US2856457A - Printing telegraph distortion indicator - Google Patents

Description

Oct. 14, 1958 H. 'r. PRlOR ET AL PRINTING TELEGRAPH DISTORTION INDICATOR Filed June l7. l953 8 Sheets-Sheet 1 F/GJ.

v U L2 1.3 u/ S l5 l0 5 O 5 10/5 (5) WWW- m)mew-90) 5) -5)?) I W i W EA/PLY LATE lnven tar H.T. PRIOR E.A. FOULKEE WwgJ Attorney Oct. 14, 1958 H. T. PRIOR ETAL 2,856,457

PRINTING TELEGRAPH DISTORTION INDICATOR Filed June l'7. 1953 8 Sheets-Sheet 2 Inventor HI PRIOR E.A. FOU L KES A ttorne y H. T. PRIOR ETAL PRINTING TELEGRAPH DISTORTION INDICATOR Oct. 14, 1958 8 Sheets-Sheet 3 Filed June 17, 1953 STOP lzo mill/secs -F/G./O.'

S .E R r o U .IR m. WIA 5 m C l s wm m m w 4 y 0 3 8 BWS .3 l. 2 w. i u 5 6 G 2 4 2 de 3 N 2 M lm 0 v F O w w Oct. 14, 1958 H. T. PRIOR ET AL PRINTING TELEGRAPH DISTORTION INDICATOR 8 Sheets-Sheet 4 Filed June 17, l953 01 m6 mm 1 o v w \k Q I wok wmmamw QWSQ Q mQmk m Or momkw Inventor H:T.R;RlOR- E.A.FOULKES Attornev Oct.- 14, 1958 H. 'r. PRIOR ETAL 2,356,457

PRINTING TELEGRAPH DISTORTION INDICATOR Filed June 17.1953 8 Sheets-Sheet 5 Invenior HJZ PRIOR EA.FOULKES Wwfi Attorney Oct. 14, 1958 H.- T. PRIOR ET AL PRINT ING TELEGRAPH DISTORTION INDICATOR 8 Sheets-Sheet 6 Filed June l7, 1953 QOKWQQEMO mmubQ Inventor H.T. PRIOR EA.FOULKES Q m Q V Attorney 8 Sheets-Sheet 7 nmpmoa E.A. FOULKES A ttorne y Oct. 14, 1958 H. T. PRIOR ETAL PRINTING TELEGRAPH DISTORTION INDICATOR Filed June 17, 1953 United States Patent PRINTING TELEGRAPH DISTORTION INDICATQR Hector Thomas Prior and Edward Archibald Foulkes, London, England, assignors to International Standard Electric Corporation, New York, N. Y.

Application June 17, 1953, Serial No. 362,304

Claims priority, application Great Britain June 26, 1952 9 Claims. (Cl. 178-69) The present invention relates to apparatus for measuring telegraph distortion in telegraph signalling systems.

Telegraph transmitters usually transmit to the line one or the other of two different current values, one of which is commonly called the marking current and the other the spacing current. The line is said to be in the marking or spacing condition, according as marking current or spacing current is transmitted. A character or other information to be signalled is represented by a group of marking and spacing periods whose durations are determined according to a pre-arranged code.

In the single current system, the transmitter sends a current of fixed value and sign to the line during marking periods and zero current during spacing periods, though in some cases these two conditions may be reversed. in the double current system, the transmitter sends a current of one sign to the line during marking periods and an equal current of the opposite sign during spacing periods.

In machine telegraph systems, a binary code is commonly employed in which each code group consists of the same number (commonly of code elements of equal duration, and each code element may correspond to either the marking or the spacing condition of the line. As a result there is transmitted to the line a series of substantially rectangular pulses the beginnings and ends of which are marked by alternate transitions from mark to space and space to mark. At the transmitting end of the line, these transitions are practically instantaneous, but after transmission over the line, the rectangular pulses may become seriously rounded, and distorted by crossfire and interference, so that the transitions are no longer instantaneous, and the times at which a receiving relay is changed over in response to the received waveform no longer correspond with the times of the original transitions, and telegraph distortion results, and the receiving machine may print the characters incorrectly.

A number of telegraph distortion measuring apparatus have been proposed hitherto in which the transitions are recorded continuously on a cathode ray tube provided with a scale on which the corresponding distortions can be read. Cathode ray tubes are inconvenient on account of the high voltage necessary for their operation, and the value of the distortion is difficult to read by unskilled operators.

The object of the present invention is to provide an arrangement for measuring telegraph distortion which does not employ a cathode ray tube, and in which the percentage distortion is indicated directly on a limited number of indicators within a specified degree of accuracy, preferably by means of a range of indicating glow lamps which form a scale of percentage distortion.

The invention accordingly provides an arrangement for measuring the distortion of telegraph signals representing information according to a binary code, comprising means for receiving an electric wave representing a group of code elements and comprising marking and spacing periods separated by transitions, a plurality of 2,856,457 Patented Oct. 14, 195.8.

'2 separate indicating devices each of which corresponds to a different range of values of telegraph distortion, and means controlled in response to a transition for operating the particular indicating device within the range of which the distortion of the transition lies.

Machine telegraph systems are generally of two types, namely the start-stop system, such as is generally used for teleprinters, and the synchronous system, for example the Baudot multiplex system. In start-stop systems, each group of code elements is preceded by a start element which is always a spacing element, while the group is always terminated by a transition to the marking condition. The start element is generally of the same duration as the code elements, and the final marking or stop condition persists for a period usually not less than one code element period, before another character can be transmitted.

Thus the idle or rest condition of the line when nothing is being transmitted, is .a marking condition, but it is necessary to explain that in some teleprinter exchange systems, a line not connected to a teleprinter is left in the spacing condition (sometimes called the false rest condition) and is changed over to the normal marking or rest condition when the connection is set up.

The embodiments which will be described to illustrate the invention are primarly designed for start-stop systems,

but can be modified in a simple way to deal with synchronous systems. For clearness it will be assumed that the speed of operation is 50 bands and that the start interval is equal to one code element period. Accordingly the total duration of each code group will be milliseconds, divided into 6 equal parts, commencing with a mark-to-space start transition and ending with a spaceto-mark stop transition, with various numbers of intervening transitions, according to the code. no telegraph distortion, the transitions occur at instants which are exact integral multiples of 20 milliseconds after the initial or starting transition; if a transition occurs 1 milliseconds before or after the prescribed instant, then which will be described are not limited to operation at. a speed of 50 bands, and it is very easy to adjust the.

circuits for other speeds, as will be explained later.

The invention will be described with reference to the accompanying drawing in which:

Fig. 1 shows a block schematic circuit diagram of .an embodiment of the invention;

Fig. 2 shows details of a portion of the indicating circuit of Fig. 1; t

Fig. 3 shows diagramsand a circuit used to explain the meaning ofcertain symbols employed to simplify the circuits of embodiments of the invention;

Figs. 4 and 5 taken together show .a detailed schematic circuit diagram of an embodiment which diflers in ,minor points from that shown in Fig. 1

Figs. 6 and 7 taken together show a detailed schematic circuit diagram of a modified form of the embodiment of Figs. 4 and 5, arranged to measure distortion to a higher degree of accuracy;

Figs. 8 and 9 show circuit details of certain elements of Figs. 4 to 7; and

Figs. 10 and 11 are timing charts used in explaining the operation of the embodiments of Figs. 4 to 7.

In describing the embodiments of the invention, some circuit symbols representing certain devices which gen- If there is.

erally occur several times will be used to make the circuit easier to understand. These symbols will be explained with reference to Fig. 3 of the accompanying drawings. In order to avoid complicating the circuits, frequently only the input and output ends of connecting conductors are shown, which ends are provided with arrows showing the direction or transmission, and the output end of each such conductor is given a designation corresponding to the input end.

In the circuits, a large number of circuit elements called gates and designated G are used. These may be circuits of the kind described and claimed in the copending application of E. M. S. McWhirter et al., Serial No. 90,326, filed April 29, 1949. Y The symbol for such a gate is shown in Fig. 3(a) and consists of a circle having a single arrow directed away from the circle and representing the output, and any number 11 of further arrows directed towards the circle representing a corresponding numberof separate inputs over which control potentials are applied. In the circle a numeral x equal to, or less than, n is inserted, and this numeral indicates that an output potential appears on the output conductor provided that potentials appear on any x of the 11 input conductors. It will be evident that when'the control potentials consist of input pulses of various durations, the output from the gate will be a pulse of duration substantially equal to the period of simultaneous overlap of all the input pulses.

' Fig. 3(b) shows a symbol, designated F, for a circuit which will be called a multi-condition device. It consists of a column of any number of on-ofi? elements or stages such as gas-filled tubes, each represented by a square, each square being separately designated by a letter or letters. The arrangement is such that the application of a suitable pulse to any stage over an input conductor, represented by an incoming arrow, switches it from off to on (for example, causes a gas-filled tube to conduct) and at the same time switches off any other stage which is on (for example, causes any other conducting tube to become non-conducting). The switching on of any stage causes a potential to appear on the corresponding output conductor represented by an outgoing arrow.

Fig. 3(a) shows details of an example of a triple stable device according to Fig. 3(b). It consists of three gasfilled trigger tubes T T T the anodes being connected tothe positive terminal of the high tension source (not shown) through corresponding load resistors R R R and the cathodes through resistor-ca acitor networks N N N to the negative terminal of the source. Each anode is connected to all the other anodes through respective capacitors Q Q Q... The input conductors are connected respectively to the trigger e ectrodes, and the output conductors to the cathodes. When any tube is fired by the application of a positive pulse to the trigger electrode, a positive potential appears on the output conductor connected to the cathode, and a negative pulse is appliedto each of the other anodes through the corresponding capacitors, thereby extinguishing any of the other tubes which may be conducting. For a bi-stable device, one tube and its associated elements is omitted, and only one of the capacitors Q Q Q is required.

A modified bi-stable device whose symbol, designated FF, as shown in Fig. 3(d) is used. Its circuit is the same as Fig. 3(a), except that the capacitor Q is replaced by a direct connection, and Q is omitted. The tubes T and T normally operate as a bi-stable device according to Fig. 3(b), but the extra tube T which operates in parallel with the tube T is used as an alternative tube, the firing of which is employed in special circumstances for extinguishing T Fig. 3( e) shows the symbol, designated FP, for a twostage singly stable device, also called a flip-flop device. It is similar to the device F shown in Fig. 3(b) except that the stage X is coupled back to the stage Y The stage Y is normally fired or on, and when X is switched on or fired by an applied pulse, it switches Y ofl, but the coupling is such that the circuit will not remain in this condition, and reverts to the normal condition after a time determined by the time constant of the coupling circuit.

A counting device having any number of stages is shown in Fig. 3(]) as a horizontal row of numbered squares, designated C. The device, may, for example, consist of a multi-cathode gas-filled tube with associated circuit, as shown for example in the co-pending application of V. Terry et al., Serial No. 235,068, filed July 3, 1951, which issued as U. S. Patent No. 2,788,940, April 16, 1957.

Triggering pulses are applied over an input conductor P to one end of the row, the arrangement being such that only one stage is fired at a time, and each triggering pulse causes the next stage to be fired, and the preceding stage to be extinguished. When the end of the row is reached, the next pulse causes the first stage (designated 0) to be fired, and so on indefinitely, so long as pulses are supplied. A resetting pulse may be applied to one of the stages over an input conductor R to restore thecounter to some specified initial condition. When any stage is fired, a positive potential appears on the corresponding output conductor.

Referring now to Fig. 1, start-stop telegraph signals, the distortion of which is to be measured, are applied to input terminals T. The telegraph speed ;will be assumed to be 50 bauds. When a mark-to-space transition occurs, a potential is applied to the conductor designated space; and when a space-to-rnark transition occurs, a potential is applied to the conductor designated mark.

When the start (mark-to-space) transition occurs, it

supplies a potential over the conductor designated space to change over a two-condition or double-stable device F1 from the stop to the start condition.

When the start stage of F1 is switched on, it supplies a potential to the gate G1. As long as an input is supplied to the gate from the start stage of F1, pulses from an oscillator or pulse generator 01 are passed to a counting circuit C1 which is shown diagrammatically as a block. This counting circuit is preferably of the type described in co-pending application No. 235,068 already referred to, which issued as U. S. Patent No. 2,788,940, April 16, 1957.

10 milliseconds after the receipt of a start transition,

a pulse is supplied from the counting circuit C1 to a twocondition device F2 similar to F1. This pulse changes over F2 into the start condition which in turn means that a potential is supplied to a gate G2 similar to gate G1. Gate G2 on being thus opened, allows pulses from a second oscillator or pulse generator 02 to pass to a second counting circuit C2. This counting circuit may comprise one or more multi-cathode gas-filled discharge devices of the type described in co-pending application No. 235,068 already referred to.

- The counting circuit C2 acts as a distributor for the pulses from scillator O2, and the frequency of the latter will be assumed to be such that potentials are applied to the twenty output leads L1-L20 from the counting circuit (or distributor) in sequence from left to right, one at a time, for one millisecond. As long as gate G2 remains open the counting cycle of C2 is continuous; that is, the application of a potential to the last output lead L20 on the right is followed by that of the first leadLl on the left.

The output leads are taken to respective pairs of tubes in the double row of gas-filled dischar 'e tubes or lamps indicated by the blocks RNl and RNZ. The tubes in these rows are preferably neon tubes, but other tubes may be used provided that the occurrence of conduction is accompanied by a visible glow.

The potential applied oyerany one of leadsL is a priming potential, and is not of itself sufficient to cause conduction (hence glow) of either of its associated pair of tubes.

The input terminals T are connected to a two-condition device F3 whose function is to produce transitional pulses .at times corresponding to the transitions of the input signals. Pulses corresponding to space-to-mark transitions are applied in common to the tubes of row RN2 and pulses corresponding to mark-to-space transitions are applied in common to the tubes of row RNl.

Considering, for a moment, only the space-to-mark transitions, each time a pulse is applied to row RNZ from device F3, it will coincide with the priming of one tube of the row over one of the leads L from the counter C2. When this occurs, the tube in question will glow for a reason now to be explained by reference to Fig. 2.

In Fig. 2 there are shown at N18 and N19 the second from last and the next to last tubes respectively of row RNZ. These tubes have only been selected at random for illustrative purposes.

It will be assumed that the distributor or counting circuit C2 takes the form of one or more multicathode tubes, two of the cathodes being indicated at K18 and K19 together with their associated circuits comprising resistors R1 and R2, respectively, and capacitors Q1 and Q2 respectively.

The cathodes K18 and K19 are connected to the trigger or priming electrodes of the tubes N18 and N19 over conductors L18 and L19 and through rectifiers X1 and X2. The conductor S/M from the device F3 (Fig. 1) is connected to the cathodes K18 and K19 through capacitors Q3, Q4 and resistors R3, R4 as shown.

It will further be assumed that when a pulse is received corresponding to a space-to-mark transition it occurs at a time when a discharge is present in tube C2 between the common anode A and cathode K19.

The junction point between rectifier X1 and resistor R3 is substantially at earth potential since no current flows to cathode K18. Consequently the trigger electrode of tube N18 remains at earth potential. The pulse applied from F3 over capacitor Q3 is absorbed in resistor R3 and tube N18 cannot be fired.

Since a discharge is present between anode A and cathode K19 of tube C2, current is flowing through cathode :load resistor R2. In consequence the cathode is at a positive potential above earth, but this potential should be insufficient by itself to cause the tube N19 to be fired. The pulse applied through capacitor Q4 then passes through resistor R4 and rectifier X2, and increases the potential already applied by cathode K19 so that the tube N19 is now fired.

In order to extinguish the fired tube after it has glowed for a long enough time for it to be properly observed, a delay device D is incorporated in the anode supply of all the tubes in the rows. This delay device is timed (by, for example, the discharge of a condenser) to reduce the H. T. supply voltage when the lowing tube has been sufficiently observed. The voltage is immediately restored, but no tubes can be re-firecl until another transition occurs.

When the tube N19 has been fired, the trigger electrode will have applied to it a positive potential due to the discharge between the anode and cathode. This would tend to charge the capacitor Q2, which would interfere with the proper operation of the arrangement. Accordingly therectifier X2 is interposed in series with the connection to the trigger electrode and is directed so that it will be blocked when the tube is fired, thus preventing the capacitor Q2 from being charged. Rectifier X1 performs the same function for the tube N18.

Referring back to Fig. l, at 150 milliseconds the counting circuit Cl emits a pulse which switches on the stop stage of the device F1. As the device Fl. changes over it removes the input from gate G1 and resets C1 to its initial condition ready for the next combination. The pulse also switches on .the reset stage of the two-condi- 6 tion device F2 which in turn removes the input from gate G2 and resets C2 to its initial condition.

The tubes in the upper row RNl of Fig. l are also arranged as shown in Fig. 2, and will be operated in response to mark-to-space transitions in a similar way.

The arrangements described provide for the examination of distortion up to :50% i. e. transitions occurring 10 milliseconds early or late. The figures in the blocks of row RNZ of Fig. 1 indicate the percentage distortion and the shaded square in this row represents the glowing of a neon tube to indicate a distortion of 25% i. e. the transition from space-to-mark is occurring 5 milliseconds early. Similarly the mark-to-space transition is shown as arriving 1 millsecond late i. e. +5% distortion.

It will be observed that the pair of tubes in the centre (i. e. O) are not supplied from the counter C2. Thesetubes may be left glowing all the time to indicate the centre of the display or alternatively they may be arranged to glow only if the distortion is exactly zero.

If desired, the oscillator 02 could be: replaced, using simple switching, by one having a higher frequency say ten times greater. The range of percentage distortion thus catered for would then be reduced to :5 percent, as shown by the bracketed figures under the blocks of row RNZ. This would enable much more accurate reading to be made.

Furthermore it is possible by simple switching to allow only the transitions in one direction, say space-to-mark, to pass to the indicator at a time to facilitate differentiation between the two types of transition.

Reference will now be made to Figs. 4 and 5, which taken together show details of the preferred embodiment of the invention. The conductors which connect the elements of the two figures are not shown complete, but only their input and output ends, which are given the same designations so that they can easily be identified. Thus for example, the conductors M and S from the output of the device QC in Fig. 5 are used in Fig. 4, where their output ends are applied to gates G12, G14 and G10, G13, and are designated M and S. Likewise, for example, the conductors +1 and 'P from the generator O3 in Fig. 4 are connected in various places in Fig. 5. The same principle is used for many of the connections between the elements in either figure alone.

Fig. 4- contains the elements concerned with the control and operation of the counting devices, while Fig. 5 contains those concerned with the indication of the telegraph distortion under the control of the counting devices.

The distortion measuring circuit is controlled by a stable pulse generator 03 (Fig. 4) which generates on separate output conductors +P and P two trains of short, substantially rectangular, pulses of duration, for example, 15 microseconds. The pulses of both trains are regularly repeated at the rate of 5,600 pulses per second (on the assumption that the speed of the telegraph signals is 50 bands). The pulses of one train are all positive, and those of the other are all negative, and positive and negative pulses occur simultaneously. The positive pulses are delivered to the output conductor +P, and the negative pulses to the output conductor P.

The generator 03 should preferably be crystal controlled, and should also preferably be provided with means whereby the frequency can be adjusted over a small range on either side of the mean or specified value.

The circuit employs a series of counting devices C3, C4, C5 and C6 which are controlled by the pulse generator O3 and are arranged on similar lines to the counting devices described in said co-pending application No. 235,068 already referred to, but modified to suit the requirements of the circuit. The counter C3 is a fivepoint counter and may be produced by means of a tencathode gas-filled tube of the type shown in the specification just referred to, by coupling the cathodes in pairs,

7 five stages apart; thus cathodes and 5, 1 and 6 and so on, are coupled.

'Twoten-point counters C4 and C are arranged in a manner to be explained, to form, in effect, a twentypoint counter operated from stage 4 of counter C3. A third ten-point counter C6 is operated from stage 9 of counter C5. Thus counter C3 divides by 5, so that the 20 stages of counters C4 and C5 come on in turn, each for a period of approximately 1 millisecond, and the ten stages of counter C6 come on in turn, each for a period of approximately 20 milliseconds. In the case of the counter C6, only the output of stage 6 is used.

Before explaining how the telegraph signals, whose distortion is to be measured, are received and dealt with, it will be convenient to explain the operation of the counters, which then need not be again referred to. When the start transition of a code group is received, the device FF is operated, and the stage ST of FF delivers a positive potential to the two-control gate G3 to which is also applied the negative pulses P from the generator 03. These pulses P then can pass to the counter C3 and step it on continuously one step at a time, until the stage 4 is reached, when a positive pulse of duration something less than 200 microseconds is delivered to the three-control gate G4, the output of which is connected to the counter C4. To this gate are also applied the pulses P, and a positive potential from the stage MA of a bi-stable device F4 which stage is normally on, the stage MI being normally off. Thus every time a pulse P coincides with a pulse from stage 4 of counter C3 it can pass through the gate G4 to operate the counter C4.

It should be explained, however, that the stages of counter C3 have associated with them, circuits with a time constant which is appreciable, so that the positive pulse generated by each stage does not immediately attain its maximum amplitude. Accordingly, the pulse P which operates the counter C4 is the one which occurs at the end of the period of the positive pulse generated by stage 4, and which at the same time steps the counter C3 back to stage 0. The previous pulse P which steps the counter C3 from stage 3 to stage 4 finds the gate G4 shut because the pulse from stage 4 of counter C3 has not had time to open the gate. Thus the stepping on of counter C4 always coincides with the stepping of counter C3 back to stage 0.

It follows that every fifth pulse -P steps the counter C4 along one stage, each stage producing a positive output potential which lasts nearly 1 millisecond, there being a slight delay in reaching the maximum amplitude on account of the time constant associated with the stage. This delay is, however, small compared with the repetition period of the pulse P. When stage 9 of counter C4 is reached, a positive pulse will be applied to a three-control gate G5 to which are also applied positive pulses +1 from the generator 03 and positive pulses from stage 4 of counter C3. The particular positive pulse -}-P which occurs at the end of the period when a pulse from stage 4 of counter C3 coincides with the pulse from stage 9 of counter C4 finds the gate G5 open and is able to reach the device F4 which is thereby changed over, thus at the same time shutting the gate G4. From what has been explained above, it will be clear that the device F4 is changed over at the same time as the counter C3 is stepped back to stage 0.

The stage MI of F4 now being on, a positive potential is applied to the gate G6 to which are also applied the pulses P, and a pulse from stage 4 of the counter C3. Counter C5 is now stepped along instead of counter C4, and in just the same way, each stage remaining on for approximately 1 millisecond until stage 9 is reached, the output of which is applied to a three-control gate G7 arranged similarly to G5. The output of G7 supplies a positive pulseto the stage MA of F4 and changes itback 8 again so that now counter C4 isstepped along as 'before. It will be seen therefore that the counters C4 and C5 operate alternately.

Stage 9 of counter C5 supplies a positive pulse to a three-control gate G8, to which are also supplied the pulses P and positive pulses from stage 4 of counter C3. When, therefore a negative pulse P occurs at the end of the period when pulses from stage 4 of C3 and stage 9 of C5 coincide, it can pass to the counter C6. It follows that since the counters C4 and C5 operate alternately, the counter C6 is stepped along at intervals of milliseconds.

Thus it will be seen that on receipt of the start transition, the counters C3 to C6 are stepped along continuously until the action is terminated at the end of the code group in a manner which will be explained later. Positive pulses are available from the stages of C3 at intervals of 200 microseconds; from the stages of C4 and C5 at intervals of 1 millisecond; and from C6 at intervals of 20 milliseconds.

It will be assumed that the telegraph system concerned is a double current system. The incoming telegraph wave is applied to an input terminal T connected to a quantizer circuit QC (Fig. 5), details of which are shown in Fig. 8. To the quantizer circuit are also supplied the pulses i-P and P from the generator 03. The quantizer circuit QC applies a positive potential to an output conductor M during marking periods, and a positive potential to a separate output conductor S during spacing periods. The change-over of the positive output potential between the conductors M and S occurs substantially simultaneously with the first pulse +P or P to appear after the potential of the telegraph wave applied at terminal T changes sign. Thus in the normal rest condition, when the potential applied to terminal T is negative, there will be a positive potential on conductor M; when the first (or start) mark-to-space transition occurs, the positive potential will be transferred to conductor S after a slight variable delay which cannot exceed 200 microseconds. On the occurrence of the next (space-to-mark) transition the positive potential is transferred back to conductor M again after a delay less than 200 microseconds, and so on.

At the moment when the positive potential is transferred from one of the conductors M or S to the other, a short positive triggering pulse is also generated in the quantizer circuit QC, and this pulse is delivered to an output conductor W. A three-point switch SW, which is diagrammatically shown, enables only the triggering pulses corresponding to mark-to-space transitions, or only those corresponding to space-to-mark transitions, to be separately selected, or alternatively enables both kinds to be selected. The details of this arrangement are shown in Fig. 8.

The purpose of the quantizer circuit QC is to bring the transitions effectively into phase with the counting chain, and does not introduce any appreciable error, since the circuit is not designed to recognise timing errors less than 200 microseconds, which corresponds to a telegraph distortion of l'percent at bauds.

The output conductor W (Fig. 5) is connected 'to a two-control gate G9 through which the bank of glow lamps RN on which the distortion is indicated is ener gised in a manner which will be explained later on.

The starting of the operation of the circuit is controlled by the two-condition device FF (Fig. 4), the stage SZ of which is normally on. This stage applies a positive potential to a three-control gate G10, to which are also connected the conductors +P and S. On the occurrence of the start transition corresponding to a code group, a positive potential appears on the conductor S. This allows a positive pulse from the oscillator 03 to pass through gate GM, and also through a single control gate G11 to the stage ST of FF, which stage is then-switched on, the stage 82 being also switched off;

9 thus blocking the gate G10. Since the stage ST is switched on, a positive pulse is applied to the gate G3, which starts up the counters, as already explained.

For the purpose of switching off the counters and restoring them to normal at the end of the code group, a five-control gate G12 is used. In the case of a fiveelement code at a speed of 50 bands, the normal time interval between the start transition and the final space-tomark or stop transition is 120 milliseconds, and if there should be as much as 50 percent distortion, the stop transition could be 10 milliseconds late, so that the maximum interval between the start and stop transitions could be as much as 130 milliseconds. The gate C12 is arranged to open at 138 milliseconds, which gives time for the circuit to be restored to normal before the next start transition which can arrive at 140 milliseconds, or later. On the opening of gate G12, a pulse +P is applied to switch on the stage SZ of the device FF, while at the same time switching olf the stage ST and shutting the gate G3. Thus besides the pulses +P there are applied to the gate G12 pulses from the stage 4 of C3, stage 7 of C5 and stage 6 of C6, and also the final marking potential from the conductor M, which commences 130 milliseconds from the start transition, or earlier. The gate G12 will be open at the end of the period of coincidence between a pulse from stage 6 of C6 (which lasts approximately from 120 to 140 milliseconds after the start transition), a pulse from stage 7 of counter C5 (which lasts approximately from 17 to 18 milliseconds after the start of the pulse from C6), and a pulse from stage 4 of counter C3 (which lasts approximately from 0.8 to l millisecond after the start of the pulse from C5). Thus the pulse +P which occurs substantially 138 milliseconds after the start transition will pass through the gate G12 and will switch on the stage SZ, provided that there is also a positive potential on conductor M.

The restoration of the stage SZ to on causes it to generate a negative pulse which is applied through a reset circuit RC to an output conductor R which is used to reset the counters C3 to C6, and the bi-stable device F4. The conductor R is connected to stage 0 of each of the counters C4 to C6 and on the restoration of SZ to on, these stages are restored to on all other stages being switched off. It is to be noted, however, that conductor R is connected to stage 1 of counter C3. The reason for this is that when the start transition arrives, and the stage ST of the device FF is switched on, there is appreciable delay in establishing the output pulse which is applied to open the gate G3, on account of the time constant of the circuit, and the first of the pulses -P is missed out. Therefore the counter C3 has to be restored to stage 1 instead of to stage 0.

The conductor R is also connected to the stage MA of the device F4, in order to ensure that at the end of the operation, the stage MA will be left on.

The extra stage LS of the device FF is provided to deal with the long space condition, which may occur, for example, if the line under test is in the false rest condition referred to above, or if there is some fault such that a more or less permanent spacing potential (positive) is applied to terminal T. In this case, the counters would be left continuously operating, and it is desirable to prevent this. Accordingly, the output of a five-control gate G13 is connected to the stage LS, which operates in parallel with SZ. The five inputs to the gate G13 comprise the pulses +P, the positive potential from conductor S, and pulses from stages 6 of counter C6, 9 of counter C4, and 4 of counter C3. From the explanation given above with reference to gate G12, it will be clear that this combination causes the stage LS to be switched on by a pulse +P which occurs substantially 130 milliseconds after the start transition. This will also switch off ST (which as already explained will have been switched on by the start potential), and so 10 the gate G3 will be shut and the counters stopped. This, then, will happen if a start condition lasts more than milliseconds, that is, more than the maximum probable duration of a normal code group.

The switching on of stage LS applies a positive potential to a three-control gate G14, the output of which is connected through the gate G11 to the stage ST of the device PP. The other two inputs to the gate G14 are +P and M. Accordingly, as soon as a space-to-mark transition occurs after a long spacing condition, the next pulse +P switches stage ST on, and at the same time switches 011 LS. Gate G3 is opened and the operation thereafter takes place in the normal manner as already described.

Having now explained the manner in which the operation of the distortion measuring circuit is started and stopped, it remains to explain the manner in which the distortion is indicated on the lamps RN shown in Fig. 5. There are 21 of these lamps, each represented by a rectangle marked with a number designating the percentage distortion represented by the lamp when it glows. The lamps are similar to those shown at N18 and N19 in Fig. 2. The trigger or priming electrodes are respectively connected to stages of the counters C4 and C5, as indicated, through appropriate resistors (not shown).

The priming potential of the zero, +50 and 50 lamps is, however, obtained through three two-control gates G15, G16 and G17. The priming potential for the zero lamp comes from the stage 0 of counter C4 when the gate G15 is opened by a pulse from stage MA of the device F4 (Fig. 4). The priming potential for the lamps +50 and 5@ come from stage t) of counter C5 when the gates G16 and G17 are opened by a pulse from stage Ml of the device F4. This arrangement is necessary since there must be an interval of 10 milliseconds between the instants of priming of the zero lamp, and the lamps +50 and -50.

It will be understood that the lamps are primed in turn, each lamp for a period of substantially 1 millisecond, but no lamp can glow unless energizing potentials are at the same time supplied from the cathode follower amplifiers A3 and A4. These energising potentials are applied in response to a mark-to-space or space-to-mark transition under the control of the two-condition device PS. The details of the circuit of the lamps and of the amplifiers A3 and A4 are shown in Fig. 9.

in the normal rest condition of the circuit of Fig. 5, the two-condition device F5 has its stage PM on, and stage F0 off. Accordingly, the gate G9 to which the output of stage F0 is connected, is blocked, and nothing can reach the bank of lamps RN. A three-control gate G18 has applied to it pulses +P, and the outputs of stage 4 and 9 of counters C3 and C4, respectively. Thus 10 milliseconds after the start transition (which, as already explained, starts the operation of the counters) the stage F0 is switched on and the gate G9 is opened, whereupon the positive triggering pulses which appear on conductor W in response to the subsequent transitions of the code group are applied to the flip-flop device PPI. The section FPA of the device FPl is normally off, and FPB is normally on. In this condition a relatively high positive potential is applied to the cathode follower amplifier A4 While a relatively low positive potential is applied to the cothode follower amplifier A3. As will be explained with reference to Fig. 9, the output of the amplifier A3 is connected to the anodes of all the lamps in the bank RN, while the output of the amplifier A4 is con nected to their cathodes. When, therefore, the stage FPB is on, the difference of potential between the anodes and cathodes of the lamps is well below the maintaining potential of the discharge so that no lamp can glow.

The switch SW of the quantizer QCis shown with the contact arm making connection with the stud des ignated M, so that only the space-to-mark transitions can be indicated on the lamps. When the first of these tran- 11 sitions occurs after the start transition, a triggering pulse appears on the output conductor W and switches on the stage FPA of the flip-flop device FPl, and switches ofi the stage FPB. This interchanges the input potentials applied to the amplifiers A3 and A4, so that the diiference of potential between the anodes and cathodes of the lamps is now increased to a value above the maintaining potential of the discharge. This will cause to glow that one of the lamps of the bank RN to which a priming voltage is applied from one of the counters C4 or C5. As already explained, the flip-flop device FPl returns to normal with the section FPA off and the section FPB on shortly after having been triggered by the pulse which occurred on conductor W. This extinguishes the lamp which has glowed, since its anode-cathode potential is thereby reduced again below the maintaining potential of the discharge. The time constant of the device FPI may be adjusted so that it remains in the operated condition for a suitable time which determines the time during which the indicating lamp remains glowing. A suitable period has been found to be about 7 milliseconds.

Each subsequent space-to-mark transition will be indicated on one of the lamps in like manner. If the contact arm of switch SW is set in the centre position, then both mark-to-space and space-to-mark transitions will be indicated, while if it is set in the lower position only mark-to-space transitions will be indicated. 7

After the whole code group has been indicated, namely 130 milliseconds after the start transition, the device F5 is restored to normal by means of a three-control gate G19 to which are applied the pulses +1 and pulses from stages 4, 9 and 6 of counters C3, C4 and C6 respectively. This switches off the stage F and shuts the gate G19. It follows that it is only possible to make an indication on the bank of lamps RN during a period between 10 and 130 milliseconds after the start transition. This pre vents confusion which might be caused if any transitions occur with an error greater than 50 percent.

In order to make clear the working of the circuit of Figs. 4 and 5, and the times at which the various operations occur, an example will be chosen at random. Let it be supposed that the waveform of a code group corresponding to the teleprinter character C is received at the terminal T. This group is shown in the diagram Fig. 10; immediately following the start transition there are two spacing periods, then three marking periods, then a spacing period, and finally the stop marking period. If the time of the start transition be taken as zero, then if there were no distortion, there would be space-tomarl; transitions at 40 and 120 milliseconds, and a markto-space transition at 100 milliseconds. However, let it be supposed that these transitions from space-to-mark actually occur at 36.4, 127.2 milliseconds, respectively, and that from mark-to-space occurs at 102.8 milliseconds. In Fig. 10 the mark-to-space transitions are shown by arrows pointing downwards, and the space-to-mark transitions by arrows pointing upwards. The marking periods are shaded, and the normal transition times are shown dotted.

It will be supposed that in Fig. the switch SW is set to accept both kinds of transitions.

At zero time, ST (Fig. 4) is switched on by the start transition, and the counters are started. At milliseconds F0 (Fig. 5) is switched on, thus opening the gate G9 and permitting the transitions to operate the indicating lamps. At 36.4 milliseconds, the first spaceto-mark transition arrives and triggers the flip-flop device FPI, which supplies energising potentials to all the lamps in the bank RN for about 7 milliseconds. At the time 36.4 milliseconds the following counter stages will be on: 1 of C6, 6 of C5 and 2 of C3. It follows that the 20 lamp of the bank RN will be primed from stage 6 of C5, and will glow, indicating that the first mark-to-space transition is about 20 percent too early.

At 102.8 milliseconds the next (space-to-mark) transition triggers the flip-flop device FPI, and this time the following counter stages will be on: 5 of C6, 2 of C4, and 4 of C3. Thus the +10 lamp of the bank RN will glow, since it is primed from stage 2 of C4, indicating that the corresponding transition is about 10 percent too late.

At 127.2 milliseconds the final stop transition occurs, and when the flip-flop device FPI is triggered the follow ing counter stages will be on: 6 of C6, 7 of C4 and 1 of C3. Hence the +35 lamp of RN will glow since it is primed from stage 7 of C4, indicating that the final stop transition is about 35 percent too late.

At milliseconds FM is switched on, and F0 is switched off, thus shutting the gate G9 and preventing any further indication on the bank of lamps.

At 138 milliseconds SZ (Fig. 4) is switched on and ST is switched off, thus shutting the gate G3 and stopping the counters.

Fig. 10 also indicates that at 130 milliseconds LS (Fig. 4) (which is shown in brackets) is operated if no mark-to-space transition has been received. As already explained, this shuts down the counters, but leaves the circuit so that it can be put in operation again on receipt of the next space-to-mark transition.

If the character C be continuously transmitted, the lamps 20, +10 and +35 (Fig. 5) will be observed repeatedly to glow. If the switch SW be operated to select only the mark-to-space transitions, then only the +10 lamp will glow; while if it be operated to select only the space-toniark transitions, then only the lamps 20 and +35 will glow.

The maximum number of transitions that can occur after the start element is five in the case of the teleprinter code (for the characters R and Y for example); and 1f the timing errors of all transitions difier by more than. 5 percent each will be recorded on a corresponding lamp, so that five of the lamps may glow in response to a single group. On the other hand if two or more transitions have timing errors which are equal to within about five percent, the same lamp may glow in response to each such transition.

Figs. 6 and 7 show a modification of the circuit of F gs. 4 and 5 by means of which the scale of percentage distortion is expanded five times, so that the accuracy of measurement is increased to :1 percent. Certain add1t1onal elements are required, as will be made clear, and the counters are rearranged. It will be evident to those skilled in the art that suitable switching arrangements may be included to enable the arrangement of Figs. 4 and 5 to be converted to that of Figs. 6 and 7, but these arrangements are not shown, since they would greatly complicate the circuits and make them difiicult to follow.

The 21 indicating lamps RN of Fig. 7 are arranged in the same way as in Fig. 5, but each is primed from the counters for a period of 200 microseconds instead of l millisecond, as before, and they accordingly cover a range of :2 milliseconds which corresponds to a range of :10 percent distortion in steps of 1 percent. Since any transition may occur at any time within a range of :10 milliseconds of the normal time, a mechanical five-point switch arrangement is used to enable each period of :10 milliseconds to be divided up into five periods of 4 milliseconds duration each of which is separately surveyed by the lamps.

In. addition to the 21 lamps of the bank RN, there are two further lamps designated EY and LT. The first of these, EY, glows when a transition occurs earlier than the period covered by the lamps RN, while the other LT, glows when the transition is later than this period. These lamps indicate which way it is necessary to adjust the five-point switch in order to bring the transition within the range of the lamps RN.

A number of elements in Figs. 6 and 7 are the same 13 as in Figs. 4 and 5 and have been given the same designations.

In Fig. 6, in order to prime the lamps of the bank RN (Fig. 7) at intervals of 200 microseconds, the counters C4 and C5, together with the bi-stable device F4, are interchanged in position with the counter C3. Counters C4 and C5 are now operated directly by the pulses P through the gates G4 and G6 controlled by the device F4 as before described, so that the counters operate alternately. The gates G4 and G6 are opened in response to the start transition by the switching on of the stage ST of the device FF, and are shut again 138 milliseconds after the start transition in just the same way as before. The arrangements for dealing with the long space condition are identical.

The device F4 is operated from the counters C4 and C5 through gates G and G21 which correspond to G5 and G7 of Fig. 4, but are two-control instead of threecontrol gates.

The counters C4 and C5 thus provide 20 steps each of 200 microseconds duration. Counter C3 is stepped on from stage 9 of counter C5 through a two-control gate G22 to which the pulses P are supplied. This counter C3 will take one step every four milliseconds. Counter C6 is operated from step 4 of counter C3 through the gate G8 as previously described, and provides ten 20 millisecond steps as before. The restoring pulses are applied from stage SZ of the device FF to stages 0 of C3 and C6; to stage 1 of C4 for the reason explained in connection with C3 of Fig. 4; and to the stage MA of the device F4.

In order that the gate G12 shall open 138 milliseconds after the start transition to operate the device FF for stopping the counters, inputs are taken from stage 6 of counter C6, stage 4 of counter C3 and stage 9 of counter C4, which produces a delay of l20+16+2=138 milliseconds. Similarly for gate G13, which has to open 130 milliseconds after the start transition, inputs are taken from stage 6 of counter C6, stage 2 of counter C3 and stage 9 of counter C4, which produces a delay of l20+8+2=l30 milliseconds.

The conductor W in Fig. 7 is connected to gate G9 and to two additional two-control gates G23 and G24, the outputs of which are connected through a single control gate G25 to a second flip-flop device FP2 similar to FPl. The outputs of stages FPC and FPD of FPZ are connected through cathode follower amplifiers A5 and A6, similar to A3 and A4, to the extra lamps EY and LT which are energised when FP2 is triggered, in the same way that the lamps of the bank RN are energised by the triggering of FP1 as explained with reference to Fig. 5.

As in the case of Fig. 5, the operation of the circuit of Fig. 7 takes place in the 138 millisecond interval between the starting of the counters in response to the start transition, and their stopping by the opening of the gate G12 (Fig. 6), as already explained. During this interval, the counter C3 is stepped along several times, so that each stage of counter C3 is on during several diflferent 4 millisecond intervals which recur at intervals of 20 milliseconds. Thus if the start transition be taken as marking zero time, the beginning of the period when stage 0 is on marks the times 0, 20, 40, 60, 80, 100, and 120 milliseconds; the beginning of the period when stage 1 is on marks the times 4, 24, 44, 64, 84, 104, and 124 milliseconds; and so on for the other stages. When C3 is stepped along for the first time after the start transition, the beginnings of the periods when stages 0, 1, 2, 5 and 4 are on mark the times, 0, 4, 8, l2, and 16 milliseconds. Thus it will be seen that by connection to appropriate stages of the counter C3, it is possible to mark times from 0 to 128 milliseconds in steps of 4 milliseconds.

Since the circuit is designed for measuring the telegraph distortion of .a start-stop five-element group up to :50 percent, the period over which transitions can occur after the start transition is 10 to 130 milliseconds, which is divided into six equal element periods of 20 milliseconds centered respectively on the normal or undistorted transition times at 20, 40, 60, 80, and milliseconds; and every transition after the start transition will occur during a corresponding one of these element periods. For the purpose of the expanded scale, each element period is divided into five equal sub-periods each of duration 4 milliseconds, and one sub-period of each element period is selected by a five-point switch for the distortion measurement. This will be understood more clearly from Fig. 11, which shows to a larger scale than Fig. 10 the first two element periods after the start transition, centred respectively on 20 and 40 milliseconds, and each divided into five similarly numbered sub-periods of duration 4 mil liseconds. When the five-point switch is set in position 1, the arrangement is such that a transition occurring in the first or 20 millisecond element period can only be recorded on the bank of lamps RN if it occurs during the first sub-period from 0 to 4 milliseconds; but a transition occurring in the second or 40 millisecond element period will also be recorded in the lamps if it occurs in the first sub-period from 30 to 34 milliseconds of the second element period; and similarly for all the other element periods. When the switch is set in position 2, transitions can only be recorded if they occur in the second subperiods of the corresponding element periods, which in the case of the first two element periods will be from 14 to 18 and 34 to 38 milliseconds, respectively; and similarly for the other positions of the five point switch.

During each of the five sub-periods of every element period the lamps in the bank RN are each primed once from the counters C4 and C5 for a period of substantially 200 microseconds, so that thereby the distortion is indicated to an accuracy of 1 percent.

Having given a general account of the operation of Fig. 7 for the production of the expanded scale, the details by which this is achieved will now be explained.

The times of energising the indicating lamps are controlled by a triple-stable device F6 (Fig. 7) which replaces the double stable device F5 of Fig. 5, and which has an extra stage designated FY. The gates G9, G23 and G24 which control the energising of the indicating lamps are opened by pulses which are generated when stages F0, FY and PM of the device F6 are respectively switched on, and shut when they are switched off.

The stages F0 and PM are switched through the threecontrol gates G18 and G119, while the stage FY is switched through an additional three-control gate G26 not represented in Fig. 5.

The times at which the gates G18, G19 and G245 open and shut are controlled by the five-point switch already referred to, which has five banks of contacts, three of which, designated SWA, SWB and SWC, are used to control the said gates, and two others, SWD and SWE, to control the gates G16 and G117 used for priming the +10 and -10 lamps of the bank RN, and their action will be explained later. The contact arms of all the five banks are shown in position 1, and the arms are all mechanically coupled and rotate in a clockwise direction from position 1. The contacts of the banks are connected to various stages of the counter C3 as indicated.

Before the arrival of the start transition, the stage FM of the device F6 is on, thus opening the gate G24, and shutting the gates G9 and G23. It the five-point switch is in position 1, as shown, gate G18 will be opened at 10 milliseconds (the start transition being taken to be at zero time), since the inputs of this gate are taken from stage 2 of counter C3 (8 milliseconds) and stage 9 of counter C4 (2 milliseconds), and from +1. Thus stage P0 of device F6 is switched on, and stage FM is switched off. However, at 14 milliseconds PM will be switched on again and F0 switched ofi from gate G19, the inputs of which are connected to stage 3 of counter C3 (12 milliseconds) through the switch bank SWE and to stage 9 of counter C4 (2 milliseconds), and to +P.

It follows that gate G9 is open during the first subperiod from 10 to 14 milliseconds, so that if a transition occurs during that period it will be indicated on the lamps RN. As has already been explained above, gate G9 will also be open from 30 to 34 milliseconds; from 50 to 54 milliseconds and so on; that is, during the first sub-period of each element period, and a transition occurring during any such sub-period will be recorded.

When the switch is set in position 2, the first operation is the switching on of stage FY of the device F6 through the gate G26 which occurs 10 milliseconds after the start transition, since contacts 2 to 5 inclusive of this bank are all connected to stage 2 of counter C3, the gate G26 also having an input from stage 9 of counter C4. The switching on of stage FY opens the gate G23, and shuts gates G9 and G24.

The switch being in position 2, the stage F will be switched on at 14 milliseconds through the gate G18, thus switching off FY, opening gate G24 and shutting gate G23. At 18 milliseconds stage PM will be switched on, and F0 switched off, so gate G9 will be shut and G14 opened. Gate G9 will thus be open during the second sub-period of the first element period; and from what has already been explained it will be seen that it will be open also during the second sub-period of each of the other element periods. Thus a transition will be recorded if it occurs during any one of these second sub-periods.

When the five-point switch is set to the three remaining positions in turn, it will be clear that the action will be as follows (a) FY is switched on, and gate G23 opened at 10 milliseconds (b) Ft) is switched on and gate G9 opened at the beginning of the sub-period corresponding to the switch setting in each element period (0) FM is switched on and the gate G9 shut at the end of the said sub-period. Thus by setting the switch to any one of the five positions, a transition occurring during the corresponding sub-period of any element period will be recorded on the bank of lamps RN.

The operation of the lamps EY and LT will now be described. Let it be assumed, for example, that the fivepoint switch is in the position 3. Before the arrival of the start transition, the stage PM of the device F6 is on and gate G24 is open. However, at the beginning of each element period, FY is switched on and gate G23 is opened instead of G24. At the same time lamp EY is primed. If therefore a transition arrives before the beginning of the third sub-period of any element period, it passes the gates G23 and G25 and operates the flip-flop device FPZ, so energising both the lamps EY and FM. Since lamp EY only is primed, this will glow and will indicate that an early transition is present, that is, one which occurs before the sub-period concerned. To pick up this transition, therefore, the switch must be set backwards to position 2 or 1.

At the end of the third sub-period, the stage FM is switched on, thus opening gate G24, and priming the lamp LT. It follows that if a late transition is present, that is, one which occurs after the end of the sub-period concerned, the lamp LT will glow, and will indicate that the switch must be set forwards to position 4 or to pick up the transition.

The first contact of bank SWC is left unconnected because it is always necessary to switch on stage F0 at the beginning of the first sub-period, and if the distortion does not exceed 50 percent, a transition cannot occur earlier than the beginning of the first sub-period and therefore the lamp BY is not required to indicate. Likewise a transition cannot occur after the end of the fifth subperiod, and so the lamp LT does not glow in the fifth position of the switch.

In order to give an example of the use of the expanded scale, the case illustrated in Fig. will be taken. Transitions occur at 36.4, 102.8 and 127.2 milliseconds. These will occur respectively in sub-period 2 of the second element period, sub-period 4 of the fifth element period, and

. '16 sub-period 5 of the sixth element period. If the fivepoint switch is set on step 1, the lamp LT will glow, showing that at least one transition is later than the first sub-period of its elementperiod, and none of the lamps of the bank RN will glow, showing that no transition occurs in the first sub-period. If the switch is now set to step 2, the transition at 36.4 will be picked up, and the lamp +2 of the bank RN will glow, indicating that the transition is 2 percent later than the centre of the sub-period, which corresponds to 20 percent. The distortion of the transition is therefore 18 percent. Lamp LT also glows because both the other transitions are later than the second sub-period.

If the switch be now set on step 3, both the lamps EY and LT will glow but none of'the lamps in the bank RN, since no transition occurs during sub-period 3. If the switch is moved to step 4, again both the lamps EY and LT will glow, but since there is a transition at 102.8 milliseconds, the 6 lamp in the bank RN will glow indicating that the transition occurs 1.2 milliseconds before the centre of the fourth sub-period which corresponds to +20 percent. The distortion of the transition in the fourth element period is therefore +14 percent.

If the five point switch be set in step 5, then the lamp EY will glow on account of the two previous transitions, but not LT, and the 4 lamp of the bank RN will glow, which indicates that the transition in the fifth sub-period occurs 0.8 millisecond before the centre of the sub-period, which corresponds to +40 percent The distortion of the corresponding transition is thus +36 percent.

It should be pointed out that the five settings of the five-point switch correspond, in order, to mean distortion values of +40, +20, 0, +20 and +40 percent, and should preferably be so marked. Then the distortion corresponding to any transition will be obtained by adding algebraically the marking corresponding to the point on which the switch is set and the marking of the lamp in the bank RN which glows.

It will be evident that if more than one transition occurs during a given sub-period, then more than one lamp in the bank RN will glow when the five-point switch is set to the point corresponding to the said subperiod.

It remains to explain the operation of the gates G16 and G17 which are controlled respectively by banks SWD and SWE of the five-point switch.

The lamps +10 and +10 of the bank RN are both primed from stage 0 of counter C5, which produces 200 microsecond pulses at intervals of 4 milliseconds, but it is essential that these lamps should only be primed respectively at the beginning and end of the sub-period during which the gate G9 is open. They are therefore primed through the three-control gates G17 and G16 which have control inputs from stages of C3 determined by the switch banks SWE and SWD respectively. The contacts of the switch bank SWD are respectively con nected to the same stages of C3 as those of bank SWB, and the contacts of switch bank SWE are connected to the same stages of C3 as those of bank SWA. It will be evident therefore that the lamp +10 will be primed from stage 0 of C5 only when F0 is switched on; that is, only at the beginning of any sub-period; likewise the lamp +10 will be primed from stage 0 of C5 only when FM is switched on; that is, only at the end of any subperiod. Apart from the introduction of the switch banks SWD and SWE, the gates G16 and G17 operate exactly as described with reference to Fig. 5, as also does gate G15 for priming the lamp 0, in which case no switching is necessary.

For the proper operation of the arrangements shown in Figs. 4, 5 and 6, 7, therepetition frequency of the pulses generated bythe oscillator 03 should be 100,

times the speed in bauds of the telegraph signals being received. Thus, for example, to measure the distortion when the speed is 60 bands instead of 50 bands the only modification necessary is to change the oscillator frequency from 5,000 to 6,000 cycles per second. Since, in practice, the telegraph speed is liable to differ slightly from the nominal value, it is desirable that the oscillator 03 should be provided with a frequency adjustment over a small range so that it can be adjusted exactly to the speed of theincoming signals. A small difference in frequency will be immediately apparent because the efi'ect will' be a systematic tendency for the distortion of the later transitions to increase, which effect will be visible on the lamps RN.

The circuits canbe easily modified to deal with synchronous signals of the baudot multiplex type instead of start-stop signals. Since in this case transitions are continuously received and there are no stop signals, the device FF and its associated apparatus are not required. A potential for opening the gate G3 at the beginning or the test can be applied through a suitable hand-operated switch (not shown), for the purpose of starting the counters, which thereafter run continuously until the test is finished. Counter C6 and gate G8 are not required.

Details of the quantising circuit QC of Figs. 4 and 6 are shown in Fig. 8. It comprises a double-stable circuit consisting of two valves V1 and V2 of which each anode is coupled to the opposite control grid in the conventional way, so that one valve is cut off and the other conducting. The circuit is triggered by pulses +P or P from the pulse generator 03 (Figs. 4 or 6) which are admitted respectivelyby two similar but oppositely polarised gate circuits comprising rectifiers X5 to X9 and X to X14, respectively.

Two direct current sources (not shown) are used, one of which has its positive terminal connected to terminal S1, and the other of which has its negative terminal connected to terminal S2. The remaining terminals of both sources are connected to the earth terminal E. Rectifiers X6 and X7 are connected through a common resistor R5 to terminal S1, and through individual resistors R6 and R7 to terminal S2. The upper ends of resistors R6 and R7 are connected through the rectifiers X8 and X9 to earth. Rectifiers X6 and X7 are directed to be unblocked in this condition, and X8 and X9 are directed oppositely to X6 and X7, and hold the upper ends of resistors R6 and R7 substantially at earth potential, when the rectifiers X6 and X7 are unblocked.

The values of the resistors R5, R6 and R7 are so chosen that in the absence of rectifiers X8 and X9 the upper ends of resistors R6 and R7 would be slightly negative to earth.

The pulses +P are supplied through a capacitor Q5 to the junction point of rectifiers X7 and X9.

Rectifiers X11 and X12 are connected to terminal S2 through a common resistor R8, and through individual resistors R9 and R10 to terminal S1. These rectifiers are also directed to be unblocked in this condition. The lower ends of resistors R9 and R10 are also connected to earth through the rectifiers X13 and X14 which hold these lower ends substantially at earth potential. The values of the resistors R8 to R10 are so chosen that in the absence of the rectifiers X13 and X14, the lower ends of resistors R9 and R10 would be slightly positive to earth.

The pulses P are supplied through a capacitor Q6 to the junction point of rectifiers X12 and X14.

The input terminal T, to which the received telegraph waveform is applied, is connected to the junction point of rectifiers X6 and X8 through the rectifier X5, and to the junction point of rectifiers X11 and X13 through the rectifier X10. Rectifiers X5 and X10 are directed so that they Will pass respectively positive and negative potentials to the corresponding gates.

The anode of the valve V1 is connected to terminal S1 through a load resistor R11 and to terminal S2 through 18 j resistors R12 and R13 in series: the anode of V2 is similarly connected to S1 through a load resistor R14 and to S2 through resistors R15 and R16 in series. The control grid of valve V1 is connected through a current limiting resistor R17 to the junction point of R15 and R16, and the control grid of valve V2 is connected to the junction point of R12 and R13 through a current limiting resistor R18. The junction point of resistors R12 and R18 is connected to earth through resistor R19. Two conductors resistor R20, and to the junction point of rectifiers X11 and X12 through an equal resistor R21.

Conductor m is connected through a capacitor Q7 and two resistors R22 and R23 to terminal S2, while conductor s is connected to this terminal through a correspond ing capacitor Q8 and resistors R24 and R25. The threepoint switch SW has its moving contact arm connected to the output conductor W, and its centre fixed contact to terminal S2 through a resistor R26, and to the junction points of resistors R22, R23 and R24, R25 through respective rectifiers X13 and X14 directed so that each of them will pass positive pulses to the centre fixed contact of the switch. The two other fixed contacts are connected respectively to the opposite ends of the rectifiers X13 and X14 as shown.

The circuit operates as follows. In the normal idle or rest condition of the telegraph circuit, a negative marking potential will be applied to terminal T, and the valve circuit will be in the condition such that the valve V2 is conducting and V1 is out 01f. A high positive potential will therefore be applied to conductor m while a low or nearly zero positivepotential will be applied to conductor .9. During this period pulses +P will be applied through capacitor Q5 to the rectifier X7, but can have no effect on the two-condition valve circuit since the rectifier X6 is conducting and holds the left-hand terminal of R20 substantially at ground potential. When the start markto-space transition occurs, the potential applied to terminal T changes sign and becomes positive, thus blocking the rectifier X6 through the rectifier X5, which will be conducting under this condition. However until the arrival of the next positive pulse through capacitor Q5, rectifier X7 will be conducting and will still hold the lefthand terminal of R20 substantially at ground potential. However, on the arrival of the first positive pulse after the potential applied to terminal T has become positive, both the rectifiers X6 and X7 will be simultaneously blocked, and this will cause an increase in the positive sense of the potential applied to the left-hand terminal of R20, which is communicated to the control grid of the valve V1, and triggers the circuit over to the other condition in which the valveVl is conducting and V2 is cut oft". A relatively large positive potential thus appears on conductor s instead of on .conductor m. The circuit having been triggered, all following positive pulses applied through capacitor Q5 can have no further. effect.

It will be understood that since the potential applied to terminal T is now positive, rectifier X10 will be blocked, so rectifier X11 will be conducting, and will hold the left-hand terminal of resistor R21 substantially at earth potential. The negative pulsesapplied through capacitor Q6 from conductor P are therefore prevented from having any effect.

On the occurrence of the next space-to-rnark transition,

the potential applied to terminal 7 changes sign again and becomes negative. This blocks rectifier X5 and prevents the positive pulses supplied over conductor +P from having any effect, and unblocks rectifier X10, thus blocking rectifier X11 and permitting the next following nega-' tive pulse applied over conductor P to block the rectifier X12. This causes the potential of the left-hand end ofresistor R21 tobecorne :negative, and. this potential is communicated to'the controlgrid of valve V1, and cuts off. this valve, thusqtriggering the .circuit back to the first condition inwhich valve V2 is conducting. A relatively large positive potential accordingly re-appears on con ductorv m and. the potential on. conductor. s. is reduced nearly to zero. The, following negative pulses applied through. capacitor;Q6 can have no effect after the circuit has been triggered.

Thus it will be seen that within.200 microseconds after thepotential applied to terminal T.has changed sign in either direction, the two-condition circuit will betriggored. Bythismeans, the leading edges of the. positive pulses generated'on. conductors -.m and s inv responserespectively to space-.to-mark,. andmark-to-spacev transie tions are effectivelysynchmnisedWiththe pulses generatedby'the, generator. O3.(Fig. 3.) and confusion which mightarise =fr.om..the occurrencecf. a transition between two pulses is. avoided.

Inorder to provide the; triggering pulses whichoperate the devices, FPI andFPZ (Figs, andl), thepositive pulses which..appear on conductors m and s are differentiated by the capacitorsi Q7 and Q8 acting-in.seriesrespectively with the resistors. R22,.R23. and-R24,-R25.. A short positivedifierential. pulseis. generatedin response to the leadingedge of each such pulse, .and.a similar negative differential pulse .in response. to the trailingedge. resistors. R22,.R23 and R24, R25. form potential dividers for reducing the amplitude of the differential pulses to.

a suitable value. The positive differential pulses arethe triggering pulses,.and the negative differential pulses. can bedisregarded. since thegates G9, G23. and G24 to. which they are supplied aredesigned to. be unatfectedlby negativerpulses.

If the contact arm of theswitch SW is on the left-hand.

contact, as shown, only, the triggering pulses. corresponding. to .spaceeto-mark transitions. can. reach the output conductor, W, since the differential pulses coming from conductor s will be blockedby the rectifiers'X13 and X14.

Likewise, if. the contact arm is.on .the right-hand contact,.

only the triggeringpulses corresponding to mark-to-space transitionstcan reachconductor W. Howevenif. the arm is .in., the. centre contac.t,-.,both kinds .of triggering pulses" can .reachithe outputlconductor W, each through the. corresponding rectifier. X13 .ancl..X14,-1and.;itwill. be noted that inthis case the negative difie'rentialpulse'which followseach .triggering pulse: will ;.be. stopped by the v-cor.-

responding one. -of; the. rectifiers X13 and. X14- It. maygbeg added; that the. resistor R26 provides the mo essary. return path for; theirectifiers.

Thegpositive: pulses which appear. alternately: on" conductorsm ands are amplifiedbytheamplifiers-A1 andA2 and are;deliveredtothe-outputconductors M and S for controlling the: counting circuits aswalready explained. T heseamplifiers are. not. required. when thecircuit is arranged for usewith synchronous-systems."

Although it has been assumed that the waveform're ceived from. thetelegraph line orcircuitwill' be applieddirectly tomterminal T, the. received waveform. may, if

the lamps ofthe bank RN in Fig. 5 or 7 are energised fromthe amplifiers A3 and A4. These amplifiers consist respectively. of. valves V3 and V4 arranged as cathode followers; with their. anodesaconnectecl directly to; the

The.

20. terminakSS :of a .p'ositive-.:.-source1 .:(I10tKSi10WII).l andltheircathodesconnected through respective.equalresistors R21 LI andRZii to the earth terminal E andtotheterminal SW of a negative source (notshown). The control; grids are' connected respectively'to the outputs of ithefstagesIEPA and FPB of the deviceFPlshownin Figs. 5 and7.

The cathodes of valvesV3 and .V4 are con'nected to conductors L21 and L22 throughresistors R29-andeR30z All the lamps of the bankRN have their anodes connected to conductor L21 and their cathodes' to conductorIL22.- In Fig. 9, only the firsttwo lamps RNA'andRNBLof the bank RN of Fig. 5 are shown, theselamps corresponding .1. to 50 and 45 percent distortion. Thespriming elec-" trodes of. the lamps RNA and .RNB are respectively connectedto the gate G17 and to' stage 1 of counter C5," as shown in Fig. 5. All the other lamps not showndn Fig.9 are connected in like manner to-conductors'L21 and L22,

and their primingelectrodes as shown in Fig.5 or 7.

As already explained, normally, stage FPB of the fiip-- flop device FPl shown in Figs-5 and 7 is: on,'-thus'a'pplying a positive potential to the control. grid of valve V4."- Arelatively large current thus flows through the resistor R28 which raises the potential of the cathod'eto-a relatively high positive value with respect to terminal S4-,"-and' l the circuit can bearranged so thatin thiscondition the cathode of V4 has a small negative potential with respect to earth. At the-same time stage FPAof'the device FPI is off, and so the potential applied to the control grid'of the valve V3 is small,'so that the cathode: will be slightly positive with respectto earth. The potential difference between the cathodes of V3 and V4 should in this con'di tion-be-well below the maintaining voltage ofthe lamps connected to conductors L21 and L22 so that none of them can glow.

When the device -FP1 (Fig. 5 or 7) istriggered in response to a transition, the stageFPAwill be switched on and FPB will be switched off. This willraise the poten tial of the cathode of valve V3 and lower thatof'the valve- V4, in such manner that the diiference 'ofpo'ten'tial" be tween the two cathodes is'increasedto' a value exceeding the firing potential of any lamp which is primed; 'so' that thatlamp will be caused to glow. When thedevice'FPl' restores itself after 7 milliseconds,'the' originalcondition in which there is only a small difference"ofpotential'b'e tween the cathodes of the valves 'V3 and V4*is"restor'e'd,..

aid the glowing lamp is extinguished."

The resistors-R29 and R30 are introduced so that'as soon as a lamp is fired, the resulting current 'reduces'the potential between the conductors" L21 and L22 substan-.

tially toithe maintaining potential. of the lamp which'is' glowing, so that no other lamp can be fired during the interval before the device FPI restores itself. This prevents more than one lamp from being caused to glow in response to any transition.

The energising of the additional lamps EY and LT'Of Fig. 5 from the amplifiers A5 and A6 is effected'in the manner indicated in Fig. 9.

Although certain particular arrangements have been described forillustrating the. inventiomit will. be. understood thatit can be carried outinother ways. For ex ample, the-devices. for indicating the distortion need not be glow lamps. For systems working atv relatively low speeds, for example, incandescent lamps might be used, or. electrorrnagnetic. relay devices. Similarly, the counting chain which. provides the priming. voltages need not be. a

gas-tube. system but. couldcbe a counting. arrangement employing hard valves,. or even anelectro-rnechantcal system.

While the principles-of the invention have been. de-iscribed above in connection .with. specificembodlment's,

and particularv modifications thereof, it is tobe clearly understood that this description is made only by way of example and notas a limitation on. thescope Of-thedtl" vention.

What we claim is:

1. An arrangement for measuring the telegraph distortion of telegraph signals of the start-stop type in which each code group of signals representing information consists of a plurality of code elements of equal duration each code element being a marking or a spacing element according to the code comprising means for receiving an electrical wave representing a code group, a plurality of indicating lamps, means for priming the lamps in succession one at a time for equal periods of time in such manner that any lamp will glow only when energized during the period of priming, said priming means coupled between said receiving means and said lamps and operable in response to the arrival of a signal transition, said priming means including time base generating means and counting means, said counting means adapted to count impulses from said time base generating means under control of said receiving means, and means controlled by a transition of the received wave between adjacent element periods of opposite type for simultaneously energizing all the lamps, whereby the particular lamp which is primed at the time of arrival of the transition will be caused to glow, said counting means comprising a chain of counting devices controlled by said time base generator which produces timing pulses of short duration and having a repetition period which is an integral sub-multiple of the duration of a code element, each lamp having a priming electrode connected to a corresponding stage of one of the counting devices, the arrangement being such that the stages are switched on in turn for a given period, and when switched on, each stage applies a priming potential to the priming electrode connected thereto.

2. An arrangement according to claim 1 in which means is provided for generating a short triggering pulse in response to each transition, the said triggering pulse synchronising with the first of the timing pulses to occur after the arrival of the transition.

3. An arrangement according to claim 2 in which means is provided for selecting for measurement transitions of one kind only.

4. An arrangement according to claim 2 in which each lamp is provided with an anode and a cathode and a priming electrode, and in which the lamps are normally prevented from glowing by polarising the cathodes positively to the anodes, comprising means controlled by the triggering pulse for polarising the anodes positively to the cathodes for a period short compared with the period during which any lamp is primed, whereby the lamp which is primed at the time of arrival of the corresponding transition is caused to glow.

5. An arrangement according to claim 4 in which means is provided for stopping the operation of, and resetting, the counting devices after each code group has been received, and means operating in response to the start 22 transition of the next following code group for restarting the counting devices.

6. An arrangement according to claim 5 comprising means for stopping the operation of the counting devices it the start condition persists for a period greater than that occupied by a code group, and means for re-starting the counting devices in response to the first transitio which terminates the start condition.

7. An arrangement according to claim 6 in which means is provided for preventing the energising of the indicating lamps except during the period beginning half an element period after the start transition and ending half an element period after the normal time of the stop transition.

8. An arrangement according to claim 7 in which the period during which any transition subsequent to the start transition can be received is divided into a plurality of equal sub-periods, comprising means for priming all the indicating lamps in turn once during each sub-period, means for selecting any one of the sub-periods in order to indicate the time of a transition which occurs during the selected sub-period by the glowing of the corresponding lamp, and means for preventing the energising of the indicating lamps except during the selected sub-period.

9. An arrangement according to claim 8 comprising two additional indicating lamps, means for priming one of the additional lamps during a period beginning half an element period after the start transition and ending at the beginning of the selected sub-period, means for priming the other additional lamp during a period beginning at the end of the selected sub-period and ending half an element period after the normal time of the stop transition, and means for energising the additional lamps in response to transitions which do not occur during the selected sub-period, whereby the glowing of one of the additional lamps indicates the occurrence of at least one transition before the beginning of the selected sub-period, and the glowing of the other additional lamp indicates the occurrence of at least one transition after the end of the selected sub-period.

References Cited in the file of this patent UNITED STATES PATENTS 1,816,621 Stacy July 28, 1931 1,845,994 Wise Feb. 16, 1932 1,867,209 Chauveau July 12, 1932 1,873,440 Locke Aug. 23, 1932 1,920,454 Weaver Aug. 1, 1933 1,972,941 Lewis Sept. 11, 1934 2,342,318 Wilkerson Feb. 22, 1944 2,425,307 Desch Aug. 12, 1947 2,482,932 Payatt et a1 Sept. 27, 1948 FOREIGN PATENTS 906,654 France Jan. 16, 1947

Images