PHILCO

BULLETIN.tolume 4 September and October, 1954._

"7; t714 ,A4ae

No. 3

Editorial John E. Remich 1

Training Slides-Homemade Bud M. Compton 2

What's Your Answer? 5

Thevenin Equivalents For Vacuum -Tube CircuitsFred Pfifferling and Richard Rojas 6

The Engineering Report Frank G. Kear 11

A Voice -Operated Relay Murray Elson 13

Servicing Synchros and Servomechanisms J J Mills 16

Solution to July -August "What's Your Answer?" 28

A New Philco Training Manual 29

PHILCO

TECHREP DIVISION

BULLETINPublished bimonthly by

The TechRep Division of Philco CorporationPhiladelphia, Pennsylvania

&litoi John E. Remich

,ittartagritg Gail W. Woodward

Yeeim ica &filo Francis R. Sherman

John W. Adams

Baldwin G. Sullivan

eafitoidal Office:

Technical Information SectionPhilco TechRep Division22nd St. and LehigliAvenuePhiladelphia 32, Penna.

If any information contained herein conflicts with a technical order,manual, or other official publication of the U.S. Armed Forces, theinformation in the official publication should be used. Views expressedin this publication are the opinions of the authors and do not necessarilyrepresent those of Philco Corporation. All technical information andcircuits are presented without responsibility as to patent infringement.

Copyright 1954, Philco Corporation

WANTED BULLETIN ARTICLES

by John E. Remich

Manager, Technical Department

The BULLETIN is devoted primarily to the type ofarticle which offers specific technical information andassistance to the field engineer, although it includes

timely news articles concerning the latest developmentsin the electronics art. In other words, the BULLETIN isyour publication. Obviously, then, most of the materialshould originate in the field-it should relate directly tothe problems encountered in the field and to the methodsby which these problems are solved. The BULLETIN isin need of a greater backlog of this typeof technicalarticle. From our conversations with visiting field engi-neers, we find that many fine ideas have been pigeon-holed simply because the engineer feels-he has no flairfor "technical writing." This should not he a deterrent,!ince the BULLETIN staff can edit the submitted materialand. if necessary. add the professional writing touch topint the aril, I, in proper form. This editorial. then, is acall for more technical articles from the men in the field.

We are ako anxious to receive your comments on themid any constructive ideas you may have to

Prol,ably v,e arc a little too close to the project for;I( .,Inektaking. We believe that your commentsAVOIdd OhirriiVr, and therefore helpful to us in makingthe MILLET IN moreour many teaderm,

it111,11111111vy and Mutt' vultrahle If)

TRAINING SLIDES HOMEMADE

by Bud M. ComptonPhilco Coordinator

A discussion of how to make your own training slides.This problem often besets the instructor, especiallywhen commercially prepared slides are not available or

local facilities are inadequate.

WHAT WE SEE is retained, accordingto some authorities, about 500% betterthan what we hear. That is why visualaids are so much desired in trainingactivities. Training slides, which are onetype of visual aid, offer an instructorthree important advantages: stimulationof class interest, saving in time, andease of storage.

Training slides stimulate interestpartly because they can be made locallyto fit the immediate requirements. Whenproperly made, they are clearer andeasier to follow than typical blackboardwork. Some of the students may volun-teer to help make the training-slide fileand thus will take personal pride andinterest in the project. Slides go a longway in dressing up the training andmaking it more "appetizing."

Slides are time savers in that they canbe used time and time again, whileblackboard diagrams must be redrawnevery time it is necessary to retrace stepsin the training. One evening of workmaking up slides will produce as muchtraining material as weeks or months ofwork preparing wall charts or black-board drawings. Since slides take upvery little space, they are much easier tostore than large, bulky wall charts.

In some instances, slides can presentinformation, such as radar scope photosor troubles, jamming, etc., which wouldbe very difficult to display to the stu-dents in any other manner. Any subjectmatter which should be shown in photo-

graphic form rather than by blackboardsketches or locally made wall charts ishandled best as a slide.

Before a program of slide making isinitiated, it is important that an inven-tory be made of the facilities that areavailable. A decision must be made asto the type and size of slides that are tobe used. Cameras and projector facilitieswill determine this. The most commonslide sizes are: (1) the popular 2" x 2"slide, which can be made with a 35 -mm.camera, and (2) the 214" x 21/4" size,which can be made on 120 or 620 rollfilm exposed in square frames. Mosttwin-lens reflex cameras are of this size."While some projectors, such as the oneshown in figure 1, can accommodateeither size slide, by using an adapter,most projectors will handle only onesize. In any event, a 35 -mm. projector(for 2" x 2" slides) 'will probably be

available to most instructors. Cameraequipment should, of course, be com-patible with the projector equipment.However, a larger camera can be used

for making smaller -size slides merely byplacing the camera at a greater distancefrom the chart or drawing to be photo-graphed, so that the image on the film is

limited to an area, in the center of theframe, corresponding to the size of slidedesired. Figure 2 shows a typical arrayof basic equipment.

In actually making the slides, it is

important to limit the amount of detailthat is included in each one. Experience

2

Figure 1. Example of a Typical Slide Projector. (The additional lens has a differentfocal length, and, while not necessary it will give the advantage of better control overprojected -picture size. Examples of glass (21/4" x 21/4" ) and cardboard (2" x 2")

mounts are shown. A choice of glass or cardboard is available in both sizes.)

Figure 2. Basic Equipment Used for Making and Projecting Slides. (The projectorshown can handle either 2" x 2" or 21/4" x 21/4" slides. The camera is equipped withextension tubes, which allow the making of close-ups. Two types of flood lamps are

shown, one with a metal reflector and the other with a built-in reflector.

3

Figure 3. Setup for Photographing Copy.(Simply place the material to be copiedagainst a wall, illuminate, and align andfocus the camera. An elevating tripod willsave considerable time in getting goodalignment. A single-lens reflex camera is

first choice for this type of work.)

will serve best in determining just howmuch detail may be shown on a given-

size slide. In case of schematic dia-grams, for a 2" x 2" slide, three stagesare usually optimum for maximum clar-ity. This number of stages enables theinstructor to show the stage under dis-cussion along with the one before it andthe one after it. With 214" x 21/4"slides, four stages can usually be clearlyshown. In case of block diagrams, agreater number of stages may be in-cluded. Factors that determine theamount of detail a slide can handle are:quality of camera, projector, and screen;technique and method of processing thefilm; and the resolution limitations ofthe film.

Material to be copied can be sup-ported against a wall by thumb tacks,gummed tape, etc., as shown in figure 3.The camera is placed on a tripod. (An

elevating tripod is ideal as it enablesrapid camera alignment with the copymaterial.) Two flood lamps, one on eachside, provide illumination, or an outsidewall in natural light will do. As mostcameras without attachments are inca-pable of making close-ups, the use ofextension tubes, extension bellows, orsupplementary close-up lenses will prob-ably be necessary.

Ordinary film can be used. Of course,such film as contrast pan, process, orlitho will give better results for black -

and -white line drawipgs. Color slidesmay be made where color improvesclarity. However, color is more expen-sive, and, if no local facilities are avail-able for processing, the time elementmay be undesirable.

The exposed and processed black -and -white film has a negative image, whilein most cases, it is desirable that theslide be positive. To make the positive,use the negative to make a contact printon slow -speed contrast film stock suchas a litho -type film. The slower the filmspeed, the easier it is to control theexposure in the contact printer. Evenwith the slowest of films, the printingtime is usually less than a second. Toobtain the correct exposure, one mustresort to trial and error. Those who arelacking in processing know-how shouldtry to get an experienc.ed person to helpwith the first few slides.

After the final positive transparencyis printed, it is ready to be mounted.Glass mounts are better than simplecardboard types and are recommendedfor those slides that will be used most.For light service, the cardboard mountshave the advantage of low cost and easymounting.

Before the time comes to show theslides, make sure the light in the roomcan be subdued without interfering withthe ventilation. It is not necessary nordesirable to have complete darkness. In

4

fact, there should be enough light topermit the students to take notes.

All the slides should be prearrangedin the correct sequence, so a trainedassistant (usually a member of the class)can operate the projector while the in-

structor stands near the screen to pointout the details as he lectures.

In conclusion, it should be pointedout that there is usually a wealth ofphotographic reference material in anymilitary or public library.

"What's Your Answer?"

This problem was recently shown tothe BULLETIN staff by John Meehanof the Headquarter's Instructional Staff.We found it so interesting that we arepassing it along to our readers.

Assume that the neon lamp in the following cir-cuit strikes at 90 volts, extinguishes at 10 volts, t ndhas negligible current drain (for the circuit shown)when it conducts.

R1

100 -VOLTBATTERY

The problem is to describe how the neon lampwill act after the switch is thrown. (Assume thatthe capacitors have no initial charge.)

(Solution next issue)

5

THEVENIN EQUIVALENTS FORVACUUM -TUBE CIRCUITS

by Fred Pfifferling and Richard RojasPhilco Government and Industrial Division

A simplified method of vacuum -tube circuit analysisusing Thevenin equivalents. The general rules and ap-proach are covered so that the reader can apply the

method to any circuit.

(Editor's Note: All rights to this article have been reservedby the authors; therefore, no portion of it may be ex-tracted, reprinted, or reproduced in any form, without

written permission.)

THE PURPOSE OF THIS ARTICLE is toderive a set of rules which will permitthe engineer to write down the equiva-lent circuits of commonly encounteredamplifiers at sight. Although the resultsderived are not original, the systematicapplication of these rules is somethingwhich has not been presented before.Armed with this set of rules, the engi-neer can quickly determine the outputimpedance and gain of amplifiers, andthus acquire new insight into circuit de-sign and analysis.

As a first approach, the general caseof an amplifier with current degenera-tion in the cathode will be examined.(See figure 1.) To derive the Theveninequivalent circuit, we proceed in theconventional manner to obtain theequivalent circuit looking into the plateand cathode. From part A of figure 1, itcan be seen that:

Egk E - lb Rk (1)

For the closed loop formed in theequivalent circuit shown in part B ofthe figure, it can be seen that:

µ,Egk -= ib (RL Rk rp) (2)

Combining equations (1) and (2), andsolving for ib, the following relationshipis obtained:

14 Eibµ+1

rp RL

Rk +

itLE(3)

rp RL (IL + 1) Rk

From equation (3) the two equivalentcircuits shown in figure 2 may be ob-tained.

The following set of rules can be de-duced from figure 2:

1. To transfer a voltage from grid toplate, multiply by µ.

2. To transfer a voltage from plate tocathode, divide by IL + 1.

3. To transfer a voltage from cathodeto plate, multiply by p. + 1.

4. To transfer a resistance from plateto cathode, divide by p. + 1.

5. To transfer a resistance from cath-ode to plate, multiply by p, + 1.

6. A resistance appearing in the gridcircuit does not appear in the equivalentcircuit because it draws no current.

7. To transfer a voltage from grid tocathode, first transfer it to the plate and

6

Figure 2. Equivalent Circuits of Figure 1A. Looking Into the Plate CircuitB. Looking Into the Cathode Circuit

then apply rule 2; if preferred, multiplythe voltage by

II, -I- 1

Either Ilk or R1, may be zero, as in thecase of grounded -cathode and cathode -follower amplifiers. The same rules willstill apply even in these special cases.

The second case to be discussed isthat of an amplifier with plate to gridfeedback as shown in part. A of figure 3.This amplifier has shown a host of usesin recent years.

In this type of amplifier the input

terminals arc grid and ground, and theoutput terminals are plate and ground.Feedback is obtained by connecting animpedance between plate and grid, whilethe external voltage is applied to theinput terminals in series with an im-pedance (source impedance). In the fol-lowing derivation it is assumed that nogrid current is drawn. From the equiva-lent circuit it can be seen that:

ant

I = I -l- = I-

147, KE,

I E0(11.)

(5)

7

K=GAIN OF AMPLIFIERWITHOUT FEEDBACK

Figure 3. Amplifier with Plate to Grid FeedbackA. Schematic DiagramB. Equivalent Circuit

Therefore, combining equations (4)and (5), the equation for the currentbecomes:

Ei (1 - K)g R2

Ei Ei

Rg R2

1 -K

(6)

From equation (6) another equiva-lent circuit may be obtained which hasthe same input current, input voltage,and output voltage as shown in figure 4.

Thus any impedance between plateand grid may be placed between gridand ground in the input circuit by sim-

ply dividing this impedance by (1-K).Therefore, another useful rule is ob-

tained. Since is usually small-2 Kcompared with Rg, and since K is large,the parallel combination is essentially

R,1 - K. The voltage which appears at

the input of the amplifier is simply:

R2

= R2 + ( 1 - K)R1 E (7)

Now using the rules previously de-rived, this voltage when transferred tothe plate is multiplied by and the cir-cuit shown in figure 5 is obtained look -

Figure 4. Alternate Equivalent Circuit of Figure 3

8

Figure 5. Equivalent Circuit of Figure3, Looking Into the Plate Circuit

ing into the plate. The plate current isgiven by the equation:

R2 p.ER+ K)111

i - (8)RL

The gain with feedback, K1, is givenby:

p, RL

(RL 1-0[R2 + (1 - K)R1] (9)

Since K > 1, and KR1 >> R2, theequation above becomes:

pAIL R2K1 =

( RL rP KRI

but since

K=RL

K1 = (11)

As an example of the use of the de-rived rules, the case of a common differ-ential amplifier, shown in figure 6, willbe examined. The advantages of theserules will become apparent.

Looking into the cathode of T1 andapplying rules 4 and 7, the equivalentcircuit shown in figure 7 is obtained.

Note that the equivalent circuit ob-tained is simply that of a cathode fol-lower. The output impedance, Ro, of thecathode follower, T1, is Rh in panillel

E0 pK -with rE , or:

1/.1 + 1

(10)

Figure 6. Schematic Diagram ofDifferential Amplifier

rp

(12)

Rh

and the voltage across Rh due to E1 is:

Ro

Eh - r+ 1) + 1

Rh

It now remains to bring Ro and Ehinto the plate of by using rules 3 and

(13)

Figure 7. Equivalent Circuit of Figure6, Looking Into the Cathode of T1

9

Figure 8. Final Equivalent Circuit of Figure 6

5, to obtain the final equivalent circuitshown in figure 8.

Note that this circuit is drawn withrespect to El. The voltage gain with re-spect to E1 is easily determined in theabove circuit and can be shown to begiven by the following equdtion:

Gain =

}RL

RL + rp2+ (µ2 + 1)r (14)

Pi rp{(p.i 4- 1) +

A similar analysis for E2 can be madeby this method, but this task is left tothe reader as an exercise in using theforegoing rules.

All the rules are summarized in thetable shown in figure 9. These resultsapply only to these cases for which thelinear equivalent tube circuit is applic-able. Many circuits lend themselves toanalysis by this method. Among othersthe conventional Wallman and thecathode -coupled Wallman circuits imme-diately come to mind. It is hoped thatthese rules will prove useful to engineersinvolved in vacuum -tube circuit analysisand design.

QUANTITY TO BETR AN SF ER ED

VALUE OF QUANTITY AFTER TRANSFER TO

PLATE CATHODE GRID TO GROUND

0 GRID11E i-l-

--,

GENERATOR E

CATHODE (p.-H)ECO GENERATOR

PLATE I

CO GENERATOR E

R GRID--MAW--RESISTOR 0 0

R-"wont--CATHODERESISTORR(1-1-1)

--MAAA/s-R

AAAM-- PLATE---A-,RESISTORR-\AAAAA-

R PLATE-TO-__AAAAAA__. GRID

RESISTORR

-NVVVV,-- I -K

Figure 9. Table Summarizing the Rules for Equivalent Circuits

10

THE ENGINEERING REPORT

by Frank G. Kear

One important link between the engineer and his pro-fessional and commercial associates is the EngineeringReport. Its significance and helpfulness are often under-estimated, and its preparation is occasionally and regret-tably neglected. Many readers of these Proceedings willaccordingly benefit from the following helpful analysisof report preparation by a Fellow of the Institute ofRadio Engineers, who is a consulting engineer and amember of Kear & Kennedy, of Washington, D. C.

(Editor's Note: This article originally appeared in theMay 1954, issue of Proceedings of the I.R.E., and is beingreprinted here with permission. The material applies sowell to the Field Engineer, that we felt our readers wouldappreciate it.)

THE TRADEMARK of the engineer is, bycommon acceptance, the slide rule, aninstrument also generally assumed to behis most important tool. There is, how-ever, a tool even more basic to an en-gineer, but one which is frequentlyneglected: this is the EngineeringReport.

From the beginning of his profes-sional education, the engineer receivestraining in the writing of reports. Manyof these are in the nature of laboratoryreports on specific tests, but a largenumber cover the investigation of moregeneral problems. Familiarity with theframework of these reports is vital toclear, organized thinking, and only bymeans of such reports can the engineeradequately present the results of hiswork. Upon completion of his formaleducation, the engineer is often prone tooverlook the continuing need for usingthe report framework as part and parcelof his thinking, with the result thatmuch effort and time expended in per-forming work is lost to industry becauseof failure to prepare an adequate and

complete presentation. Unless an engi-neer can convey to his management orhis client a correct and adequate pre-sentation of the work which he has per-formed, he, has failed in his obligations.

Since the Proceedings of the 1.R.E.constitutes the medium whereby engi-neering information is disseminatedwithin the profession, it seems appropri-ate that it include in its pages a re-minder of the importance of the properuse of the engineering report.

The basic report outline is simpleenough, having only five subdivisions-(1) the statement of the problem, (2)the facts bearing upon the problem,(3) an analysis and discussion of thesefacts, (4) the conclusions drawn there-from, and (5) recommendations. Let usconsider these.

(1.) Formulation of a correct state-ment of the problem presented is obvi-ously fundamental, yet frequently over-looked. The engineer must take ampletime to determine exactly what is re-quired of him in order that no time is

:11

wasted on issues which are not relevant.The statement should be reduced to writ-ing before any work is begun.

(2) The facts bearing upon the prob-lem should be carefully marshalled, andgreat care taken to include every per-tinent fact which the engineer will use.By so doing it becomes a simple matterat a later date to look back and deter-mine the reasons for a recommendedcourse of action. Frequently an engineerfinds it advisable to change his recom-mendations materially because of thediscovery of facts not originally con-sidered. Reference to the original reportwill immediately disclose the absence ofthose new facts, the inclusion of whichwould justify the change in recom-mended action.

(3) The analysis and discussion ofthe facts will vary greatly in size andformat, depending on the type of prob-lem. It could run into several notebooksfull of pertinent information and data,or it might be a mere paragraph. If theanalysis is lengthy, it should be reducedto summary form for inclusion in theexposition.

(4) The conclusions reached aftercompletion of the analysis should bestated concisely and clearly, pointingout the line of reasoning employed inreaching these conclusions and, ofcourse making certain that this reason-ing is substantiated by the precedingportions of the report. This subdivisionmust be carefully limited to the conclu-sions of the engineer making the report.It should not contain any new materialpreviously undisclosed, nor should itduplicate material which will be in-cluded in the fifth subdivision. Many en-

gineering reports will close at this point.The average paper published in theProceedings customarily contains thesefour subdivisions.

(5) The last and most important sub-division, although (as in the case of apaper for the Proceedings) it is notalways required, is the one containingrecommendations. The person or groupto whom this report is made is usuallyinterested only in this last paragraph. Itmust contain clear and concise recom-mendations as to the proper course ofaction to be taken as a result of theengineering work which has been per-formed. Brevity is desirable. The rec-ommendations should not be confusedby quotations from previous portions ofthe report. The ability to present thisparagraph properly represents the divid-ing line between the technician and theprofessional engineer. When an engineeris assigned a task, the presumption isthat he is competent to perform thistask. Accordingly, he should not wastetime and space in establishing this com-petency. The recipient of the reportmerely wants to know what he shoulddo and how he should do it. The recom-mendations, therefore, should be limitedto simple, clearly worded sentences,avoiding multiple or conflicting recom-mendations unless the writer clearly in-dicates the relative merits of each.

If the engineer follows carefully theforegoing outline in each problem pre-sented to him, he will find his work issimplified, his reports present a perma-nent record of the project, and manage-ment or the client will have the neces-sary information on which to base theiraction.

12

A VOICE-OPERATED RELAYby Murray Elson

Philco Field Engineer

A construction item that can perform a multitude ofuseful functions in communications systems.

O CCASIONALLY, the field engineer findsit necessary to construct a voice -oper-ated relay (VOR) in order to provideautomatic control, by voice, of impor-tant circuit functions. Such a relay hasseveral applications. It may be used inconjunction with a receiver to turn on awarning light, to start a recorder at thebeginning of a radiotelephone transmis-sion, to switch a series of alarms, and toperform other functions suitable forcontrol by voice.

APPLICATION

The VOR described in this article wasspecifically designed to control the oper-ation of a recorder used to record radio-telephone transmissions, and in this ca-pacity, it has given excellent service. Itssensitivity, stability, and overall per-formance has been very satisfactory.The unit is relatively simple to con-struct, and with minor circuit modifica-tion and adjustment, it can easily beadapted to other applications. To aidthe field engineer in making such adap-tations, a brief description of the theoryand operation of the circuit is included.

THEORY OF OPERATION

In most applications of the VOR thelength of time required for the unit toaccomplish its switching function afterthe input signal is applied or removedis very important, and must be care-fully evaluated in consideration of thecircuit that is to be controlled. In gen-eral, the time required for energizingthe VOR should be as short as possiblecompatible with the requirements of the

controlled circuit, and the holding timeafter the input signal is removed shouldbe somewhat longer than the time be-

tween successive utterances. In addition,the VOR should be positive acting, andit should be easily adjusted to acceptdifferent level input signals.

Figure 1 shows a simplified block dia-gram of the setup for controlling therecording of radiotelephone transmis-sions using the VOR. Figure 2 showsthe schematic diagram of the VORcircuit.

Referring to the schematic diagram,it can be seen that the VOR consists ofa clamping circuit (V1), a bias supply(V2), and a thyratron relay control cir-cuit (V3)r

The clamping circuit, employing bothsections of a 6H6 duodiode, V1, clamps

Figure 1. Simplified Block Diagram ofSetup for Automatic Control of a

Recorder with a Voice -Operated Relay

13

the negative extremity of the input-

signal waveform at the fixed -bias levelestablished for the thyratron relay con-trol tube, V3, so that the input signalvaries only in a positive direction withrespect to the bias level. Thus the peak -to -peak voltage of the signal, rather thanmerely the peak voltage, is effective intriggering the thyratron, thereby ren-dering the circuit more sensitive thanit would be if clamping were not used.The input signal is applied to the cath-odes of clamping tube V1 through theinput circuit consisting of coupling ca-pacitor C1, resistor R1, connected fromthe cathodes to the plates of the tube,and capacitor C2, connected from thejunction of the plates and resistor R1 toground. The clamped output is fed di-rectly from the cathodes of V1 to thecontrol grid of V3.

The bias supply, using both sectionsof a 6X5 rectifier, V2, and connected

for half -wave rectification, provides thenecessary bias voltages for the thyra-tron tube, V3. A -c voltage for the supplyis obtained from the 115 -volt, a-c powersource to which the VOR is connected.After the negative rectified output isfiltered by the R -C filter consisting ofresistor R4 and capacitors C3 and C4 aportion of the voltage is suppliedthrough potentiometer R3 to the platesof V1 and thence through resistor R1 tothe control grid of V3. This potentiom-eter is usually set for a voltage, at theplates of V1 slightly greater than minus5 volts, which is the firing point of thethyratron.

The thyratron relay control circuitemploys a type 2050 gas tetrode, V3 tocontrol the operation of a double -pole,single -throw plate relay, K1, which ac-complishes the switching functions pro-vided by the VOR. A -c plate voltage forthe thyratron is supplied from the sec-

RECORDING( 0-0-MOTOR

STARTING 0-0-0 0

AF OUT00 0AF INO

-I-TS!

ALL CAPACITORS ARE INJJ.F

* MERIT 3045 OREQUIVALENT

**3000 -OHMSENSITIVE RELAY WITH

NORMALLY OPENDPST CONTACTS

C21041.01

R5

Nam CI

.01

V32050

Figure 2. Schematic Diagram of Voice -Operated Relay

14

ondary of transformer T1 through therelay coil and current limiting resistorR5. With no signal applied to the con-trol grid, V3 is held below the firingpotential by the bias voltage, and therelay is unenergized. When the thyra-tron is triggered by an input signal, thetube ionizes, and the flow of rectifiedplate current through current limitingresistor R5 rapidly charges capacitorsCr,, Cc, C7, and C8, connected in parallelwith the relay coil. After a very shortdelay, the charge on the capacitors be-comes great enough to provide the 1 -ma.current required to energize the relay,and after the additional delay inherentin the relay, the relay is activated, giv-ing a total elapsed time of approximately2 milliseconds following the firing ofthe thyratron. As the relay activates, oneset of contacts applies power to the re-corder motor, and the other set connectsthe audio output of the communicationsreceiver to the recorder circuits.

Since the thyratron plate voltage isa.c., the length of time required for thetube to fire after the application of aninput signal will vary, depending uponthe amplitude and polarity of the platevoltage at the instant the input signal isapplied. Assuming that the VOR is con-nected to a 60 -cycle source and that thethyratron fires at the positive peak ofthe a -c cycle, the maximum possible de-lay time before firing occurs is approxi-mately 17 milliseconds, the period ofone a -c cycle. It can be seen therefore,that the maximum possible delay timefrom the titne of arrival of the inputsignal to the time the relay is activatedis approximately 19 milliseconds (17 -1-2). Of course, more rapid operation ispossible if the VOR is connected to a400 -cycle power source, but a 19 milli-second delay cannot be detected by thehuman ear.

When the input signal is removedfrom the control grid of V:1, the thyra-Iron becomes deionized on the follow-ing negative half cycle of primary power

and ceases to conduct. Relay K1, how-ever, is maintained in an activated con-dition by the slow discharge of capaci-tors C5, C8, C7 and C8 through the relaycoil. With a total capacitance of 160 to200 ILL in the circuit, the relay will re-main activated for 3 to 4 seconds beforethe current decreases below the 0.5 -ma.drop -out value of the relay. The lengthof time that the relay holds after V3stops conducting is a function of thetime constant of the four capacitors andthe relay coil, relay characteristics andthe leakage of the capacitors. It shouldbe noted that the holding time will de-crease with aging of the capacitors;hence, the capacitors should be replacedabout every 18 months.

In case it is desirable to use a sensi-tive relay that has characteristics dif-ferent from the one described in thisarticle, it is merely necessary to alterthe value of shunt capacitance to a valuethat results in the desired holdinginterval.

Careful adjustment of the bias withpotentiometer P3 is required to preventunwanted-, closure of the relay whennoise is present at the input of the VOR.This is especially true with intermittentnoise such as the ignition type.

CONCLUSION

Operation of the VOR may be initi-ated not. only by audio signals but alsoby higher -frequency signals and evenby a d -c voltage. In the case of therecorder application described above,this fact can be used to advantage toprevent loss of the first part of thetransmission during the interval (about1/1 second) required for the recordermotor to ranch recording speed. 13y con-necting the V011 to the receiver so thatit is carrier -operated rather than voice -operated, the recorder motor will beturned on as soon as a carrier appears,and the recorder will be, ready for re-cording by the lime the transnrissiostarts.

15

SERVICING SYNCHROS ANDSERVOMECHANISMS

PART 3 OF A SERIES

by J. J. MillsPhilco Field Engineer

The third installment of a popular series devoted tosynchro and servomechanism practices. Part 3 exploresthe theory and operation of dual -speed positioningsystems as well as the electronic circuitry associated with

them.

THE FIRST INSTALLMENT of this series(May -June, 1954, BULLETIN) coveredthe theory, operation, and constructionof synchro transmitters, followers, anddifferentials. Part 2 of the series (July-

August, 1954, BULLETIN) continuedwith a study of the synchro controltransformer and its operation as a nulldevice in single -speed servomechanisms.In this installment, the author dealswith dual -speed servomechanisms and

explains the importance of crossovernetworks, anti-stickoff voltages, andother circuit refinements necessary forsatisfactory servomechanism perform-ance.

DUAL -SPEED SYNCHROPOSITIONING SERVOMECHANISMS

The combined static accuracy of thesingle -speed positioning servo, describedin the last installment, is limited chiefly

0II5V

0

DIAL

RI

R2IGIX

136 :I

IG36X

SI SI RIS2

S3S2 ICT

S3

SI

IX R2

36 I - - --I

SI RI

S2 S2 ICTS3

CROSS-OVER

NETWORK

36XS3 R2

MFI

MF2

DIAL

LOAD

CF2

SERVOAMPLIFIER

CFI

SERVO MOTOR

Figure I. Components of Simple Two -Speed Servo System

16

by the static accuracy of the synchrosinvolved. The combined static errors ofa 1G and a 1CT run as high as 1° ormore. In order to reduce this staticerror, dual -speed data transmission sys-tems are frequently used.

A typical dual -speed positioning servois shown in block diagram form in fig-ure 1. It may be seen that the onlycomponents added to the single -speedservomechanism are:

1. A 1G, geared to the data transmit-ter dial and the single -speed 1G ata 36:1 ratio.

2. A 1CT, geared to the follow-upsystem dial and to the single -speed1CT at a 36:1 ratio.

3. A crossover network, which deter-mines whether the single -speed(1X) rotor voltage or the 36Xrotor voltage is fed to the servoamplifier.

If the combined static error of a 1Gand a 1CT is 1° of mechanical rotation,and if these synchros are geared to theirrespective dials and mechanisms at a36:1 ratio, they will still have a com-bined static error of 1° of mechanicalrotation, but 1° of their mechanical ro-tation now means only 1/36th of a

degree of dial error. Hence, the dial -indication accuracy of this system isconsiderably greater than that of asingle -speed system.

Another advantage of the high-speedsystem is that it increases the "gain" ofthe data transmission system (in voltsper degree of error angle), and there-fore the "stiffness" of positioning themechanism. For example, for an errorangle of .25° (neglecting static errors),a voltage will be induced in the 1X CTrotor equal to 55 sin .25°, or about .25volt. However, the voltage induced inthe 36X CT rotor will be equal to 55sin 9°, or about 9 volts. Thus, the 36Xsystem will provide more sensitive, more

accurate, and more rapid positioning ofthe mechanism because its higher rotorvoltage, when amplified, will drive theservomotor and associated CT's to alower error -angle position than could beattained using the 1X system alone.

The 36X synchros make a completerevolution, and therefore have true zeropositions, for every 10° rotation of theirdials. A 36X positioning system by it-self, then, would have 35 false zeros.Therefore, the 1X system must be re-tained in order to ensure that the re-mote mechanism will not be displaced10° or some multiple thereof from thedata transmitter.

The Crossover Network

The crossover network determineswhether the 1X or the 36X CT rotorvoltage will be fed to the servo amplifierand, therefore, which of these voltageswill control the positioning of the mech-anism at any given error angle. Whenthe error angle is large, the 1X systemmust be in control so as to prevent themechanisnl from locking in on one ofthe 36X false zeros; when the errorangle is small, it is desirable to have the36X system in control in order to posi-tion the mechanism stiffly and accu-rately. In general, in a 1- and 36 -speedsystem, the crossover network (some-times called a synchronizing circuit) isdesigned so that the 36X system is putin control for error angles less than 2.5°(an error angle of 2.5° corresponds toa maximum 36X CT induced rotor volt-age, the closest maximum to zero -degreeerror angle).

A widely used crossover system isillustrated in figure 2. The relay is nor-mally de -energized and the 36X systemis in control. Voltage induced in the 1XCT rotor, caused by an error angle, isamplified and fed to the vacuum -tuberelay circuit. The gain of V1 or thebias of V2 is set so that when the errorangle between the output shaft and the

17

FROMDUAL- SPEED

SYNCHRO -TRANSMISSION

SYSTEM

0 SI

S2

S3

I CTIX

36X

IX

RI

R2

36.1

SI RI

S2

S3 R2

CHANGE -OVERRELAY

1-- -1

I

L

-105V

-4" TOSERVO

AMP

Figure 2. Change -Over System, Relay Type

data transmitter exceeds 2.5° (cor-responding to an induced 1X CT rotorvoltage of 55 sin 2.5°, or about 2.5volts), the relay will be energized, plac-ing the 1X CT in control of positioningthe mechanism until the error angle be-comes less than 2.5', at which point therelay drops out, permitting the 36X sys-

tern to position the mechanism accu-rately to a zero -degree error angle.

Another type of crossover networkfrequently used is shown in figure 3.This system eliminates the relay and issmooth and trouble -free in its perform-ance characteristics. In operation, the

fnz

ccowI-1-

>-Nr01

0WW 0a-W

--1

0

0ccu.

S2

SI

0---0-

II5V60''LREF.EXC.

I CTIX

RI

-2.4 V+4.4 VO

1 X

ERRORVOLTAGE

LOAD

0 DIAL

VI6H6

100K

V2 I m" 0.511F -,6SN7

36XERROR

VOLTAGE

ANTI- -HUNT

VOLTAGE

INDUCTIONGEN

.12 5

0 0 0

REFFIELD

SERVOMOTOR

2.2K

40UF

T

+330V

Figure 3. Change -Over System, Continuous Type

18

current through the first section of V2and the total cathode resistance aresuch that a 1 -volt bias appears at R2 ofthe 1X CT. A twin diode, V1, is con-nected to the grid circuit in such a man-ner as to limit grid signals to valuesbetween plus 4.4 volts and minus 2.4volts. Neglecting the damping voltageintroduced into the circuit by the induc-tion generator, the 1 -volt bias of thesignal source (1X CT) means that lim-iting occurs for source voltages in ex-cess of 3.4 volts. As stated previously,the r -m -s value of the 1X CT error volt-age is related to the error angle by theequation:

Er.m.s. = 55 sin 0

Where 0 is the error angle between theremotely positioned mechanism and thedata transmitter.

The maximum instantaneous (peak)value of the error voltage is, then,

ei = V2-(55) sin 0

The error voltage of the 1X CT reachesa value of 3.4 volts when the errorangle, 0, is such that

ei = AT2-(55) sin 0 = 3.4Then,

3.45)

sin 0 = = 2.5°-0-(5In solving the above equation and ob-taining a value for 0 of 2.5°, it is seenthat limiting does not occur for errorangles less than 2.5°. With no limiting,the cathode and grid signals are identi-cal, and negligible output appears at thefirst anode (pin 2) of V2. About onefourth of the 36X CT rotor voltage,introduced at the second cathode (pin6) of V2, is amplified and will be incontrol of positioning the mechanism.

When the error angle exceeds 2.5°,the signal at the first cathode (pin 3)can increase to a maximum r -m -s valueof 55 volts (at 90° error angle). Thesignal at the first grid (pin 1) how-ever, is constrained to the values pre-

viously stated. A voltage difference be-tween grid and cathode, therefore, isestablished during those portions of thesignal excursions when the amplitudeexceeds plus or minus 3.4 volts. Suchportions of the signal are amplified andappear at the second grid (pin 4) witha gain of 10.

It is interesting to note the behaviorof this system at various error anglesbetween 0° and 10°. When the errorangle is less than 2.5°, only the 36Xsignal is amplified. For error anglesbetween 2.5° and 5°, both the 1- and36 -speed signals are amplified in thesecond section of V2, but they are bothof similar phase and tend to drive themotor in the same direction. At anerror angle of 5°, the 36X rotor voltageis zero, and only the 1X CT rotor volt-age is amplified and fed to the controlfield of the servomotor. At error anglesbetween 5° and 10°, the phase of the36X CT rotor voltage is such (180° outof phase with the 1X CT) as to drivethe servomotor to an error angle of 10°,whereas the 1X CT rotor voltage is stillof a phase that tends to drive the mech-anism to a 0° error angle. Consequently,it might be said that for error anglesbetween 5° and 10°, the 1- and 36 -speedsystems fight for control of the directionof servomotor rotation. This effect ismost pronounced at an. error angle of7.5°, since the induced rotor voltage ofthe 36X CT is at a maximum at thaterror angle and its phase is in the wrongdirection. It is necessary, therefore, thatthe combined voltages at the secondanode of V2 be such that the 1X rotorsignal may override the 36X rotor sig-nal when the latter is in the 7.5° posi-tion. Since the gear ratio is 36:1, if alinear relationship is assumed (for smallangles), the 36X rotor voltage will riseat a rate 36 times faster than that of the1X rotor voltage. Therefore, it appearsthat the 1X signal gain must be mademore than 36 times that of the 36X sig-nal (if it were exactly 36, the input to

19

Vg would be zero volts at a 7.5° errorangle). In the system described, the 1Xsignal is amplified by a factor of 10,while the 36X signal is attenuated by afactor of 4. Effectively, then, the 1Xsignal has a gain of 40, which fulfillsthe requirements satisfactorily.

The foregoing analysis is importantto the field engineer as well as to thedesign engineer; in systems using thistype of crossover network, the possibil-ity that a mechanism may lock out by a7.5° error angle must be considered.The field engineer must accomplish thesystem checkout with this fact in mind.If the synchros in the system are notquite mutually zeroed, or if the circuitcomponents have changed somewhat invalue, it frequently happens that themechanism will lock out 7° or 8° awayfrom zero error angle. This possibilitymay be checked by turning off theequipment and setting the remote mech-anism to an 8° or 10° error angle, thenturning on the equipment again andchecking to see that the mechanism doesnot encounter too much difficulty ingetting past the 7.5° error angle on itsway to 0° error angle. Another goodcheck is made by turning the data trans-mitter through at least 20° of rotationto a zero dial reading from each direc-tion since, if the synchros are not mu-tually zeroed,.the mechanism will tendto lock out more easily in one directionthan in the other.

It will be noted that the power outputstage of the servo amplifier is connectedas a cathode follower, the output loadconsisting of the control field of theservomotor, shunted by 5.7 of capac-itance. This arrangement eliminates theneed for an output transformer, pro-vides some damping, and smooths theoutput waveform so that the harmoniccontent of the signal at the grid of V3has negligible effect on the servomotor'soperation.

Further damping is provided by the

induction generator, the rotor of whichis mounted on the servomotor shaft.One coil of the generator is excited by82 volts, 60 cycles. Voltage is inducedin the other coil by the rotating actionof the induction -cup rotor, an outputvoltage of about 6 volts being generatedat a speed of 1000 r.p.m. One disadvan-tage of a large damping factor is thatit limits the top running speed of themechanism. In this system, this disad-vantage is overcome by feeding backthe induction generator voltage to theservo controller at a point where it doesnot take effect except at small errorangles, where the top running speed isnot of prime importance. The phase ofthe induction generator output voltageis controlled in design by the phase-

shifting network in series with the exci-tation winding of the generator, so thatthe induction generator output is 180°out of phase with the 1X CT rotorvoltage. It is introduced across the 10Kresistor between the grid and cathodeof the first section of V2. When theerror angle is greater than 2.5°, limitingoccurs at the grid of the first section ofV2, and the grid -cathode potential dif-ference is caused principally by the 1XCT rotor voltage at the cathode. Forthis condition, the induction generatorvoltage has little effect on the signal atpin 1 of V2. When the error angle isless than 2.5°, the grid -cathode poten-tial difference may be attributed chieflyto the induction generator output, be-cause the rotor voltage of the 1X CT isless than 3.4 volts. Since the signal atpin 1 of V2 may now effectively be con-sidered as the output of the generator,and since this damping voltage is ofsuch a phase as to tend to reverse thedirection of the motor, it will, whencombined with the 36X order in the sec-ond section of V2 limit the speed of themechanism as it approaches zero errorangle, and aid considerably in bringingthe mechanism to a smooth stop withno overshoot. Thus, it is seen that the

20

function of the crossover network is toeliminate the possibility that the mech-anism will lock out at some multiple of10° error angle.

Anti-Stickoff VoltagesAnother ambiguity that must be over-

come in dual -speed positioning servo-mechanisms is the possibility that themechanism will lock out at a 180° errorangle. At this error angle both the 1Xand 36X CT rotor voltages are zero. The1X CT rotor is at a false zero; the 36XCT rotor is at a true zero, since it goesthrough a true zero every ten degrees.If the 1X system were in control, anytransient would cause the mechanism toposition itself quickly to zero errorangle. However, if the error angleshould be within 2.5° of 180° when thesystem is first turned on, the 36X systemis in control, and remains in controlunless a large transient should cause the

mechanism to lag more than 182.5°, atwhich time the crossover network wouldallow the 1X system to take control (forexample, this might occur when slewingthe data transmitter). However, it canbe seen that with the 36X system in con-trol, the mechanism will remain 180°away from its correct position.

This ambiguity can be overcome byplacing an anti-stickoff voltage in serieswith the rotor voltage of the 1X CT.The amplitude of this voltage must belarge enough to operate the crossovernetwork, but not so large that the 1Xfalse zero must be displaced enough .toapproach the next stable null of thehigh-speed (36X) system. In a 1- and36 -speed system, the r -m -s value of theanti-stickoff voltage is usually about 2.5volts.

Figure 4 illustrates the use of an anti-stickoff voltage in a dual -speed system.

R

115VA -C

OR2

RI

SI

16 S2

IXS3

1:36

IGX}2 S3

SI

S

TO REFERENCEFIELD OF

IP- SERVO MOTOR

fcs

s.1

EASY

SI

S2 ICTIX

S3\4\ R211:36

S2 ICT36X

S3\

RI

R2

IX ERRORVOLTAGE=ER + EASV

IX

36X

TOCROSSOVER

NETWORK

EASV- ANTI-STICKOFF VOLTAGE (USUALLY ABOUT 2.5 VOLTS INI AND 36X SYSTEMS)

ER IS THE INDUCED VOLTAGE DUE TO OFFSET OF CT FROMELECTRICAL ZERO PLUS THE ERROR VOLTAGE INDUCED DUETO AN ERROR ANGLE.

AT ZERO ERROR ANGLE, ER=EAsy AND IS lee OUT OF PHASEWITH IT ER+ EARN./= 0 VOLTS.

AT 180° ERROR ANGLE (THE FALSE ZERO OF THE IX ICT),ER' EASV AND IS IN PHASE WITH IT ER+ EARN/ 5 VOLTS,

WHICH IS ENOUGH TO OPERATE THE CHANGE-OVERSYSTEM,PUTTING THE IX ICT IN CONTROL

Figure 4. Introduction of Anti-Stickoff Voltage in Two -Speed System,

21

The 1X CT must be offset from its elec-trical zero position by an amount nec-essary to make the sum of the anti-stickoff voltage and the 1X CT rotorvoltage zero when a zero error angleexists between the remote mechanismand the data transmitter. The directionof offset, of course, must be such thatthe two voltages are 180° out of phase.If the anti-stickoff voltage is 2.5 volts,the 1X CT must be offset approximately2.5°, since for small error angles the CTerror voltage is about 1 volt per degreeof error angle. If, at a zero degree errorangle, these two voltages are made equaland 180° out of phase, they will beequal and in phase at an error angle of180°. Their sum at this angle will be 5volts, enough to permit the one -speedsystem to take over and position themechanism to zero error angle, thuspreventing the mechanism from lockingout at an error angle of 180°.

Lead Reversals

If the remote mechanism rotates inthe wrong direction, the reversal of astator lead to one or both CT's is indi-

.cated. If the mechanism is positioned180° from its correct position, severaldefects are possible:

1. The 1X synchro generator or con-trol transformer may be zeroed180° out.

2. The 1X CT rotor leads may be re-versed.

3. The leads to the reference field ofthe servomotor may be reversed.

4. If no anti-stickoff voltage is pres-ent, the mechanism may be lockedout on a false zero of the 1Xsystem.

Zeroing Dual -SpeedCT -Positioned Mechanisms

Instructions for zeroing the synchrosand the remote mechanisms in a dual-

speed servo system generally call for thefollowing procedure:

1. Secure the system.

2. Set the data transmitter on zero,and set the synchro generators orDG's geared to the data transmit-ter to their electrical zero positions(using the jumper method previ-ously described) by loosening thestator clamps and moving the syn-chro barrels.

3. Set the remote mechanism so thatits dial reads zero.

4. Block the mechanism in positionby means of soft wooden blocks.

5. Zero the control transformersgeared to the mechanism by thejumper method.

If the system can be energized, a sim-pler and more effective zeroing pro-cedure can be used. After setting thedata transmitter to zero, check its syn-chro generators or DG's for electricalzero by using a test synchro in conjunc-tion with a voltmeter, using the latterfor a fine check of the S1 -S3 voltage (orthe R1 -R3 voltage, in the case of a DG).

Remembering that the 1X CT returnsthe mechanism to within about 2.5° ofzero error angle, and that the 36X sys-tem then takes over -to position themechanism to a very small error angle,it is only necessary to perform the fol-lowing:

1. Disconnect the ungrounded rotorlead of the 36X CT, and turn the 1X CTbarrel until the remote mechanism dialreads zero.

2. Reconnect the 36X CT rotor lead,disconnect the ungrounded 1X CT rotorlead, and turn the 36X CT barrel untilthe dial reads 0°, 10°, or some multipleof 10°.

The method outlined above has manyadvantages:

22

1. It takes considerably less time thanany other method.

2. It provides a separate "power -on"check of the 1X system and the 36Xsystem. Each speed is thus checked tosee that it will do its part in positioningthe mechanism to zero error angle.

3. The operation of the crossover net-work can be checked while the 1X CTrotor lead is disconnected. This may bedone by securing the system, and set-ting the remote mechanism at an errorangle of 10° or more. With the 1X CT

rotor lead disconnected, and with thepower again turned on, the mechanismshould turn to the nearest multiple of10° error angle. Touching the 1X CTrotor lead to its terminal should thenoperate the crossover network, causingthe lX system to take over until theerror angle decreases to 2.5°, at whichtime the 36X system should again take

over for the stiff positioning of themechanism to a zero error angle.

4. The direction of rotation producedby each system can be checked sepa-rately.

5. No special arrangement need beused in the zeroing procedure when ananti-stickoff voltage is used in the sys-tem. (In some equipments, accuratelyspaced scribed marks are used to indi-cate the offset required to compensatefor the anti-stickoff voltage.) The onlyrequireawnt is that the l X CI' be offsetby an amount such that ale SUM of theIX CI' rotor voltage and the anti -stick -off voltage is equal to zero at zero -

degree error angle. Since, with thepower on, the CT is driven 10 a "zrrofour!" ii0=j1 ion, Ihe offset is !WI in autor11:111etIlly by the "power -on" el ofZeroittr.

6. 11 makes sill excellent demonstra-tion fra- onthejol, traininc. TIhe elec.Ironic re/11,661in can 11111(11 more rend.ily iwturiv, the op,trition of the vittiomcomponents of the ,,y7,tem by p.eeing

method in use than he can by seeing amass of jumpers, wooden blocks, etc.,which mean nothing to him so far as theactual operation of each major compo-nent is concerned.

A few qualifying statements shouldbe made concerning the "power-on"method of zeroing. When zeroing the1X CT with the 36X CT rotor lead dis-connected (step 1), it will he observedthat the stiffness of positioning is verypoor. Moreover, if a relay -type cross-over system is used, and set to operateat a 2.5° error angle, obviously therewill be no torque at all for 2.5° eitherside of zero error angle. This slight dis-advantage can be overcome by turningthe 1.X CT until the dial is close tozero, setting the remote mechanism tozero dial reading by hand (making surethat it stays there), and slightly adjust-ing the CT to a zero rotor voltage, or toa "zero sum" voltage if an anti-stickoffvoltage is present.

The Positioned Load

The load which is positioned byservomechanisms can, of course, he any-thing that is rotatable. In most simpleelectronic systems, the only use of aservomechanism is in positioning suchelements as a radar antenna or a sonartransducer. In the more complex andintegrated systems, however, servomech-anisms are required to position manytypes of components, such as resolvers,linear potentiometers, synehro transmit-ter, one side of a mechanical differen-tial, etc. When working with such sys-tems, it is helpful in onthe-joh trainingIn emphasize strongly the difference be-tween the positioning elements and theirmitioned elements of the mechanism,since, in many instances, this is not looapparent. In most systems, for example,cool rot I raipifortncrm arc used in con-junction svith a servo amplifier and mo-tor Its the positirming elements ,,f thesystem, hut many other types of mill

GYROEXC.

GYRO

IG36X

GYRO RELAY TRANSMITTER

SERVOCONTROLLER

I CTIX

I- - - -ICT36X

C.F

R.F.

SERVOMOTOR

COMMUTATORTRANSFORMERS

8FOLLOWERS

SYNCHRO AMPLIFIER

SERVOCONTROLLER

I CTIX

ICT36X

R.F.

ONE -LINE DIAGRAM

OE

SERVOMOTOR

. 4R FOLLOWERS

S

Figure 5. Synchro Amplifierdevices are in use today. In some com-puters, a resolver is used as a null posi-

positions synchro transmitters, while inone computer (now obsolete) a controltransformer is employed in the usualmanner as a positioning null device ina mechanism which positions anothercontrol transformer as the load. Theseand other computing mechanisms willbe discussed in detail in a future install-ment of this series.

Synchro Amplifiers

A control transformer presents amuch more desirable load to a synchrotransmitter than does a synchro fol-lower, because the CT has a muchhigher impedance than a follower, andtherefore draws much less current fromthe transmitter. Furthermore, no voltageunbalance can be fed back from thecontrol transformer to the synchrotransmitter, as it could be from a fol-lower whose dial, for example, wasbinding mechanically. For these rea-sons, servomechanisms called synchro

amplifiers are frequently used aboardship.

Figure 5 illustrates the use of a syn-chro amplifier to increase the allowableload of followers on the synchro trans-mission system of a gyrocompass. As-sume that the gyro transmission systemwas designed to accommodate 12 re-peaters, but that some time later, it wasfound necessary to install four addi-tional units. The 1X and 36X CT's ofthe synchro amplifier present negligibleload to the gyro transmission system. Itis seen that the synchro amplifier isnothing more than a dual -speed posi-tioning servomechanism which positionsas its load a size 5 synchro transmitter.The 5G will position, without beingoverloaded, the four additional follow-ers required in the system.

The synchro amplifier can be zeroedeasily in the following manner:

1. Set the gyrocompass to zero.

2. Check the voltage at the synchroamplifier's CT stator terminals to

24

be sure a zero order is actuallyapplied.

3. Disconnect the 36X CT rotor lead,and turn the 1X CT barrel untilthe dial reads zero.

4. Reconnect the 36X CT rotor lead,disconnect the 1X CT rotor lead,and turn the 36X CT until the dialreads zero or some multiple of 10°.

5. With the dial on zero, zero the 5Gwith a test synchro, or by thejumper method.

Synchro amplifiers are also frequentlyused as "converters." It is sometimesdesired to feed information from a 60 -cycle data transmission system to a 400 -cycle servomechanism. In this case, thesynchro amplifier consists of 60 -cycle1X and 36X CT's, a 60 -cycle servoamplifier and motor, and 1X and 36X400 -cycle synchro transmitters.

Troubles in Dual -SpeedPositioning Systems

All the troubles that occur in single -speed systems are encountered in dual -speed systems. In addition, other trou-bles, most of them having to do withinaccurate positioning, are encounteredin dual -speed systems:

1. Non -mutual zeroing of synchroscan cause hunting. If, when the 36X CTrotor voltage is zero, the 1X CT rotorvoltage is 4 volts, the 4 volts will ener-gize the change -over relay, putting the1X CT in control. The 1 -speed CT willposition the mechanism to a point cor-responding to approximately zero 1XCT rotor voltage, at which time the re-lay will drop out, placing the 36X CTagain in control. The hunting cycle willthen start again. In other types of cross-over systems, non -mutual synchro zero-ing frequently produces erratic inaccu-racies in positioning.

2. Depending on design (i.e., on therelative gain of the 1X and 36X sys-

tems, the type of crossover netwo0cused, etc.), stator lead reversals willproduce positioning inaccuracies of va-rious types. Stator lead reversals canbe detected easily by rotating the datatransmitter with first the 1X CT andthen the 36X CT rotor lead discon-nected, in order to determine whethereach speed system positions the mech-anism in the proper direction.

3. Malfunctioning crossover networkscan be detected by methods previouslydescribed.

All of these possibilities should bechecked in a complete synchro systemcheckout, especially in those systemswhich have only an index mark for acalibrating dial. Even if a dial is pro-vided, errors often occur at odd dialreadings at which checks would notordinarily be made.

MOTION SYSTEMS

It is frequently desired to rotate cer-tain remote mechanisms at a given ro-tational rate. Since positioning ambi-guities ark of no consequence in suchapplications, a 36X system only canadequately provide the required motion.In general, motion systems require nozeroing.

A basic aided -tracking system for tar-get bearing is illustrated in figure 6.The rate -of -change of target bearing isdependent on several quantities, such asbearing, range, ship's course and speed,and target course and speed. Some ofthese quantities are fed to the computer,others are calculated by the computer.In addition, the computer determinesthe rate -of -change of bearing; in a link-age computer, the final element in thiscomputation is usually a mechanicalintegrator. The output shaft of the in-tegrator turns at 36 times the rate ofbearing change, and geared to this shaftis a synchro transmitter. As seen infigure 6, the synchro transmitter (No.

25

INDICATOR CONTROL UNIT COMPUTER

I .}.(11\IIMECHANICAL DIAL

7 DIFFERENTIAL I)(-J-T

36:1

ICT36X

SERVOAMPLIFIER

SERVOMOTOR

(16

II5V

NO I0

SERVOMECHANISMPOSITIONING

RADAR ANTENNA,SONAR TRANSDUCER,

OR OTHERINFORMATION DEVICE

TARGET

RANGE

OWN SHIPSCOURSE

OWN SHIPS

SPEED

BEARING INFORMATION IX

115/78 VOLT ELECTRICALZERO TRANSFORMER

36 X AIDED TRACKING INFORMATION

ICTIX

MECHANICALINTEGRATORTIME

NO21

IG36X

O

AIDEDTRACKING

SWITCH

FID

I I

AIDED TRACKINGRELAY

OII5V60%

Figure 6. Aided Tracking System

2) feeds its changing stator voltages toa 36X CT which is geared to one sideof a mechanical differential in thebearing data transmitter. The CT willcause motion of the data transmitter inexact synchronism with the integratorshaft (i.e., at a 36:1 ratio with theintegrator shaft). Once the bearing ratecomputation is completed, the operatorcan theoretically sit back and relax untilthe target changes course and speed.Note that aided tracking provides aclosed loop between the data transmitterand the computer. Bearing data andother information are fed to the com-puter, the computer solves for rate -of -change of bearing, feeds it back to thedata transmitter, which re -positions thecomputer bearing mechanism, etc.

Figure 6 shows that an electrical zerolock is applied to the CT stator ter-minals when the aided -tracking switchis off. The application of this lock isnecessary in order to ensure positivepositioning of the data transmitter withthe knob; otherwise, at least part of theknob motion would back up throughthe differential. However, the use of thezero lock produces an important oper-ating defect in the system of which theoperator should be aware. Assume thatthe attack has proceeded far enoughfor the computer to provide a stable so-lution for rate -of -change of bearing, andthat the aided -tracking switch is thrownat the instant the stator voltages oftransmitter No. 2 represent a rotationof 180 electrical degrees (5 mechanical

26

O

r -u. 36X POSITIONING

SYSTEMI CT I G

36X 36X

I I5V 0--6OrL 0

I 0I j

1

I

I

-1.---0-11

I

FIXED

0 I I

C.F. FIELD

I I 0 1

SERVOAMPLIFIER

I;1 i1 RADAR

I II

SERVO ANTENNAMOTORII

1111111111111

II I 1

0 ----1-I

1. 0 I1

I SLEWI SWITCH

tSLEWI RELAI `1")

I I- - - - -JI- - 1 _ _ _ _I

,

()DIALI IX INDICATING

SYSTEM

I

410ill

Figure 7. Positioning Mechanism for Radar Antenna

degrees) from electrical zero. Underthese conditions the data transmitterwill jump 5° and nullify the computersolution. Therefore, the operator shouldthrow the aided -tracking switch early inthe attack, so that in case the computersolution is nullified, there will be suffi-cient time for the computer to reachanother stable solution before zero timeto fire.

Another frequent use of the motionsystem is illustrated in figure 7. Theradar antenna is positioned by a 36Xmotion system which requires no zero-ing. The position of the antenna is indi-cated to the operator by the dial gearedto the 5F.

An obvious requirement of the servo-mechanism in a motion system is thatits top speed and response be adequate

to follow fast rates; otherwise, themechanism will drop behind in 10°steps. In the system illustrated in figure7, for example, when the knob is spun,unless the response is very good, the CTrotor voltage will go through a completecycle before the mechanism really gets.moving. Therefore, a slewing voltage isprovided for faster positioning rates, asshown.

The step-by-step system, a discontin-uous d -c type, is sometimes used inmotion transmission. This system is

shown in figure 8. The step-by-steptransmitter is simply a three-way switch,usually of the cam -operated type, withthe cam shaft driven by the motion -producing device. The step-by-step mo-tor consists of six coils mounted arounda soft -iron armature, connected as

27

CAM DRIVEN BY RATEINFORMATION DEVICE

CAM SWITCH

STEP -BY-STEPTRANSMITTER

DIALLOAD

STEP-BY-STEPMOTOR

Figure 8. Step -by -Step Motion System

shown in the figure. When the voltage\is applied to the No. 1 coils, the arma-ture will align itself with these coils.When the switch reaches the No. 2 posi-tion, the armature will align itself with

the No. 2 coils. etc. Consequently, themotor armature will follow the motionand run at the same rotational rate asthe cam shaft. The step-by-step systemis gradually falling into disuse.

(to be continued)

Solution to . . . July -August "What's Your Answer?"

The extra power results from an increase in the plate efficiency of the r -f ampli-fier during modulation. The plate efficiency of a class -B linear amplifier is givenby the formula,

miPlate efficiency = -77 ( 1 En4 EB

where Emi n is the instantaneous minimum plate voltage during an r -f cycle. and Enis the power supply voltage. From this formula it can be seen that if the instan-taneousEmminimum plate voltage is decreased, the fraction n becomes smaller and

EBthe plate efficiency increases. For 100% modulation, the maximum amplitude ofthe modulated carrier is double that of the unmodulated carrier. Hence, whenmodulation is applied, the instantaneous minimum plate voltage of the amplifier isreduced to only one-half the value that exists for the carrier alone. Therefore. theamplifier plate efficiency is increased, and since the average input power to theplate is constant, extra output power is salvaged from the power that is dissipatedin the tube with an unmodulated signal. That this is true is shown by the fact thatthe red glow often visible around the plates of a linear class B amplifier diminisheswhen modulation is applied.

28

A NEW PHILCO TRAINING MANUAL

A new training manual, SYNCHROS ANDSERVOMECHANISMS, has just come off the press

and is now being distributed to the field.

'We feel that this manual fills a long standing

need and that it will prove to be a valuable addition

to the technical series. Many favorable comments

have already been reported by people who have

examined copies of the advance printing.

This 139 -page manual contains 168 illustrations,

a glossary of terms. and a comprehensive index.The subject is approached from the field engineer'sstandpoint and mathematical analyses are kept at aminimum. Thus the manual should serve as a refer-

ence volume as well as an instructor's guide to any

individual concerned with such subjects as guidedmissiles, computers, automation, and remote con-trol. In addition, considerable information on main-tenance is presented so that the reader will find a

well-rounded coverage of such synchros and servo-mechanisms as he will be likely to encounter in the

field.

SYNCHROS AND SERVOMECHANISMS isbeing made available to BULLETIN readers for$1.74. Requests for the publication should be ad-dressed to:

Philco Technical Publications Department18th and Courtland StreetsPhiladelphia 40, Pennsylvania

Remittances should be in cash, check, or moneyorder, made payable to Philco Corporation.

A listing of other publications in the technicalseries appeared in the July, 1953, issue of theBULLETIN, and additional copies of the list canbe obtained at the above address.

29