central point.
apart. Assume as an illustration that a copper wire having a
resistance of 2.05 ohms per mile, being such a wire as is now used
for long-distance telephone circuits, is to be devoted solely to the
purpose of maintaining synchronism. The theoretical resistance
of 500 miles of this wire is 1,025 ohms, and allowing 200 ohms ad-
ditional for the resistance of instruments, the total becomes 1,225
ohms for the longest circuit. The polarized relays which are used
as receivers at the various stations are adjusted with their arma-
tures in the neutral position, so as to remain on whichever side
they are placed; and in this condition they are very sensitive
to current.
The reason for choosing such heavy wire is to permit of
working with the heaviest line leakage which is likely to occur.
A leak of 200,000 ohms per mile is a value so large that Morse
wires are considered unworkable when the leakage exceeds this
amount. Assuming the leak to have this value, its approximate
effect on the above circuit may be easily calculated. The
approximation will be sufficiently close if it is assumed that
the leakage all occurs at three points of the wire, instead of
being evenly distributed. The leakage for 500 miles of line
at 200,000 ohms per mile is 400 ohms total, and applying this at
three points, B, C, and D, of the line, Fig. 24, each leak resist-

50
MULTIPLE TELEGRAPHS.
ance is 1,200 ohms. If a current of 20 milliamperes is received
by the distant relay under the most severe conditions of leakage,
the operation of the system is not likely to be interfered with.
With 20 milliamperes in the branch DE, it is simple to calculate
A
ww--|·|·|·|·|
R=256 B R=256
R=256
www
wwww
1=0.0559 1 -0.0366 1=0.0251
www
R=100
43 volts
R=1200
C
R=1200
wwwwww
1=0.01931=0.0115
R=256
www
1=0.020
GEOR
R=50
R=1200
1=0.0051
GE
R=50
E
لیا
FIG. 24.-Example of a wire with leakage.
backward toward the source of power to find the electromotive
force and current required at the transmitting end, A. The
resistances of the several branches and the current in amperes
in each are given in the diagram, showing that the required
current at A is about 56 milliamperes, and the apparent resist-
ance of the line measured from A is 770.5 ohms. The real
resistance of the wire and instruments, 1,225 ohms, is almost
twice as great as the apparent resistance, and the received
current, 20 milliamperes, is little more than one third of the
current entering the wire at A. The electromotive force
required to drive this current is 770.5 X.056, or about 43 volts.
If the same value of the electromotive force is used in all kinds
of weather without regard to the state of line leakage, the current
in the receiver will rise from 20 during the worst leakage to 35
when there is no leakage, and the current entering at A will
fall from 56 to 35 as the leakage gradually disappears.
The more complicated circuits shown in Fig. 23 will not be
considered further than to say that one transmitter is capable
of supplying current to several such lines without undue sparking
at the brushes under the most severe conditions of leakage. In
practice an electromotive force would be chosen great enough
to operate the longest circuit, and resistance might be inserted
in the shorter circuits so that the same electromotive force may
be used for all. If the transmitter becomes loaded to its full

OPERATION OF SYNCHRONOUS MOTORS.
51
capacity, two or more of them may be geared together if required,
and part of the wires supplied by each transmitter.
Instead of employing a transmitter of the type shown in
Fig. 22, in which the two batteries alternately supply the
line current, an alternating-current generator giving a sinu-
soidal wave might be employed; but when speeds as low as
four revolutions of the trailer per second, that is, an alternating-
current frequency of n=4, are employed, which is about the
speed desired for operating printing-telegraphs, it requires a
special generator to obtain sufficient electromotive force to
operate the line. Since the polar-relay armature remains upon
one side until the current is reversed, it seems better to main-
tain the current at a large value up to the time of reversal and
then reverse it suddenly, as is the case with the battery cur-
rent, rather than to reduce the current gradually to zero. There
is then no chance for interfering induction currents on the line to
make any variation in the regular periodicity required in the
times at which the reversals take place. In other respects and
where a higher frequency of reversals is employed an alter-
nating-current generator would be preferable to the form shown.
in Fig. 22.
In using a polar relay in the manner described, there is
always a current in one direction or the other through its coils
to hold the armature firmly against its contacts until the instant
of reversal. The reversals occur with regular periodicity, and
the duration of the current may be long compared with the
time occupied by the reversal. The frequency of reversals may
be anything which seems preferable, and this is determined by
the particular kind of telegraph system employed, a subject
considered elsewhere.
4
Due to the long distances over which the waves travel, the
reversals of the current in the various relays connected with
the system do not occur simultaneously. They occur later in
Philadelphia than in New York, and later in New York than in
Albany, if the circuits are as represented in Fig. 23. It will
appear that this is not of importance, as the chief considera-
tion is the speed, irrespective of the phase difference, that is,
the time interval between consecutive reversals received in
Chicago must be precisely the same as in Philadelphia and at
all other points.
In operating a telegraph system, however, attention must

52
MULTIPLE TELEGRAPHS.
be given to this time lag in transmitting from one point to
another, and it is shown elsewhere that such a system of rotors
once obtained enables other wires between any two of these
stations, by whatever route, to be used for transmitting several
messages at once, either by a printing system or by the Morse
system. For example, the circuits illustrated in Fig. 23 enable
Pittsburgh and Chicago to use a wire by the same or by a different
route from that shown, to send several printed messages either
from Chicago to Pittsburgh or from Pittsburgh to Chicago. Simi-
larly New York and Boston may use wires running directly
between these cities for multiplex messages, employing the
synchronism obtained by the system of Fig. 23.
METHOD OF CONTROLLING SYNCHRONOUS MOTORS BY RELAY
CONTACTS.
A method of connecting a motor to the local contacts of
the polar relay is shown in Fig. 25, in which PR represents the
B
wwwww
титри
D
M
A
PR
S
↑
D. C. Mains
E
T
Gle
C₁ C2
FIG. 25.-Method of controlling synchronous motors by relay contacts.
=
relay, connected in the main line, having a tongue, T, adjusted
so that when there is no current in the relay it remains in con-
nection with either one of the two contacts, C1 or C2, against
which it is placed. The synchronous motor, M, is connected
between the middle point, D, of the resistance BE and the
tongue, T; and a source of constant electromotive force is
applied at the points A and E to the resistances AB and BE in

OPERATION OF SYNCHRONOUS MOTORS.
53
series.
The resistances BD and ED are about 400 ohms each,
and AB serves as a protective resistance, and also may be used
to regulate the current if the apparatus should be employed
with some different electromotive force. The non-inductive
resistance shunt, S, on the motor prevents sparking at the
relay contacts, and in practice no sparking whatever can be seen
at these points when the shunt is used.
When the tongue of the relay is against the contact C₁ a
current passes through the motor, M, in the positive direction,
say, from D to T, by the path DTC1B; but when it is against*
the other contact, C2, a current passes through the motor in
the opposite direction, from T to D, by the path TDEC2, thus
reversing the current in the motor at each passage of the tongue
from one side to the other. These reversals of the motor current
are therefore in synchronism with the main-line impulses which
actuate the relay.
Fig. 26 shows the general appearance of one form of these

FIG. 26.-Special form of two-pole synchronous motor adapted for
synchronous telegraphs.
motors as mounted in its case and adapted for use with syn-
chronous telegraphs. It is a magneto-machine with a permanent
magnet field and a two-pole armature, which when started

54
MULTIPLE TELEGRAPHS.
runs in synchronism with the trailing brushes illustrated in
Fig. 22. The motor is not self-starting, but is easily started
by hand, the handle for this purpose being seen at the top of
the instrument. By pressing a lever at the left the handle
becomes geared to the shaft, and is thrown out of gear after
the motor is started by releasing this lever. The operation of
starting occupies but a few seconds, and the sense of feeling
tells when synchronism is obtained, as the armature begins to
drive itself.
The resistance of the motor winding is about 160 ohms,
and it requires a current of 120 milliamperes when running
at the ordinary speed. If the motor is stopped, the current
is higher, about 160, the exact value when running depending
upon the speed.
The only mechanical work required of these motors, besides
overcoming the friction of their bearings and the air resistance,
is to drag the trailing brushes so as to make good contact with
their sunflowers; and the chief difficulty in operating them is
the well-known effect of "hunting," which is experienced by
most synchronous motors. The principle described many years
ago of using a viscous medium to damp out these oscillations is
used in this motor. At the lower end of the shaft which carries
the trailing brushes is a cast-iron fly-wheel weighing about three
pounds, and shown in cross-section in Fig. 27. The outside
diameter is 5 inches, and the hollow portion, A, contains 1
pounds of mercury, which circulates freely around the wheel.
A
FIG. 27.-Fly-wheel and mercury cup for damping oscillations of the
motor shown in Fig. 26.
If the fly-wheel starts from rest and finally attains a per-
fectly uniform angular velocity, the mercury it contains does
not accelerate its velocity as rapidly as the iron wheel, and at
first appears to move backward; but because of the friction it
is accelerated until it revolves at the same speed as the wheel.
The mercury finally comes to rest relatively to the wheel, but

OPERATION OF SYNCHRONOUS MOTORS.
55
its surface is deformed, due to the action of the centrifugal force
and gravitation together. There is a certain amount of work
expended upon the mercury by the wheel which is not stored in
the form of kinetic energy of motion, but is dissipated in the
form of heat; and it is this energy that makes the mercury
cup of value in damping the oscillations. An iron fly-wheel
alone opposes any change in the angular velocity of a motor;
but the energy stored up by the fly-wheel is given back again,
and does not prevent the motor from oscillating out of synchron-
ism, while that expended upon the mercury and dissipated as
heat, not being returned again, uses up the energy of the motor
oscillations and thus prevents them from attaining a magnitude
great enough to be harmful.

CHAPTER VII.
THE OPERATION OF SYNCHRONOUS MOTORS ON WIRES
WHICH ARE ALSO USED FOR TELEGRAPHS.
THERE are several good methods for the operation of syn-
chronous motors on wires employed for sending telegraph
messages at the same time which deserve some description.
Three of these plans will be considered, the first being one in
which the wire is used as a single Morse wire, the second one
in which the wire is used for duplex Morse, and the third one
in which it is used for the synchronous multiplex.
SINGLE MORSE WIRE.
The principle upon which the first plan depends is to make
use of one side of the duplex-diplex system (see Part I) for the
synchronism, and the other side for the operation of the telegraph.
Fig. 28 shows such a line with several stations, A, B, D, and E,
on the direct-current side, and with two synchronous motors,
one at each terminal station, A and E, operated by a trans-
mitter at the station C, which is purposely located near the
middle of the line. The motor and local circuits at stations A
and E are supposed to be exactly alike, as described in Figs. 25
and 26, but in place of the tongue and contacts of the polar
relay in Fig. 25 is substituted the lever, T, and contacts, c1
and C2, of the repeating-instruments, J. These instruments are
controlled by the transmitter K at station C, which is operated
by a continuously rotating trailer, H. The trailer is in a local
circuit, and serves to make and break the circuit of the magnet,
M, of the transmitter at regular intervals. The distant relays,
PRA and PRE, respond to the current from the generator, G,
which is a 150-cycle alternating-current generator. At the
rate at which the motors are to be operated, say four revolu-
tions per second, each closure of the transmitter K sends about
56

OPERATION OF SYNCHRONOUS MOTORS.
57
FIG. 28.-Diagram showing method of operating synchronous motors on a single Morse wire.
-OF
nineteen complete waves to line, and each open period omits
sending waves for the same interval of time. In effect the relays
respond as quickly as though they were operated by the direct
current.
Addolololo
H
GO
PRE
wwwwwwww
四
​Z-
S
HM
F
K
D
Ο
JA
OF
PRA
wwwwwwww
N
B
Y
16
By placing the generator in the middle of the line, as shown,
the synchronism can be obtained on a longer line than by placing
the generator at either end.
It is obvious that in a similar manner the direct-current side
of the system may be employed for operating the motors, and
the alternating-current used for transmitting messages.

58
MULTIPLE TELEGRAPHS.
HIGH-FREQUENCY SYNCHRONOUS MOTOR.
Before giving the method of obtaining synchronism as
applied to a duplex wire a description of a high-frequency
synchronous motor is introduced here as the synchronism
on the duplex circuit is shown with such a motor. Heretofore
the trailers were supposed to revolve at a comparatively slow
speed, which is especially adapted to printing-telegraphs; but
this is not necessary if the synchronous circuits are operated
as Morse circuits. In the motor the general appearance of which
is shown in Fig. 29 the trailer is on the same shaft as the arma-
6
$5
FIG. 29.-Special form of high-frequency synchronous motor adapted
for synchronous telegraphs.
ture, which revolves at a very much higher speed than that
of the motor previously referred to, a working speed for this
motor being 1,125 revolutions per minute, as compared with
240 revolutions per minute in one special case of the other
motor.
The construction of this motor is given in detail in the plan,
Fig. 30, and elevation, Fig. 31. It is provided with a hand

OPERATION OF SYNCHRONOUS MOTORS.
59
FIG. 30.-Plan of motor shown in Fig. 29.
starting-device similar to the motor of Fig. 26, which drives
the gear A by the crank B, the gear engaging the pinion C on
the armature shaft D only when the lever E is pressed against
a stop, F. A spring, G, releases the gear from the pinion when
O
O
O
r
K
Ꮎ
ㄨ​ˋ
וד.
Эдо
w
C
K
ㄨ​ˊ
www
D
D
Φ
O
the lever, E, is released. The ratio of the gear to pinion is 120
to 10, and as the normal speed of the armature is 1125, the
starting-handle makes only 94 revolutions per minute, or about
1 per second.
The motor has 16 poles, and operates at a frequency of

60
MULTIPLE TELEGRAPHS.
150 cycles per second. There is no wire wound upon the
rotor, H, mounted upon the shaft, D, which carries in addition
the mercury cup, I, and trailing brush, J. There are, there-
fore, no brushes of any kind required in the operation of the
K
B
FE типий
C.
R
1:03
G
A
10=>
M
H
T
D
So
7808
J
L
บ
On
FIG. 31.-Elevation of motor shown in Fig. 29.
R
motor, so that practically all of the mechanical work which
the motor is called upon to do is to drive the trailers. This
motor is of peculiar construction, being in effect a magneto-
machine in which the permanent magnets revolve. The rotor,
H, consists of two interlocking spider-arms, L and L', each
having eight poles and being permanently magnetized by the

OPERATION OF SYNCHRONOUS MOTORS.
61
magnet, K, although they revolve at a high speed. There are
eight electromagnets, R, with sixteen poles, as shown, corre-
sponding to the sixteen poles of the spider-arms, L and L'.
The magnetic circuit may be traced from the pole M through
a narrow air-gap to the upper spider, L, which is separated by
a brass centre-piece, P, from the other spider, L'; and thence
through a second air-gap to one side of the eight electromagnets,
R, through the magnets and through a third air-gap the
lower spider, L', passing across a fourth air-gap to the opposite
pole, T, of the permanent magnets, K. When the rotor, H,
revolves, the magnetism in the magnets, R, is evidently reversed
when it moves from one pole to the next; and if driven mechan-
ically, an alternating electromotive force is generated in the
coils of the magnets, R; and conversely, if a periodically reversed
current is sent through these coils, the rotor will revolve if it
is started so as to run at the proper speed.
The mercury cup, I, is 24 inches in diameter, made of wood,
with a row of tacks, U, inserted to increase the friction.
DUPLEX MORSE WIRE.
A wire which is at present equipped with standard polar
duplex Morse instruments may be used at the same time for
obtaining synchronism by a method, shown in Fig. 32, which
operates well without disturbing the Morse working or being
disturbed by the Morse working. The generator, G, is here
shown in the middle of the wire, and there are two synchronous
motors, M1 and M2, one at each terminal station. A polarized
relay, PR1, controls the synchronous motor, M1, in the local
circuit operated through the tongue, T1, and contacts, C1
and C2.
A constant-potential power circuit is connected through a
protective resistance, R1, to the two contacts, C1 and C2, of the
polar relay, and also to two condensers, A and B. The motor
last described, see Figs. 29, 30, and 31, is connected from the
middle point between the condensers A and B to the tongue,
T1, of the relay, and receives an alternating current as the
tongue vibrates between the contacts C1 and C2, from the dis-
charge of first one condenser, A, and then the other, B. The
non-inductive resistances R₂ prevent sparking at the contacts
C1 and C2.

62
MULTIPLE TELEGRAPHS.
The main-line receiving-circuits are arranged so that the
coils, E and F, of the alternating-current relay, in series with
balanced condensers C' and C", from shunts to the two sides
of the direct-current polar relay, PR2. This arrangement is
very free from any interfering disturbance, because the direct
current is prevented from entering the relay PR₁ by the con-
densers, and the impulses which do get through are balanced
414-
H4040400
C"
L
C3
PR2
PR₁
G
SES
Line
A
R₂
R₂
B
M,
R
Artificial Line
G
---
FIG. 32. Diagram showing method of operating synchronous motors on
a duplex Morse wire.
SYNCHRONOUS MULTIPLEX WIRE.
in the two windings. The extra condenser and coil, C3 and L,
are connected from apex to ground as an additional precaution,
so as to prevent the direct-current impulses from being so
sudden at the moment of reversal of the pole-changing trans-
mitter.
One method will now be given by which motors may be
operated on the same wire which is used for synchronous tele-

OPERATION OF SYNCHRONOUS MOTORS.
63
R
graphs.* For this purpose the motor shown in Fig. 26 and
operated by the contacts of a polar relay, as in Fig. 25, is
taken as an example. Suppose that with the arrangement in
Fig. 22 the receiving-circuit does not pass directly through
the relay, but is first connected to the trailer driven by the
same synchronous motor which it is intended to operate,
as shown in Fig. 33. If this trailer, T2, makes contact with
A
VH
T2
It:
J'
U'
10
SIF
115
S
KU
O'
M.
B
K HVI
M2
P
unc
GP
PR
FIG. 33.-Diagram showing method of operating synchronous motors on
a synchronous multiplex wire.
a continuous ring, M2, which is connected to ground through
the polar relay, PR, similar to that in Fig. 22, and for the
purpose of operating the motor carrying the trailer, T2, then
the relay must operate in precisely the same manner as if the
line were connected to the relay itself, as in Fig. 22. The
trailer, T2, when the motor is running, will, under these con-
* The method of running a synchronous motor on the same wire with
synchronous telegraphs is intimately connected with these telegraphs,
which forms the subject of Part III, and some of the points referred to
will seem more clear if read after Part III.

64
MULTIPLE TELEGRAPHS.
ditions, always be situated at one of two diametrically opposite
points, O or O', at the instant when the current reverses in
the relay, except for a very slight permissible oscillation of
the receiving-motor backward and forward about these two
points, which is because the synchronism can never be perfect.
Although the current in the relay is, say, in a positive direction
when the trailer, T2, is on the half of the ring ORO', and in
the negative direction on the half O'PO, yet the relay, being
adjusted so as to remain on whichever side its armature is
left, would behave in precisely the same manner if it could
receive the current just long enough before and after the instant
of reversal to cause the tongue to move from one contact to
the other at the same time that it would if receiving current
all the time.
If portions of the ring, HK' and H'K, were cut away so
as to leave segments, HK and H'K', sufficiently long to include
a short space before and after the imaginary line OO' so as to
permit of some oscillations of this imaginary line, as before,
without ever causing it to pass beyond the limits of the seg-
ments HK and H'K', then the operation of the motor will
remain as before. The transmitting ring, M1, which is really
at the other end of the telegraph wire, is here represented within
the ring M2, to show the relative positions of the transmitting-
and receiving-rings. In the transmitting-ring the openings F
and G correspond to the openings in Fig. 22, the half A trans-
mitting a positive and the half B a negative current, the instants
of reversal occurring at the points F and G at the transmitting
end and at the points O and O' at the receiving end a little
later, because of the time required for the wave to travel from
the transmitter to the receiver.
In a similar manner, at the transmitting end it is not neces-
sary that the current should be continuously transmitted
during the whole of the halves, A and B, in order to operate
the relay in exactly the same manner. If a short segment,
FJ, preceding the point F, transmitted a negative current,
and another segment, FS, following F, transmitted a positive
current, the instant of reversal of the relay PR would occur
at the same time as before.
In this manner the time during which the trailer is passing
over the remaining portions of the rings, not included by these
short segments, may be utilized for operating synchronous

OPERATION OF SYNCHRONOUS MOTORS.
65
telegraphs, for which purpose the portion of the rings outside
of these segments is divided up into other segments connected
to the various instruments in any desired manner.
The whole of the receiving-ring, M2, outside of the seg-
ments HK and H'K', is not, however, available, for it is evident
that the negative current transmitted by the segment JF
may reach the receiver before the trailer has passed the point
K. It is necessary, therefore, to provide a short segment, KU,
connected to ground, so that the current designed to operate
the motor shall not disturb the currents operating the syn-
chronous telegraphs. The segment HK cannot be extended
up to the point U because the relay might then receive an
interfering current from the operation of the synchronous
telegraphs, which would move its armature over before the
trailer arrived at O, and thus eventually stop the motor. In
a similar manner the segment FS at the transmitting end,
following the point of reversal, must be long enough to make
it certain that the relay PR receives a positive current as long
as the trailer remains upon the segment HK, after passing
the point 0; and to make sure of this the point S is situated
later than H, corresponding to some point V on the receiving-
ring, and a short segment, HV, is connected to ground. These
two short grounded segments, HV and KU, on either side of
the motor segment, HK, thus protect the motor from all out-
side disturbances, and the portions of the receiving-ring VRU'
and V'PU may be divided into segments for operating syn-
chronous telegraphs.
In using this arrangement it is evident that the motor can-
not be started as readily as if there were no trailer at the
receiving end, as in Fig. 22, for to start the motor there must
first be a reception of periodic reversals by the relay, and these
reversals are interfered with when the motor runs at any speed
different from that intended. This offers no particular difficulty
in practice, for, since the synchronous telegraphs cannot be
operated while the motor is not running, a switch may be em-
ployed to connect the receiving-relay substantially as in Fig. 22
until the motor is started, and to return it to the working
position after the motor is running.


PART III.
SYNCHRONOUS TELEGRAPHS.
T


INTRODUCTION.
HAVING described some of the means for obtaining syn-
chronism, we shall next consider, as an independent problem,
the subject of how to use synchronism for operating telegraphs.
There are certain general properties or characteristics com-
mon to most if not all synchronous systems, which should be
understood before undertaking the study of any particular
system. For this reason, and because synchronous telegraphs
are so little used in America, it has seemed desirable to treat
this part of the subject more fully than would otherwise be
required. That several independent messages may be sent over
the same wire at the same time by means of the synchronous
revolution of two rotors, one at each station, is quite generally
understood; but the difficulties which exist, and the limitations
to which such systems are subject, are not so generally appre-
ciated.
The line wire is an important factor in determining the
limit of message-carrying capacity with any system, and, in
general, long lines are harder to work than short ones, the
message capacity being smaller the longer the line. There
are many short lines, say less than 100 miles in length, which
must be operated, and it is important that the same system
used on the short lines should also be adapted to long lines.
This is so important from a practical standpoint that the sub-
ject is here considered with particular reference to the longer
lines.
The common Morse telegraph employs the line wire to
convey the electrical waves or impulses from the transmitting-
key to each receiver, so that every receiver on the line receives
the impulses of all the messages transmitted. The speed at
which an operator can manipulate a key is so slow that the
69

70
MULTIPLE TELEGRAPHS.
electrical waves on the wire with this system are comparatively
long. To increase the speed of transmission many different
systems have been devised, on the principle of operating the
Morse key by a machine which can work faster than any human.
operator. Some of these systems give a printed record, and
others receive the message in dots and dashes on paper tape;
but the principle of sending all of the messages that go over
the wire through each receiver is common to them all, and
distinguishes them from synchronous systems.
In synchronous telegraphs that portion of the waves which
forms one message is received in one receiver, and a second
portion forming a second message in a second receiver, a third
portion in a third receiver, and so on. Because of this fact
the synchronous system, from some points of view, has several
conspicuous advantages over the others mentioned.
Suppose, for example, that some system of those referred to
is required to transmit 200 words per minute over the wire.
From the nature of the case it is evident that no human operator
can manipulate a keyboard at this speed, and the messages
must therefore be prepared beforehand in some way, say per-
forated on a tape. By the synchronous system the 200 words
per minute are transmitted over the wire by dividing up the
messages among, say, four operators, each being required to
operate at only 50 words, a speed well within his ability to
operate a keyboard, so there is no necessity to resort to machine
transmission with its attendant delays.
It is as if each operator has control of an independent wire;
and this analogy holds true in another important respect, for
the different receivers in the synchronous system may be located
at different points and the messages designed for those places
alone sent there, whereas in the other systems the messages de-
signed for other places all have to pass through each receiver.

CHAPTER VIII.
MEASUREMENTS IN SYNCHRONOUS TELEGRAPH CIRCUITS.
To give a clear conception of the operation of synchronous
telegraphs some measurements have been taken, using actual
apparatus, to enable the reader to approach the subject from
a practical standpoint, and to perceive where the difficulties
lie, and learn what may be expected when using special arrange-
ments other than those considered. These measurements may
easily be repeated by those having the instruments, the only
special pieces of apparatus required being the two "sunflowers."
The principal observations consist in making readings of
the current in the various circuits leading from the segments of
the commutator under varying conditions. Absolute synchron-
ism is obtained by mounting the trailers of the two sunflowers at
the opposite ends of the line upon the same shaft. An artificial
line was used for convenience, though a real wire connected
as a loop might have been employed. It is understood that
measurements made on an actual line will differ slightly from
those made with the artificial wire, but the approximation is so
close that the measurements on the artificial line illustrate the
points which it is intended to show as well as if they were made
on an actual line.
THE INSTRUMENTS REQUIRED.
The apparatus required is illustrated in the diagram Fig. 34,
which shows an artificial line made up of five equal resistances,
R₁ to R5, of 400 ohms each, and four condensers, C1, C2, C3,
and C4, of approximately 1.48 microfarads each; two commu-
tators or "sunflowers," mounted in such a way that the same
shaft, S, carries the trailing brushes for each; four transmitters,
KA, KB, KC, and KD, of the continuity or circuit-preserving
type, one connected to each of four wires, A1, B1, C1, and D1,
71

72
MULTIPLE TELEGRAPHS.
leading from sunflower M₁; four polarized relays, A2, B2, C2,
and D2, connected to the corresponding segments of the sun-
flower M2; a direct-current generator, G; and four protective
resistances, rA, TB, YC, and rD.
The sunflowers illustrated are eight-segment commutators,
having diametrically opposite segments connected together, so
that in effect they are four-segment commutators with four
wires, A, B, C, and D, leading from each.
With this arrange-
ment the trailer makes only half as many revolutions per minute.
TA
13
2
Го
3
G
KA
KB
Kc
Ko
E
=2
A,
B₁
C
D
R
S
U
V2
U2
T₂
T
M,
M₂
R₁
R2 R R R
mahahala
C
AGTE
Gr
Look
B2
D₂
DE
E₂
ليات
E
FIG. 34.-Diagram of experimental apparatus for studying the currents in
the various circuits of a synchronous system.
to pass over a given number of segments per minute as it would
if there were four segments in the revolution, and 1125 revolu-
tions per minute makes the trailers pass over 150 segments
per second, while if they were four-segment commutators it
would require 2250 revolutions to give the same number of
segments per second. The chief reason for using the eight
instead of the four-segment sunflower is one of mechanical
convenience, as the lower speed is preferable. The rate at
which the trailers pass over the segments is one of the most
important factors to be taken into account in synchronous
telegraphs, and it does not matter whether we use a four-, eight-

SYNCHRONOUS TELEGRAPH CIRCUITS.
73
or sixteen-segment sunflower to obtain the required frequency
of segments per second.
The artificial line is made up to represent a copper wire
of 5.14 ohms per mile, having a distributed capacity of about
1.50 microfarads per hundred miles. Being made of five
sections of 400 ohms each, the total resistance is 2,000 ohms,
and corresponds to a length of 2000/5.14 about 390 miles
of wire. This length requires 390X1.50=5.85 microfarads
total, or 5.85/4=1.46 microfarads per section, when divided
into four sections. One end of this line is connected to the
trailer T1, bearing on the commutator M1, making connection
with it through the continuous ring P1 and brush U₁; and
the other end is connected to trailer T2 of the commutator M2
in a similar manner. The four terminals, E, of the condensers
must be connected together and to the ground- or return-wire,
E2. It is important that this connection should be correctly
made, for if it is neglected the artificial wire will not give
results even approximately close to the real wire.
The four polarized relays are connected with one common
terminal to ground at E2, and one relay is connected to each
diametrically opposite pair of segments of the sunflower M2,
the trailer passing in succession from the segment A2 to B2,
and from B2 to C2, and so on until it completes a whole cycle
after one-half a revolution.
The four continuity-transmitters, KA, KB, KC, and KD,
are connected to the four wires A1, B1, C1, and D1, in such
a manner that the latter are normally connected to ground at
E₁ without resistance. When, however, one of the trans-
mitters is closed the circuit from the segment containing that
particular transmitter is altered, and passes through the pro-
tective resistance, r, and generator, G, to ground. These
resistances, rA, TB, TC, and rp, are 200 ohms each.
THE METHOD OF MEASURING CURRENT.
Before any measurements are taken some consideration
should be given to the method of measuring currents; for it
must be remembered that it is desired to measure the currents
which circulate in the various circuits while the trailers are
revolving, and these currents have a pulsating character, flow-
ing in each of the polar relays, for example, twice every revolu-
tion of the trailer; each relay remaining on open circuit three-

74
MULTIPLE TELEGRAPHS.
quarters of the time, and closed circuit one-quarter of the
time.
In practice we can measure and compare these currents
with sufficient accuracy by the use of a direct-current milam-
meter. In the measurements which follow, a Weston milli-
voltmeter of the permanent-magnet direct current-type was
employed, as it is more sensitive than a milammeter. One
division deflection of the millivoltmeter is equivalent to .625
milliampere, and the millivoltmeter readings may be expressed
in milliamperes by multiplying by this factor.
To assist in understanding the readings obtained the follow-
ing experiment was made, using the sunflower M₁ alone. The
line in Fig. 34 was disconnected from the trailer T₁, and a
non-inductive resistance of 2,000 ohms, the millivoltmeter
above referred to, and an ammeter of the double-coil type,
suitable for measuring either direct or alternating currents,
were connected instead, in series from T₁ to E1.
Before the trailer was started in rotation one of the trans-
mitters, Kв for example, was closed and the trailer moved
around until it rested upon the segment B1, when the deflection
of the pointer of the alternating-current instrument indicated
59 milliamperes and the direct-current millivoltmeter indicated.
97.5 divisions for the same current. A standard instrument
with which each of the meters was compared indicated 60.9
milliamperes as the true current which produced the above
readings, so that the readings of the millivoltmeter must be
multiplied by a factor .625 to convert them into milliamperes.
While one transmitter, KB, only remained closed the trailer
was started rotating, and it was observed that the pointers
of the ammeters pulsated slowly at first, and more and more
rapidly as the speed was increased, until at a speed of about
60 segments per second (450 revolutions per minute) the vibra-
tion of the pointers of the instruments was not more than about
one division of the scale, this being steady enough to admit
of fairly accurate readings. When the speed was increased,
however, the deflection did not change, but became more
free from vibration, showing that the deflection was independent
of the speed, within certain limits. The millivoltmeter pointer,
at a speed of 61.6 segments per second, oscillated between
the 24th and the 25th divisions of the scale, and the alternating-
current instrument indicated 29.5. Taking 24.5 as the reading

SYNCHRONOUS TELEGRAPH CIRCUITS.
75
of the millivoltmeter, it is evident that this value is almost ex-
actly one-quarter of the reading 97.5, when the trailer was not
running, one-quarter of this being 24.4. On the other hand,
the alternating-current instrument indicated exactly one-half,
29.5, of the indication, 59, when not running, though the
currents in the two instruments were identical.
The explanation of this is that the alternating-current in-
strument, being a double-coil variable-field instrument, indi-
cated the square root of the mean square value of the current,
whereas the millivoltmeter, being a single-coil permanent-
magnet-field instrument, indicated the average value of the
current.
Assuming that the current attains its full value, I, imme-
diately after the trailer enters the segment B1, and remains
at this value while the trailer is on this segment, falling imme-
diately to zero when it enters the next segment, C1, and remain-
ing so until it arrives again at B1, it is evident that the average
value of the current taken throughout the four segments-that
is, for a complete cycle, which is the same as the average for an
indefinitely long time-is just one-quarter of the current when
the trailer is not running. The millivoltmeter indicated almost
exactly one-quarter of the full deflection, which shows that the
supposition is approximately correct.
The reading of the alternating-current instrument is explained
by taking the square of the current throughout the four seg-
ments, averaging the square of the current, and then taking
the square root of the result.
root of the result. The current in the first segment.
is I, and its square 12, and the square of the current in the next.
three segments is zero. The average of the squares over the
four segments is therefore 12/4, and the reading of the alternating-
current instrument should be V12/4, or I/2, which is the observed
value.
It might be supposed that the alternating-current instru-
ment would be the more desirable of the two instruments to
employ for these measurements because the same actual cur-
rent indicated double the number of milliamperes in the alter-
nating-current instrument, which would therefore be more
sensitive. This would be true were it not for the fact that most
of the readings which are of importance are very small, and
the millivoltmeter shows them very distinctly because its
deflections are proportional to the current, whereas no deflec-

76
MULTIPLE TELEGRAPHS.
tion at all is noticeable on the alternating instrument, as its
scale divisions are very much cramped near the zero value.
When the speed of the trailer was increased to 150 segments
per second the deflections of the instruments in the same situa-
tion as above described began to show a slight falling off, the
millivoltmeter indicating 23 instead of 24.5. This is probably
due to the fact that there was some inductance in the circuit,
viz., the millivoltmeter coil and the dynamo-generator, which
begins to have an effect at the higher speeds. It was also
noticed that the mechanical pressure of the brushes upon the
commutator segments made some slight difference in the read-
ings at these speeds, indicating that another cause for a de-
crease in the reading may be a slight abridgment of the con-
tact which the trailer makes with the segment.
THE MEASUREMENTS.
The observations are arranged for convenience in tabular
form, and consist of twenty series of readings made under
different conditions, the number of the series being given in
the first column of the table (see pp. 78-83). The apparatus
used was as shown in Fig. 34, and the first series was taken
with no leakage on the line; with 130 volts measured in the
direct-current generator, G; four polar relays having 50 ohms
and 4 henry each; and with the transmitter KB, only, closed.
The trailers revolved at a speed of 462 revolutions per minute,
causing them to pass over 61.6 segments per second. Before
the trailers were started the commutators were adjusted around
the shaft, S, so that trailer T1 entered segment A1 at the
same moment that trailer T2 entered segment A2, and the
readings of the scales, V1 and V2, were noted. The millivolt-
meter was inserted in turn in the circuits A1, B1, C1, D1, A2,
B2, C2, and D2, and it was noted that in the four receiving-
circuits the largest reading occurred in the circuit B2, corres-
ponding to the transmitting-circuit B1, which contained the
closed transmitter. The following segment, C2, however, showed
a reading nearly half as great, while the two remaining seg-
ments showed almost nothing.
The sunflower M2 was next adjusted around the shaft, S,
in the same direction that the trailer revolved, without moving
the sunflower M1, and with the millivoltmeter in circuit C2 it
was observed that as the sunflower was moved slightly the

SYNCHRONOUS TELEGRAPH CIRCUITS.
77
reading decreased, and reduced to zero if it were moved far
enough. At the same time the reading in B2 increased and
passed through a maximum value, and a current began to
appear in A2, increasing as that in C2 diminished. The series
of observations No. I was taken after moving the sunflower M2
until the readings in circuits A2 and C2 were equal to each
other, when B2 showed a maximum of 21.0 divisions, and A2
preceding and C2 following each gave 1.2 divisions. The fourth
circuit, D2, showed only a trace, about o.1 division.
It is important to observe that with the sunflowers set as in
series No. 1 a position has been found where practically all of
the current transmitted by the transmitter KB that reaches
the receiving end of the line passes to ground via the segment
B2, and does not therefore interfere with instruments which are
connected to the adjacent segments.
When these readings were taken the setting of the commu-
tator M2 showed that it had been moved .29 of a segment
forward, so that the trailer T2 arrived at the beginning of
segment A2 some time after trailer T₁ had arrived at the begin-
ning of segment A1. The amount of lag was in this way measured
with a fair degree of accuracy; for one segment corresponds to
1/61.6, or .01625 second, and .29 segment to .0047 second.
The readings reduced to milliamperes are given in columns
A and B of the table, column A being obtained from the milli-
voltmeter readings by multiplying by the factor .625, and
denoting the average value of the current in milliamperes
taken through a complete cycle of four segments. Column B
is obtained from the millivoltmeter readings by multiplying by
the factor .625 X4 or 2.5, and denotes the average value of
the current in any particular segment, in milliamperes, while
the trailer is passing over that segment. It is evident, there-
fore, that the average current through the relay in B2 during
the time the trailer was upon the segment B2 was 52.5 milli-
amperes, whereas the current in each of the adjacent segments
was only about 3.0. It is this interfering current in the adjacent
segments, however, which finally limits the length of line it
is practicable to operate; and one of the objects in making
these measurements is to study the causes which tend to affect
this interference.
The readings taken at the transmitting end of the wire
are also important, as they show that there is no position of

78
MULTIPLE TELEGRAPHS.
TABLE I.
NOTE.-Current in milliamperes. The figures in Column A are obtained from the millivoltmeter reading by multiplying by the
factor 0.625, and denote the average value in milliamperes of the current flowing in the corresponding segment during a complete cycle,
revolution of the trailer. The figures in Column B are obtained from the millivoltmeter reading by multiplying by the factor 0.625 X
4=2.5, and denote the average value in milliamperes of the current flowing in the corresponding segment during the passage of the
trailer over the segment.
Leak- Speed,
Number age,
Seg-
of Ohms ments
Series.
per
per
Mile. Second.
I
2
8
8
61.6
150
Lag of
Commu-
tator M2
behind
M1, Frac-
tion of
Segment.
.29
.50
Circuits
from
Seg-
ments.
HAURRUS RUDRA
A₁
C₁
D,
A2
A₁
A2
Milli-
volt-
meter
Read-
ings.
Current in Milli-
amperes.
-
0.0
0.0
0.0
O. I
+28.0 +17.5 +70.0
5.4 3.4
-13.5
O. I 0.3
+1.2 + 0.8 + 3.0
+21.0 +13.1+52.5
+1.2 + 0.8 + 3.0
+ O.I + O.I
+0.3
-
A.
B.
O. I
O.I
-12.0
- 0.3
+32.3 +20.2 +80.7
7.5 30.0
0.5 0.3 1.3
+ 2.0 +1.3+ 5.0
+16.3 +10.2 +40.7
+ 2.0+ 1.3+ 5.0
+0.3 + 0.2 + 0.8
――
Circuits Leading from Commu-
tator M1.
Contain Trans- Contain Receivers
mitters.
Closed through
Open.
Closed,
with
E.M.F.,
Volts.
O
130
O
129.5
O
Circuits Leading from Commu-
tator M2.
Contain
Trans-
mitters
Closed,
with
Ohms. Henrys. E.M.F.,
Volts.
Contain Receivers.
Open.
Closed through
Ohms. Henrys.
50
ཤྩ
50
엉엉​엉엉​:
50
.4
.4
.4
.4
.4
.4
.4
4

SYNCHRONOUS TELEGRAPH CIRCUITS.
79
3
4
Сл
5
6
8
8
8
8
61.6
61.6 0.0
150
.29
61.6
.50
.29
HRUDHRSD RUTRUS
A₁
D2
HAURRUR ARRA
.0
0.0
0.0
+13.0 + 8.1 +32.5
+ 9.0+ 5.6 +22.5
0.2
O.I
+ O.I + O.I + 0.3
0.5
+28.0 +17.5 +70.0
6.0 3.8
-15.0
0.8
0.3
0.2
-
-
+ O. I +o. I
+ 0.3
+16.0 + 10.0 +40.0
+
+ 7.1 4.3 +17-7
+ 0.2 + 0.1 0.5
+29.3 +18.3 +73.2
6.3
3.9-15.8
0.2
O.I
-
-
-
-
111
..
-
-
0.0
0.0
0.0
+ 2.1 + 1.3
+5-3
+15.8 9.9 +39.5
+ I.I + 0.7 + 2.8
0.4
0.3
I.O
+32.5+20.3 +81.3
-13.2 8.3 -33.0
0.8
0.5
2.0
-
O.I
O.I
-
-
-
0.5
0.3
-
-
+30.0 +18.8 +75.0
+1.3 +0.8 +3.3
+23.0+14.4+57.5
+ 5.8 +3.6 +14.5
+ 0.6+0.4 + 1.5
Open
Open
Open
138
50
9,9,8,9,
50
50
50
%%%%
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
128
OO
135
0 0 0
O
129.5
006
50
50
50
.4
.4
.4
..4

80
MULTIPLE TELEGRAPHS.
Number
of
Series.
7
8
9
Leak-
age,
Ohms
per
Mile.
8
8
8
Speed,
Seg-
ments
per
Second.
150
61.6
150
Lag of
Commu-
tator M2
behind
M1, Frac-
tion of
Segment.
.50
.29
.50
Circuits
from
Seg-
ments.
HAUDHRA
ABCAABCD
ARUDARSD
Milli-
volt-
meter
Read-
ings.
Current in Milli-
amperes.
—
TABLE I-(Continued).
+32.8 +20.5 +82.0
+ 2.8 + 1.8 + 7.0
+16.8 +10.5 +42.0
+9.0+ 5.6 +22.5
+3.8+ 2.5 + 9.5
-0.3
A.
0.2
O. I
0.5
+22.0 +13.8 +55.0
14.2 - 8.9-35.5
0.3
+ I.Io.7
-
+5.5 3.4
+ 0.3 + 0.2
-
0.2
B.
- 0.2
0.8
2.8
+13.8
+ 0.8
—
- 0.8
+33.0+20.6+82.5
21.313.3-53.3
I. I
1.8
4.5
+1.6 + 1.0 + 4.0
7.4 4.6 +18.5
+0.4 + 0.3 + 1.0
Circuits Leading from Commu-
tator M1.
Contain Trans-
mitters.
Open..
Open
Open
Open
Closed,
with
E.M.F.
Volts.
132
0
138
O
132
O
Contain Receivers
Closed through
Ohms. Henrys.
Circuits Leading from Commu-
tator M2.
Contain
Trans-
mitters
Closed,
with
E.M.F.,
Volts.
Contain Receivers.
Open.
Open
Open
Closed through
Ohms. Henrys.
%%%%
50
585
50
50
50
50
.4
.4
.4
.4
.4
.4
.4
.4
.4

SYNCHRONOUS TELEGRAPH CIRCUITS.
18
IO
II
12
13
8
8
8
8
150
150
150
150
.50
.50
.50
.61
A₁
C₁
「ABCD
C2
A₁
B₁
C₁
D
A₂
ABCD AB
+1.8 1.1 4.5
+16.2 +10.1
I
+40.5
1.2 + 4.8
C2
+1.9
D2 + 0.2 + O.I + 0.5
C2
D2
A₁
AUD
S
+2.6+ 1.6 +6.5
+17.3 +10.8 +43.3
1.4 + 5.8
+ 0.5
D2
+ 2.3
+ 0.2 + O. I
+ 0.8+ 0.5+ 2.0
+12.0+ 7.5 +30.0
+2.6+ 1.6 + 6.5
+ 0.4 0.3 + 1.0
+1.6+ I.O + 4.0
+12.1 7.6 +30.2
+1.7 1.1 4.3
+ 0.2+ O.I
+ 0.5
O
132
O
O
132
O
132
。
132
O
000
50
நம்ழ்ம்
50
200
200
200
200
200
200
200
200
OOOO
.4
.4
.4
.4
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6

82
MULTIPLE TELEGRAPHS.
Number
of
Series.
14
15
16
Leak-
age,
Ohms
per
Speed,
Seg-
ments
per
Mile. Second.
8
150
520,000 150
520,000 150
Lag of
Commu-
tator M2
behind
M1, Frac-
tion of
Segment.
.53
.53
.46
Circuits
from
Seg-
ments.
A₁
CRURIRUR AURUS HARSA
C₁
C2
D 2
A₁
D₂
A₁
C2
D,
Milli-
volt-
meter
Read-
ings.
TABLE I-(Continued).
Current in Milli-
amperes.
A.
B.
+1.7 + I. I
+ 4.3
+15.4+ 9.6 +38.5
+1.9 +1.2 + 4.8
+ O.I + 0.5
+ 0.2
+1.4
+ 0.9 + 3.5
+II.I + 6.9+27.8
+ 0.8 + o.5+ 2.0
+o. I + 0.1 + 0.3
+ 1.0 + 0.6+ 2.5
+II.1+ 6.9+27.8
+ I.I + 0.7+ 2.8
O. I +༠.༢
+ o. I +
Circuits Leading from Commu-
tator M1.
Contain Trans-
mitters.
Open.
Closed,
with
E.M.F.
Volts.
O
130
128
O
128
Contain Receivers
Closed through
Ohms.
Henrys.
5
Circuits Leading from Commu-
tator M2.
Contain
Trans-
mitters
Closed,
with
E.M.F.,
Volts.
Contain Receivers.
Open.
Closed through
Ohms. Henrys.
%%%%
50
50
50
50
50
50
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4

SYNCHRONOUS TELEGRAPH CIRCUITS.
83
17
18
19
20
520,000
520,000 150
8
150
8
150
150
46
-55
-55
.63
ARCDARCD HAURRA AURAUA HARRUA
C2
D2
A2
D2
A₁
D2
+ 0.5 + 0.3 + 1.3
+ 7.8+ 4.9 19.5
+1.5 + 0.9 + 3.8
+ 0.1 + 0.1 + 0.3
+ 0.9 +0.6 + 2.3
+8.1+ 5.1 +20.2
+1.0 +0.6+ 2.8
+ O.I + O. I + 0.3
+ 0.9 +0.6 +2.3
+11.6 + 7.3 +29.0
+2.6+ 1.6 + 6.5
+ 0.4 0.3+ 1.0
+1.6 1.0+ 4.0
+11.6 + 7.3 +29.0
+1.8+ I.I + 4.5
+0.3 + 0.2 + 0.8
O
128
OO
0000
128
128
OO
0 0 0 0
O
128
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6

84
MULTIPLE TELEGRAPHS.
the commutator at which there is not serious interference
with an adjacent segment. The average current in the line
while the trailer was passing over the segment B1, with the
transmitter closed, was +70.0, while in the following segment
it was as much as 13.5, the negative sign indicating that
the line current was reversed. The next two segments were
practically free from current, but it is evident that the interfer-
ing current is so large that no relay can operate satisfactorily
in a segment immediately following an operating transmitter.
The practical effect of this is of importance, since this cause
alone is sufficient to prevent using each pair of segments,
B1 and B2 for example, for transmitting messages in each direc-
tion; and it will appear from other measurements that there
is another reason which will make working in both directions
on a single pair of segments impracticable.
The second series of readings was taken to show the effect
of revolving the trailers at a higher speed, everything else
remaining as in series No. 1. The trailers were run at 1,125
revolutions per minute, or at 150 segments per second, and
the sunflower M2 again adjusted until the readings in segments
A2 and C2 equalled each other, and B2 showed a maximum.
In this position the lag of segment A2 behind A₁ was increased
from .29 to .50 of a segment, which reduced to time is equivalent
to .0033 second, and is less in time than the lag for the slower
speed, though a greater fraction of a segment.
-

CHAPTER IX.
EXPLANATION OF THE MEASUREMENTS.
THE calculation of the current received at the distant
end of a grounded land line due to a constant source of electro-
motive force impressed at the sending end is a matter of con-
siderable difficulty, and no working formula suitable for calcu-
lation exists for this case, as far as we are aware. The nearest
approach to it is a formula * which has been applied to sub-
marine cables by Lord Kelvin, and curves, Fig. 35, have been
calculated from this formula for a copper line 390 miles long,
grounded without any receiver at the distant end, having
5.14 ohms per mile, and capacity .015 microfarad per mile,
with no leakage. The formula takes no account of inductance,
and probably for this reason gives such incorrect results. The
curves are shown merely to serve as a guide to aid in finding
from the measurements the true character of these curves.
Curve I shows that the current remains nearly zero for about
.0005 second and then abruptly rises and finally approaches a
fixed value at 65 milliamperes, if the electromotive force of
130 volts remains applied, and the line has a total resistance
of 2,000 ohms. If the line is connected to ground with the
electromotive force removed after .0066 second, the curve
II descends rapidly to zero as shown, and makes a kind of
*The formula from which curve I, Fig. 35, has been calculated is
2 T
I=
2 V [~ + w ³°³ + 2 + 2 ¹² + ... 2 (² + 1) + . . . ].
a212
in which v=e 4t, p=resistance per unit length, a=p (where y=
capacity per unit length), l-length of the wire, V, the electromotive
force applied at the origin, and t=the time. In the particular case
2 Va
assumed, Vo=130, and
PVT
.00794, and log10 v=-
π
1
.00127
t
85

86
MULTIPLE TELEGRAPHS.
wave having a sharp peak corresponding to a single closure
of the circuit for .0066 second.
That these theoretical curves do not correctly represent
the received current on a real line appears from a consideration
of the velocity of propagation.
There is no known wave
propagation having a greater velocity than that of light and
electric or magnetic waves in the free ether, which is known
to be approximately 186,000 miles per second. If these waves
One Division One Milliampere
70t
60
50
40
30
20
10
II
10 20 30 40 50 60
80 90 100 110 120 130 140 150
One Division=0.0001 Second
FIG. 35.-Theoretical curves of received current, calculated by a formula.
Curve I, received current with electromotive force applied for an
indefinite period. Curve II, received current when electromotive
force is removed and circuit put to ground after .0066 second.
70
travelled along the wire at this speed, it would require at least
390/186,000 or .0021 second for any trace of the disturbance
to reach the receiver; that is to say, the curve of Fig. 35 should
have remained at zero for fully .0021 second, whereas it indi-
cates that the current attains about 65 per cent. of its full
value by this time.
The effect which the line wire has in general upon an
electrical wave of this kind, caused by suddenly impressing
a constant source of electromotive force upon the wire for a
brief time and then removing it and putting the line to ground,
thus making points of abrupt changes in the current at the
transmitting end of the line, is to smooth out these abrupt
changes more and more the farther the wave travels along the
wire, until, on a very long wire, the curve rises gradually to
a maximum and falls again to zero. The length of this wave,
measured from zero to zero of current, bears almost no rela-

SYNCHRONOUS TELEGRAPH CIRCUITS.
87
tion to the period of closure of the electromotive force, since,
theoretically, the decreasing current reaches the zero value
only after an indefinite time.
In a synchronous system where the same length of seg-
ments is used for both transmitting and receiving commutators,
the transmitting segment determines the period of closure of
the electromotive force; and it is desired to receive the whole
of the current in a segment on the receiving sunflower which
corresponds to the same length of time as the period of closure
of the electromotive force. Since this received current wave is
spread out, however, and occupies a longer time than the
period of closure of the electromotive force, it is evident that
some of it must lap over, so to speak, and be received in seg-
ments adjacent to the one intended to receive the wave.
Figs. 36 and 37 illustrate in a general way single current
waves, the measurements for which are given respectively in
series numbers 1 and 2 of the table. The upper curve in each
figure represents the current in the line at the trailer T1, Fig.
34, and the lower curve represents the current received at
the trailer T2. The horizontal axis is divided at T, 2T, 3T,
and 4T, into four equal units, each representing the time
that the trailer requires to pass over a single segment. The
horizontal lines PQ and MN, at 56.8 milliamperes, represent
the limits which the current curves approach if the electro-
motive force remains steadily applied.
In Fig. 36 the line RS is drawn at a height of 70 milliamperes,
and extends over the period covered by the segment B₁ where
the transmitter is closed, the time of this period being .0162
second. This represents the average value of the current
during the period, which is 13.2 milliamperes greater than the
final value of the current, and is evidently greater because
the normal current is increased at the time of closure by the
current required to charge the line capacity. This current
rises rapidly from the origin O by the curve OA2S2 to a maxi-
mum at some point S2 above the line RS, and falls to the line
PQ at some point L before the time T, as will appear from
other measurements. The curve is drawn so as to make its
average value approximately equal to OR. This may be done
roughly by making the sum of the areas B2STL and ORA 2
equal to the area A2S2B2.
At the time 7 the line was put to ground through the follow-

88
MULTIPLE TELEGRAPHS.
ing segment C1, with electromotive force removed, when it
was observed that the average current was 13.5 and in the
opposite direction. This is represented by the curve
120f
Current in Milliamperes
100
'80
60 P
56.8
40
20
10
56.8
60 M
40
20
O
RA2 B
Ο
O
S₂
1
B
!!
D
0047
1
1
1
B,
ų½
010
L
B?
IT
XH
--2910
Z2
——
X2
H-3
E
-020
N
3
W.030
1
1
1
HY
1
1
_
-ö
2T
Time in Seconds
D
N
K.040
.0487
.050
31
?
AI
060
-065
4T
.070
FIG. 36.-Curves showing the transmitted and received currents on the
artificial line of Fig. 34 at a frequency of 61.6 segments per second,
corresponding to a period of .0162 second. The upper curve repre-
sents the current flowing into the line at the transmitting end, while
the trailer passes over segment B₁, and the discharge of the line
through the following segment, C₁. The lower curve represents
the received current-wave a portion of which, from C to E, only
is received in segment B₂, this segment lagging behind B, by the
amount OC, equal to .0047 second, or .29 segment.
TZ2X2Y2Z, which shows an average value in the negative
direction of approximately 13, the line XY being drawn 13.2
units below the axis. That the current is practically zero
during the next two segments, D1 and A1, is shown by the
readings in series 1.

SYNCHRONOUS TELEGRAPH CIRCUITS.
89
The curve of the current received through the trailer T2
is drawn as starting at some point A later than the time of
closure of the electromotive force, and rises by a curve ADUV
somewhat similar to the theoretical curve of Fig. 35, practically
reaching the final value at some point U before the time T.
Current in Milliamperes
120-
100
80
60
56.8
40
20
0
60 M
56.8
20
RA2 B2S
T
Pl
Sz
O
T
40 BL
B₁
D
19900
A/.0033
C
T
F
H
0198
W.020.
K 3T
Time in Seconds
2T
FIG. 37. Similar to Fig. 36 except that the speed of the trailers is greater,
150 segments per second, corresponding to a period of .0066 second.
The lag of segment B, behind B, is represented by the distance
OC, which is .0033 second, or one-half segment.
0132 -
.010E
D₁
1
이
​.0264
4T
.030
In a similar manner the point V, when the current begins to
descend to the zero value again, by the curve VFW, occurs
as much later than the time T as the point A was later than 0.
The line BC, at a time .0047 second, corresponding to the
lag of the commutator M2 behind M1, represents the time
when the trailer T2 enters the segment B2; and the line GE
at a time T later represents the time at which the trailer leaves

90
MULTIPLE TELEGRAPHS.
segment B2 and enters segment C2. The line JK corresponds
to the end of segment C2.
The average value of the current in segment B2 was 52.5,
and the line BG represents this average. The curve shows that
the area included between that portion of it DUVF and the
axis CE is approximately equivalent to the area of the rectangle
BCGE; that is, the small triangular areas BDU2 and FGV 2
together are approximately equal to the portion UVV 2U2.
The line HJ is drawn at a height of three milliamperes, and
represents the average current during the following segment
C2. The area FEW included between the curve and the axis
is approximately equal to the rectangle HEJK. Similarly
there is an area ACD corresponding to the current in the seg-
ment A2 which precedes B2. The adjustment of the commutator
M2 back and forth until the readings in the segments A2 and C2
were equal is evidently equivalent to the shifting of the lines.
BC and GE together backward and forward until the areas FEW
and ACD are approximately equal, and CDUVFE a maximum.
The curves of Fig. 37, corresponding to a frequency of 150
segments per second, and the measurements in series 2 of the
table, make evident still other effects which the low frequency
of series No. I did not show. The period of closure of the elec-
tromotive force is so short in this case that neither the trans-
mitted current reaches its final value OP nor the received
current its final value OM. These curves, however, in Fig. 37,
are made identical with those of Fig. 36 up to the time T, equal
to .0066 second, when the line in this instance is connected to
ground. Referring to the upper curve, the current at the
transmitting end, it is evident that the average value of the
curve OPA2S2B2T, taken to .0066 second, is considerably
greater than the average taken as in Fig. 36 to .0162 second;
and the reading in segment B1, 80.7 milliamperes, represented
by the line RS, confirms this, showing that the average inflow-
ing current is considerably greater with the higher frequency.
In a similar manner the current in the following segment, C1,
- 30.0, is represented by the line XY, though the curve itself
is quite similar to that in Fig. 36. A portion of this current
curve is now observed to lap over into the following segment,
D₁, as indicated by the observations, the value being -1.3.
Referring to the curve of received current, it is made identical
with that of Fig. 36 up to the point U, which in this case is below

SYNCHRONOUS TELEGRAPH CIRCUITS.
91
the line MN, and from which point it rapidly falls to zero by
the curve FW. The line BG is drawn to represent the average
observed current of 40.7, and represents fairly well the average
of that portion of the curve DUF. The line HJ represents the
average of 5 milliamperes in the following segment, C2. The
reason why the reading in the working segment B2 decreases
while that in the adjacent segment C2 increases as the trailers
increase in speed is evident from the curves; for in Fig. 36 the
current remained at its maximum value for a considerable time,
from U to V, thus increasing the average value for that segment,
whereas, in Fig. 37, the curve ascends to a sharp peak and evi-
dently reduces the average throughout the segment considerably.
In the following segment, C2, the average is increased because
the point J has moved so much nearer to the point H that it
evidently increases the average throughout the segment.
It has been stated that there will be differences between
measurements made upon the artificial line of Fig. 34 and a
real line which it is designed to represent, and one of these
differences is apparent in the observations taken. On a real
line of 390 miles length it is certain that the received current
cannot begin (at the point A) before the electromagnetic wave
has had time to travel this distance, and this requires at least
.0021 second. The curves illustrated have been drawn beginning
somewhat earlier, as this seems to agree with the observations
better; and it is reasonable to suppose that an artificial line of
the character described, which more nearly approximates the
condition assumed by the formula of Lord Kelvin referred to,
does not require so long a time to transmit the wave as is required
by a real line. If this supposition is correct, we should expect
to find that the setting of the commutator M2 would lag some-
what more on a real line than it does on the corresponding
artificial line constructed in the manner described.

CHAPTER X.
DIFFERENT CONDITIONS OF INSTRUMENTS AND LINE
LEAKAGE.
THE measurements in the preceding chapter have shown
that there is a certain position for the receiving-commutator
relative to the transmitting-commutator which permits of
sending current impulses to any one of several different circuits
without sending more than a very small amount of the cur-
rent into any other circuit than the one selected. It is im-
portant to know whether any interference can be produced
with the same setting of the commutators by anything that
might happen to one of the transmitting circuits, and it is
also important to inquire what causes, if any, will require a
different adjustment of the commutator. An examination of
the extent to which changes in receiving-instruments and in
leakage upon the line affect the commutator adjustment forms
the subject of the present inquiry.
EFFECTS OF OPENING CIRCUITS LEADING FROM SEGMENTS.
If the circuits leading from the sunflowers in Fig. 34 are
to be used for transmission of messages, it cannot be said that
each pair of segments forms a circuit independent of the others,
for it is true that if any one of the circuits from the sunflowers
is opened, that is, disconnected from ground, it immediately
has a serious effect upon some of the other circuits; whereas
there is no serious effect provided all circuits are kept continu-
ously connected to ground, either directly, through a resistance,
or through an electromotive force.
To show the effect which opening some of the circuits has
upon others, the four series of measurements, Nos. 6, 7, 8 and 9,
of the table were taken, Nos. 6 and 7 having the transmitter B1
closed so as to impress an electromotive force upon the line,
92

DIFFERENT CONDITIONS OF INSTRUMENTS.
93
and the circuits from all other segments of sunflower M1, viz.,
A1, C1, and D1, open; and also having all receiver-circuits con-
nected to ground, each through a relay having a resistance of
50 ohms and an inductance of .4 henry. The difference between
series No. 6 and series No. 7 is in the speed of the trailers, No. 6
being at 61.6 segments per second and No. 7 at 150 per second.
There was no leakage, and the receiving-commutator was set
in each case to give minimum interference before the trans-
mitter-circuits were opened. In series No. 6 the current in the
receiving segment, C2, immediately following the working seg-
ment, was increased from about 3 to 14.5 milliamperes, and the
currents in the other segments were not much affected.
The explanation of this is made evident by reference to the
curves Fig. 36. Instead of putting the line to ground at the
transmitting end at the time T, the line was freed at this time
because the circuit C1 was open, and hence the charge in the
line could not flow out at the transmitting end and was com-
pelled to pass out through the receiver, thus increasing the
current in the segment C2, following the working segment, as
observed. This effect is so large that it would interfere with
the use of the segment C2 for an independent circuit.
The series No. 7, which is given merely to show the effect of
an increased speed of the trailers, indicates that the current, 22.5
milliamperes, in segment C2 was in this case more than half as
great as the working current, 42 milliamperes, and that the
line was not completely discharged, even after the passage of
the trailer over segments D2, having a current of 9.5, and A2,
having a current of 7.
The measurements in series Nos. 8 and 9 were taken to show
the effect of opening a single one of the receiver-circuits, B2,
all of the other receiver-circuits, A2, C2, and D2, as well as all of
the transmitter-circuits, being connected to ground, and the
electromotive force being applied in segment B1, as before.
With the exception of these changes series No. 8 was taken
under the same conditions as series No. 6; and No. 9 under
the same conditions as No. 7. No reading is shown for the
working segment B2, which was open, but the current in the
following segment, C2, in series 8 is 13.8, and in series 9 still
larger, 18.5. This is a cause of serious interference with seg-
ment C2, due to opening receiver-circuit B2.
This may be explained by considering the charge of the

94
MULTIPLE TELEGRAPHS.
line. When the receiver segments are connected to ground
and an electromotive force applied at the transmitting end and
allowed to remain long enough for the current to reach its final
value, the potential falls steadily from the transmitting end
to the receiving end of the line from, say, 130 volts to zero.
The density of the charge upon the wire under these conditions
is greatest at the transmitting end, and falls steadily to zero at
the receiving end, and the centre of mass of this charge is there-
fore one-third of the length of the line from the transmitting end,
and two-thirds of the length from the receiving end.
As a
result nearly all of the charge of the line escapes by the trans-
mitting end, only a small proportion of it going by the receiving
end; and this is still further assisted by the fact that the line
is put to ground without resistance at the transmitting end,
while it must pass through the receivers at the receiving end.
When, however, the working receiver, as B2, for example, in
series 8 and 9, is opened, the line is freed at the receiving end
and receives the full potential throughout its entire length,
thus accumulating a greater charge on the wire than ever occurs
under normal conditions. This is noticed in the measurements,
the current flowing from the line in opposite directions at the
same time, going to ground through segments C1 and C2.
INFLUENCE OF THE RECEIVER-COILS UPON THE RECEIVED
CURRENT.
Four series of readings, Nos. 10 to 13 inclusive, were taken
to illustrate the effect which a change in the resistance and
inductance of the receiver has upon the received current.
The three series Nos. 10, II, and 12 were taken under the
same conditions of speed and commutator setting, there being
no leakage; the only difference between the series being that
series No. 10 had no receivers at all except the millivoltmeter,
while No. 11 had four receivers, each of 50 ohms resistance
and 4 henry inductance, and No. 12 four receivers, each of
200 ohms resistance and 1.6 henrys inductance, all transmitter-
and receiver-circuits being closed. One hundred and thirty-
two volts were applied by the transmitter KB, and it is observed
that the interfering currents 6.5 in A2 and 5.8 in C2 in series
No. 10 when there is no receiver are somewhat greater than
they are with the 50-ohm receiver in series No. II.

DIFFERENT CONDITIONS OF INSTRUMENTS.
95
The setting of the commutators in either of these conditions
shows that they are approximately in the position for minimum
interference, the readings in A2 and C2 being nearly equal.
In series No, 12, however, having the 200-ohm and 1.6-henry
receivers, the setting of the commutator was not good, as the
reading in C2, 6.5, is in excess of the reading in A2, 2.0. The
working current in B2 is observed to decrease from 43.3 with
no receiver to 40.5 with the 50-ohm receiver, and to 30.0 with
the 200-0hm receiver. It is evident that to improve the com-
mutator setting in series 12 the commutator M2 should be
moved forward in the direction of rotation of the trailer still
more, in order to reduce the reading 6.5 and increase the read-
ing 2.0. This adjustment was made and the readings of series
13 give the result, showing that the commutator M2 was
moved so as to lag .61 segment instead of .50 segment behind M1.
It is important to observe that the windings of the receiver
cause such a difference in the commutator setting, and in
general experience has shown that it is more desirable to
use receivers with small resistance and inductance than with
large values. The setting of the commutators has an important
bearing upon one method of operating the line in both direc-
tions, as is explained elsewhere; and anything tending to change
this setting should be carefully noted.
LINES WITH LEAKAGE.
To show the effects of line leakage three separate leaks
of 4,000 ohms each were introduced in the line of Fig. 34 at
the condensers C1, C2, and C3. This value of leakage is equiv-
alent to 4000/3 ohms total, or to 520,000 ohms per mile on
the line of 390 miles. The series of readings No. 14 was taken
without any leakage, and with commutator set at .53, using
50-ohm receivers. The adjustment for non-interference is
seen to be fairly good, the current in A2, preceding the working
segment, being 4.3, and in C2, following the working segment,
being 4.8, the working current being 38.5.
Series No. 15 is obtained under the same conditions except
that the leak is inserted, causing the working current to fall
from 38.5 to 27.8, and it is seen that the adjustment of the
commutator, which has not been changed, is still fairly good,
the current in A2 being 3.5, and in C2 2. In series No. 16 the

96
MULTIPLE TELEGRAPHS.
only change made was to adjust the commutator for best
minimum interference, and the setting was then .46, only .07
segment less than it was when there was no leakage on the line.
The tendency of leakage is therefore to reduce slightly
the lag of the receiving-commutator M2 behind the transmitting-
commutator M₁, and also to reduce the value of the received
current. It can hardly be said that leakage has such a harmful
effect on synchronous telegraphs as it does upon the closed-
circuit Morse system, this being a fairly heavy value of leakage
with which to operate a line.
As a matter of still further interest four additional sets
of readings, Nos. 17 to 20, were taken, using the 200-0hm relays.
The only difference between series No. 16 and series No. 17
is that the 200-ohm form of relay is substituted for the 50-0hm
form. This was observed both to reduce the received current
from 27.8 to 19.5 and to alter the commutator setting for best
adjustment, the reading in C2 indicating 3.8, and in A2 1.3.
The commutator M2 was then moved forward so as to increase
the lag to .55, in series No. 18, and this nearly equalized the
readings in C2 and A2. This changed the setting .09 of a seg-
ment in the direction of increased lag. Series No. 19 was
obtained under the same conditions as series No. 18, except
that leakage was removed. This was observed to make the
reading in C2 again larger than that in A2, which, as before,
shows that leakage slightly decreases the lag of the commutator
M2. Series No. 20 is taken under the same conditions as No.
18, except that the leakage is removed and the commutator
set for minimum interference.

CHAPTER XI.
WORKING IN OPPOSITE DIRECTIONS ON THE SAME WIRE.
IT has been shown in the preceding pages that the arrange-
ment of apparatus in Fig. 34 permits any one of the receivers
to be operated by its corresponding transmitter without undue
interference with any of the other receivers; but it was also
evident that no receiver could be placed in any of the trans-
mitter-circuits and no transmitter in any of the receiver-cir-
cuits-without causing the receiver to be interfered with by the
operation of the transmitter in the preceding segment at the
same end of the line. This is due to the discharge of the line
capacity, which passes to ground through the segment imme-
diately following an operating transmitter. The four circuits
provided by the arrangement of Fig. 34 for sending messages in
one direction cannot, therefore, for the reason just stated, all
be used at the same time to send messages in either direction.
There is still another reason why no two of the circuits of
Fig. 34 can be used to send messages at the same time in opposite
directions, and this is because the commutator setting in general
must be changed to find the position of minimum interference
when attempting to transmit from the sunflower which was
previously used to receive; for it has been shown that the
receiving segment must always lag behind the transmitting
segment by some fixed value, depending upon the speed, line,
and instruments, and if sunflower M2 lags behind M1, it is clear
that sunflower M₁ precedes M2 instead of lagging behind it.
The series of observations Nos. 3, 4, and 5 of the table werc
taken to show the effect of attempting to work backward as
explained. Series No. 3 is taken under the same conditions of
line, speed, and commutator setting as series No. 1, previously
described (page 76), except that the positions of the transmitters
and the receivers were interchanged, the receivers being con-
97

98
MULTIPLE TELEGRAPHS.
nected with the segments of sunflower M1. The transmitter KB
in segment B2 was closed and the received current in the corre-
sponding segment B₁ was 32.5, while the current in the following
segment, C1, was 22.5, the current in the remaining two seg-
ments, D1 and A1, being almost zero. Thus it is evident that
receivers in segments B1 and C₁ respond almost equally well
to the transmitter KB.
If the line wire is very short, say 100 miles or under, the
lag of the commutator M2 behind M₁ would not be so much.
as .29 segment, and in such cases it might be possible to com-
promise by not setting the commutators in the best possible
position for receiving, but placing them in such a position that
they receive equally well from either end of the wire. This
position is evidently found by giving no lag to either commu-
tator behind the other, and the series of readings No. 4 was
taken to show that this compromise adjustment does not
answer for a line of so great length as 390 miles. The segment
B₁ in this case received 40.0 instead of 32.5, as before, and
C₁ received 17.7, which is somewhat less than 22.5, but is yet
too large to admit of good operation.
The amount of this interfering current depends chiefly
upon the length of the line and the speed of the trailers, and
series No. 5 was taken to record the values of these interfering
currents when the speed was 150 segments per second and the
setting of the commutator as in series No. 2, namely, the best
position for receiving in commutator M2. The transmitters were
placed as in series Nos. 3 and 4. The current received under
these conditions in segment B1 was 5.3, as compared with 40.0
in the preceding series, No. 4, and the current in C₁ was 39.5,
which is about the same as the working current, 40.7, in segment
B2 in series No. 2, while the current in D1 was 2.8.
It is now
seen that commutator M₁ is set, without change, in nearly the
best position for receiving from M2, as the interfering values,
5.3 and 2.8, are nearly the same; but it is to be noticed that
the working current, 39.5, is now received, not in the correspond-
ing segment B1, but in the segment C₁ immediately following.
This result was to be anticipated, for if segment B2 lags one-
half a segment behind B1, it is evident that segment C₁ must
lag the same amount, one-half a segment, behind B2.

WORKING IN OPPOSITE DIRECTIONS ON SAME WIRE. 99
METHOD OF WORKING IN BOTH DIRECTIONS IN WHICH THE LINE
IS ADJUSTED TO THE SPEED OF THE TRAILER.
These measurements serve to illustrate a general principle
which may be utilized to enable messages to be transmitted
without interference over a long line in opposite directions at
the same time. A single pair of segments, however, cannot be
used for sending messages in opposite directions without resetting
the commutators between each transmission, and this is evi-
dently impracticable. Fig. 38 illustrates an application of this
principle showing that messages may be transmitted without
interference over a wire the length of which is adjusted to
the speed of the trailers in such a manner that each segment
lags one-half a segment behind a corresponding segment of the
other commutator. The figure is a development of two sun-
flowers, M1 and M2, each having six segments, A, B, C, D, E,
and F. A2 is supposed to lag one-half a segment behind A1,
and consequently B₁ lags one-half a segment behind A2, B2
lags one-half a segment behind B1, C1 one-half a segment behind
B2, and so on.
-tr.
ri
-tr3
MARG
rz-
-tra-r4
M₁
A₁
-tr₂-2C,
B₁
D₁
E
FI
M2
A2
B2
C2
D2
E2
•r tr
-tr₂
r2
3-tr3.
tra
TA
2
-
F2
FIG. 38.-Diagram showing method of working in both directions when
line is adjusted to speed of trailer. The segments of the commu-
tators are displaced, relatively to each other, one-half segment.
If a transmitter tri is connected in segment A1, and a
receiver r₁ in B1, and both a receiver ri and a transmitter
tri in A2, it is evident that transmitter tri at M₁ may transmit
to receiver r₁ at M2 or that transmitter try at M2 may transmit
to receiver r₁ in B₁; but both transmitters tri cannot be operated
at the same time. If the transmitter tr₁ and receiver r₁ at M1
are assigned to a single operator and the transmitter tr₁ and
receiver r₁ at M2 are assigned to another operator, these two
operators will be supplied with instruments equivalent to

100
MULTIPLE TELEGRAPHS.
those on a single Morse wire, in which one message only can be
transmitted in either direction.
The operation of the transmitter tr₁ at M₁ does not affect
any instrument at the receiving commutator M2 except that
in the corresponding segment A2, which is desired. It does
affect, however, the instrument in the segment B₁ immediately
following A1, but this instrument, r₁, is the home receiver
for the transmitting operator at M1, and it is desirable to
have this instrument respond to the operator's signals. The
transmitter tri at M2 will operate the relay r₁ at M₁ without
interfering with any other instruments at M1. It would,
however, interfere with a receiver located in the following
segment, B2, and for this reason no receiver is connected in
this segment; but a transmitter tr₂ is connected there, and
the transmitter tri cannot interfere with this.
The first independent circuit involves the segments A1,
A2, and B₁, and is indicated by the dotted lines, with arrows
showing the direction of transmission. The second independent
circuit involves the segments B2, C1, C2, and the third the
segments D1, D2, E1, and the fourth E2, F1, F2, there being
four wedge-shaped sets of dotted lines, fitting into one another,
so to speak, with the points of two of them on the right side,
and of the other two on the left side. Thus four independent
circuits may be obtained by the use of six sets of commutator
segments, and in general the number of independent circuits
thus obtained, each of which is capable of being worked in
either direction, but not in both at the same time, is equal to
two-thirds of the number of sets of commutator segments.
The application of this method to a wire depends upon
maintaining a fixed relation between the speed of the trailers
and the length of the wire, but this should not offer
any serious
difficulty in practice. The higher the speed of the trailer the
shorter is the line to which it may be applied, but a limit to the
speed of the trailers, measured in segments per second, is soon
reached, from a practical standpoint, and this limit seems to
lie somewhere between 150 and 200 segments per second, above
which speed it does not seem advisable to operate a practical
system. Assuming that we arbitrarily set some limit, as
150 segments per second, for example, above which speed
we do not care to operate any line, then there is a fixed length
of wire, when given receiving-instruments are used, which

WORKING IN OPPOSITE DIRECTIONS ON SAME WIRE. IOI
corresponds to this speed. The length of artificial wire on
which the measurements were taken was 390 miles, but, as
explained elsewhere, it is probable that the lag on a real wire
is slightly greater, and consequently the length of a real wire,
corresponding to this speed, is something under 390 miles.
The number of segments per second which the trailer passes
over determines the number of circuits which it is practicable
to work, in a manner to be explained. It is usually not prac-
ticable with a given set of instruments to change the working
speed from one value to another according to the number of
circuits which it is desired to operate, and as a rule the same
speed which would be used on the limiting length of line is
also desired for lines shorter than this limit. Some other
means than adjustment of the speed should therefore be
employed to set the commutators in the best working position
for shorter lines. This is afforded by connecting the real line.
in series with an adjustable artificial line, so that its effective
length may be increased to equal the standard length required.
When this adjustment is once made it will remain permanent,
for it has been shown that leakage, which is the chief cause of
the variable condition of lines, has little effect upon the com-
mutator adjustment, and the artificial line will enable any
slight changes to be made when desired.
METHOD OF WORKING IN BOTH DIRECTIONS BY MEANS OF TWO
INDEPENDENTLY ADJUSTABLE SUNFLOWERS.
A different arrangement from that just described for
obtaining circuits which will work without interference in
opposite directions is sketched in Fig. 39, which shows the
development of two sunflowers, M1 and M2, one at each end
of the line wire, and indicates the direction in which the trailers.
T1 and T2 move over the segments. A complete cycle or
revolution between the dotted lines is seen to consist in com-
mutator M₁ of six segments, A1, B1, G1, C1, D1, G3, which are
all fixed relatively to each other and are all of equal length
except G1; and in the sketch this segment is made 1.5 times
as long as each of the others. The commutator M2 consists
of two parts, one part containing the segments A2 and B2,
fixed with relation to each other, and the other part the segments
C2, D2, G2, fixed with relation to each other. These two parts

102
MULTIPLE TELEGRAPHS.
are, however, separately adjustable relatively to each other
and to the commutator M1. In the sketch the segments A2
and B2 are set so as to lag segment behind A1 and B₁ respect-
ively, and segments C1, D1, and G3 lag segment behind.
C2, D2, and G2 respectively.
M,
Two transmitters, tr₁ and tr2, are connected, one to each
segment, A and B1, and will
operate the receivers r₁ and r2
respectively at the opposite end
of the wire. Similarly two trans-
mitters trg and tr4 are connected,
one to each segment, C2 and D2,
and will operate the relays rз and
14 respectively, connected in seg-
ments C₁ and D1. The remaining
1 segments, G1, G2, and G3, are con-
nected directly to ground.
1
1
tri tr₂
[
M2l
1
1
F
A, B G₁ T₂
A2 B2
T3
C₁
C2 D2
F 12 trz tr4
G
r4
D₁
G2
G3
I
These commutators are sup-
posed to be adjusted in the
FIG. 39.- Diagram showing position for minimum interfer-
method of working in both ence in the manner explained,
directions by means of two and if the trailer runs at a
independently adjus- speed of 150 segments per second,
table sunflowers, M₁ and
M 2.
this arrangement permits of op-
erating a line somewhat longer
than the maximum line which could be worked on the plan
last described; and it does not require any adjustment of an
artificial line, since the parts of the commutator M2 may be
separately adjusted, within limits, to the proper position for
a long or a short line.
G
The arrangement described is the equivalent of the common
quadruplex circuit, as far as the handling of messages is con-
cerned, for only four messages are transmitted at a time, two
in each direction. With the arrangement last described only
four messages could be transmitted at once, but it had the
advantage of being equivalent to four independent circuits,
since three messages could be transmitted in one direction.
and one in the opposite, or all four in one direction if desired.
More instruments were required, however, with that plan than
with this, there being eight transmitters and eight receivers
involved, counting both ends of the wire, in the previous plan,

WORKING IN OPPOSITE DIRECTIONS ON SAME WIRE. 103
and but four transmitters and four receivers with the arrange-
ment of Fig. 39.
That this is a possible working arrangement will be seen
by examining each circuit in detail. Transmitter tr₁ in A₁
evidently operates the receiver ri in A2 without disturbing
any other receiver connected to commutator M2. The only
segment in which there is any interference is B1, immediately
following A1, but this segment contains a transmitter, and
not a receiver. Similarly tr₂ connected to B1 operates r2
connected to B2 without disturbing any other receiver con-
nected to M2, and the segment G1, immediately following B1,
is connected to ground and contains no instruments. A study
of the observations in the table shows that the open space
between segments B2 and C2 can have no harmful effect, since
it is followed by a transmitter trз, and the segment, G1, cor-
responding to it is grounded without a transmitter. The
transmitter tr3, connected to C2, operates receiver r3 without
disturbing any other receiver in M₁, and since the segment
following it, D2, contains no receiver, it does not interfere
with anything. The transmitter tr4 evidently operates the
receiver 74, and, being followed by a grounded segment G2,
does not interfere with the receiver in A2 immediately follow-
ing, although there is an open space of 1 segments between
G2 and A2.
Another method of working in opposite directions on the
same wire, by the use of artificial lines, is shown at the end
of the following chapter,
THE USE OF GROUNDED SEGMENTS.
It has sometimes been proposed to provide a means for
working a line in opposite directions by the device of inserting
extra segments, which may be shorter than the working seg-
ments, connected to ground for the purpose of discharging
the line to get rid of the interference which is caused by the
line charge. This device is perfectly effective if the grounded
segments are long enough, but on long lines, used at as great
a speed as 150 segments per second, it is evidently a very
wasteful principle to employ; for in this case the grounded
segments must be at least as long as the working segments,
to discharge the line, and it is necessary to sacrifice fully one-

104
MULTIPLE TELEGRAPHS.
half of the number of circuits which would otherwise be avail-
able.
It will be shown in the following chapter that four Morse
circuits are about all that can be obtained satisfactorily on a
line of this length, and consequently if grounded segments
were employed there would not be enough circuits obtained
to pay for using synchronism at all. In view of the methods
described above it is questionable whether there is any need
for the employment of grounded segments, even on shorter
lines.
Oce

CHAPTER XII.
MORSE CIRCUITS: GENERAL CONSIDERATIONS AND
PARTICULAR SYSTEMS.
THE principles of operating synchronous telegraphs given
in the preceding pages may be applied to almost any of the
particular systems of such telegraphs which now exist or may
be devised; and it should be stated that these systems are
not confined to those which have for their object the manual
operation of Morse circuits in which one operator sends with
a key and another receives by a sounder, but include systems
in which the operator manipulates a keyboard and the message
is received in page form on a printer, as well as systems in
which the operators prepare perforated tape and the messages
are sent by automatic machines into the several circuits.
When machine sending is resorted to, however, the quantity
of matter which can be sent over the wire by the synchronous
system seems to be no greater than, if as much as, may be
sent by other automatic systems, and it is an open question
whether the advantages offered by affording independent
circuits for the transmission of messages to different points are
sufficient to compensate for the advantages offered by other
systems.
The superiority of synchronous systems in general is that
there is no necessity to resort to automatic transmission, since
the speed of each individual circuit is lowered so as to come
within the sending capacity of a single operator, and the wire
is thus, in effect, divided up into several independent wires,
in which business may be transacted as though there were in
reality a number of actual wires. Consideration will here be
given especially to manual Morse working employing keys
and sounders.
105

106
MULTIPLE TELEGRAPHS.
USING ONE WIRE FOR ALL MESSAGES IN ONE DIRECTION, AND A
SECOND WIRE FOR THOSE IN THE OPPOSITE DIRECTION.
When there is a large amount of traffic to be handled,
and there are two or more wires available, the simplest way
is to use one wire to send messages in one direction, and another
to send in the opposite direction, as this avoids the use of
special devices such as those previously described. This
plan is illustrated in the diagram Fig. 40, in which the com-
plete set of instruments for Morse working in one direction is
shown.
Although all of the measurements in the table were taken
using a single source of electromotive force, the use of two
generators of opposite polarity is preferable in practice, because
it permits the use of polar relays adjusted in the neutral or
most sensitive position, which has certain advantages as regards-
the receiving-apparatus. The measurements made with cur-
rents of one polarity, however, apply, as far as the question.
of interference is concerned, to the case where double currents
are used; for, the proportion of interfering current in adjacent
segments remains about the same.
The instruments respond to the interfering currents less
easily because they are always acted upon by a strong current
of one polarity or the other, depending upon whether the
transmitting-key is open or closed. The arrangement of
apparatus shown in this figure operates better than any of the
many other arrangements which we have tried.
It is possible
to operate a system with currents of a single polarity, in which
case the relay armature vibrates when the transmitting-key
is closed so rapidly that it has the effect of opening the circuit
of a repeating-sounder. In the arrangement of Fig. 40, how-
ever, there is no vibration of the polar relay armatures, A2,
B2, C2, and D2, since each is adjusted in such a way that the
armature will remain against whichever of its two contacts
it is placed, when no current is flowing. With this adjustment
the slightest current will move the armature from one side to
the other, the direction depending upon the direction of the
current.
When the commutator M2 is adjusted as explained to
the position of minimum interference for the line being oper-
ated, a large current is always received in segment A2, of one

MORSE CIRCUITS.
107
polarity or the other, every time the trailer T2 passes over
the segment, the polarity being determined by the position
of the lever of the transmitter KA, which, when on its back
contact, di, sends, say, a negative current to line, from battery
messages over a single wire in one direction only, from M, to M₂.
FIG. 40.-Diagram of four-circuit synchronous Morse system capable of sending four
E
Q
•////////
E
Hololololol
G2
Ehhe
K
D₁
TEK
00
C
TKB
h
B₁
D
00
8
TEKA
d2
M,
T.
F
M₂
T₂
A,
I
G₁; and when on its front contact, d2, sends a positive current
to line from battery G2. In a similar manner the relays in
segments B2, C2, and D2 are acted upon every cycle by strong
currents from their respective transmitters KB, Kc, and KD.
Each of these relays is controlled by these strong currents
in preference to the small interfering currents from adjacent
segments all the time during which the trailer remains upon

108
MULTIPLE TELEGRAPHS.
the segment connected to the relay; and as soon as the trailer
passes off from the segment the relay has one terminal imme-
diately freed, so that no current whatever can pass through
it, either interfering current or working current. The relay
armature will therefore remain in the position in which it was
last left by the working current.
Suppose that the operator closes the key KA at the time the
trailer is on the segment B1, for example. The wire A1 imme-
diately receives a positive charge from battery G2, but, being
open at this time, no current can flow to line until the trailer
has passed over the segments B1, C1, and D1, when current
will flow to line as soon as the trailer enters A1. The trailer
T2 arrives a little later at the beginning of segment A2, and
the armature which was previously against the contact c1
moves over to contact c2, where it will remain until a negative
current is again received by the relay. A positive current
will contine to be received by the relay every time the trailer
passes the segment A2 as long as the operator holds the key
KA closed.
There is thus a firm contact established at c2 which will
make a sharp and clear signal upon the sounder S very shortly
after the operator closes the key KA. In a similar manner,
when the operator opens key KA, thus closing contact di to
the negative battery G1, the relay A2 will receive a negative
current the next time the trailer T2 arrives upon the segment
A2. This will move the tongue of the relay back to the contact
C1 and terminate the signal, restoring the apparatus to the
normal condition.
Disregarding the time required for the electrical wave
to travel over the wire, there will be in general an interval
between the time of closing of the transmitter KA and the
time when the relay armature responds to the signal. This
time interval varies according to the position which the trailer
T₁ happens to occupy when the operator closes his key. It
may chance that he will close the key KA just as the trailer
is entering the segment A1, in which case there will be no
delay in delivering the signal more than is ordinarily required
on a Morse wire. But the worst case must be considered, and
that is when the trailer happens to be in the position just
entering the segment B1, for here the signal will not be delivered
at A2 before the trailer has passed over all the intervening

MORSE CIRCUITS.
109
segments. The number of intervening segments is always
one less than the number of Morse circuits being operated by
the sunflower, and in this case happens to be three. So it
appears that with the trailer passing over a given number of
segments per second the time interval between the making and
the delivering of a signal is greater the more circuits there
are operated.
In a similar manner it may happen that the operator upon
terminating his signal raises the key KA when the trailer is
again just entering segment B1. In this case the signal will
not be terminated until three segments later, and the length
of time from beginning to end of the received signal will then
be the same as the transmitted signal. If, however, the
trailer is just entering segment A1 when key KA is opened,
then the signal is immediately terminated, and the duration of
the signal is in this case shortened by the length of time cor-
responding to three segments. If the next closure of the
key happens to occur at the beginning of segment B1, the space
between signals is evidently lengthened by the time of three
segments. This effect is not noticeable if the time interval
of three segments is very short in comparison with the dura-
tion of the signal, since a slight lengthening and shortening
of the signals is not perceptible at the receiver; but if the
length of a signal is very short, so that the time of three seg-
ments is any considerable proportion of it, then the received
signals will be perceptibly distorted.
The fastest Morse operators under the most favorable
conditions provided in the tournaments which have been
held in this country have been known to transmit something
over 50 words per minute, including punctuation and all signs.
This is extremely fast sending, as becomes evident when it is
considered that this speed is equivalent to the transmission of
dots at the rate of something like 16 per second, including
the time of the dot itself and the space following.
One thirty-second, or .0312, second is therefore the very
shortest time that could ever be required for a period of closure
of the key KA for a dot. If the rate of the trailer is 150 segments
per second, three segments correspond to .02 second. This
value is about 64 per cent. of .0312 and proves that a dot
signal would be shortened by 64 per cent. if an operator should
send at 50 words per minute, with a four-circuit sunflower.

IIO
MULTIPLE TELEGRAPHS.
This would distort the received signals to such an extent that
they could hardly be read on a sounder. All of the signals
would be legible, however, if a Morse ink-writer or embosser
were substituted for the sounder, since the trailer is sure to
pass over the working segment once at least during the period.
of closure, .0312 second. It requires .0266 second for the
trailer to pass over four segments, which is less than the .0312
second.
It is necessary, if Morse signals are to be received on a
sounder, to make sure that the period of closure for a dot
will at least include two passages of the trailer completely
over the working segment. With the arrangement of Fig. 40
this means that the time of closure for a dot must be at least
equal to the time occupied by the trailer in passing over nine
segments, in order to be sure to include any one segment com-
pletely twice. At 150 segments per second the time of nine
segments is .06 second, and if .0312 second is the period of
closure when sending at 50 words per minute, it is evident
that the permissible speed of sending for good Morse working
with the arrangement described is 50 X.0312/.06, or 26 words
per minute in each circuit, or at the rate of 104 words per
minute on the wire.
To obtain higher speeds than this per circuit it is necessary
either to increase the speed of the trailer or decrease the number
of circuits. The variation in the period of closure of the sounder
for a dot will be between the time of eight segments and twelve
segments, depending upon the time when the period of nine
segments above referred to begins. If it begins immediately
after the working segment, then the received signal is eight
segments long; but if it begins immediately before, the received
signal is twelve segments long.
If there were but three circuits instead of four, and the
commutators had six segments instead of eight (the diametrically
opposite ones being connected together), then in order to be
sure that the dot covers at least two working segments the
time of closure of the dot would be the time of seven segments,
which at 150 segments per second is .0467 second. This would
allow the operator to work at a speed of about 33.5 words
per minute, and would send almost as many words (100) over
the wire as with the four circuits. At this speed the minimum

MORSE CIRCUITS.
III
time of closure of the sounder is six segments and the maximum
time nine segments for a dot.
In some cases the four-circuit arrangement, working at
the lower speed, is preferable; and in other cases the higher-
speed operators, sending on a smaller number of circuits,
may be desired, according to circumstances.
While it is not absolutely necessary for working, yet it
is preferable in practice to introduce between the sounder, S,
of Fig. 40 and the polar relay, A2, a repeating-relay, RR, which
is also a polarized relay, wound for a local circuit, having three
terminals, e, f, and g, the terminal f being connected to the
middle point between the two spools and to the local battery,
LB. When the relay-tongue, ti, is against the contact c₁ a cir-
cuit is established through the local battery LB, the wire f,
the wire e, and contact c1, causing the armature t2 of the repeat-
ing-relay to remain against the stop c4. When a main-line
current actuates the tongue t₁ and brings it into contact with
c2 the current in the repeating-relay, RR, flows from the local
battery, LB, and f through the other coil to g, and contact
C2, thus bringing the armature to against the contact c3 to close
the circuit of the sounder.
A METHOD OF DISTRIBUTION OF CIRCUITS.
It has been stated that the circuits obtained by the syn-
chronous principle may be distributed to different localities,
and Fig. 41 illustrates a good method of accomplishing this,
using the principles previously described. Two line wires,
L1 and L2, serve as two trunk lines connecting the cities X and
Y, at each of which there are two sunflowers, M1 serving for
the transmission, and M2 for the reception of messages. These
are represented as eight-segment sunflowers, and are supposed
to have diametrically opposite segments connected together,
as in Fig. 40; but these wires are omitted to simplify the dia-
gram. At the central stations, X and Y, all of the apparatus
except the keys and sounders, which are to be placed at the
outlying stations, may be located. The equipment at each
trunk-line station therefore consists of four transmitters,
KA, KB, KC, and KD, four polar relays, A2, B2, C2, and D2,
with their four corresponding repeating-relays, RR1, RR2,

112
MULTIPLE TELEGRAPHS.
لا
E
E
ليا
E
X'
X2
X3
X4
X6
尚
​wfo
pf
of
&
L'
2
X
To
@HHHHHH-
2-|·|·|·|·|·|·|·|-
6111-
$13
els
RR
Lete
RRA
B2
GO
Glo
M2
Li
L-2
M2
M
22
GOL
Y
or
go
이이​아
​15
19
19
Y2
*7
19
FIG. 41.-Practical method of using the plan illustrated in Fig. 40 for sending four messages in one direction on one wire, L₁,
and four messages in the opposite direction on a second wire, L₂. This plan does not require the use of artificial lines.

MORSE CIRCUITS.
113
RR3, and RR4, two main-line batteries, G1 and G2, and one
local battery, LB.
From this station four pairs of local wires are shown as
leading to four different series of places. One pair of wires
is shown as containing three stations, X1, X2, and X3. These
wires may be long or short according to circumstances, and
contain as many stations as desired. One of the wires, 71,
contains nothing but a series of keys, and the other wire, l2,
nothing but a series of sounders or relays, and no battery is
required at the various stations. If more than one station
is required in each pair of wires, the keys must be provided with
circuit-closers, and worked on the closed-circuit plan. With
only one station, however, the closed-circuit plan is unneces-
sary. The diagram is so complete that it is not necessary to
describe it more in detail.
It is evident that this arrangement may be used as the
equivalent of four single Morse circuits, in which the messages
may be transmitted in either direction. When employed
at its full capacity, however, it is used as four duplex circuits,
for at any station, X2, two operators may be working at the
same time, one sending a message to Y2 and the other receiving
from Y2, for example. The arrangement illustrated is the
equivalent of two quadruplex circuits, which gives a capacity
of eight messages on the two wires. It is a very different
way of accomplishing the same result, and it is not intended.
to discuss here the merits of this synchronous Morse system
as compared with the quadruplex.
USE OF ARTIFICIAL LINES.
The system last described does not involve artificial lines
for obtaining a duplex balance, and if use is made of them
the same result just described as being obtained with two
wires may be obtained with one wire. In this case the capacity
of the line for handling messages becomes the equivalent of
two ordinary quadruplex circuits, allowing the transmission
of eight messages simultaneously, four in each direction.
The manner in which this is accomplished is shown in
Fig. 42, where there are two sunflowers, M1 and M2, at each
station, X and Y, as in Fig. 41. The circuits of the receiving-
sunflower, M2, are controlled over local wires, instead of by

114
MULTIPLE TELEGRAPHS.
ū
-|········
+·········
E
KB
KD
G
GFOA
Leta B
LGTS
LotS D2
a
Bl
IF
X
PR
Tib T21
Σ
M₂
L'
L
_ Ľ
M₂
Σ
Y
PR
M₁
F
TI
a
ド
​T2
BI
Cr
KA
Kc
KD
ALGI
B2
eta
acto
D2
ww
-·|·|·|·|·|·|-·|-·|·
-|·|·|·|·|·|·|·|·|-
E
ū
HilLE
LB₂
FIG. 42.-Octoplex Morse telegraph employing one main-line relay at each station and the common duplex balance
by means of artificial lines.

MORSE CIRCUITS.
115
a main-line wire as before, and the circuits of the transmitting-
sunflower, M1, are connected to the wires leading to the apexes,
a, of the two main-line polar relays PR, one at station X and
one at station Y. Commutator M₁ in each instance may be
fixed and all the adjustment required may be made in the
receiving-commutator, M2, to set it in the best position for
no interference. Each key K at X will then control its cor-
responding relay at Y in the following manner:
Before any key is closed current passes to line from battery
G₁ through the back contacts of each of the four keys at X,
say, to all of the segments of commutator M1 and to apex a
by the trailer T1. This current is not interrupted or reversed
when the keys are all open, and passes to line as a steady current,
but does not affect the home main-line polar relay PR, since
this is balanced for currents from the home battery by means
of the artificial line L¹. It operates the relay PR at Y, how-
ever, and holds its armature constantly against the contact c1,
thus closing the local circuit through battery LB1, each of
the polar relays, A2, B2, C2, and D2 in turn, the trailer T2,
and tongue t to contact C1. This local current through the
relays is in such a direction as to keep the sounders open.
When, however, one key, KA, at X is closed the battery G2
of the opposite polarity is connected to one pair of segments of
the commutator M1, whereas the other three pairs of seg-
ments still remain connected to battery G1. The current
at the apex a at X and in the line will therefore be reversed
at the time the trailer T2 passes over the particular segment
connected to KA, and after the wave has had time to travel
to the other end of the wire the relay PR at Y will move its
tongue t from contact c₁ to contact c₂ during the time of the
passage of the trailer over one segment.
The commutator M2 at Y is so adjusted that the trailer T2
enters the segment connected to A2 at the same instant that
the tongue of the relay responds to the main-line current impulse.
During the time that the tongue remains against the contact c2
the local battery LB2 sends current in the opposite direction
through relay A2 and its segment upon which trailer T2 is just
entering. This opposite current from local battery LB₂ moves
the tongue of relay A2 over against the opposite contact so as
to close the sounder circuit, where it will remain until the
relay again receives a local impulse from the battery LB1.

116
MULTIPLE TELEGRAPHS.
&
This will not occur as long as the key KA continues to send
reversed impulses from battery G2, that is, until the operator
releases the key and terminates his signal.
Assuming that each operator sends at an average speed
of 25 words per minute, the total message capacity for the line
is 200 words, and this result is obtained without resorting to
machine transmission.

INDEX.
Adjusting independent sunflowers.
Adjusting line to speed of trailer.
Advantage of balancing condenser in duplex-diplex.
Advantages of duplex-diplex.
Alternating and direct current in duplex-diplex.
Alternating and direct current in quadruplex.
Alternating current,
applied to polar duplex..
•
frequency in duplex-diplex.
generator.
limitations in closed-circuit working.
receiver in duplex-diplex. .
repeater.
in duplex-diplex..
value..
A
decrease along wire. . .
division in duplex-diplex receiving apparatus.
effect of line capacity.
effect on telephones. ...
Battery
Balanced condenser, polar diagram.
Balancing coils..
construction.
at each end of wire.
at one end of wire only...
..
transmission on closed-circuit plan.
Apparatus for experimenting with synchronous telegraph circuits.
Applications of duplex-diplex system..
Artificial lines in synchronous telegraphs.
Automatic compared with synchronous systems.
B
..
•
..
.
..
PAGE
102
IOI
10
25
5
37
30 00 00 38
37
19
10
8
40
8
32
20
6
30
20
71
27
113
105
00400
9
14
9
19
18
117

118
INDEX.
N
Charge of line, escape of. ...
Closed-circuit working
Calculation of received current by Kelvin's formula.
Capacity of telegraph wire.
Central station with subscriber's stations.
C
for direct-current side and open circuit for alternating-cur-
rent side. ...
in duplex-diplex. .
limitation on alternating-current side.
transmission on alternating-current side.
Comparison between
closed-circuit and open-circuit working.
polar duplex and duplex-diplex.
synchronous and automatic systems.
Compensator in duplex-diplex.
Condenser
Current
...
...
balanced, impedance of. .
balanced, polar diagram.
discharge, effect on Morse relay.
in duplex-diplex, value of.
unbalanced, polar diagram..
Construction of special form of two-pole synchronous motor.
Continuity-preserving transmitter in duplex-diplex.
Control of synchronous motors by relay contacts..
Current, effect of leakage on
with divided battery.
with single battery..
•
• ".
..
in long wire having maximum leakage.
in sunflower segments in synchronous telegraph.
in various branches, duplex-diplex receiving apparatus. . . .
Curves showing
received current.
single-current waves.
...
D
Damping out oscillations by mercury-cup...
Decrease of alternating current along wire. ...
Difference between closed-circuit and open-circuit working..
Difference between real and artificial lines..
Differing instrumental and leakage conditions..
Difficulties in synchronous telegraphs.
•
•
..
...
PAGE
85
73
33
94
..
– ទីម ៦ ៖ ៖៦៨
14
IO
IO
16
9
53
24
.52, 61
II
19
18
033340
49
73
IO
85
87
54
19
18
92
69
Direct- and alternating-current quadruplex.
Direct current, effect of leakage on...
37
18
Direct current, receiver in duplex-diplex.
6
Distance of working increased by load coils. . . .
36
Distinction between synchronous and non-synchronous telegraphs.. 69
Distribution of synchronous circuits.
III
Divided line battery.
19

INDEX.
119
Double currents in synchronous telegraphs..
Duplex-diplex
adaptation to two single Morse wires...
adapted for operating synchronous motors.
alternating-current receiver.
..
applications of system.
closed-circuit working.
closed-circuit working, alternating-current side.
comparison with polar duplex. .
complete working station.
continuity-preserving transmitter.
difference between closed-circuit and open-circuit working...
diplex working only, for very long-distance working.
direct-current receiver.
division of alternating current in receiving circuits.
duplex system only, with way stations.
frequency of alternating current.
limiting line.
...
Earthing segments.
Effect of
low-frequency system..
means for obtaining..
most probable disposition of apparatus..
open-circuit systems.
receiver on alternating-current side.
receiver on direct-current side.
...
..
..
receiving apparatus.
reception of signals.
resistance branch of receiving apparatus.
resonant circuit, avoidance of...
resonant circuit on alternating-current side.
station for direct current only...
subscribers' stations..
transmitting apparatus.
two independent circuits..
value of balancing coils.
value of condenser.
way stations on direct-current side and through alternating-
current circuit...
E
***
Duplex-diplex system adapted for operating synchronous motors..
Duplex Morse wire with synchronous motors. ...
alternating current on telephones.
condenser discharge on Morse relay.
.
instruments and leakage in synchronous telegraphs.
line capacity on alternating current. ...
line leakage in synchronous telegraphs.
line leakage on direct current.
opening circuits from segments.
PAGE
106
23
56
15
27
16
20
IO
22
24
18
36
6
IO
23
8
35
40
4,5
25
23
6
6
5
15
. 10, 13
7
7
.16, 17
33
16
33
IO
9
25
56
61
103
40
16
92
..8, 19
95
18
92

I 20
INDEX.
v
Eight circuits on one wire. .
Escape of line charge..
Experimental apparatus for synchronous telegraph measurements.
Explanation of measurements.
Finding the circuit. . . .
Fly-wheel with mercury to damp out oscillations.
Formula of Lord Kelvin for received current.
Frequency of alternating current in duplex-diplex.
Impedance of
F
General considerations on synchronous Morse circuits.
Generator, alternating-current, for duplex-diplex.
Grounded segments..
G
High-frequency synchronous motor..
Hunting of synchronous motor damped out by mercury-cup.
ing..
Inductance coils in duplex-diplex
H
balanced condenser in duplex-diplex.
150-ohm Morse relay.
at receiving end..
at transmitting end.
I
Independent circuits in duplex-diplex.
Independently adjustable sunflowers.
Independent Morse circuits on same wire.
Inductance coils as load coils, to increase possible distance of work-
.
to balance receiving condensers. . .
to prevent flow of alternating current.
Influence of receiver coils in synchronous telegraphs.
Installation of duplex-diplex on railroad wire. .
Instrumental conditions, effect on currents in synchronous tele-
graphs.
Interfering current in adjacent segments
K
Kelvin's formula for submarine cables.....
L
Lag in current arrival due to length of wire.
Lag of receiving commutator. .
.
PAGE
113
94
71
85
...
co court to
48
54
85
105
32
103
58
54
17
IO
33
IOI
3
36
9
.21, 24
94
27
92
77
77
85
....51, 77
89

INDEX.
121
Leakage
effect on alternating current..
effect on current in long wire. .
effect on direct current..
effect on synchronous telegraphs.
maximum permissible..
of lines, in synchronous telegraphs..
Leased wires, two obtainable from duplex-diplex.
Limiting distance of working..
Line adjusted to speed of trailer...
Line capacity
Line charge, escape of..
Line leakage
effect on alternating current.
of overhead wire. .
..
****
effect on alternating current.
effect on direct current.
in synchronous telegraphs.
Load coils to increase distance of working..
Long-distance working.
by use of load coils.
by using diplex only..
Low-frequency duplex-diplex system.
Low-frequency synchronous motor...
Morse circuits
M
Magneto-generator, eight-pole...
Magneto-motor, two-pole..
Maintenance of synchronism on wire used solely for this purpose.
Maximum allowable leakage.
Maximum speed for circuit in synchronous telegraphs.
Measurements in synchronous telegraph circuits.
Measurements of synchronous telegraph currents.
Mercury-cup to damp out oscillations of motor.
Method
advantages of single.
duplex.
..
of controlling synchronous motors by relay contacts.
of distribution of synchronous circuits..
of measuring synchronous telegraph currents.
of obtaining synchronism.
in synchronous telegraphs..
two independent on same wire.
Morse 150-ohm relay, impedance of.
Morse operator, speed of....
Morse system
•
Morse wire with synchronous motors..
Motor operated on same wire used for telegraphing.
Motor, two-pole synchronous, construction of.
Multiplex wire with synchronous motor on same wire.
•
PAGE
25
49
18
..
92
49
95
33
36
99
8, 19
73
94
..92, 95
36
25
18
www w
36
36
40
53
32
53
47
49
IIO
71, 76, 85
73
54, 61
3333
.52, 61
III
73
45
105
3
IO
109
3
3
56, 61
56, 61, 62
633
53
62

122
INDEX.
س
Number of circuits in synchronous telegraphs..
N
O
Open circuit for alternating-current side and closed circuit for direct-
current side....
Open circuits from segments..
Open-circuit working in duplex-diplex system.
by relay contacts.
from a central point.
Operating diplex only.
Operation of sounder in synchronous telegraphs.
Operation of synchronous motor
on duplex Morse wire. .
on single Morse wire.
on synchronous multiplex wire.
on wires also used for telegraphs.
on wires devoted solely to this purpose.
Quadruplex,
P
Particular systems of synchronous Morse telegraphs. .
Pennsylvania railroad installation of duplex-diplex.
Pittsburgh-Toledo wire with duplex-diplex system.
Polar diagram for balanced condenser..
Polar diagram for unbalanced condenser..
Polar duplex compared with duplex-diplex.
Polar duplex with alternating current added.
Polar relay as repeating sounder..
Polar relay in synchronous telegraphs..
Polar relay used to control synchronous motor.
Problems involved in synchronous telegraphs.
on wires having heavy leak...
Opposite transmissions on the same wire in synchronous telegraphs. 97, 99
Oscillations of synchronous motor prevented by mercury-cup.
..
·
calculated by Kelvin's formula.
Q
means for obtaining, with alternating and direct current.
alternating and direct current.
R
Railroad wire equipped with duplex-diplex.
Rate of transmission in synchronous telegraphs.
Reasons for divided line battery.
Received current
..
...
...
curves.
Receiver coils, influence of, in synchronous telegraphs...
Receiving apparatus in duplex-diplex system.
Reception of signals in duplex-diplex.
PAGE
IIO
..
fotograf EUNOR
25
105
27
27
IO
II
33
37
III
106
52
45
4
37
27
IIO
19
85
85
94
5
15

INDEX.
123
Reduction of alternating current along wire. ...
Relay contacts used to control synchronous motors..
Repeater for alternating current..
Repeater station in wire equipped with duplex-diplex.
Repeating polar relay.
Repeating sounder in duplex-diplex.
Resistance branch, duplex-diplex receiving apparatus.
Resonant circuit in duplex-diplex. .
**
Reversed currents in synchronous telegraphs.
Segments
for grounding line. . .
with open circuits...
Signals in duplex-diplex.
Single-current waves shown by curves. .
Single Morse wire with synchronous motors.
Sixteen-pole synchronous motor.
Special form of two-pole synchronous motor..
Speed
S
Separate wires for transmission in opposite directions.
Separation of alternating and direct current in duplex-diplex...
Side-line repeater in wire with duplex-diplex. .
of Morse operators. ..
per circuit in synchronous telegraphs.
Subscribers' stations, in duplex-diplex..
Sunflowers independently adjustable.
Superiority of synchronous systems.
Synchronism at distant points, methods of obtaining.
Synchronous compared with automatic systems.
Synchronous motors
controlled by relay contacts.
for high frequency.
for low frequency....
on duplex Morse wire.
on single Morse wires..
on synchronous multiplex wire.
on wires also used for telegraphs.
on wires having heavy leak.
on wires used solely for this purpose.
oscillations prevented by mercury-cup.
sixteen-pole..
advantages of..
limitations of..
measurements in circuits.
problems involved..
systems of circuits..
two-pole, construction of special form of..
Synchronous multiplex wire, with motors on same wire.
Synchronous telegraphs
•
PAGE
19
...52, 61
.
..
15
...10, 13
..
30
27
III
7
106
2005 555
103
92
106
27
15
87
56
58
53
109
IIO
33
IOI
105
45
105
...57,61
58
53
61
56
62
56
49
47
54
58
53
62
105
69
.71, 76
45
105

124
INDEX.
T
Telegraphing and operating motors on same wire.
Telephones, effect of alternating current on...
Theoretical curves of received current.
Through working with alternating current on way-station wire. . . ..
Time lag due to length of wire. ..
Time necessary for transmission of signal.
Transmission in opposite directions on same wire in synchronous
telegraphs.....
Transmission of alternating current on closed-circuit plan...
Transmission on separate wires in opposite directions.
Transmitter, continuity-preserving, in duplex-diplex.
Transmitting apparatus in duplex-diplex system. .
Transmitting station for direct current only, in duplex-diplex.
Two independently adjustable sunflowers.
Two-pole synchronous motor, construction of.
U
Unbalanced condenser, polar diagram.
Use of
Waves of received current
at frequency of 150.
at frequency of 61.6.
from measurements.
theoretical.
artificial lines in synchronous telegraphs.
double currents in synchronous telegraphs.
duplex-diplex to obtain synchronizing circuit.
grounded segments.
load coils in wires.
.. ..
W
V
Value of balancing coils in duplex-diplex. .
Value of resistance branch in duplex-diplex...
Voltage employed in duplex-diplex working installation.
·
on same wire.
on separate wires.
Working installation of duplex-diplex.
Working long distance by diplex...
• • ..
..
points.....
Wire with maximum value of leakage..
Working by means of independently adjustable sunflowers..
Working in opposite directions in synchronous telegraphs
•
25
.51, 77
... 109
•
.
...97,99
Way-station wire, with through alternating-current circuit. ....
Wires devoted to maintenance of synchronism between distant
.
20
106
24
.16
.16, 17
IOI
53
.
PAGE
56
40
85
.
.
II
113
106
56
103
36
IO
IO
32
2555 49
25
IOI
.97,99
106
27
30

Sug li
UNIV. OF MICH.
MAR 30 1906
RECEIVED
Engin. Library
TK
5535
.C91
145222
UNIVERSITY OF MICHIGAN
3 9015 07508 5947
Crehore
Synchronous and
other Multiple
Telegraphs