US1926169A - Phase and frequency control - Google Patents

Description

P 12, 1933- H. NYQUIST PHASE AND FREQUENCY CONTROL Filed March 11, 1931 3 Sheets-Sheet 2 Sept. 12, 1933. H NYQUlST PHASE AND FREQUENCY CONTROL Filed'March 11; 1931 3 Sheets-Sheet 5 Position, 22'

INVENTOR EJV 6 BY yguw ATTORN EY Patented Sept. 12, 1933 ATENT; orF cr.

1,926,169 PHASE AND FREQUENCY; CONTROL -Harry Nyquist, Millburn; N. 3., 'assig'nor to American Telephone and Telegraph Company, a corporation of New York Application March 11, 1931. Serial No. 521,851 1 21 Claims. (01. 250-2) This. invention relates to an arrangement for maintaining distant sources of current substantially in phaseandfat approximately the'same frequency. One embodiment of the invention is described herein inconnection with an arrangement for controlling the phaseandfrequency of the carrier waves of a chain of broadcasting stations.

When two or more distant radio stations are connected by telephone lines in order to broadcast the same program,; it is desirable that all broadcasting stations'use the same carrier frequency so that the numberof frequency channels required for one program may be a minimum.

'- Unfortunately, slight changes in. the relative phase of the radio frequencyfcarrier currents from the various broadcasting stations afiect the field strength in certain intermediate, localities,

thus causing received'signals to vary in amplitude and in seine cases to vanish completely.

If two radio broadcasting stations are to operate ,on the same nominal carrier frequency, .11; is

' necessary, if the service furnished is to be satisfactory and at the same time reasonably economi-- cal, that the nearness the actual frequencies come tobeing'the same be considered with some care.. It-would obviously be ideal, from the point of viewof the service rendered, to requirethe actual'frequencies tobe identical. Ifthis' were the case thejphase relation would be invariable and at any given receiving point there would be a fixed phase relation between the fields. received from the two broadcasting stations.

In practice, however, itwould be sufficient to merelyspecify that the phase angle between the carrier waves of the two stations should not change by more than a certain number of degrees,

1 and to specify in addition that the rat'e of :phase it is not so important to limit the maximum phase drift should not exceed a givenjvalue. A slow phase drift merely causes gradual fading atthe receiver, which maybe taken care. ofby again adjustment in the receiving set itself. Rapid phase drift,thowever', would obviously. ruin the reception,'eitherby causingfading which is too rapid to, be compensated for at the receiver or by causing such a rapid fluctation .of the signal strength thatthe signal would be distorted. It'is absolutely essential, therefore, that the rate of phase drift be limited. If, however, the rate of phase drift be held within sufficiently small limits,

anglewhich may be permitted between the stations. 1

f It has heretofore been proposed to maintain the desired constancy of frequency and phase relaimportant.

transmitted over a line, the impulses at the rebe manifested as changed in'rphase.

tion by transmitting a control current from a master station over wire lines to the various radio stations. At each radio station the carrier wave from the station is generated by means of a local generatonsuch as a high frequency oscillator. 0 The frequency of this oscillator is automatically adjusted whenever the normal relation between the generatedcarrier wave and the received control wave is changed more than a'given amount, and thisadjustment is automatically made regardless of the cause of the variation. 7

In such a system differences in the phase and I frequency of the carrier waves of different radio stations will depend upon'two general sourcesof error: (1-) errors due to the oscillators usedfor generating the carrier waves at the radio stations or errors due to the oscillator used for generating the control wave at the master station, and (2) errors due to variations in the transmission line used for transmitting the master control wave.

The errorsof oscillators are usually classified for convenience into (a) the longtime error in the rate or frequency of the oscillator and (b) the short period errors or differences in period of successive cycles. The' characteristics of oscillators of course vary, but inmost cases particular attention is paid to reduce. the long timeerror because it is cumulative, and itis assumed that the short period errors are less important becausein general they are not cumulative. However, the short time errors in some in-,- stances may be so .relatively'large as to become Where suchis the case, it is now possible, by the use of modern methods, to so design crystal-controlled oscillators that both the long time error-and the short time error may be kept verysmall. I

As to the errors due to the line itself, it should be noted that if a current of given frequency is ceiving end as compared with those at the sending end will, of course, show no long time error. Short time errors may occur, howevenand will These are due to-variations in the elements going to .make up the line caused by such things as tempera--. ture changes, repeater steppings and (if this particular channel is amplified in repeaters common to other channels) overloading of -repeat. ers by signals on other channels. Q 7 It should also be noted that the phase shift be tween two oscillators which is caused by difference between their long time frequencies, is cumulative with the first power of the time. The

phase shift caused by short period errors, if these no are random, is cumulative with the square root of the time. The short period errors causing phase shifts between the two ends of a line are not entirely randon'nhowever, so that although for a short time the phase shift may accumulate as the square root of the time, over a long period it will practically accumulate to zero. In

other words, the phaseshift due to changes in.

the line will not accumulate over a long period but in general the phase will be advanced or retarded withrespect to an average phase relation, and the amount of this phase advance or 7 phase retardation with respect to this average will not exceed a given maximum for a given line.

With these conditions in mind, it becomes-evident that a master control system such as above outlined will suppress long; period diiferences in frequency by-not allowing the phase shift to accumulate. Also, the short period difference.

between the resultant of the short period errors inthernaster oscillator and the controlled oscillator, and the error-inthe line, is taken up by varying the controlled oscillator. Such a method of control will be satisfactory where the re sultant of the short period errors in the two oscillators is of the same order of magnitude or greater'than the short period errors in theline. However, if oscillators are available which are constructed to be so steady in their operation that their short period fluctuations are substantially negligible when compared with the corresponding fluctuations in the line, this method forces the controlled oscillator to conform to the vagaries of the line and wastesthe valuable property of the steadiness in the oscillators.

In accordance with the present invention it isproposed to take advantagelof the'fact that the frequency of the oscillators may be held within narrow limits so that the cumulative long period error may be relatively small as compared 7 with the error due to fluctuations in the line. To this end it'is proposedin effect, to loosen the coupling between the receiving end of the line "and thecontrolled oscillator at the radio station so-lthat the oscillator will not be adjusted .in

response to the short period variations of the line but will be adjusted when long time variations occur.

This is accomplished by introducing between the line-and the control element a phase compensating device by which changes 'in'ph'ase within limits corresponding to the phase changes introduced by the line itself will'be cornpensated for without causing any adjustment of the controlled oscillator. When the phase change exceeds this limit, however, as will be the case when cumulative long time erroris introduced, by anydiiference of the frequencies of the two oscillators, the mechanism for adjusting the oscillator will be brought into playtobring the frequency of the oscillator back to its normal relation to the control wave.

7 More specificallythe inventioninvolves-transmitting a' control frequency (preferably a voice frequency) from a master stationover a line to I the radio station and at'the radio station stepping 'up the control frequency to a frequency equal to that of the local oscillator for generating the carrier wave of the radio station. Some of the energy from the local generator or oscillator at the radio station is applied to a push-pull detector circuit, together with'energy from the telephone line. The currents in the two output branches of the push-pull detector will be balanced at a certain optimum phaserelation of the two input waves. Any departure from this phase relation produces an unbalance in one of two senses, depending upon the direction of the phase shift. Thisprinciple is utilized to control two separate mechanisms one of which is a phase changer which may be connected in the telephone line and the other of which is a variable condenser or other device fo'r'changing the frequency of the controlled oscillator. The control exercised over these two devices is such that for small changes of phase of the order of those produced by variations 'in the telephone line itself, the phase changer may be operated to compensate therefor without causing any change in the frequency of the oscillator. The variable condenser of the oscillator is connected to the control mechanism by a lost motion device to permit of this operation With this arrangement the phase shifter takes up the short shifts (except those so short as not to be able to effect a change in the setting of the device) between the receiving end of the line and the controlled oscillator on the assumption that the latter is more'steady. If these phaseshifts start accumulating to an appreciable amount, however, an adjustment is made on the frequency of the controlled oscillator on the assumption that it is the latter whichisbeginning to drift.

Thisis accomplished by means of the lost motion device. The optimum adjustment of the lost motion device depends upon the relative errors -to be'expected-from the line. and the oscillators.-

If the line is bad the lost motion should be large;

if it is good the lost motion should be small.

if the adjustment of the lost motion device is such as to take care of the maximum phase shift due to the line, it is evident that an adjustment of-the controlled oscillator will never occur unless cumulative drift due tochange in frequency of the oscillator, occurs. If, however, the adjustment of the lost motion device isnot sufiicient to take care of the maximum phase changes which may occur in the line, occasionally the oscillator will be adjusted'due to a large phase change in the line. This, however, will soon be corrected, 'as'the resultant cumulative drift in phase between the oscillator and-the control frequency will cause re-setting of the oscillator.

One more point should be noted. With this arrangement the maximum phase change between any two radio stations may exceed the maximum phase shift introduced bythe line (which is cared for by the lost motion device).

This is due to the fact that the phasecompensator' is adjusted simultaneously with the adjustment of the oscillatorin additionto the adjustment introduced'to take care of the line shift. The amount of this additional adjustment of the phase compensator may be kept small, however,

by introducing suitable gearing or other means so arranged that a'very slight change in the control mechanism over and above that necessary to take up the lost motion will bring the adjustment of the oscillator back to normal.

The invention may now be more fully underphase changer and the mechanism for adjusting the phase. changer and the variable condenser for frequency control; Fig. 2 (a) is a sectional'view of part of the apparatus of Fig; 2; Fig.3 illustrates a the electricalelements of the frequency changer;

Fig. 4 is a vector diagram corresponding to the arrangement of Fig. 3; Fig. 5 is adiagram similar to Fig. 3 but showing various positionsof adjustment'of the'phase controller; While Figs. 6, 7 and 8 are vector diagrams corresponding to definite positions of thephase' controller of Fig. 5. r Figure 1 shows the essential features necessary for maintaining distance :sources of current inphase and at the same frequency. In "these drawings an 'oscillator O, which may generatea voice frequency of, say, 1,000 cycles per second,

is connected to an harmonic generator l-lGi to step up they frequencyto correspond with thewave length of a' local broadcasting station. ,In the case assumed,.the frequency is 50,000 cycles per second, and a filter F1 is'provided to select the desired harmonic from the -other' harmonics in the output of theharmonic generator FIG-1. Some of theenergy of the control wave from the oscillator O is also connected through parallel repeating coils to various telephone linessuch as Ln leading to other5broadcastirig stations. The apparatus associated with the .distant terminal of line Ln only is shown, it being understood that the apparatus associated with the other lines will be similar. In the case of the line Ln the 1,000 cycle control frequency-passes through a trans former into a selecting filter F'n and, is impressed upon an harmonic. generator HG-n to generate the carrier wave for the broadcasting station associated with the distant end 'of the line Ln. r

The carrier wave in this instancealso has a frequencyof 50,000 cycles per second. Accord-- ingly, the harmonic of this frequency is selected by means of the filter Fn and is passed through the phase shifter PSn, symbolicallyindicated, and

by means of a transformer Tn is applied tothe grid circuits of a push-pull detector circuit comprising detectors Dn and Dn whose output cir cuits are oppositely connected through opposing windings of a polar relay Rn. The carrier wave supplied to the broadcasting station is generated by means of a local oscillator On. In order that the phase and frequency of this oscillator may be controlled with respect to other similar, oscillators at other broadcasting sta tions. associated with. the control station,= some of the energyfrom the output of theoscillator On is connected to the common branch of the grid circuitsof thepush-pull detector-arrange.- ment through a transformer T 1 V The relationship of the various elements of the push-pull circuit is such thatwhenthe waves transmitted through transformers-Tn and T'n are 90 degrees. out of phase with respect to each other, equal opposing currents will flow through the two windings of the polar relay Rn and its armature. is held in neutral .posiiton. A phase shift of the two Waves in one direction results in an increase in the direct current flowing in the output of the one detector and a decrease of thedirect current flowingin the output of the otherdetector, thereby producing an unbalance which causes the polar relay to attract its arthe polar relay will attract bolically indicated at lEPn, whereby the phase shifter PSn may be adjusted and a frequency control arrangement such as the variable condenser associated with the oscillator On may be adjustable.

Thedetails of the mechanical parts MPn and the phase shifterPSn and their relationship are shown in Fig. 2. Before explainingthese details,

it is desirable, however/to have an understanding of what is required in order that the circuit shall function properly. Bearing in mind that the push-pull arrangement of Fig.1 is so arranged that a very small departure from the proper phase relation of the two input waves results in an operation of the polar relay, and hence, of the control mechanism, it is desirable that some provision should be made whereby short time changes in phase,.arising from abnormal conditions .(such as temperature changes, repeater steppings, repeater overloading, etc.) of the tele phone line overv whichthe control frequency is transmitted, should be compensated for without adjusting the frequency control arrangement for the oscillator On. Furthermore, since any change in'the frequency of the oscillator results in a cumulative phase shift of the carrier wave with in response to such a' cumulative phase shift the oscillator On will be adjusted to bring its fre quency to normal. 1

These requirements are met in the present instance by connecting the phase shifter PSn so that it may be adjusted by the motor Mn whenever a. shift in phase occurs between the control wave and. the carrier wave regardless of the amount of said shift. ,The variable condenser Cn for controlling the frequency of the local oscillator On, however, is connected to the motor drive through a lost motion device. This permits the. phase shifter PSn to be adjusted over a small w range to compensate; for the small changes in phase which arise in the telephone line itself without disturbing the frequency of the oscillator. When the frequency of'the 'oscillator does change with the consequent cumulativeshifting in phase between the control wave and the carrier wave, the lost motion is taken up and the phase shifter'PSn and the variable condenser Cn are adjusted together until the normal phase difference between the control wave and the car-J rier wave is again established. When this occurs the frequency of the oscillator will be back to normal. As has been previously pointed out, the phase shift due to the line is advanced or retarded with respect to an average phase rela-' tion. The lost motion is as proportional with respect to the maximum change of phase due to the line as to cause no adjustment of the oscillator in response to ordinary line changes. A phase change greater than that for which the loss of motion is" designed is assumed to be'due to cumulative error in the oscillator and causes ad- J'ustment of the phase shifter PStuntilthe loss -with two armsb and b.

'g and h. which are normally 90 degrees apart,

as shown in Fig. 2a. 'Theshaft S is also provided A corresponding arm a is carried by a shaft S and a pin upon the arm aniay be engaged by. one or the otherof the arms'b'and b when the shaft S is rotated through a suitable arc. The shaft S drives the variable condenser Cn (through a suitable reduction gear, if desired) and the arrangement of arms a--b b provides a lost motion arrangement so that'the' cranks g and It may be driven through a suitable range of adjustment without any change'of the variable condenser Cn.

Two drive rods 0 and d are connected from the' crank throws g and h to insulated portions of suitable sliders p-and (1 which move over a pair of guides c with 'a'travel approximately proportioned tothe sine and cosine of the angular positions of their respective crank throws. These sliders make contact with the uniformly Wound resistances 2R and also with the metallic guides which are connected to the terminals of the primary of the input transformer Tn. The resistances 2R are connected in'two parallel circuits bridged across the terminals of the telephone line. In the series branch with the one resistance are two capacities X0, onearranged .on either side of the resistance. In series with the resistance in the other branch are two inductances XL, the inductances being arranged on either side of'the resistance. By making XL;X0=R at the frequency 1 of transmission, which in this case is50,000 cycles, the resultant voltage across the primary on the input transformer can be shown to have a substantially constant amplitude, with phase shifted at either direction, depending upon the direction of rotation of the motor Mn. As shown on the drawings, the sliders will pass each other moving in opposite'directions at approximately 0.707R from the middle of the resistance 2R when the connecting rods are long.

* In order to understand the operation of the phase. shifter PS1, let us refer to Figs. 3 to 8, inclusive. Fig.3 shows the circuit arrangement of the phase shifter in simple form. If we imagine a voltageEo applied to the terminals of the network in Fig. 3,v currents I1 and I2 will flow Since the elements R, X0 and XL have impedanceswhose numerical values are equal, the numerical values of the through. the branches.

currents I1 and I2 will be equal. If We consider the impedances of'the'two capacities inthe one branch as being combined in one, and the impedances of the two inductances in the other branch likewise combined, and represent the drops through the elements of each branch due to the currents I1 and I2 flowing therethrough,

respectively, the resultant vector diagram will i be as represented in Fig. 4:. ,The vectorial drop through the capacities and resistance 'must be equal to the applied voltage E0. The drop 2 I1 R through the resistance. 2R will be degrees out of phase with the drop 2 I1 X0 through the total capacity of the branch and the drop through the resistance will be numerically equal to that through the capacity of the branch since the-impedances of the resistance and capacity are numerically equal and the same current iiows through both. Applying the same principles to thedrops through the resistance and inductance of the second branch, we get the two vectors 2 IZ'XL and 2 12 R.

With the mechanicalarrangement previously,

described the two contacts will be simultaneousdirection of movement of the contact as it approaches each'contact point. Corresponding contact positions on the two resistances are indicated by the same numeral, primes being added in the case of the resistance in the right-hand branch. I

First, let us suppose'that the contacts are in positions 1 and 1. Let us see what is the effective drop between these-two points. Referring tothe vectorial diagram of Fig. 6, the vectorial drops through two branches are as shown in full lines. Here the drops through each of the two condensers are shown as separate vectors, and likewise, the drops through each of the two inductances are shown by separate vectors. The result is that the applied voltage E0 instead of being subtended by two vectors for each branch so as to form two triangles as shown in Fig. 4, is subtended by three vectors in such manner that the resultant vector diagram forms two pairs of triangles.

points 7 and 1 of the second branch. The drop between points 1 and 1 will be the vectorial resultant of these drops; It will be noted that the circuit traced from the point 1 to the point 1 passesthrough the uppercondenser of the first branch in the direction opposite-to the direction of current flow so thatin' taking the vectorial sum of the drops, the dropthrough the condenser must be reversed- Asshown in full line in Fig. 6, the vectorial drop through this condenser has the value I1 Xc. Let us draw the drop throughthis condenser in the opposite direction as shown by the dotted vector .-I1'Xc. The drop through the upper inductance of the second branch will be I2 X1, as shown in full line inFig/e. If We obtain the resultant of this drop with the drop .-I 1 Xofwe obtain the vector Va. It remains to combine this vector withthe vectorial drop through the resistance R between the points ,7 and 1'. This drop through the resistance R will have a value of one-half of the vector 2 I2 R of Fig. 6 and will be in the same direction. Drawing this vector in dotted lines at I2 R and obtaining the resultant of 50 the vector thus drawn'and the vector. Va, we have as the vectorial drop from the point 1 to the point 1 the dotted line vector V1. This vector, it will be noted, is numerically {equal and in phase with various other positions which the two sliding con-.

tacts may assume ll'1'Flg. 5, it will be found that the resultant vector is always of the same numerical value (that is, has the same length): but rotates about the'point O of Fig. 6. This will be 1 clear by considering two other positions ofthe sliding'contacts. i Let us first consider the case,

where'the sliding contacts are in positions 2 and 2', respectively, which are the next positions in order of rotation of the crank shaft. Due to the sine relation that exists between the motion of the cranks and the motion of the sliding-contacts, the

left-hand slidingcontact only passes over about .3 R in passing from position 1 to position 2. The

right-hand sliding contact, .on the other; hand, passes over approximately .7 R. in going from point The method of obtaining the vectorial drop between points 2 and 2 is shown'in Fig. 7.] Here we have reverse drops through .3 Rbetween 2 -and .l and alsothrough the upper condenser of the left-hand branch, these drops being in series withthe normal drops throughthe upper induct-y Y ance of the right-hand branch, through the resistance R and a resistance ofapproxim'ately .7 R, the latter. resistance being between points 5' and 2. 1 First let us draw the vectorial drop through theresistance .3 Has shown in dotted line. This drop .is oppositeginydirection to the drop 2 I1 R, in Fig. '7' and has the value -3 11 R as shown in dotted lines. Since this vector is opvectors may be combined to give the resultant posite in direction to the vector I2 XL representing. the drop through the upper inductance, these two vector Vb. Next let us draw the vector represent. ing the drop through the upper. condenser aswe did-in Fig. 6. Thisvector is shown at -11 Xo of Fig; '7, and whencombined with the vector Vb gives'a resultant vector Vd. There remains to be considered the drop throughthe resistance between -points '7 and 2. The vector representing this drop will be in the same direction as the drop? I2 R inFig. '7 but its numerical value will be only 1.? I2 R and this vector is shown indotted lines at 1.7 I2 R. of Fig. '7'. Combining this vector with the vector Va we obtain the resultant vector V2, which is indeed theresultant of all the vectors considered. Numerically, the vector, V2 is equal'to the vector V1 of Fig. 6 but is rotated in phase degrees withrespect thereto, so that as'the result of the 45-degree rotation of the crank shaft, there is a shifting in phase in the drop between the slidingcontact's of 45 degrees l without any change" in amplitude;

Now let us consider thedrop'for the last positions 8-8 of the two sliding contacts before coming back to their starting positions. It will be I noted that the two contacts are now directly op posite each other but are moving in opposite directions .as shown by the arrows in Fig. 5. The drop betweenthese two points is the resultant of reverse drops throughyfi B, through the upper condenser, and normal drops through the upper inductance and the. resistance 3R in the righthand branch.

This is shown vectorially in Fig. 8. First let'us draw the vector. .3 l1 R'for the reverse drop through the resistance in the right-hand branch.

"also :beclear that if the crank shaft is rotated continuously there will be a continuousrotation This vector is oppositein direction'to,thevector I2 XL for the upper inductance of the right-hand branch and upon combining these two vectors we get the resultant vector V e shown in dotted lines.

Next, let us draw the reverse drop through the upper condenser of the left-hand branch as shown at I1 Xo. Finally, let us draw the vector .3 I2 R 'representingthe drop'throughthe resistance between 7 and 8" in the rightvha-nd branch. This vector is opposite in direction to the ,vector I1 Xe and when these two vectors arecombined we get, a resultant vector Vr. If we now combine vectors Ve and Vr'we obtain the resultant or shift in phaseintroduced by the phase shifter With this understanding of the operation of the phase shifter in mind, let us assume that'the arm a of Fig. .2 carrying the pin is in'some. position between the arms '2) and b and thatgdue tosome cause such as change in temperature, the phase of the telephone line is shifted by a small amount. The output circuit of the push-pull detector arrangement will be unbalanced and relay R1 will set; the motor, M1 into rotation in one direction, thereby adjusting the-:phase shifter PSn until a compensating .phase shift is introduced. into the line, thus bringing the two wave components entering the detector again into proper relation, whereupon the armature of the polar relay R11 is restored. to neutral position and the motor ceases rotating. The wave entering the detector'from' the'line is now restored to'its original phase. Obviously, small changes in phase carrierwave withrespect to the control wave coming from the telephone line. :Again the pushpull circuit is unbalanced and the motor. Mn is set into rotation. The actual working of. the mechanism will depend-upon the position of the pin carried bythearm a. with respect to the. arms I) and b. ,Letuus suppose that 'atthe time the motor starts. to'rotate the-pin carried by the arm a is midway between the arms I) and b. As the'shaft S rotates,.the phase shifter may tend to momentarily bring about a balance of. the push-pull circuit, the adjustment of the phase shifter tending to compensate forv theaprier wave andithe control wave due tothe error in the oscillator. Since where the two waves are parent existing phase difference between the carnot at exactly thesame frequency,the effect of the error in the oscillator is cumulative, however,

the push-pullcircuit is immediately unbalanced 1 and the shaft S continues to rotate until the pin has been similar to its action where the difference the arm I) or b as the case may be,

carried by the arm a is brought into contact with V Unto this point the action of the mechanism in phase was due to an error in the telephone line itself. The action is one of attempting to" compensate by means of the phase shifter for the apparent-difference in; phase of the car-' rier wave and the control wave. As soon, however, as the pin comes in contact with the'arm b or b,-the cumulative phase difference due to the error of the oscillator causes the shaft S to continue to rotate and the shaft S begins to rotate with the shaft S,'thereby bringing about an adjustment of the variablecondenserCn.

There is a simultaneous adjustment of the phaseshifter PS, and these two actions continue together until a balance is again obtained. The

I result is that the frequency of the oscillator at the arm a resting against one of the arms 1) or the radio station has been changed slightly, tendto bring it back to the same frequency as the control wave.- The result of this action is, of

, c'ourse,to bring about an adjustment of the phase shifter PSn, which was not necessary in order to bring the carrier wave back to the same frequency as the control wave. This adjustment of; the

phase shifter cannot be avoided, however, be-

cause the "mechanism forv operating, the phase shifter cannot distinguish between a phase differ'ence due to the telephone" line and a cumulative phase difference due to an error in the'oscillator. w

As a consequence of thisadjustment J of the phase shifter, itwill be noted that the phase relation between the radio station at which the condenser Onwas adjusted and any other radio station in the chain will-to some extentbe a function of the correcting condenser. This' can be tolerated, however, so long as the shift does not occur too frequently, and it is assumed that a crystal-controlled oscillator of the type now known in the art, is good enough to avoid the necessity: of; adjusting 1 the correcting condenser too frequently. .After' the adjustment the mechanism has come to rest with the pin carried by b as the case may be, and with the shaft S shifted an amount corresponding to onehalf the lost mo tion plus the additional amount necessary to bring about the desired adjustment'of the correcting condenser 011., If, with the apparatus in this condition', amaximum phase change should occur in the line in suchdirection as to cause the shafts to rotate in the same direction, both the phase 55.

again operated until the condition of balance-was shifter and the correcting condenser would be restored. Of course, the adjustment of the variable condenser tends to assist the phase shifter in-compensat'ing for' the phase shift in the 6 u line, but the adjustment of the condenserresults in an over-correction of the frequency of the oscillator... This'will result ina cumulative phase difference between'the carrier wave and the control wavein the opposite sense to that which caused the previous correction of the oscillator, and the shaft S will: becaused 'to-rotateiin the opposite direction until thelost motion'is taken up in the opposite direction and the variable con-i denser and phase shifter simultaneously adjusted until the frequency of the oscillator is corrected.

The mechanism will now come to rest with the pinch the arm a resting against the other of the? arms bor b as the case may be. 'lf,in this condition a maximum phase shift'in a direction oppo site to that previously considered should'occur in I Y 1,926,169 a i V the line itself the push-pull circuit will be again unbalancedandthe shaft S will'be rotated in the same direction as it was last moved until the adjustment of the variable condenser ,Cn and the phase shifter'PSn compensate for the phase shift in'the line. This would also result in over-correctionof the variable condenser in the opposite sense to the former over-correction, so that a cumulative phasediiference would ultimately result which would cause the mechanism to correct the error as before.

Of course when the pin carried by the arm a lies in an intermediate position between the arms band b any phase shift in the line can be corrected by adjustment of the phase shifter PS1 so long as thephase changes in the line (which compensate to zero over along-period) do not require the pin carrled'by the arm a to move out of the angle between arms 1) and b. p 1

It will be apparentfrem the foregoing description, that the'phase relation ofthe carrier wavev of the'station illustrated with respect to that of the carrier wave of another station in the chain,

may shift by an amount corresponding toa rotation of the shaft Ssufficient to take up the complete loss of motion between the arms by and 1) plus twice the additional rotation of the shaft S necessary for the variablescondenser and phasesary to rotate the shaft S over and above the amount required to-take up theloss of motion, in order to keep thev possible phase difference between the station in question and any other 'staion in the system as small as possible. This may be accomplished by using a gearing to control the operation of the variable condenser On, this gearing being so arranged that the maxi'mumadjustment of the variable. condenser necessary to compensate for any phase change Which brings it into play, will be accompanied by a simultaneous adjustment of the phase shifter PS, which'is very small as compared with the adjustment of the phase shifter represented by a movement of the shaft sufii'cient to take up the complete loss of motion. The result is that the maximum of adjustment of the phase shifter which 'could normally be expected to occur would be but little greater than the adjustment corresponding to a movement of the shaft S sufficient totake up the complete loss of motion between the arms bandbfl Since the maximum adjustment of the phase shifter PSI; in any given direction" from its normal position'is a measure of the actual shift in phase of the generated carrier wave of the station with respect to the control oscillator, the carrier wave ofthe station can only get out of phase with respect to the carrier wave of any other 1 station whose phase relation has not changed, by an amount equivalent to this maximum adjustment of the phase shifter, since the 1 other station may be presumed to be in normal synchronous relation with respect to the common control oscillator. Now the maximum possible adjustment which maybe expected of the phase shifter PSn under any condition. beingbut little corresponding to one-half the angle greater than the angle 1 between the arms b and b, it becomes evident that even if another station in the chain should have its adjusting apparatus adjusted to the maximum phase, shift of its phase shifter in one direction, and the station here illustrated should haveits phase shifter adjusted the maximum amount in the opposite direction, the two stations would only be ou'trof phase by an amount equal to the angle between thearms b and la plus twice the additional rotation of the shaft in either direction that would be necessary for the maximum expected adjustment of the condenser On.

-In other words, ifboth stations had their carrierwaves in normal phaserelation; with the pin carried by the arm a at both stations just half- Way between the arms b and b", an error causing one of t e stations to move its shaft S to'its maximum expected adjustment in one direction would merely adjust its phase shifter anamount etween the arms 22 and b, plus the additional angle in the same direction corresponding to the maximum adjustment of the variable condenser Cu at that station. If, at the same time an equal but opposite error should occur. at the other station, causing its apparatus ,tobe adjusted in the opposite direction, the shaft S at the latter station would a be adjusted in the opposite direction by an amountequal to one-half -the angle between the" arms I) andb? plus the additional angle in the same direction corresponding to the maximum adjustment of the condenser On.

From the foregoing it will be seen that'the ar rangement of the invention not only tends to keep the'carrier waves of; all of-the stations of the chain at the'same frequencybut it limits the maximum amount that the carrier waves-of any two stations inayget out ofphase with respect to each other By means of the provision of the lost motion device, it provides for a compensat ing adjustment of the phaseshifter-alone with H out any necessary adjustment of the variable condenser controlling the oscillator, where there has been-a change in phase due entirely to the line transmitting the controlfrequency. This renders it unnecessary to make the. highly stable oscillator which generates the carrier wave follow the, relatively large phase'variations due to the line itself. I l

t will be obvious that the general principles hereindisclosed may be embodied in many other organizations widely difierent from those illustrated without departing from the, spirit of the invention as definedin the following claims.

What is claimed is:. Y I g 1; In a frequency control system, means to generate a wave whose frequency -is to be controlled, means to supply a control wave, relaymeans responsive to certain changes in thenormal phase relation of the control and controlled Waves, means controlled .hy said relay means to adjust the frequency ofjthe controlled'wave in response to'sueh changes; and means to defeat such adjustment when said relay means oper-. ates, in response to phase changes of the controlling wave only. "1 q 2. Ina frequency control system, means to generate a wave whose frequency is to be controlled, means ,to supply a control wave, relay means responsive to'certain changes in the normal phase relation of the controland controlled waves, means controlled by said relay means to adjustthe frequency of the controlled wave in response to such changes", and means to defeat suchadjustrnent when said relay means operates in response to small changes from the normal phase relation of the control and controlled waves.

3. ma frequency control system, means to generate awave whose frequencyis to be controlled, means to supply a control wave, a push- -pull detector to which both waves may be applied, meanscontrolled bysaid detector to adquency in response to phase'changes of the con-.

trolling wave only. I I

A. 1111a frequency control system, means to generate a wave whose frequency isto be 0on -v trolled, means to supply acontrol wave, apushpull detector to which both Waves may be ap-. plied, means controlled by said detector to adjust the frequency of said control. Wave in response. to certain changes from the normal phase relation of the control and. controlled waves applied to said detector, and means controlledby said detector to prevent adjustment ofv frequency in response to small changes from the normal phase relation ofv the control and controlled wave. V r

5. In a frequency control system, means at a in response to certain changes in the phase rela-' tion of the waves applied to said controlling device,v and. means to defeat such adjustment when said relay operates in response to phase changes in said line.

, 6. In a-frequency control system, means at'a station to generate a Wave whose frequency is to be controlled, means to generate a control:

wave atla distantv point, aline to, transmit said control wave from said distant point to said s'tation, a push-pull detector at said station to which said control and controlled wave may beapplied, means con rolled thereby to adjust the frequency of said;contro1led wave in response 'to certain changes in the phase relation of the waves applied to said detector, and means to prevent such adjustment in response to phase changes in said: line.

7. In a frequency control system, means at a station to generate a wave whose frequency is to be-controlled, means to generate a control wave at a distant point, a line to transmit said control wave from said distant point to'said station, a controlling deviceat said station to which said control and controlled waves maybe applied,

means controlled thereby to adjust the frequency ofsaid controlled wave in" response to certain changes in the phase relation of ,thewaves applied to said controlling device, a phase changer in said line, said phase changer being controlled by said controlling device, the connections between said controlling device and said frequency adjusting means and said phasechangerbeing: such that in response to phase changes in said line said phase changer will be adjusted to compensate therefor without operating said frequency adjusting means.

8. Inafrequency control system, means-at a,

station to generate a wave whose frequency is to be controlled means to generate a control wave at a distant point, a line to transmit said control wave from said distant'point to said station, a push-pull-detector at said station to which said 7 control and controlled waves may be applied,

, as 'ofjsaidrelay, an adjustable, phase shifter in said I telephone line, a frequency controller associated push-pulljdetectorat said station, means toapply andrsaidfrequencyadjustingw means and said phase changerbeingsuch 'that'in response to phase changes,in'said 'line said phase changer will be adjusted to compensate therefor without operating sai'd'fre'quency adjusting means.

9.- In a frequency control system,means at a station to generate a' wave Whose frequency is to be controlled," means to. generate a control ave at a distant point, a line to transmit said :control wave from said distant point to said sta;

, I tion; a push-pull detector at said station, means to apply said control and controlled waves to .1

said detector so that they will be differential with respect to the one and cumulative with respect to the'other unit oi said push-pull detector,

; a polarnrelay having an armature, said relay n being so connected 'tothe outputof said detector as to be neutral when the two waves are applied to said detector in quarter phase relation but to attract its armatureto one side or the other depending upon the direction in which said two waves depart ,from quarter phase relation, an operating mechanism controlled by the armature iththe generator whose Wave is to be controlled,

A andmeans operatively connecting said operating mechanism to said phase shifter and said frequency controller. V i I 1 10. In afrequency control system, means at a station to generate a wave whose frequency-is to be controlled, means to 'generate a control wave at a distant -point,-a line to transmit saidcontrol Wavefrom said distant point to said, station, a.

said control and controlled wave to said detector so that they will be differential with respect to the oneand cumulative with respectto theother unit-of said-push-pull detector, a polarjrelay having an armature, said relay being so cona nected to the output of said detector as to be frequency controller;

neutral when'the twowaves are applied to said detector in quarter phase relation but to attract:

its armature to one side or the other depending upon the direction inwhich said two waves depart from quart'er phase relation, an operating j mechanism controlled by the armature of said relay an adjustable phase shifter in said tele-' phone line,'a frequency controller associated with the generator whose wave is to be controlled, and

' 'means op eratively connecting said operating mechanism,to'said phase shifter and'said frequency-controller, said last mentioned means be-- ing 'so arranged as topermit limitedadjustment of said phase changer without, operating said llf'Infrequency control system, a phase changer, a frequency controller, a drive shaft,

connections from said shaft'for' directly driving said phase changer,'and a lost-motion drive between said shaft and said frequency controllen 12. In a frequency control. system, a phase changer, a frequency-controller, a driveshaft,

connections from said shaft for directly driving said phase changer, and means connectingsaid shaft and said frequency controller, said -means including means to delay the operation of said frequency controller afterthe beginning of rotation of said shaft.

13, In a frequency control system, means to generate a wave whose frequency is to be controlled, means to supply a control wave, a relay jointly controlled by said control wave and controlled. wave means operated by said relay to determine the frequency of said controlled wave, and means to defeat a change in frequency of the controlled wave when said relay operates in response to certain changes in the arriving phase of the controlling wave.

i l. In a frequency control system, a controlling source of cyclic vibrations, a controlled source of cyclic vibrations, each of said sources having small short period errors, a transmission medium connecting said sources; said medium having relatively large short period errors but substantially no long'period error, and means discrimi-' substantially no 'long period error, means discriminating between short period errors and long period errors for associating said transmission medium and said controlled source to prevent the phase angle of said controlled source from "varying more than a predetermined amount, and

for preventing the rate of phase drift from exceeding a predetermined value, and a phase compensator associated with said discriminating means and controlled thereby'to compensatefor short period errors in said medium.

16. In a frequency control system, a controlling source of cyclic vibrations, a controlled source of cyclic vibrations, a transmission medium connecting said sources, said medium hav-,- ing short time errors but substantially no long time error, and said controlled source having cumulative long time error, and means discriminating between'short period errors and cumulative long period errors for associating said transmission medium and said controlled source to prevent the phase angle of said controlled source from varying more than a predetermined amount,

and for preventing the rate of phase drift from exceeding a predetermined value.

' 1'7. In a frequency control'system, a controlling source of cyclic vibrations, -'a controlled 'sourc e of cyclic vibrations, atransmission mediumrconnecting said sources, said medium having short time errors but substantially no long time error,han'd said controlled source having cumulative longtime error, means discriminating between short period errors and cumulative long period errors, said means being interposed between said transmission medium and said controlled source to prevent the phase angle of said controlled source from varying more than predetermined amount and for preventing the rate of phase" drift from exceeding a predetermined value, and a phase compensator associated with said discriminating means and controlled thereby to compensate for short period errors in saidmedium. a

18. In a frequency control system, a controlling source of cyclic vibrations, a controlled source of cyclic vibrations, a transmission medium connecting said sources, said transmission medium having substantially no long time error and havingfshort time errorscumulating practically to zero over a comparatively long; period of time, said controlled sourcehaving errors'cumulating as the firstpower of time, and means disc'riminating between short time errors cumulating to zero and errors cumulating as the first power .of

time for associating ,said transmission medium and said controlled source to adjust the'lfrequency of said controlled source; said means preventing any adjustment of the frequency unless the phase drift accumulates to a predetermined angle.

19. In a frequency control system, a controlling f;

" source of cyclic vibrations, a controlled source of cyclic vibrations, a transmission medium connecting said sources, said transmission medium having substantiallyrno long time error and hav ing short time errors cumulating practically to zero over ,a comparatively long period of time, said controlled source having errors cumulating as the first power of time, a phase shifter to compensate for short time errors compensating the phase drift accumulates to an appreciable angle.

e 20. In a frequency control system, means to generate a wave whose frequency is to becontrolled, means to supply a control wave, means responsive to certain changes in the normal phase relation of the control and controlled waves to adjust the frequency of the controlled wave and readjust the phase relation between the control and controlled waves, and means to defeat said frequency adjustment while permitting said' phase adjustment in response to phase changes of the controlling wave only.

21. In a frequency control system, means to: generate a wave whosefrequency is to be controlled, means to supply a control wave, means responsive to certain changes in the normal phase relation of the control and controlled waves'to adjust the frequency of the controlled wave and to readjust'the phase relationof the control and controlled waves,'and means to "defeat said frequency adjustment while permitting phase adjustment in response to small changes from the normal phase relation of the control and controlled waves.

, HARRY NYQUIST.

to zero, a frequency adjuster to adjust the fre-

Images