Copyright©2012bySkyhorsePublishing,Inc.

AllRightsReserved.Nopartofthisbookmaybereproducedinanymannerwithouttheexpresswrittenconsentofthepublisher,exceptinthecaseofbriefexcerptsincriticalreviewsorarticles.AllinquiriesshouldbeaddressedtoSkyhorsePublishing,307West36thStreet,11thFloor,NewYork,NY10018.

SkyhorsePublishingbooksmaybepurchasedinbulkatspecialdiscountsforsalespromotion,corporategifts,fund-raising,oreducationalpurposes.Specialeditionscanalsobecreatedtospecifications.For

details,contacttheSpecialSalesDepartment,SkyhorsePublishing,307West36thStreet,11thFloor,NewYork,NY10018orinfo@skyhorsepublishing.com.

Skyhorse®andSkyhorsePublishing®areregisteredtrademarksofSkyhorsePublishing,Inc.®,aDelawarecorporation.

Visitourwebsiteatwww.skyhorsepublishing.com.

10987654321

LibraryofCongressCataloging-in-PublicationDataisavailableonfile.

ISBN:978-1-61608-373-1

PrintedintheUnitedStatesofAmerica

PREFACETHISBOOKwaspreparedforuseasahomestudycourseinhorologyandasa

referenceworkforthestudentinthehorologicalschools.Initspreparation,theauthor has kept in mind the practical and theoretical knowledge required bythoseorganizationsthatprepareexaminationsandthestateboardexaminationsoftheseveralstateswhichrequirealicensetopracticehorology.

Theauthorhasconsultedoriginalsourcesandespeciallytheworksofthosewho have made horological history. In addition, much of the work has beendrawn fromscientific publications, particularly those concernedwith themorerecentdevelopments.

Mostbooksonwatchworkgivenospacetothetheoreticalanalysis.Onthispointadeparturehasbeenmade,andthereaderwillfindthat,inadditiontothepracticaltreatment,thereisageneralsurveyofhorologicaltheory.

It isrecognizedthat thepracticeofhorologyis largelyoneofmanualskill,theacquisitionofadelicatetouch,andmuscularcontrolofthewristandfingers.Practical work is, then, the starting point, and thematerial in this book is soarranged. One must perform, at the outset, a satisfactory job of cleaning andrepairingaclockorwatchbeforethemoreintricateworkofisochronal,position,and temperature adjusting can be achieved. When this plan of attaining ahorological education is followed, the student will find himself developingprogressively toward a successful career in a very interesting and pleasantsemiprofession.

HAROLDC.KELLY

263MayAvenue

Monrovia,California

CONTENTS

PrefaceIntroduction

PartI:PRACTICALHOROLOGY

1.IntroductoryRemarkstoBeginners2.PracticalClockRepairing3.PracticalWatchRepairing4.RepairandAdjustmentProblemsTakenFromPractice

PartII:THEMECHANICSOFHOROLOGY1.WheelWorkofPendulumClocks2.WheelWorkofBalanceWheelClocksandWatches3.Gearing4.EscapementsforPendulumClocks5.EscapementsforBalanceWheelTimepieces6.StrikingMechanisms7.AutomaticWatches8.StopWatchesandChronographs9.CalendarMechanisms

PartIII:THEORETICALHOROLOGY1.AStudyofthePendulum2.AStudyoftheBalanceandBalanceSpringBibliographyIndex

INTRODUCTION

TIMEANDTIMERECKONINGWhat is time? It is a something that has baffled thoughtfulmen since the

beginning of civilized history. Innumerable astronomers, mathematicians,scientists, and philosophers have investigated this thing called time and haveadvancedvarioustheoriesandincidentallyworkedoutsystemsbywhichithasservedmankind.

OldandNewConceptsofTimeIn 1687, when Sir Isaac Newton published his “Principia,” certain

philosophical concepts of timewere given the world that have been acceptedeventothisday.Infactthewholescienceofphysicsandmechanicsisstillbasedupontheseconceptsandissotaughtinourschools.

AccordingtoNewton,timeisabsoluteandwithoutanyrelationtobodiesortheir motions. It is conceived as unbounded, continuous, unchangeable in theorder of its parts and divisiblewithout end.Time existswithoutmatter. If theuniverseweredestroyed,timeandspacewouldremain.

On the other hand Dr. Albert Einstein conceived a different idea of time.Einsteinsaidthat timeisrelative.Itcannotexistwithoutbeingassociatedwithspace. And since objects in space have three dimensions (length, breadth andheight),timeisaddedtothese,whichgivesallobjectsfourdimensions.

Suppose,forexample,weareinthecapitolbuildinginWashingtontodayat12:00a.m.Returningtheretomorrowat12:00a.m.,thespacewillhavechanged.Theearthhasrotatedtotheeastonitsaxis;ithasrevolvedtotheeastaroundthesun; the sun hasmoved around theMilkyWay. Thuswhen a given period oftime has returned, we cannot return to the same given space. Suppose, foranotherexample,theobserverwasplacedonMarsorVenus.Theunitvalueoftimewouldbedifferent.

The dimensions of an object, its shape, the apparent space it occupies,depend, according to Einstein, upon the object’s velocity— that is, upon thetimewhich theobserver takes to travelagivendistance, relative to theobject.Henceitisimpossibletodefinespacewithouttime.Thatiswhywenowsaythat

the earth on which we live has four dimensions and that time is the fourthdimension.*

*Avery interesting andhighly instructive discourse onRelativity is found inTheUniverse andDr.Einstein,LincolnBarnett,WilliamSloaneAssociates,NewYork.Everyhorologistshouldreadthisbook.

Thusthescientistshavegivenustwowidelydivergentconceptsoftime.Weashorologistsshallleavethattothescientiststodiscussamongthemselves.Ourinterest is concernedwithobtaining themeanof solar time inorder to setourclocksandwatches.Thismuchwearecertain:thatmanhasinventedamethodof dividing successive rotationsof the earth on its axis, in relation to the sun,into days, hours,minutes, and seconds for the purpose of regulating his dailyactivities.Themethodbywhichtimeisreckonedwillnowbeconsidered.

TheEarthasaClockTheearth isourmaster clock.The rotationon its axismaybe regardedas

practicallyinvariable.Thefrictionofthetidesontheoceanfloorissaidtoslowdown the earth’s motion, but calculations show a retardation of less than asecondin100,000years.

Togetaclearpictureoftheearthclockandhowitworks,letussupposethatwecanplaceourselvesinspacedirectlyabovethenorthpole,say250,000milesfrom the earth. From this position wemay see the earth from the north polenearly to theequator.Figure1 showshow itwill appear.Wemayobserve theearth’srotationtotheeastattherateof18½milesasecond.Furtherobservationshowsthathalfoftheearthislightandtheotherhalfisdark,andwemayseethesun’srayscreepoverthenorthpoleasthesummerseasonapproaches,andalsobecomedarkenedwhenwintercomes.

Figure1.

Our earth clock is to have a 24-hour dialwith two sets of figures readingfrom1to12toshowa.m.andp.m.Meridiansoflongitudewillprovideforthelocation of the figures. The meridians are 15 degrees apart and the primemeridian, which passes through Greenwich, England, indicates 12 a.m.Greenwichtime.Ourearthclockdiffersfromordinaryclocksin that the“hourhand,”whichmaybeanymeridianforlocaltime,turnscounterclockwise.

Theequinoxes.Supposeweareobservingtheearthat12a.m.onMarch21orSeptember23.Thesedatesarecalledthevernalandautumnalequinoxesandindicate aperiodwhenboth the sunandan imaginary location in the sky (theequinoxes,Figure2)fallonagivenmeridianatpreciselythesametime.Duringtheperiodsoftheequinoxes,theequatorliesontheplaneofthesun’srays,or,toputitinanotherway:bothpolesareanequaldistancefromthecenterofthesunandnightanddayareofequallengtheverywhereontheearth.

Asiderealdayisshorterthanasolarday.Now,atabout11:56a.m.,meansolartime,onMarch22orSeptember24,theearthclockhasmadeonecompleterotationonitsaxisandtheabovegivenmeridianagainliesontheplaneoftheequinoxes.Thedurationofonerotation—thatis,arotationof360degrees—iscalledasiderealday.Inthemeantime,theearth,initsrevolutionaroundthesun,hasmovedtoanewpositionasshowninFigure3.Fromthislocationitcanbe

seenthatfurtherrotationisrequiredbeforeasolardayiscompleted.Thusasolardayisalmost4minuteslongerthanasiderealday.

Figure2.

Figure3.Thearrowspointtowardthevernalequinox.

Sidereal time is determined by observing the more distant stars and is,therefore,consistent,whereasapparentsolartimeisnot.Siderealtime,however,isnotusefulforcivilpurposesbecauseourdailyactivitiesaregovernedby thesun’spositioninthesky.Siderealnoon,forexample,comesatnightduringpartoftheyearandasiderealyearcontains366¼daysinsteadofthe365½daysofasolaryear.

SolardaysarelongestinDecemberandJanuary.Apparentsolardaysarenotuniformfortworeasons.Thefirstcanbeexplainedbynotingthattheearth’sorbitisanellipseandthesunissituatedneartooneendoftheellipseasshowninFigure4.Whentheearthisnearestthesun,whichoccursaboutDecember22,itsrevolutionaroundthesunisfastestandthedaysarelongestbecausetheearthmustrotatefarthertobringagivenmeridiantosolarnoon.

Thesecondreasonwhyapparentsolardaysarenotuniformhas todowiththe inclination of the earth on its axis. During the latter parts of March andSeptember,theearth’srotationtotheeastislessthananyothertimeoftheyear,sincepartofthemotion,asobservedattheequator,issouthinMarchandnorthinSeptember.When therotation iseastwardas in the latterpartsofDecemberandJune,furtherrotationisrequiredtocompleteasolarday.Thesolardaysare,therefore,longestinDecemberandJuneandshortestinMarchandSeptember.

Figure4.

Theeffect of the eccentricityof the earth’s orbit and the inclinationof theearth’s axis, when combined, tends to increase the length of the solar day onDecember 22. The effect of the two phenomena, separately and combined, isshowninTable1.

Mean time or civil time. For practical reasons, it is necessary to arrange the

dayssothattheirlengthsareequal.Suchanarrangementiscalledmeantimeorciviltimeandisarrivedatbyaveragingtheapparentsolardaysforawholeyear.Civil timebeginsatmidnightandcontinues through24hours to the followingmidnight.

Equation of time. For the same reasons that the length of the apparent solardaysarenotuniformwefindthatthesuniseitherfastorslowtotheclock.Solartime andmean time agree only four times a year. The difference for all otherperiodsiscalledtheequationoftime.Table2showstheratesofapparentsolarnoonasrelatedtomeansolarnoon.

ComputingMeanTimebytheStarsOur observations of the earth from the distant point on the celestial north

pole are now completed. We shall again come “down to earth” and placeourselvesontheprimemeridianatGreenwich,England,foradifferent“pointofview.” We shall visit the Greenwich observatory and possibly take a “peek”through a special kind of telescope called the transit instrument, Figure5. Bymeansofthetransitinstrument,Greenwichciviltimeisgiventotheworld.

Figure5.Thetransitinstrument.

The observatory has a sidereal clock. By consulting sidereal time, theastronomercanpointthetransitinstrumenttoastarofknownpositionwithinafew minutes of its appearance. According to the sidereal clock, the stars arealwaysinthesameplaceatanygiventime.Thesuniseitherfastorslowmostoftheyearascomparedwiththeclock,butthestarsareneverfastorslow.

It isessential that the transit instrumentbesecured ina rigidposition.Theaxisliesinaneast-westpositionandispivotedsothelineofvisionisalongtheplane of the prime meridian. The astronomer has previously adjusted theinstrumenttothepathofafixedstarthatliesonthemeridian.

It ispossiblethattheastronomerwill letthereaderlookthroughthetransitinstrument. Looking through the eye piece, the observerwill see a number oflines,seveninnumber,asshowninFigure6.Theseareforgreateraccuracyinfollowingthestaras itcrosseseachof thelines.Oneor twostarsnot intendedforaparticulartimecalculationmaybeseen.

Figure6.

It is now about the time for the fixed star to cross the meridian. Theastronomerseatshimselfbeforetheinstrumentandlooksthroughtheeyepiece.The observer has his finger on an electric button. When the star passes thesecondline in theeyepiece,hepresses thebutton.Hedoessoagainwhenthestarpasses thecenter lineandagainon thesixth line.The threepressesof thebutton cause three notches to be cut on a moving tape of a recordingchronograph.Inthismanner,timeisrecordedtoafractionofasecond.

There is, however, the personal equation, a possible error in which theobservermayrecordthearrivalofthestaralittletoosoonortoolate.Thiserroris avoided in the best modern observatories by replacing the reticle with amovingwirewhichkeepspacewith thestar’s image.Whenthewirereachesagivenpoint,electricalcontactismadeautomatically.

Variouscountriespublishastronomical almanacswhich list a largenumberof stars, giving the right ascension anddeclination. From the right ascensiongiven,theastronomercancalculatethetimewhenthestarcrosseshismeridian.Bycomparingthesiderealclockwiththeobservedtime,theclockerror,ifany,canbenotedandaccordinglycorrected.

It isobserved thatwehavebeen takingaccountof sidereal timeonly.Starobservationsaremucheasierandmoreaccurate thansolarobservations.Meansolartimecanafterwardsbecomputedbyreferringtotablesinthealmanacs.

In theUnitedStates,correct time isbroadcastdailyby telegraphand radiofromthenavalobservatoryatWashington.Signalsaresentoutatnoon,Easternstandard time, by consulting an astronomical clock. These are rebroadcast bydifferentagenciestothefourdifferenttimezones.

Timereckoningwithoutspecialequipment.Oneneednotpossessatransitinstrumenttoobtaintimefromthestars.Infact,nospecialequipmentisneededat all. Choose a window of your home, preferably one to the south where achurchsteeple,radiotower,ortelevisionaerialcanbeseen.Securealargepieceofcardboardand,withasharp instrumentasapencil,makeasmallhole in it.Place thecardboardover thewindowand, looking through thehole,watch theprogressofaparticularstaruntilitdisappearsbehindthedistantobjectselected.Atthemomentthishappens,signalanassistantwhoisobservingaclock,fortheexacttimeastardisappearsmustbenoted.Onthefollowingeveningthesamestarwill vanish behind the fixed point 3minutes and 56 seconds sooner thanindicatedbythecorrectciviltime.Iftheclockdoesnotshowthisdifference,thetimeisincorrectandmustbemaderight.

Ifthetimeisnottakendaily,merelymultiply3minutesand56secondsbythe number of days that have elapsed since the first observation. Subtract theresults of the above calculation from the time of the first observation. Theansweristhetimetheclockoughttoshow.

Newstarsmustbeselectedafteramonthortwoastheymovetothedaylighthoursandbecomeinvisible.

Shouldyouobserveaplanetinsteadofastar,timereckoningwillnotwork.Planetsmoveindependentlyasrelatedtothestars,andafewnightsobservationwilldeterminewhichisastaroraplanet.

TimeZonesSince15degreesof longitude represent1hour change in time, timezones

have been arranged throughout the world. The inconvenience of continuallyresetting our watches when traveling great distances is avoided. The standardmeridiansshowmeansolartimeandthezonesextendroughly7½degreeseastandwestofthemeridians.

TheUnitedStateswasdividedintofourstandardtimebeltsin1883.Itwasthebuildingoftranscontinentalrailroadsthatencouragedtheirestablishment.In1884,thepresidentoftheUnitedStatessuggestedaninternationalconferencetoestablish a system of time zones for the entire world. The conference, withtwenty-six nations represented, met in Washington. The majority of themfavored making the meridian of Greenwich zero longitude. Meridians of 15degrees each were accordingly fixed with an international date line at 180degrees.

TheInternationalDateLine. Thus by general consent a line through thePacificOceanbecametheplacewherenewdaysbegin.Bygoingeast,hoursarelostuntilthedatelineisreached.Thenadayisadded.Intravelingwest,hoursaregaineduntilthedatelineisreached,whenadayissubtracted.

Thehoursoftheday,however,continuewithoutabreakineitherdirection.The necessity for a date line is seen in supposing that we travel from

Greenwichtothewestasfastastheearthrotateseast.Thesunwouldnotmoveduringtheentiretrip.Time,accordingly,wouldnotchangeandwewouldreturntoGreenwichonedaylateifthedatelinehadnotbeencreated.

DaylightSavingTimeDaylightsavingtimewasfirstputintoeffectintheUnitedStatesonMay21,

1916.Theexperimentwassuccessfulandhasbeenapprovedbyamajorityeversince.Clockshavebeensetaheadonehourduringthesummerinmostcountriesinrecentyears.

Those opposed are of the opinion that time is of a divine nature andmanshouldnot tamperwithwhat theycall“God’s time.”On thecontrary,manhasconsistently tamperedwith time long before the sundial becameobsolete.Thetruthis:civiltime,whichthesepeopleaccept,variesfromapparentsolartimeasmuchas17minutes in aday.And, furthermore, thedividing linesof the timezonesdifferbyatleast½hourfromthetimeofmeridianchosen.

PARTI

PRACTICALHOROLOGY

PARTIChapterOne

INTRODUCTORYREMARKSTOBEGINNERSWHENTHEhorologicalstudent firstexamines the toolandmaterialcatalogs

of those companies who cater to the needs of horologists, he is, no doubt,bewildered because of the vast array of tools, materials, and supplies. It iscustomarythattheworkmanownhisowntoolsanditisadmittedthatacompletesetrunsintoconsiderablemoney.Thebeginnerneednotbedisturbedregardingthispoint,asit ispossibletostartworkingwithafewessential tools.Later,asnewtechniquesare learned, therequired toolsmaybepurchased.Infact,mosthorologists build up their tool equipment gradually over a period of years.Horology,itseems,isthattypeofsemi-professionwheretheunfoldingofskillsdevelop inorder fromoneprerequisite to thenext.This explainswhy the toolrequirement is one of gradual expansion. Horologists often make, or have amachinistmake,toolsoftheirowninventionordesignasspecialneedsarise.

Tools,Materials,andSuppliesIn this section we are concerned with the tool and material needs of

beginners and themanner inwhich these items areprepared for use.We shallexplain,also, thereconditioningof toolsandsupplieswhichbecomenecessarybecauseofordinarywear.

Screwdrivers.Inclockwork,threeormorelargescrewdriversofvaryingwidthsareneeded.Aspurchased,thebladesaretoothickandneedtobegroundnarrower to fit the slot in most screws. The student needs a Carborundumgrindingwheel2½inches indiameter,andmountedonanarbor foruse in thelathe. The grinding wheel will prove to be a most essential part of thehorologist’s equipment, being useful in innumerable ways besides that ofreconditioningtools.

Watchscrewdriversespeciallymustbekeptingoodcondition.Ingrindingthese,useaneye loupe.Hold theblade to thesideof thewheel,andfinishbyslightlytouchinguptheendofthebladeonthefrontfaceofthewheel.Watch

screwdriversaresoldinsetsoffourorfive.Twoadditionalscrewdrivers,calledjewelscrewdrivers, areneeded to remove thevery small screws that hold thecapjewelsinplace.

Gravers. Gravers for turning in the lathe are ordinarily made of fineSwedishsteel,hardenedandtemperedreadyforuse.ThesearesharpenedonanArkansasstonebyaddingasmallamountofPikeoilstoneoil.Holdthegraverasonewouldgraspapenorpenciland,withasweepingmotion,grindthegravertotheshapeshownincatalogsoftooldealers.

Inrecentyears,anextrahardalloycalledcarbidehasbeenusedforgravers.Itcanbesharpenedonlyonadiamond lapand isuseful incutting thehardeststeel.Carbideisverybrittle,chipseasily,andcannot,therefore,begroundtoalongpoint.

Tweezers. The supply of tweezers should be quite large. Long, heavytweezersareneededforclocks,andsmallerdelicatetweezersforwatches.Onepairwithextrafineandthinpointsareusedforbalancespringsandshouldnotbeusedforanyotherwork.

Tweezers, likescrewdrivers,needtobereconditionedoccasionally.Spreadthepointssothattheylieonbothsidesofthegrindingwheel.Resttheotherendofthetweezersonthebedofthelathe.Nowbringtheinsideofonepointtothegrindingwheelandgrindthesurface.Reversethetweezersandgrindtheotherpoint.Thistakessomepracticetodoagoodjob,butitispossibletorestorethetweezerstoaconditionasgoodasnew.

Eyeloupes.Forgeneraluse,a3-inchglassisneeded.Forbalancepivotsofwatches,useapivotloupeforexamination.Anextrastrongloupe,½-inchfocus,is often desirable.Thematerial catalogs show awide selection fromwhich tochoose,anditisadvisabletopurchaseonlythebettergrades.

Other essential tools. Many uses will be found for flat-nosed pliers, andcuttingpliers.ANumber1 file is required forgeneraluse, and for specializedrequirements we need a set of needle files. The beginner’s equipment alsoincludestwosmallhammers,onewithasteelheadforrivetingandoneofbrassforusewith thestaking tool.Thereare,ofcourse,manyother tools,but thesewillbementionedwhenthevarioustypesofrepairjobsaretakenup.

Materials for clocks. In order to startwork at all,we shall need cleaningsolutions.High-testgasolineorwhitegasolinemaybeusedonalarmclocks,butinthebettergradeofclockswherethemovementsaretakencompletelyapart,itis preferred to use the especially prepared cleaning solutions sold bymaterial

dealers.Tocleanaclockmovement,aratherstiffwashoutbrushisrequired.Usepegwoodforcleaningtheholesintheplatesandchemicallytreatedsawdustfordryingtheplates,wheels,andotherparts.

Asupplyofclockpins,andclockwashersisneeded,forthesearesometimesmissingintheclockorpossiblylostinthecourseofrepairing.Abottleofclockoilisobviouslyneeded,togetherwithanoilerwhichmaybepurchasedormadefrom a large broach. An assortment of clock bushings is desirable, althoughthesecanbemadewhen required.Pendulumrodsandmainspringsareusuallypurchasedasneeded.

Materials forwatches. Thematerials forwatches are, of necessity,muchmore numerous than those for clocks. In addition to cleaning and dryingsolutions, watch oil, oilers, oil cups, and pegwood, we need a stock ofmainsprings, balance staffs, balance jewels, plate jewels, roller jewels, timingwashers,watch papers, jewel cement, crystal cement, and pith. It is desirable,too,tohavealargecollectionofoldwatchmovementsfromwhichpartscanbetaken.

GRINDINGANDPOLISHINGPOWDERSThe grinding and polishing of steel parts, train wheel pivots, and balance

pivots require the use and preparation of a variety of abrasive and polishingpowders. These include Carborundum powder, oilstone powder, crocus,saphirine, and diamantine. The above items are listed in the order of theirabrasivequalities.Carborundumcutsrapidlyandproducesacoarse,greyfinish.The powder ismade in several grades, but only the finer grades are useful inhorology.Oilstone powder is quite fine, but leaves a grey finish. It is used inreducingsquareshoulderpivotsoftrainwheelsofsmallclocksandwatches.Forthefinalpolishsaphirine,ordiamantinemaybeused.Thepreparationofanyofthesepowdersisasfollows:Secureathree-sectionpolishingblockandusethelowestsectionfor theoilstonepowder.Placeasmallamountof thepowderonthemetaldisk,addalittlewatchoil,andmixwiththebladeofapocketknifetoathick paste. The other powders are prepared in the samemanner. The secondsectionisusedforcrocusandthetopsectionforsaphirine,ordiamantine.

GRINDINGANDPOLISHINGSLIPSTheabrasiveandpolishingmediumsaspreparedonthepolishingblockare

used inconnectionwith iron,bellmetal, andboxwoodslips. Iron isusedwiththeoilstonepaste,bellmetalwiththecrocus,andboxwoodwiththesaphirineor

diamantine.Someworkmenusethesaphirineordiamantineinitsdryconditionbymerelyapplyingitdirectlytotheboxwoodslip.

Ironandbellmetal slipsareprepared foruseby filing thepiececrosswisewith a medium file. The abrasive material becomes imbedded in the grooveswhenused.Whenthegroovesbecomeworn,thefilingisrepeated.

In the treatment of balance pivots, the Arkansas slip may be used forreducing, the jasper slip for additional smoothing of thework, and finally theboxwoodslipwiththepolishingmaterialforthefinalpolish.Steelburnishersaresometimesusedtoproduceahardsurfaceontrainwheelandbalancepivots.

Arkansasand jasper slipsbecomewornwithuse.Thesecanbe restored totheiroriginalconditionbyslidingthesidesonapieceofgroundglasspreviouslycoveredwithamixtureofCarborundumpowderandoil.

TheWorkshop,theBench,andWorkingConditionsHavingconsideredsomeoftheessentialpointsrelativetothebeginner’stool

equipment, letusnowtakealookat theshop.Ifwindowsfacingthenortharenotavailable,largedouble-tubefluorescentelectriclampsshouldhangoverthebenchatadistanceofabout8feetfromthefloor.Inadditiontothis,thebenchshouldbefittedwithadouble-tubefluorescentlampwithalongadjustablearmso that the light can be moved in any position. The best possible lightingequipmentisimportanttosavetheworkman’seyesfromanyextrastrain.

Thefloorshouldbecoveredwithalight-coloredlinoleumsoastoaidinthefinding ofwatch parts thatmay be dropped.Right here the student should becautionedalwaystohavetheapronofthebenchpulledoutwhileworking.Allwatch-repairbenchesarefittedwithaprons,andthenewerbenchesareprovidedwith an extra wide apron which helps greatly in catching the parts that aredroppedorflippedoutofthetweezers.

Thetopofthebenchshouldbeofwood,finishedinalightcolororpaintedwithawhiteenamel.Abenchplatefinishedinwhiteenamelmaybepurchasedfrommaterialdealers.Thisprovidesanexcellent surfaceuponwhich towork.Thetopofthebenchshouldbekeptclean.Washthebenchplateeverymorningbeforestartingwork.

Themodernwatchbenchiswellsuppliedwithdrawerspace.Thereshouldbeafixedplaceforeverytoolandeverytoolinitsplace.Theonlysmalltoolsthatmaybeleftonthebenchwhileworkingare:thebenchblocks,eyeloupes,tweezers,screwdrivers,oilcups,andoilers.Oilersare inserted inpith tokeepthetipsclean.

Thestakingtoolboxisplacedinahandypositionontherear,rightcornerofthebench.Thelatheisfittednearthefrontoftheleftside.Manyhorologistsbeltthelathedirectlytothemotor,butacountershaftinadditionismuchpreferred.

Theothertoolsusedquitefrequentlyshouldbearrangedindrawersthatarehandyandaccessible.Toolsused lessfrequentlymaybeplacedindrawersnotquitesohandybutwithineasyreach. Inotherwords, thearrangementof toolsshould be so planned as to make for pleasure and efficiency. An orderlyarrangementavoids thepossible inconvenienceofendlesslyhunting for thisorthattool,therebycausingconfusionandfrustration.

Thehorologistshouldbesuppliedwithaposturechair,withcushionandanadjustable back. Since the workman spends long hours at the bench, it isessential that satisfactory comfort is provided if peak efficiency is to beexpected. The type of chair is important if a tired and aching back is to beavoided.

Theaboveanalysissuggeststhatthehorologist,tobesuccessful,mustworkunder ideal conditions — that is, conditions wherein one can remain calm,peaceful,and relaxed.Watch repairingcannotbe rushed,andworking tosomedeadlineshouldbeavoidedasmuchaspossible.

Theshopshouldbewellventilated,coolinsummerandcomfortablyheatedinwinter.Andmore important still, the shop shouldbe locatedapart from thestoreproper,wheretherewouldbelittlenoiseandcommotion.

Now that the basic considerations have been outlined, we are ready toexamine, clean, and repair a clock. This is the subject with which the nextchapterhastodeal.

PARTIChapterTwo

PRACTICALCLOCKREPAIRINGTHEREARESOmanydifferent typesandgradesofclocksthatananalysisof

repairmethodsmust,ofnecessity,bequitegeneral.The treatment,however, isentirely practical and, we believe, sufficiently complete to cover all populartypes of clocks— the clocks that are found in most homes and offices. Theanalysis to followdealswith the cleaning and oiling of the popularAmericanmantel clock with counting wheel striking mechanism. The methods outlinedapply just as well to the cleaning of French, English, and German clocks,regulator,chime,andgrandfatherclocks.

Repairs apart fromcleaningaregiven special treatment in later sectionsofthischapter.

The 400-day clock and the American alarm clock are given separateconsideration,inasmuchastheseclocksareinaclassbythemselves.

CleaningandOilingtheAmericanStrikingClockRemoving movement from the case. Removing the movement from the

caseisgenerallyasimplematterand,inmostcases,tooobviouseventocallforan explanation.Americanmantel clocks inwooden cases are so designed thatthehandsandgongmustberemovedfirst.Thisdone,removefourscrewsthatholdthemovementinplaceandliftoutthemovement.Aproblempresentsitselfifoneofthemainspringsisbroken.Inthiscasewindthegoodspringfullyandpull out first the side that contains the good spring. The movement will, ofcourse,comeoutwithdifficulty,buttakecaretoavoidbendingthewheels.

Dismantling themovement. Before taking the movement apart, examinethe striking train and see if there are marks on the wheel and pinion whichperform the functionof lockingandwarning. If therearenone,markwith thegraver the point where the wheel gears with the pinion. If this is done, thestriking mechanism may be assembled by noting the marks, and the timing

remainscorrect—thatis,ifitwasrightbeforedismantling.Themainsprings are let downwith bench keys, but before doing so place

springclampsovereach spring.Newmainspringsareenclosed inclamps, andseveralsizesmaybeacquiredwiththepurchaseofsprings.Thesearesavedforfutureuse.Placethekeythatfitsonthesquareofthewindingarbor,releasetheclick,andallowthespringtorundownslowlybylettingthewoodenhandleslipbetweenthethumbandfingers.

Afterthespringsareletdowncompletely,examinethefreedomofthepivotsintheholesintheplates.Notingtheconditionofwear,removethetopplateandlift out the timing train, keeping the severalwheels together.Next remove thestriking train and levers and keep these parts together and separate from thetimingtrain.Thepartsmaynowbegivenfurtherexamination.Seeifthepivotsarewornorinneedofpolishing.Nextexaminethepallets.Theseusuallyneedpolishing, and theworkmay be donewith aNumber 0 emery stick. If badlyworn,thegroovesneedgrindingoutonanemerywheelorCarborundumwheel,afterwardspolishedonahardfeltbuffingwheelchargedwithrouge.

Determineifthewheelsaretightontheirarbors.Bentescapewheelteetharecommon faults, and every tooth should be examined. The locking pin on thewarningwheelisoftenfoundtobeloose.Brokenteethandbrokentrainwheelpivotsaresometimesfoundasaresultofamainspringbreaking.Themethodofrepairingtheseandotherfaultswillbeconsideredinanotherpartofthischapter.Assuming that the clock is in good condition, we shall proceed with thecleaning.

Cleaning themovement. Some of the cleaning solutions intended for thewatchcleaningmachinesaresatisfactory,providedthesolutionisofatypethatmaybemixedwithwater.Fillalargebasinwiththecleaningsolution.Placeinthe solution the plates, wheels, and other parts, and allow them to stand forseveralminutes. Brush the parts thoroughlywith a rather stiff brush. Sharpenpeg-wood to a point andworkvigorously in theholes in theplates.Rinse thepartsinwarmwaterandshakeinwarmboxwoodsawdustuntildry.

Thepartsmaybedriedwithoutthesawdustifonehasasmallelectricheater.Place the parts on a large tablewith the heater nearby.Make certain that theparts,especiallythemainsprings,aredrybeforeassembling.

Assemblingtheclock.Restthelowerplateonawoodenframeand,startingwiththefirstwheelsandsprings,assemblethewheelsintheirproperplaces.Itismostconvenient,usually,tofitthewheelsfromthebottomofthemovementandworktowardthetop,fittingtheflywheel,theescapewheel,andthepalletslast.

Thetailofthestrikinghammershouldbejustfreeoftheliftingpin.Thepinonthewarningwheelshouldstandabouthalfa turnawayfromthestoppieceonthewarninglever.

Nextplacethetopplateinpositionandguidethecenterwheelarborandfirstwheel arbors to their holes. Now press the top plate closer and, with thetweezers,guidethepivotsofthelargerwheelsand,lastly,thesmallerwheelstotheirrespectiveholes.Pinorscrewdownthetopplateandtestboththestrikingandtimingtrainforfreedom.

Windthemainspringpartlyupandoiltheseveralcoilswithmainspringoil.Oil with clock oil the pivots, the pallet, and every other tooth of the escapewheel.Thecrutchisoiledatthepointwhereitenclosesthependulumrod.

Fittheminutehandandtestthestriking.Turnthehandslowlyalmosttothehour, and see if thewarning takesplaceproperly.Thewarningconsistsof thereleaseofthewarningwheel.Theflyturnsaboutonecompleteturnandstops.Whenthehandisadvancedtothehour,thestrikingbegins,andstopswhenthecounting leverdrops intoadeepnotch in thecountingwheel.Test thestrikingforallhoursfrom1to12and,ifitissatisfactory,setthemovementinawoodenframeorshelfandtesttherunning.Ifitremainssatisfactoryafterafewdaysofrunning,fit themovementtothecase.Adjustthestrikinghammertothegong.Listentotheticktodeterminethebeat.Adjustingthebeatisattainedbybendingthecrutchwiresothattheticksarespacedequally.Nowfitthehandsandsettotime.

Theregulatingthatfollowsordinarilyrequiresabouttwoweekstime.Theaboveaccountofthecleaningandsettinguptheclockassumesthatwe

encounternoparticulardifficulty.Moreoften it isnotquite sosimple,and thenextseveralpagesdealwithmanyoftherepairproblemsusuallyrequiredinthecourseofpracticalclockrepairing.

MainspringProblemsThe better quality clocks have going barrels for the reception of the

mainspring. In thegoing-barrel typeof powerunit, the spring iswound at thearbor and the barrel carries the first wheel. Mainsprings enclosed in barrelsshould always be removed for inspection and cleaning. Unwind the springslowly,allowingthespringtouncoilwithinthepalmof thehand.Mainspringsthat are set should always be replaced with new ones, and although themainspringwinderispreferred,itispossibletowindthespringintothebarrelbyhandwithoutanynoticeabledistortion.Placeasmallamountofoilonthespring

afteritiswoundintothebarrel,andreplacethecap.Most clockswith going barrels have a riveted hook in the barrel, and the

outerendofthespringhasapear-shapedholethatfitstothehook.Occasionallythehookispulledoutandanewoneisrequired.Infittinganewhook,broachouttheholeinthebarreltoaclean,straighthole.Selectapieceofsteelwireandturn the hook to the same shape as the old one. Fit the hook from the inside.Placetheinsideofthebarrelonananvilinsuchamannerastoholdthehookinplaceand,witharound-endhammer,rivetthehookinplace.

Thepear-shapedholeintheendofthespringoftenpullsout.Iftherestofthespringisingoodcondition,anewholemaybedrilledintheend.Breakofftheouterendclearof theoldholeandheat theenduntil it turnsblue.Drillanewholeandfiletheopeningtofitthehookinthebarrel.Finishtheendofthespringwithaflatfile,andbendtheendbeyondtheholetothecurveofthebarrelwall.

RepairingtheTrainClosingholes.Inasmuchasclocksruntoolongbetweencleanings,itisquite

commonthat theholesintheplatesbecomebadlyworn.Sincethepressureonthepivotisawayfromthewheelthatdrivesthepinion,itfollowsthattheweartakes placemostly on one side of the hole. In popular-priced clocks, the holemaybeclosedwithahollowcenterpunchwithhalfoftheendgroundaway.Byfittingthepunchtothesideoftheholewherethewearisgreatest,andtappingthe punch sharply with a hammer, the metal is worked toward the hole. Thepunchleavesacrescent-shapedmarkontheplate,butthisdoesnoharm.Someclockrepairersmayobjecttothismethodasbeingrathercrude.Weagreethatitshouldnotbepracticedonclocksofhighqualitybutseenoobjectiontousingitonthecheaperwork.Theholeshouldbeclosedsufficientlysothatthepivotwillnot go in. Broach out the hole to a free fit for the pivot and try the wheelbetweentheplatesforfreedom.

Fittingbushings.Thewornpivotholes ingoodqualityclocksarerepairedbyfittingbushings.Bushings,madeupinassortments,maybepurchasedfrommaterial dealers, or theymay bemade from a piece of brasswire. Secure thewireinalathechuck,faceofftheend,andturnasmallcenterinit.Selectadrill,slightlysmallerthanthepivot,placeitinthepinvise,anddrillaholeabout½inch deep. Next, broach out the hole in the plate from the inside and to adiameteraboutthreetimesthesizeofthepivot.Turnthebrasswiresothatitjuststartsinthehole.Polishthegraverona4/0emerypaperandbright-cuttheendofthebrasswire.Turnandoilcuptotheholealreadydrilledinthewire.Now

cutoffthebushing,takingnoteofthelengthrequiredtomatchthethicknessoftheplate.Drivethefinishedbushingintotheplateflushtotheinsideandbroachouttheholetofitthepivot.Theburrmadebythebroachmayberemovedwiththewheelcountersinks.

Pivoting.Therepaircalledpivotingisrequiredwhenpivotsarebrokenofforbadlyworn.Center thewheel in the lathe chuck.Removewith the graver theportionof thebrokenpivotandturnasmallcenter.Selectapivotdrill slightlylarger than theoldpivotandplace in thepinvise.Thedrillmaybe lubricatedwith turpentine formoresuccessfuldrilling.Thehole shouldbedrilleddeeperthanthelengthofthepivot.Selectapieceofstaffwireorpivotwireandturnitdown to fit the hole. The wire should be tapered very slightly. When it fitshalfwaytothebottomofthehole,drawfileitcarefullyandcuttolength.Nowfiletheendflat.Replacethewheelinthelatheandinsertthepivot.Afewblowswith the hammerwill fasten the pivot securely in the arbor.The pivot is nowturnedtosize,burnishedandpolished,andtheendroundedslightly.

FittingTeethtoWheelBroken teeth are often found in clock movements but the repair is not

difficult.Fileaslotbelowthemissingtoothtoadepthequaltotheheightofthetooth.Selectapieceofbrassofthesamethicknessasthewheelandfiletofittheopeningjustprepared.Insertthepieceandsolderwithsoftsolder.Aconvenientmethod of soldering is to lay the wheel on a copper plate and hold over analcohol lamp.Cautionshouldbeexercisedsoasnot toheat thesteelpivots.Asectionofanoldwheeloftenservesasmaterialwithwhichtorepairthebrokenwheel.Afterthesoldering,thetoothisfiledtothepropershapewithneedlefiles.

EscapementRepairs*Theescapewheel isoftenwornat theendsof the teeth, anddeepgrooves

maybefoundinthepallets.Consideringfirsttheescapewheel,theworntipsoftheteethmayberestoredtogoodconditionbygrindingdowntheteeth.Centertheescapewheelarborinthelathechuckand,whilerunningthelatheatahighspeed, lightly grind down the tips of the teethwith anArkansas slip. Smoothfurtherwithasteelburnisher.Thistreatmentreducesthediameteroftheescapewheelslightlyandmayrequireanadjustmentofthepalletstofitthewheel.Anexcess of inside drop may result, and the correction consists in closing thepallets.Thepalletsmust firstbe tempered toabluecolorand thenclosed ina

vise.Make thecorrectiongradually, testing thepallets to thewheel frequently,untilthedrops,bothoutsideandinside,arefoundtobeequal.Havingfoundthecorrection satisfactory, heat the pallets to a cherry red color and plunge intowater to restore theoriginalhardness.Thewornsurfaceof thepalletsarenowgroundoutandpolished.

*TheconstructionandactionofclockescapementsareexplainedinChapterFourofPartII.

Insomecasesthepalletsorescapewheelmaybeshiftedsothattheteethofthewheelslideonaportionofthepalletsthatarenotworn.

RepairingtheStrikingMechanism*Counting wheel system. Failure of the counting wheel system to stop

striking at the proper time is usually due to incorrect timing.Turn theminutehandtothehourandallowthestrikingtorunslowly.Whenthecountingleverdropsintoadeepnotchinthecountingwheel,thedroplevershouldlockinthecamwheel.Ifthecamwheelisnotintheproperposition,thedroplevercannotlock in the notch. Let the mainspring down or insert a watch screw driverthroughthesecondwheelandplatestolockthetrain,andseparatetheplatesonthestrikingside.Separatethecamwheelfromthewheelsnexttoitandturnthecamwheeltoalterthetiming.Assemblethestrikingtrainandtryagain.

*TheconstructionandactionofstrikingmechanismsareexplainedinChapterSixofPartII.

Often the several levers need bending so that the drop leverwill drop thenecessarydistance to engage thecamwheel.Thereare,of course,manyotherreasons for striking difficulties. In fact, most jobs will bring up innumerableproblems. There are no fixed rules or exact methods. One must have a fairdegreeofmechanicalaptitudetosucceedalongwithplentyofexperience.

Rack and snail system. Striking errors in the rack and snail system areusuallyattributedtowornparts.Theholeforthegatheringpalletpivotmustfitclosely.Alargeholeatthispointoftencausesthegatheringpallettohangupontherackorfailtogatherupthepropernumberofteeth.

Awornstoppieceonthewarningleverisshownbythefailureofthelevertodropat thehour.Thewearmark iseasily filedandstonedoutand thestriking

willagainfunctionproperly.Theseveralleversoftherackandsnailsystemshouldbefreeonthepivots

thatcarrythem.Thisismentionedbecauseafaultyactionmayprovetobeverypuzzling,inasmuchasatightleverisnoteasilydetected.

Anotherfaultyadjustment,occasionallymetwith,istheliftingofthestrikinghammer when the clock warns. This results in a lot of useless run after thestriking is completed.There should be very little run after the striking andnoliftingofthestrikinghammerduringthewarning.Correctadjustmentlessenstheprobabilityoftheclock’sfailingtostrikewhentheoilgetsthick.

RepairingBrokenHandsHandsoftenbecomebrokennearthecenter,butthispresentsnoproblemin

thenewerstandardmakesofclocks,asnewhandscanbepurchased.However,hands for French clocks and antique clocks cannot be purchased andmust berepairedormadecompletely.Thefollowingmethodforrepairingissuggested:

In theFrenchclocksabrasscollet is fitted to thecenterof thehands.Thiscolletisremovedfromthebrokenpiece.Nowselectasmallpieceofflatsteelofaboutthesamethicknessasthehand.Drillaholeinitandenlargetheholetofitthecollet.FilethesteelpiecetotheshapeshowninFigure1.Thewidthofthenarrowpart shouldbe left a littlewider than the hand and about half as long.Nextfitthecollettothepiece,butmakecertainthattheendpointstoward12toinsurecorrectstriking,andrivetsecurely.Nowapplyalittlesolderingfluxandsoft solder to the steel piece.With the tweezers, hold the steel piece and thebrokenhandtogetherfirmlyandheatoveranalcohollampuntilthesoldermelts.Finish by filing or grinding the surplus steel away.Clean the hand and,whenfittedtotheclock,therepairedhandwillbefoundtoworkaswellandlookasgoodastheoriginalhandbeforeitwasbroken.

Figure1.

Figure2.Movementofthe400-dayclock.

Repairing400-DayClocksThe400-dayclockisoftenreferredtobyothernames,forexample:annual

windclocks,andanniversaryclocks.Insteadofapendulumithastheequivalentofabalancewheelsuspendedat

theendofaflattenedwireofsteelorphosphorbronze.Designedastheyaretorun a year between windings, it will be noted, Figure 2, that there are threeintermediate wheels. The barrel makes one turn in 80 days and the balancemakes8vibrationsaminuteor7½secondsforeachvibration.Exceptionsmaybefoundinsomeoftheoldermodels.

Figure3.Escapementof400-dayclock.

Theescapementisofthedead-beattypesimilartothedead-beatescapementdescribedinPartII.SomearemadewithadjustablepalletsasshowninFigure3,whileothermakesusepalletsofone-piececonstruction.Lyingabovethepalletframeisaroundwireabout1½incheslongcalledthepalletpin,theupperendofwhichplaysfreelybetweentheforkF,Figure4.

TheforkissecuredtothesuspensionwireS.Whenimpulseisdeliveredtothe fork, thesuspensionwire is twistedand the force iscarried to thebalance.When the wire is twisted, the balance is raised slightly; thus two forces arepresent:(1)thetensionduetotwistingand(2)aslightpullofgravity.

Thecleaningandoilingdoesnotdiffergreatlyfromanyothertypeofclock,but theadjustmentof theescapement, thefittingofnewsuspensionwires,andregulating do present newproblems.Suppose the clock is set up and running.The locking of the escapement should be equal and this may be noted byobserving the sliding action of the pallets along the face of the escapewheelteethimmediatelyafterlocking.Ifthelockingisnotequal,itmaybemadesobyturningtheblockthatsupportstheupperendofthesuspensionwire.ThispieceisfittedfrictiontightintheplateP.However,beforeadjustingtheblockB,itis

well to inspect the suspensionwire for bends or kinks thatmay lie about thefork.

Figure4.

Supposeaclockforrepairhasabrokensuspensionwire.Thesewiresmaybepurchased frommaterialdealers.Usingametricmicrometer, selectawire thatmeasurescloselytotheoldone,ifavailable,andfittheendpieces.Withpliersholdtheupperendandfitthebalancetothelowerend.Givethebalanceaslightturnand,whileholdingtheassemblyabovethebench,countthevibrationsfor½minute.Fourvibrationsshouldbecountedand,ifnot,continuetochangewiresuntiloneisfoundthatvibratesclosely.*Whenawirewithcloseratingisfound,setuptheclockandregulatetotime.

*Someoftheoldermodelsmaynotvibrateto8beatsaminute.Todeterminethenumberofvibrationscount the teeth of the centerwheel, the leaves of the escape pinion, and the teeth of the escapewheel.Dividethenumberofleavesoftheescapepinionintothenumberofteethofthecenterwheel.Multiplythequotientbytwicethenumberofteethoftheescapewheel.Theansweristhenumberofvibrationsofthebalanceperhour.

The 400-day clocks are becoming increasingly popular of late. They are adelightful novelty, the movement is exposed, and the entire operation can beobserved.Unfortunately,theclocksarenotgoodtimekeepers.Aratewithin2or3minutesaweekisaboutthebestthatonecanexpect.

TheAmericanAlarmClockRepairmen generally do not like to work on popular-priced alarm clocks.

Owingtotheirlowcost,itisnotpracticaltogointoextensiverepairs.Thereareafewpoints,however,thatshouldbeobservedand,inmostcases,thepopular-pricedclockcanberestoredtogoodrunningorderataminimumcostoftime.

Dismantling the clock. The first act is to remove the time-winding andalarm-windingkeysthatarelocatedatthebackoftheclock.Arrowsindicatethedirectionthekeysareturnedtowind.Toremove, turnthekeysintheoppositedirection,asmostkeysarescrewedon.Inafewcasesthekeysarefittedfrictiontighttoasquareonthearbor.Ifthisisthecase,insertthejawsofcuttingpliersunder the keys and pry off.Usually themovementmay be removed from thecasewithoutdisturbingthetime-settingandalarm-settingknobs,astheholesinthebackplatearelargeenoughtopermittheknobstogothrough.Removingthemovement from the case differs according to themake of the clock. A quickexamination shoulddetermine this. Inmost alarmclocks thebezel ispriedoffandthreeorfourscrewsat thebackareremoved.Themovementis thenliftedoutforexamination.

Placeamarkon theoutercoilof thebalancespringat thepointwhere thespringissecuredtothestud.Removethebrasswedgetoreleasethespring.Withthepliersbackoutoneofthethreadedsteelcupsthatcarrythebalancestaffandremovethebalancefromthemovement.Examinethepivotsofthebalancestaff.Theendsofthepivotsshouldfeelsharpwhentouchedwiththefinger.Iftheyarenot sharp to the touch, theymustbemade so.Assuming that the correction isnecessary, firstmark the balance showing the position of the outer end of thebalancespring.Nextremovethebalancespringbyforcingthe taperedsidesofthetweezersunderthecollet.PlacethestaffinasplitchuckandgrindthepivotstoasharppointwithanArkansasslipandoil.Finishwithasteelburnisher.Nowreversethestaffandpreparetheotherpivotinlikemanner.

Cleaning the clock. With the balance out of the movement, it is anopportunetimetoclean,butfirstremovethehandsanddial.Whitegasolineorbenzine is satisfactory for cleaning. Place enough in the jar to cover themovement and place the movement in the jar. Allow to stand for 10 or 15minutes. Next select a brush with rather stiff bristles and, dipping it in thegasoline,brushthemovementthoroughly,particularlyaroundthepivotsandthecircumference of the escape wheel. The escape wheel often needs additionalattention,asthereisatendencyforthepalletpinstoforcedirttothebaseoftheteeth.Thisdoesnotapply to thoseclocks fittedwithbankingpins,but thepinpalletsarenotusuallyequippedwithbankingpins.Thedirtis,therefore,builtupandlessenstheangularmotionofthelever,possiblytothepointofeliminatingtheguardfreedom.Ordinarycleaningmaynotremovethisdirt,which isoften

cakedonhard;soscrapethebaseoftheteethwithasharpinstrumentthatcanbeinserted between the plates. Rinse the movement up and down in cleaningsolution;dryitonacleansheetofpaper.

Now clean the balance with the brush, dipped in gasoline; rinse in warmwater,dipinalcohol,anddryinwarmsawdust.Thebalancespringiscleanedinlikemanner,except,ofcourse,itcannotbebrushed.

Afterthepartsaredry,replacethebalancespringtothecorrectpositiononthe balance. The mark previously made on the balance shows the properposition.Adjustthethreadedsteelcupstopermitasmallfreedomforthestaff.Place the spring between the regulator and wedge in the spring at the stud,notingthemarkpreviouslymade.Iftheescapementisnotexactlyinbeat,inserta screw driver in the slot of the collet, and turn the balance in the desireddirection.Itisunderstoodthattheescapementisinbeatwhenthebalancespringisatrestandtheimpulsepinpointsdirectlytowardthepalletcenter.

Oilingthemovementisnextinorder.Windthetimeandalarmspringsfully,andplaceclockoilormainspringoilonthecoils.Oilthetrainwheelpivots,thebalancepivots,andeveryothertoothoftheescapewheel,alsothealarmpallets.Thebalanceshouldnowstartwithabriskmotionofnotlessthanoneturn.*Ifitdoesnot,wemustlookfurtherforotherfaults.

*Seepage258forananalysisofbalancemotion.

Adjusting the escapement. If the balance does not take a satisfactorymotion,theescapementmaybeatfault.Turnthebalanceslowlyuntilsuchtimeastoallowonetoothtopassapalletpin.Theescapewheelwillturnandanothertoothwillengagetheoppositepalletpin.ThereshouldbeadefinitelockingasshowninPartII,Figure27,butwedonotalwaysfinditso.Insteadwemayfindanerrorwhereinthepalletpinfailstolock,and,ifthisisthecase,thepalletpinwill engage the impulse face of the tooth. Using long, flat-nosed pliers, thecorrectionconsistsinbendingin,towardtheescapewheel,thenarrowprojectionofbrassinwhichthepalletarborisfitted.Bendingthepalletarbortongueputsthe pallets slightly out of upright, since the opposite end usually has no suchmeans of adjustment. This, however, is not objectionable in clocks of thisquality.Theonlyadditional adjustment isbending lever to lineup fork slot toimpulsepin.

Fittingpalletpins.Ifthepalletpinsareworn,newonesmustbefitted.Letthemainspringdownandremovethe lever fromthemovement.Since thepins

arefittedinbyfriction,itisonlynecessarytolaytheleveronthebenchblockanddriveoutthepins.Ordinarysewingneedlesofthecorrectdiameterarewelladaptedfornewpins.Cuttheneedlestotherequiredlength,finishtheends,anddrive in place. Since the needles are highly polished, no further finishing isnecessary.

Adjusting the alarm. On the dial side of the movement lies a long flatspringwithabent-overend that reaches throughahole in theplate.Thebent-over end locks or releases a lever (locking lever) which permits the alarm tofunction at the desired time. When turning the alarm setting knob, one canreadilyseewhathappensandthenecessaryadjustmentcanbemade.Ifthebent-overenddoesnotrisehighenoughtoreleasethelockingleveratthemomentthealarm should ring, bend the locking lever back slightly. If the alarm ringsconstantly,moveitforward.

Replacing the hands. Turn the alarm setting knob slowly and stop theinstanttheclockalarms.Fitthealarmindicatorhandto12o’clock.Fitthehourandminutehandsto12o’clockalso.Next, turnthetimesettingknobforward,clockwise,andcheckthetimewhentheclockalarms.Ifthealarmdoesnottakeplaceat12o’clock,readjustthehandsuntiltheclockalarmsatthepropertime.

It sometimes happens that the alarm setting knob and hand turnwhen theclockissettotime.Thisshouldnottakeplaceandindicatesalackofsufficientfriction in the alarm settingmechanism.Many clocks have a nut thatmay betightenedtocorrectthefault.Insomecases,thealarmsettingknobneedstoberemovedtomaketheadjustment.

PARTIChapterThree

PRACTICALWATCHREPAIRINGMODERNhigh-gradewatchesareamarvelousexampleofman’sskill in the

makingofaprecisioninstrument.Wemustadd,also,thattobecomeanefficientrepairman, the same skill is required. It takes years of hard work to performsatisfactorily the many delicate operations. Moreover, as the student masterseach theoretical principle and technical problem, he will find that the workbecomesmoreandmorefascinatingandpresentsnewopportunitiesofspreadingoutintootherfieldsofintellectualinterest.Thereisprobablynootherinstrumentthatperformsasinglefunctionwhichmakesuseofsomanylawsofphysicsasdoesthecompactlittlemechanismthatiswrappedupinawatchcase.

CleaningandOilingRemoving movement from case. The best method we believe, to learn

watch repairing is to start on thewatch itself, taking it apart, piece by piece,examining each part carefully and noting the function and purpose of everypiece.Obviouslythefirststepistoremovethemovementfromthecase.Itlookssimpleenough,butcertainproceduremustbeobservedtoavoidthebreakingofparts.

Thetoolsneededfortheworkwillbementionedasweproceed.Attheverystartweneedapairoftweezers,a3-incheyeloupe,asetofwatchscrewdrivers,amovementholderandacaseopener.

Theconstructionofcasesvariesconsiderablyandweshallconsiderthemorepopulartypesandthemethodsofremovingthemovementfromthecase.Mostcases,bothpocketandwrist,haveabackandabezel(frontwhichcontainsthecrystal) that snap on. These are pried off in the manner shown in Figure 6.Releasetheclick,Figure7,andletthemainspringdownbyallowingthecrownto slip slowlybetween the thumband first finger.Next, place thewatch, trainsideup,onthebenchandbackoutthescrewSS,Figure5,acoupleofturns.Thisreleases thestem,whichmaynowbepulledout.Nowremovethecasescrews

andpushoutthemovement.

Figure5.Swisswatchmovement.S—stem;SS—settingleverscrew;BB—barrelbridge;DS—dialfootscrew;TB—trainbridge;EB—escapewheelbridge;CW—crownwheel;C—click;RW—ratchetwheel;B

—balancebridge;R—regulator;PB—palletbridge.

SomeoftheolderAmericanwatchesarenotdesignedinthismanner.Insteadof backing out a screw, pull the stemout to the position of setting the hands,remove the case screws and push out the movement. The crown, stem, andsleeve remain in the case exceptwhen the replacement of these parts becomenecessary.

Manymodernwristwatch cases are of the two-piece construction.Pry offthebezelwith thecaseopener.Then liftout themovementwitha largewatchscrew driver by twisting the blade.Make certain that the screw driver is notinsertedatthepointwherethebalanceislocated.

Recent years have seen thedevelopment ofwater-resistant cases and theserequireanassortmentofwrenchesandspecial toolstoopenthecases.Aswiththe three-piece case, the stemmust be removed before the movement can betakenfromthecase.

Anothertypeofwater-resistantcaseisopenedbypryingoffanunbreakablecrystal.Inthese,thestemispulledoutwithcuttingpliers.

Thereareother types, too,whichhaveacaseopenedbypressingdownon

thecrystal,therebyreleasingtheback.Stillothersareheldtogetherbyscrewsorslides.Alittlestudywillusuallydeterminethemethodrequiredtoopenthecase.Intheeventofdoubtastothemethod,proceedwithcaution.

Dismantling the movement. Place the movement, train side up, in themovementholder,replacethestem,andtightenthescrewSS,Figure5.Removethebalancefromthelowerplateandreleasethebalancespringfromthebalancebridge.Placethebalanceononeoftheholesofthebenchblock.Neverlaythebalanceonthebench,eitherbeforeoraftercleaning.Placingthebalanceonthebenchblockavoidspossibledamagetothebalancepivotsandafterthebalanceiscleaned,thepivotsremainclean.Removetheseveralpartsofthewatchinthefollowingorder:(1)thehands,usingthehandremover;(2)thedial;(3)thehourwheel and minute wheel; (4) the cannon pinion using the cannon pinionremover;(5)thepalletbridgeandlever,butmakecertainthatthemainspringhasalreadybeen letdown; (6) the ratchetandcrownwheels; (7) the trainbridges;(8) thebarrelbridge;and, lastly, the trainwheelsandmainspringbarrel.ThesepartsareshowninFigures8and9.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure6.Openingcasewiththecaseopener.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure7.Releasingpowerofthemainspring.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure8.Watchmovementshowingthedialtrain.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure9.Removingmainspringbarrelcap.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure10.Typicalbalanceassembly.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure11.Watchpartsstrungonwiresforcleaning.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure12.Properandimproperoiling.

Laythebarrel,capsideup,onthebenchblock(youneedtwobenchblocks)and,witha small screwdriver,Figure9,pryoff the cap.This shouldbedone

with caution, as the arbor andmainspringmay fly out suddenly and youwillspendtherestoftheday,orperhapstomorrow,lookingfarandwideforabarrel,a barrel cap, a barrel arbor, and a mainspring. Trusting that you are not sounfortunate,carefullyremovethearborwithheavytweezers.Turnthearborsoastounhookthemainspringandremovethearbor.Nowremovethemainspring.

Thebalanceissupportedoneachendbyaholejewelandacapjewel,Figure10. This arrangement is necessary to reduce friction. Cap jewels are alwaysremovedincleaningandifthesearefittedtotheleverandtrain,itispreferredtomark the jewel settings and the plate before removing so that they may bereplacedtotheiroriginalpositions.

Cleaning thewatch. The lower plate and bridges are strung on a loopedwire,Figure11,andemersed ina jarcontainingawatchcleaningsolution thatmaybepurchasedfrommaterialdealers.Allowthepartstostandafewminutesinthesolution;thenbrushthoroughlywithasoftbrushpreviouslydippedinthecleaning solution.Rinse inwarmwater, anddip inalcoholor a specialdryingsolution that may be purchased at material dealers. Shake the parts in warmboxwood sawdust until dry. Thewheels are strung on thewire separately andcleanedinthesamemanner.Themainspringmaybeusedagainifit isnotset.(Thetermsetimpliesthatthespringhaslostitselasticity.)Aspringshouldopentoadiameteraboutthreetimesthatofthediameterofthewatchmovement.Ifitdoesnot,anewspringshouldbefitted.Cleanthemainspringaswiththeotherpartsandwipedry.

Thesmallpiecesthatcannotbestrungonthewire,suchasthelever,jewels,andscrews,areplaced inashallowcontainer filledwith thecleaningsolution.Theseareremoved,oneatatime,withspecialtweezerssoshapedthatthesmallpieces such as the jewelswill not fly away.With the same tweezers, hold thejewelsagainstapieceofhardpithandcleanwithabrushpreviouslydippedinthecleaningsolution.Clean the leverandothersmallparts in likemanner,dipeachpieceinadryingsolution,andplaceonasheetofpapertodry.Thejewelsarefurthercleanedbytwistingapointedpieceofpegwoodinthehole.Thecapjewelsarecleanedbyrubbingwithapieceofpegwoodthathasbeenflattenedattheend.

Thebalancemustbecleanedseparately.Workthepivotsintoapieceofhardpith and examine with a ½-inch eye loupe. If they are smooth and highlypolished,proceedwith thecleaningbydipping in thecleaningsolution,nextadryingsolution,andfinallyshakinginsawdustuntildry.Ifthebalancepivotsaretarnished,cyanideofpotassiummaybeused,butcleanthoroughlyafterdippinginthecyanide.WARNING:Cyanideisadeadlypoison.If thepivotsarebadly

corroded,polishinthelatheaccordingtotheprocedureexplainedinanotherpartofthischapter.

Assembling the movement. The watch has now been cleaned and anexaminationshowsthatthepartsareingoodcondition.Replacethecapjewelsandoileachholejewelwithwatchoil.Insertasmallwire,Figures12and13,intheholejewelsoas toforcetheoil throughtothecapjewel.Examinethecapjewelwithapivoteyeloupeandnotethesizeoftheoilbubble.Usuallymoreoilneeds tobeaddedto increase thebubble to therequiredsize.Figure14showsthecorrectsize.Theoilbubbleisnotalwaysconcentrictotheholeinthejewel.AslighteccentricityasshowninFigure15doesnoharm.IfthebubbleisbadlyeccentricasinFigure16,wehaveaconditionwherethecapjewelorholejewelisoutofflatandmustbecorrected.Inafewcasesanoilbubbleof thepropersizewillnotform.Thisindicatesthat thejewelsaretoofarapartandthis, too,calls for correction. Too much oil is shown if the bubble fills the entire capjewel.Inthiscase,removethecapjewel,andcleanandreplacethejewel.Thecorrectamountofoilisimportantifthewatchistofunctionforanormalperiodbetweencleanings.

Figure13.Oilpusher.Apivotbroach,groundtofitfreelyintotheholeofasmallholejewelandfittedtoapieceofpegwood,makesanexcellentinstrumenttopushtheoilthroughtothecapjewel.Useful,also,to

testescapementfreedoms.

Figure14.

Figure15.

Figure16.

Next fit thebalance to the lowerplateandsee if thebalancespring is flat,level, and concentric to the staff. If it is not satisfactory, correct according toinstructions given in another part of this chapter. Assuming that the spring iscorrect,removethebalanceandbridgeandplaceonthebenchblock.

Oil all frictional surfaces of the winding mechanism. Using a mainspringwinder,fitthemainspringtothebarrel.Oilwithwatchorclockoil.

Force thepinion endof the escapewheel into a pieceof hardpith andoileveryother toothof theescapewheel.Placethebarrelandthe trainwheelsonthelowerplateandfitthebridges,beingcertainthatthewheelpivotsareintheholejewelsbeforethebridgesarescreweddown.Tryallwheelsforendshake.Thefourth,third,andcenterwheelsshouldshowmoreendshakethantheescapewheel,butnoneshouldbeexcessive.Nowfitthelevertothemovementandoilthetrainjewels.Windthemainspringfullyandtesttheleverforfreedom.Thisisdonebyliftingtheleverawayfromthebankingpinagainstwhichitleans.Ifitisfree, itwill jumpquickly to theoppositebankingpin.Place thebalance in thewatch and, if there are no basic faults present, the balance will take a good

motion.

THEWATCH-CLEANINGMACHINEThewell-equippedshopnowusesthewatch-cleaningmachine.Thereisno

doubt that the machine does a superior cleaning job and has the furtheradvantageofeliminatingtheuseofsawdust.Themachineconsistsofanelectricmotortowhichisattachedabasketforthewatchpartsandthreeorfourjarsforthecleaninganddryingsolutions.

Thewatchisdismantledintheusualmannerandthepartsareplacedinthebasket, the larger parts in the bottom and the smaller parts in several smallspacesinthetop.Afteryouhavecleanedthepivotsinhardpith,thebalanceisplacedinoneofthesmallspacesinthetop.Toassurenoinjurytothebalancespring,a smallpieceof tissuepaper iscrushed intoaballandplacedover thebalance.

Lowerthebasketintothejarcontainingthecleaningsolutionandallowthemotor to run for severalminutes at amoderate speed.Throwoff the cleaningsolutionbyraisingthebasketsufficientlytoclearthesolution.Lowerthebasketintoajarcontainingarinsingsolutionandrinseoffthecleaningsolution.Raisethebasketsufficientlytoclearthesolution,asbefore,andthrowoffthesolution.Next, lower thebasket intoa second rinsing solutionand repeat theoperation.Finally,allowthebasket tospin inareceptaclecontainingasmallheaterforaquickdryingofthewatchparts.Thusthecleaningjobiscompleted.

MainspringProblemsMainspring gauges. There are two systems in use for measuring

mainsprings.These are theDennison and themetric. It is ouropinion that themetric is more satisfactory, as it permits closer measurements. A metricmicrometermaybeused todetermine the thickness, and theverniergauge forthewidth.

Swissmainsprings. Swiss mainsprings are generally fitted with a tongueendasshowninFigure17.Thesespringssometimesfail tohookinthebarrel.The difficulty can usually be corrected by shaping and filing the end of thetongueas shown in the illustration. If this fails,undercut thestep in thebarrelwithasharp,long-pointedgraver.Seethatthebottomofthestepisnotrounded.Often there isapunchmarkon thebarrel to indicate theposition inwhich thecap is snapped into place. Taking note of this factormay save the horologistconsiderabletrouble,forthebarrelmaynotruntrueintheflatunlessthecapis

properlyfitted.Fittinghooktobarrel.Thehookinthebarrelissometimesfaultyorbroken

andneedsreplacing.Inmakinganewhook,therearecertaindetailsthatshouldbestrictlyobserved.Theseare:(1)Thehookmustbestrongenoughtoresistallstrainwhenthespringiscompletelywound.(2)Thehookshouldholdthespringsecurelyagainst thebarrelwall. (3)Thehookshouldbeasnarrowaspossible,permittingtheuseofanarrowholeinthespringandtherebyrealizingagreaterstrengthatthatpoint.

Thereareseveralmethodsformakingahook.Weshalldescribeonewhichwebelieveisassatisfactoryandassimpletoexecuteasany.Theprocedureisasfollows: Drill a hole in the barrel at such a distance from the cap as to beperfectlycentered to the insideof thebarrel.Thread theholewithasmall tap.Next,usingapinvise,fileapieceofsoftsteelwiretoaslighttaper.Onthiswiremakeathreadwiththescrewplatetothesamesizeasthethreadintheholeinthebarrel.Thethreadingwilltakeplaceatthelargerpartofthetaper.Continueuntilafullthreadiscut;thencutoffthewire,leavingsufficientheighttofilethehook toshapewhile thewirestill remains in thescrewplate.Having filed thehook,removethewirefromthescrewplateandscrewintothebarrelfromtheinside.Tightensecurelybymeansofthatpartofthewirethatsticksthroughtheoutsideofthebarrel.Cutoffthewireandfinishleveltothebarrelwithasmallescapementfile.

Figure17.

Theholeintheoutermostendofthespringshouldreceivethesamethoroughattention as given to the hook.Donot blue the end of the spring farther backthan necessary. Punch the hole as near as possible to the hard part and leaveabout2millimetersofthespringbeyondthehole.Thispartbeyondtheholeiscurved slightly to conform to the circleof thebarrelwall.Thehole shouldbe

justwideenoughtofitthehookfreely,andthatportionoftheholeneartheendof the spring should be beveled to keep the outermost coil tightly against thebarrelwall.

Fittingcaptobarrel.Thereare timeswhenthebarrelcapwillnotstay inplace. For this problem, the following is suggested: Lightly file thecircumferenceofthecap.Theglazedsurfaceisthusroughenedandthecapwillusuallyremaininplace.

FittingStemsWhenthefactorystemisavailableforaparticularmodelofwatch,nofitting

is required other than shortening the stem to bring the crown to the correctheight.However,thecorrectstemisnotalwaysavailableandmoreorlessfittingisrequired.Therearecases,too,wheretheplatesarebadlywornandstemswithoversizehubsareneeded.

Having selecteda stem that approximates the required size, firstdeterminethesizeofthepivot.Thiscanbefoundbyinsertingasmalldrillinthatbearingintendedforthepivot.Thepivotisnowturnedtotherequiredsizeandthehubisnext turned to size. Reduce the square with an escapement file. This isaccomplishedbyholding the stem in apinvise andusinganeye loupeas thefilingtakesplace.

RemovingBrokenScrewsTheremovingofbrokenscrewsisoftenveryannoying.Sometimesthepiece

remainingcanbepushedaroundwithawell-sharpenedgraver.Therearetimeswhenaslotcanbefiledwithascrewheadfileand thepieceremovedwith thescrewdriver.Ifthesemethodsfail,makeasolutionofalumwithaboutonepartalum to fivepartswater.Stir thoroughly.Removeall steelparts and submergetheplateorbridge in thesolution,allowing thepiece to remainsubmergedforabouttwelvehours.Thescrewwillbedissolvedbuttheplates,whicharemadeof brass or nickle, are not injured. In some cases twenty or more hours arerequired.

RepairingtheMainTrainClosingholes.Closingholesintheplatesandbridgesofwatchesandclocks

isanalmostdailytaskofthehorologist.Iftheholesarenotbadlyworn,thiscan

be done satisfactorilywith the staking tool. Select a flat-faced stump that fitsfreelyintherecessoftheplateorbridgeandplaceitinthedieofthestakingtoolframe.Usingaround-endpunchthatfitstheoilcup,taptheholejustenoughsothatthepivotwillnotgoin.Broachouttheholetoafreefit.Nextfitthewheelbetweentheplatesandobservecarefullytheendshake.Theendshakeshouldbeaboutonethirdthediameterofthepivot.Supposetheendshakeistoogreat.Ifthewatch is fittedwithfrictionbushings, thesecanbeshiftedbymeansof thefriction jewelingmachine.Themannerofusing thisdevice isexplained in thesectioncoveringfrictionjeweling.Iftheplateisnotfittedwithfrictionbushings,themostsuitablemethodofalteringexcessiveendshakeistomakeorremodelastump by turning a slight hollow on its face. When the hole is closed asdescribedabove, themetal in theplatewill sink into thehollowof the stump,thusdecreasing the end shake.Anumberof such stumpsof varying sizes anddepthswillbefoundconvenient.Ontheotherhand,iftheendshakeiswanting,thereverseprocedureisrequired.

Thewheelshouldspinfreelybetweentheplates,andthisfreedomshouldbeequalinboththedial-upanddial-downpositions.

Oftentheholesfortheescapewheelandleverarewornbeyondrepairand,since the horologist of today can take advantage of friction jeweling, it is asimple matter to broach out the holes to the required size and push frictionjewelsinplace.Frictionjewelingisdescribedindetailinthenextsection.

FRICTIONJEWELINGThe use of jewels as bearing for watches is, without question, one of the

most important achievements to the attainment of precision timekeeping.Nicholas Facio, an Italian residing in London, successfully applied jewels towatches about the year 1704. The system used by Faciowas not the same asemployed inmaking the jeweledbearingof today. Insteadofaholepiercingajewel,aV-shapeddepressionwasgroundintothejewel.Thepivotwaspointedandworkedintothedepressedjewelinmuchthesamemannerasinthepresent-dayalarmclock.TheSwisswerequicktorealizetheadvantagesofjewelingandbegan experiments which finally resulted in themaking of jewels as we findthemtoday.

Formanyyears,jewelsweresetinarecessandburnishedintomakesecure.Thesewerecalledbezel-typejewels.Since1920friction-type jewelshavebeenusedandatthepresentdaythebezeltypemayberegardedasobsolete.

Replacing a bezel-type jewel with a friction jewel. In fitting a friction

jewel, the first act is to determine the depth the jewel is to be set to give theproperendshake.Thisisaccomplishedbyusingamachinemadeespeciallyforthe job. There are many different makes on the market, but in all makes theoperationisthesame.Restthepusheronthebrokenjewelandadjustthemetricscrewnear the topof themachine so that thenew jewelmaybe forced to thesame depth as the broken one had been.Next push out the broken jewel andreamoutthehole,Figure18,withthesmallestreamerthatwillcutawayenoughmetaltogivetotheplateaclean,straighthole.Nowselecttheproperjewel,theoutsidediameterofwhichis1/100ofamillimeterlargerthantheholeintheplate.The size of the hole is shown by the stamping on the reamerwhichwas lastused.Remove the burr left by the reamerwith thewheel countersinks. Lastlyplacethejewel,bottomsideup,overtheholeintheplate.Itshouldenterslightlywithouttheaidofthemachine.Lowerthepushertothejeweland,withafirmpressure, push the jewel in place, Figure 19. The metric stop, previouslyadjusted, will arrest the movement of the pusher when the proper depth isreached.

Figure18.

Figure19.

Replacingafrictionjewel.Toreplaceajewelinawatchthathadafrictionjewelinitbefore,itisnecessaryonlytomeasurethesizeoftheholeintheplate.This may be done by inserting the reamer or using a special gauge that isavailable for the purpose. Having determined the hole size, select the jewelrequiredandpushintotheproperdepth.

Fittingjewelinremovablesetting.Ifwewishtofitajewelinasettingthatmayberemovedfromtheplate,asinthecaseofabalanceorcapjewel,weneedspecialtoolstoholdthesettingwhilethereamingtakesplace.Therearevarioustypesoftoolsonthemarket,allofwhichareusedinconnectionwiththefrictionjewelingmachine.

POLISHINGPIVOTSTrainwheelpivots.Crackedjewelsordirtypivotholesoftencutthepivots

sobadlythatanewwheelisrequired.Insomecases,aswithacenterwheel,thepivot may be reduced and repolished. This usually requires the fitting of asmallerholejewelorbushingor,perhaps,closingthehole.

Suppose the short pivot of the centerwheel is in need of repair. Place thepinioninasplitchuckandmakecertainthatthepivottobecorrectedrunstrue.Withawell-sharpenedgraver,turndownthepivotuntilitisstraightforitsentirelength. The point where the shoulder and the pivot meet must not develop aroundcorner.Thismaybepreventedbykeeping thepointof thegraversharp.Next, cross-file on two sides an iron slip and, using oilstone powder and oil,

grindthepivottoagreyfinish.Finishfurtherwithasteelburnisher.Cross-fileabell-metalslipandpolishwithdiamantineandoil.

Balance pivots. In repairing a damaged balance pivot, the procedure isslightly altered. Suppose a balance pivot is burred. Reduce the cylindricalportionwithajasperstone,touchuptheendofthepivot,androundthecornerslightly.Burnishthepivotwithasteelburnisher,particularlytheendwhichwillneednoothertreatment.Finishtheburnishingbyrollingtheburnishedfromthecylindrical part around to the end.This avoids producing a burr at the corner.Nowbringthecylindricalportionofthepivottoahighpolishwiththeboxwoodslipanddiamantine.

If abalancepivotneeds considerable reducingas is sometimes required infittinganewstaff,theArkansasslipwillworkfaster.Reducethepivotuntilthejeweljustfits.Reducetoafreefitwiththejasperslip,and,lastly,finishthejobasalreadyexplained.

Material dealers stock sapphire grinders and polishers that someworkmenmay findmore to their liking.There isnoobjection for the repairman todoalittleexperimentingonhisown.However,tryanynewmethodonolddiscardedpinions or staffs until you are satisfied that the method is all right to use oncustomers’watches.

BalanceAssemblyRepairsUnder the heading of balance assembly repairs we shall include such

operationsasarenecessaryinfittingafactory-madebalancestaff;infact,itwilltake account of practically all repairs which involve the balance and balancespring.

FITTINGANEWBALANCESTAFFRemovingthebalancespring.Laythebalanceonthebenchblock,withthe

rollertableprotrudingthroughoneoftheholes,andholdsecurelywiththefirstandsecondfingers.Nextinsertawell-sharpenedscrewdriverunderthecolletata point opposite the pinning point of the balance spring, Figure20. Twist thescrewdriverandthespringisreadilyremoved.

Removingtherollertable.Theremovingoftherollertableisnextinorder.For this we use the roller remover, of which there are many varieties on themarket.Mostoftheseworkonageneralplan,wherebytwoprongssupporttheundersideof theroller table.Byusingahollowpunchfittedtothelowerpivot

andtappinglightlywithabrasshammer,therollerisremoved.

Removing the old staff. There are two types of balance staffs: (1) theriveted type,Figure21,and(2) thefriction type,Figure22.Differentmethodsarerequiredintheremovalofthetwotypesfromthewheel.

Consider first the riveted type. It is important that the staffbe removedbysome method that will not in any way damage the hole in the wheel. TwomethodsareshowninFigure23.Thefirstshowstherivetedportionbeingturnedoff, and the second shows the removal of the hub.The secondmethod is lesslikelytoenlargetheholeinthewheel,butitisoftenmoredifficulttoexecute.Ifthefirstmethodisused,punchout thestaff in thestakingtoolbysecuringthewheelwiththestaff-removingtool.Twotypesofstaff-removingtoolsareshowninthetoolandmaterialcatalogs.

Inremovingastaffofthefrictiontype,seethatthehubrestsonthedieofthestakingtool.Otherwisethehubwillbepushedoutwiththestaff.

Selecting and fitting a new staff. Selecting a new staff for any standardmakeofwatchisasimplematter,asthewatchfactorieshaveprovidedexcellentmeansforidentifyingtheproperstaffforaparticularmodel.Themorepopularmakeshave themodelnumber stampedononeof thebridgesoron the lowerplate under the balance. Inmany of the Swisswatches,wemust examine thesettingbridge to identify aparticularmodel.The size is determinedby anoldFrenchmeasurementcalledlignes.SomeAmericancompanieshaveadoptedthesystem,butAmericancompaniesgenerallyuseasystemwherebyawatchsizeisdesignatedbythewordsize,as:12size;8/0size;14/0size,etc.

If a staff for aparticularwatchneeds tobeordered, determine the sizebymeansofagaugeshowinglignesandsizes.IfthewatchisSwiss,comparethesettingbridgewiththoseshowninthematerialcatalogs.Whenthecorrectsizeandsettingbridge is located,note themovementmodelnumberonyourorderenvelopeandsendsampleifavailable.Genuinefactorymaterialshouldbeusedwheneverpossible.Itpaysdividendsinthesavingoftime.

Havingselectedthecorrectstaff(therivetedtype)foragivenwatch,proceedasfollows:Selectaholeinthestakingtooldieinwhichtherolleraxisfitsfreely,A,Figure24.Selectaround-facedhollowpunchthatfitsoverthecolletaxisandrivetthestafftothewheel.Twoorthreelighttapswithabrasshammershouldsuffice.Turnthewheelsothattherivetingtakesplaceevenly.Finishthejobbyusingaflat-facedhollowpunch.

In fitting a staff of the friction type, no fixed rules for procedure can begiven,astherearemanydifferentkinds.Themethodofreplacingthestaffwould

besuggestedbytheprocedureobservedinremovingtheoldstaff.

Adapting the balance to thewatch. The fitting of the pivots to the holejewelsisnextinorder.Balancejewelsshouldfitfreely,butwithaminimumofsideshake.MakecertainthatthepivotsarelongenoughtoprotrudethroughthejewelsasinFigure25.Replacethecapjewelsandtrythebalanceinthewatch.Adjust the end shake, either by shortening the pivots or bending the balancebridge, butmake certain that the balance bridge remains parallel to the lowerplate.

Fittingtherollertable.Placetherollertableonthestaffandnotethat thetable stands near enough to the hub so as to stake on without any danger ofsplittingthetable.Thisdistanceshouldbeabout.5millimeterinpocketwatchesand about .2 millimeter in wrist watches. Grinding the roller axis with theArkansasslipissatisfactoryinmostcases.Iftherolleraxisneedsconsiderablereducing,usethegraverandfinishwiththeArkansasslip.

SETTINGTHEROLLERJEWELOften the roller jewel is broken or lost. Again, the jewel may be loose,

sometimesveryslightlyso.Forthisreason,averythoroughexaminationshouldbegivenwiththepivoteyeloupe.

Letusassumethattherollerjewelismissing.Placetheleveronapieceofhardpithandtryanumberofjewelsintheforkslot.Thefreedomofthejewelshouldbesuchthatitmaytiptoanangleofabout15degrees.Ifstillindoubtasto the proper freedom, examine a number of newwatches. Place the balanceupsidedownonapieceofpith.Grasptheguardfingerwiththetweezersandtrythefreedomoftheforkslotaroundtherollerjewel.

Havingselectedaroller jeweloftherequiredsize,proceedwiththesettingas follows: Insert the roller table between the jaws of the combination tool,Figure 26, and heat the end of the tool over an alcohol lamp. The greatestamount of heat occurs near the top of the flame. When the roller table issufficientlyheated,placeasmallamountofshellaconthebackoftherollertableat the spot directly over the hole for the jewel. The hole is thus filled withshellac.While theshellac isstillsoft, insert theroller jewelandagainheat theroller table.With the tweezers, straighten the jewel to an upright and parallelpositionwiththestaffbeforetheshellaccools.Theflatfaceofthejewelstandsawayfromthecenteroftherollertable.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure20.Removingbalancespringfrombalancestaff.

Figure21.Balancestaff.Thetypethatisrivetedtothewheel.

Figure22.Two-piecefrictionfitstaff.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure23A.Removingdamagedbalancestaff.

Figure23B.Two-piecefrictionfitstaff.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure24.Replacingbalancestaff.

Figure25.

Figure26.

REPRODUCEDBYTHEPERMISSIONOFTHEDEPARTMENTOFTUBARMYFigure27.Balancetruingcalipersshowingpositionofindicatorfortruingintheflat.

TRUINGTHEBALANCEWHEELTheguidesonmostcalipersaremade inamanner thatwouldsuggest that

theguidebeplacedtotherightofthebalanceandthatthetopsideofthebalancerim face the guide. This,we believe, is illogical for reasons thatwill become

apparentaswecontinue.Instead,letusaltertheguideandplaceittotheleftofthebalance,Figure27.Placethebalanceasshownsothatthetopsideoftherimfacestheguide.

Truingthecompensatingbalance.Intheactoftruing,thecalipersareheldin the left hand and the fingers of the right hand are used to manipulate thewheel.Itwillbeobservedthatthethumbnailwillpressonthelowersideofthebalancearmsandrimsand,shouldanymarksbemade,thesewillnotshowafterassembly.

Startbylevelingthearmsintheflat.Comparewiththeguide,notingthatthedistancefromthearmstotheguideisequal.Nowtruetherimfromthemiddleofeachhalftothelooseends.Nextobservetheround;settheguidetotheouterrim,andpushthelooseendsoftherimsinorpullout,asthecasemayrequire.Itis usually necessary to return to the flat and re-true, after which the round isagainchecked.Oftenitisnecessarytogothroughtheprocessseveraltimes.

Truingthesolidbalance.When thestakingofasolidbalancewheelonanewstaffiscarefullydone,truingisoftenunnecessary.Particularlyisthistrueofthenewerwatches.Therivetingmustbedonelightly,foritistheheavytappingofthehammerthatspringsthebalancewheeloutoftrue.Iftruingisnecessary,theprocedureissimilartothatalreadygiven.

Whenmakingthebends,seethatthecalipersaretightlyclosedsoastoavoidanydangerofbreakingpivots.Alltruingisdoneinthecalipers,preferablywiththefingers.Itisveryrarelythatoneneedstousethewrenchthatissuppliedwitheachcaliperbythemanufacturer.

POISINGTHEBALANCEThethree-leggedpoisingtool,havingtwolegsthatareadjustable,istheonly

kind of tool suitable for the purpose. The tool fitted with a level has oneadvantageinthattheinstrumentcanbeadjustedlevel.

Placethebalance,withrollertableattached,onthepoisingtableandnotetheheavypoint.Thepoising is accomplishedby reducing theweightwith specialcutterson theheavyportion,oraddingbalancewashers to the lightportion.Atool for reducingweight has a cutter surrounded by a hollow tube. These aremadewithseveralsizesoftubestofitthemanydifferentsizesofbalancescrews.In order to use theweight reducing tools conveniently, it is first necessary tosecurethebalanceinthebalanceclamp.*Bypressingthecutterdownonthetopofthebalancescrewsandturningaswithanywatchscrewdriver, thescrewishollowedoutintheformofaV.

*Thebalanceclamp,aspecialtooldesignedbythewriter,hasnotyetbeenplacedonthemarket.

Insomecaseswemayreduceweightbysawingtheslotinthescrewdeeper.An excellent poising sawmaybemade froma single-edge safety-razor blade.Lay theblade lengthwiseonamedium file.Give thebladea sharp tapwith asmall hammer. Teethwill be formed on the blade, and its capacity to cut thebalancescrewslotwillbefoundtobeasgoodasanypoisingsaw.

Suppose, now, we wish to add weight. Loosen the screw slightly with ascrew driver. Then,with a balance screw holder, remove the screw. Place thewasher on the rim of the balance and replace the screw.Never place under abalance screw more than one washer. It is essential that one has a largeassortmentofbalancewashersforthemanysizesofbalancescrews.

Thebalanceisconsideredinsatisfactorypoisewhenitcanbestoppedatanypointandremainsstationary.

TRUINGBALANCESPRINGSTruing balance springs is one of the most important branches of watch

repairing,andtherepairmanisjudgedbythequalityofhisbalancespringwork.Since it takes years of practice to become efficient,we advise the beginner topractice—plentyofpractice—ateveryopportunity,onthebalancespringsofoldmovements.

Tweezers forbalance spring truingmusthave finepoints. It isdesirable tohavetwopairwithstraightpointsandoneormorewithcurvedpoints.

Truingintheround.Placethebalance,withbalancespringattached,inthebrasscalipersof lightconstruction.Thesecaliperscanbemademoreeffectivebyfittingfrictionjewelsintheends.Rotatethebalanceandusinga2-incheyeloupe, followthecoils that liearound thecollet. If thecoilsappear to jumpinandoutfromthecollet,thespringisoutoftrueintheround.Threeexamplesareshown in Figure 28. The springs at A show the error. The arrows show themanner in which the errors are corrected andB shows the corrected springs.Usuallyatwistingmotionwiththetweezersisrequired.Inafewcases,apushorpull with a sharp-pointed instrument does the work. Others requiremanipulationsattwoormorepoints.Correctionsnormallydonotextendbeyondthefirst¼coilfromthepinningpoint.Theillustrationsareselfexplanatoryastothemanipulationsrequired.

Figure28.

Truing in the flat.Balance springsoutof truth in the flat are tobe foundeitherhighorlow90degreesfromthepinningpoint,orhighorlowoppositethepinningpoint.Tocorrecttheerror,raisethespringonthelowside,orpushdownthehighside.Oftenitisnecessarytoremovethespringfromthebalance,placeitonabroachor toolespeciallymadefor thepurpose,andpulldownthehighside.Whentheflatismadetrue,examineagaintheround.Oftenacorrectionintheflatwilldistortthespringintheroundandtheoperationsinboththeflatandtheroundmayneedrepeatingseveraltimes.

Adaptingthespringtothewatch.Placethebalancebridgeupsidedownonthe bench. Lay the balance thereon and secure the stud to the balance bridge.Nowplacethebalanceassemblyinpositiononthelowerplate.Sincethebalanceassemblyonlyissecuredtothelowerplate,wehaveanexcellentopportunitytoview thebalancespring fromallanglesand thus realizeanaccuratepictureofthe needed corrections with regard to leveling and centering. Sometimescorrectionscanbemadewiththebalanceinthewatch.Moreoftenitisnecessaryto remove thebalance. If so, remove thespring fromthebridgeandalso fromthe balance and lay the spring on a piece of ground glass*.The light shiningthrough the glass gives amuch clearer picture of the spring than is otherwisepossible,andtheneededcorrectionsaremoreeasilyexecuted.

*Thegroundglassispreparedasfollows:Securetwopiecesofplateglassabout4inchessquare.PlaceasmallquantityofmediumCarborundumpowderononeglassandaddalittleclockoil.Laytheotherpieceof glass over the first piece and rub the two pieces togetherwith a circularmotion.Continue until bothglassesarethoroughlyfrosted.

Correctingthecoilatthestud.Themanipulationsof theoutercoilof theflatspringortheovercoiloftheBreguetaresonumerousthatadetailedaccountcannot be given. A few general statements, therefore, must suffice. Thecorrectionsneeded in the flat springconsistof: (1)centering, (2) leveling,and(3) forming the coil that passes between the regulator pins to a circle whoseradiusisthebalancecenter.Torealizeasatisfactoryadjustmentofcorrection3,itisoftennecessarytobendthespringinwardatapointneartothestud,Figure29.

In addition to the above corrections, the Breguet spring must show (1)concentricvibrationsaroundthebalancestaff,and(2)anovercoilthatstandstoasafeheightsothatthereisnorubbingofthecoilsorcontactwiththestud.

If the overcoil is notmade to the correct shape, the concentricmotion, asstatedabove,willnottakeplace.Ifcorrectionsareneeded,theworkconsistsofmanipulatingtheovercoilaccordingtothefollowingrules:

Figure29.

(1)If the eccentricmotion takesplaceopposite the regulatorpins (as inaflatspring),bringinpartoftheovercoiltowardthecenterofthespring.

(2)Iftheeccentricmotiontakesplaceonthesamesideastheregulatorpins,movepartoftheovercoilbackintothebodyofthespring.

Other corrections of errors relative to the Breguet overcoil consist of (1)raisingtheovercoil;(2)loweringtheovercoil;(3)leveling;(4)centering;and(5)removing kinks. The studentmust create amental picture of the form desiredbeforestarting,andthenproceedwithcaution.Therearenofixedmethods.Themanipulationsaretwisting,bending,pushing,pulling,andpinching.Theupwardbendandthelevelingbend,atthestartoftheovercoil,aremadebygraspingthespringwith strong tweezers at the desired point and pushing the spring into apieceofsoftwood.Trythisonoldbalancesprings.

Tangled balance spring. It occasionally happens, even to the mostexperienced horologists, that a balance spring will become tangled due to anaccident in handling. From all appearances it would seem that there is nosolutionbuttoorderanewspring.

Itisoftenpossible,however,tountanglethespring.Notethepointwheretheouter coil is looped in and through the coils of the spring. Using a smallinstrumentsuchasthetoolshowninFigure14,workthetangledpart,bymeansofacircularmotion,towardtheouterportionofthespring.Itcannowbeseenwhere the outer coil is looped through the several coils. Next work the studbetweenthecoilsthroughwhichitwastangledandthespringwillbeinstantlyrestoredtoitsoriginalshape—provided,ofcourse,thatthespringhasnotbeenbentduringthemanipulations.Thelastoperationrequirestwopairsoftweezers;itispreferredthatonepairhavecurvedpoints.

It is sometimes desirable to remove the stud.By sodoing the tangled coilmaybepulledoutfrombetweentheseveralcoilswithlessdifficulty.

Untanglingbalancespringsrequiresconsiderablepractice,anditisadmittedthatthejobisnotalwayssuccessful.

Puttingwatch inbeat. Thewatch is in beatwhen the roller jewel pointsdirectlytowardthepalletcenteratthetimewhenthebalancespringisatrest.Toaccomplishthisproceedasfollows:Placethebalancebridgeonthelowerplate.Using an eye loupe, center thebalance above the jewels in thebridge in suchmanner as to cause the roller jewel to point toward thepallet center.Note thepositionofthestudholeinthebridgeandassociatethispositionwithsomepointontherimofthebalancewheel.Thenstakethebalancespringonthestaffinthisposition.Repeattheaboveproceduretotesttheaccuracyofthework.

EscapementAdjusting*The importance of understanding the escapement cannot be overestimated.

Largepocketwatchesoftenfunctionquitesatisfactorilywithfaultyescapements,butwithsmallwristwatchesitisverydifferent.Inthesetheescapementmustbepracticallyperfect.Watcheswithpoorlyadjustedescapementsstoppersistently,thoughperhapsonlyoccasionally.Erratic rates, too, canbe traced todefectiveescapements.

*TheactionoftheleverescapementisexplainedinChapterFiveofPartII.

WheelandPalletActionTherelationoftheescapewheeltothepalletsiscalledthewheelandpallet

action.Under thisheading,weshallconsider thefunctionsof:draw,drop, lift,andbankingtothedrop.Instudyingtheabovefunctions,thewatchisassembledcompleteexceptforthedialandthebalance.Itispreferredthatthestudentselecta16-sizeAmericanwatchforpractice.

DRAWDraw isnecessary tokeep theguard finger from too frequent contactwith

thesafety roller.Asudden jarwill cause the lever tobe thrownaway from itsbank, and the guard fingerwillmomentarily touch the safety roller.The drawwill instantly return the lever to its bank and no particular harm is done.However,ifthedrawisineffective,theguardfingerwilldragonthesafetyrollerandinterferewiththefreemotionofthebalance.Stoppagesometimesresults.

Testingthedraw.With themainspringpartlywound,weshouldexpect tofindtheleveratrestagainstabankingpin.Withthetweezers,lifttheleverawayfromitsbank,butnotsofaras tounlockthepallet,Removethetweezers,andthelevershould,ifthedrawissatisfactory,returninstantlytoitsbank.Movethelevertotheoppositebankingandtryagain.Iftheleverfailstoreturntoitsbank,thepalletstonesmaynotbeclean.Sometimesthelockingfaceofthepalletmaynot be sufficiently oiled.Again the escapewheel and levermay not have thenecessaryendshake.Ifalloftheaboveconditionsaresatisfactory,wantofdrawwouldbeduetoimproperangleofthepalletstone.

Correcting the draw. Assuming, now, that the want of draw is due to afaultyangleofthepalletstone,weproceedasfollows:Removetheleverfrom

thewatchandsecureitupsidedownonthepalletwarmer.Holdthetooloveranalcohol lampuntil theshellacmeltsand,with the tweezers,presson thepalletwantingindrawinsuchawayastoincreasetheangle.

Sometimesthejewelfitstheslotsotightlythattiltingisimpossible.Inthiscaseanarrowerjewelmaybesubstituted.Inchangingjewelsbesuretheangleoftheimpulsefaceisthesameasthatoftheoldjewel.

DROPAfteratoothoftheescapewheelpassestheliftingfaceofapallet,thewheel

turnsmore quickly, not restricted in anyway, until another tooth locks on theoppositepallet.Thisfreemotionofthewheeliscalleddrop.Whenatoothdropsoff the receiving pallet, it follows that another tooth locks on the dischargingpallet.Sincethedropofthetoothoccursbetweenthepallets,itiscalledinsidedrop.Likewisewhenatoothdropsoffthedischargingpallet,anothertoothlocksonthereceivingpallet.This,incontrasttotheformer,iscalledoutsidedrop.

Dropshouldbeequalbutwedonotalwaysfinditso.Errorsinthedroparecalledcloseinsideorcloseoutside.

Correcting the drop. Drop, like draw, is corrected by tilting the palletjewels.Ifcloseinside,spreadthepallets;ifcloseoutside,bringthepalletsclosertogether.Examinethedrawaftereachalteration.

LIFTTheliftintheclub-toothescapementisdividedbetweentheteethandpallets.

Proportions vary in the different makes of watches but the latest trend inescapementdesignistoshowalargerportionoftheliftonthepalletsasshowninFigure30.Thepurposeistoavoidaparallelmeetingoftheimpulseplanesofbothteethandpalletswhichinvariablytakesplacewhentheimpulseanglesaremoreevenlydivided.Considering thefact that the teethandpalletsarealwaysoiled,thecohesionissufficienttoresisttheseparationofthepartsincontact.

Figure30.Club-toothleverescapement.

Itwillbefurtherobservedthattheliftinganglesofthepalletsaredifferent,beinggreateronthedischargingpallet.Inthisinstancetheangleisdeterminedfrom the locking face of the pallets.Wemention this so that the studentwillnevermake themistakeoffittingapallet jewel to thewrongsideof thepalletframe.

BANKINGTOTHEDROPThetermbankedtothedrop impliesaparticularadjustmentofthebanking

pins.Thisisrealizedwhenthebankingpinsareadjustedinsuchamannerastopermita tooth todropoff thepallets.Tobank to thedrop,proceedasfollows:Turnthebankingpinsinsotheleverwillnotpermittheteethtodropoffeitherpallet.Then turn thepinsout slowlyand stop the instant a toothdropsoff thepallets*.Figure28,PartII,showsanescapementsoadjusted.Notetheextentofthe lockon thepalletsandnotealso that the roller jewelbarelyclears the slotcorner and that theguard finger shows a similar clearance to the safety roller.Thelockonthepallets,whenanescapementisbankedtothedrop,iscalleddroplock.Itisavaryingquantitydependingonthepositionofthepalletjewelsinthepalletframe.

*InSwisswatchesandsomeAmericanwristwatchesitisnotpracticaltobanktheescapementtothe

dropbecauseofthefactthatthebankingpinsarenotsuppliedwitheccentricscrews.Inthiscasetheusualpracticeistomovetheleverslowlyuntiltheescapetoothdropsandatthesameinstantceasemovingtheleverandtakenoteoftheextentofthelockonthepallets.Aslightadditionalmotionofthelevershouldberequiredbeforetheleverwillreachitsbank,whichis,ofcourse,beyondthatofdroplock.Theadditionalmotion is called slide and will be considered further in the later portion of this section on escapementadjusting.

Bankingtothedropservestounravelmostescapementdifficultiesandformsa basis for making important tests to determine the true condition of theescapement.Thesetestsweshallconsiderpresently.

ForkandRollerActionWe have, up till now, considered the functions of: draw, drop, lift, and

bankingtothedrop.These,itisobserved,constitutethewheelandpalletaction.We are now ready to investigate the relationship between the fork and rollerjewel,whichistermedforkandrolleraction.This,however,mustshowaproperrelationshiptothewheelandpalletaction,whichisdeterminedbywayofthreetests:(1)guardsafetytest,(2)cornersafetytest,and(3)thecurvetest.

Figure31.Showingtheprocedureinmakingtheguardsafetytest.

GUARDSAFETYTESTThe escapement is banked to the drop and the balance is fitted to the

movement.Rotatethebalancesothattherollerjewelstandsoutsideoftheforkand,with the first finger, hold thebalance in thisposition.Now,with the toolshowninFigure15,lifttheleverawayfromitsbank,therebycausingtheguardfingertocomeincontactwiththeedgeofthesafetyroller,Figure31.Withthelever held in this position, examine the remaining lock on the pallet. Thisremaininglockiscalledthesafetylockandshouldrepresenthalfofdrop lock.Thetestshouldnextbetriedontheoppositepalletandasimilarlockshouldbefound.

CORNERSAFETYTESTStartingwiththerollerjewelintheforkslot,rotatethebalanceslowlyuntil

suchtimethatonetoothpassestheletting-offcornerofapalletandanothertoothcomesincontactwiththelockingfaceoftheoppositepallet.Aslightadditionalmotionapplied to thebalancewillbring the roller jewel inapositionoppositetheslotcorner.Withthebalanceheldinthisposition,lifttheleverawayfromitsbank, thereby causing the slot corner to come in contactwith the roller jewel,Figure32.Withtheleverheldinthisposition,examinetheremainingorsafetylock.Try this teston theoppositepallet and, if the safety lock is the sameonboth pallets, the lever’s angular motion from the line of centers will bepracticallyequal.

The safety lock shownby the corner test shouldbe the sameas the safetylockshownbytheguardtest,which,aswehavealreadylearned,ishalfofdroplock.

CURVETESTThepurposeofthecurvetestistodeterminethecorrectcurveofthehornsof

thefork.Thehornsshouldbesoshapedthattherollerjewelwillenterwithoutcatching,while,atthesametime,theguardfingerisrestingonthesafetyroller.Whentheguardfingerentersthecrescent,thesafetyactionistransferredtotheroller jewel and the horns of the fork. It is important, therefore, that asatisfactorysafetylockisstillmaintained.

Figure32.Cornersafetytest.

Figure33.Curvetest.

Toapply the curve test, proceed as follows:Rotate thebalance so that theroller jewelstandscompletelyoutside thehornsof the fork.Next lift the leverawayfromitsbank,therebycausingtheguardfingertocomeincontactwiththesafetyroller,Figure33,and,whiletheleverisheldthus,turnthebalancesothatthe roller jewelwill move toward the fork slot. If the roller jewel passes thehornsoftheforkandenterstheslot,theescapementissatisfactoryasfarasthistestisconcerned.Iftherollerjewelcatchesonthetipsofthehorns,anumberoffaultyconditionscouldbepresent.Themostcommonare:guardfingertooshort,orrollerjeweladvancedtoofar.

EscapementErrorsThe functions and tests of the escapement having been considered,we are

readytomakesuchcorrectionsasappearnecessary.Several toolsandsuppliesare required.Theseare:apalletwarmer;alcohol lampandalcohol;2-incheyeloupe;jewelcement,andleverholdingtool.

Tomovethepalletjewels,placetheleverupsidedownonthepalletwarmer,andtightenitsecurelybymeansoftheclampprovidedonthetool.Heatoverthealcohollamptosoftentheshellacand,usingtheeyeloupe,shiftthejewelsinthedesireddirectionwithasmallscrewdriver.Caremustbe takennot tochipthefaceofthejewels.Themannerinwhichthejewelsarealteredisdeterminedbythenatureoftheerrorspresent.Inthefollowingparagraphs,thefaultsarenotedandtheproperproceduresuggested.

Lockingtoodeep.Lockingtoodeepisshownbytoomuchguardandcornerfreedom.Afterbanking to thedrop, apply theguard safety test and thecornersafety test as already explained and determine which side shows the greaterguardandcornerfreedom.Thisisourguideastowhichpalletjeweltopushin.Obviously the pallet opposite to the side showing the greater freedom is thepallettopushin.Aftereachalteration,re-banktothedropandrepeatthetests.

Lockingtoolight.Whenthelockingistoolight,someoftheteethmayfailtolock.Insamecasesthereisnolockingatall.Whentheescapementisbankedto the drop, the roller jewelmay not leave the fork slot, or perhaps the rollerjewel may leave the fork slot only on one side. The adjustment consists inadvancingthepallet jewelonthesidewheretheangularmotionoftheleverislonger from the line of centers. Test for guard and corner freedom after eachalteration.

Lever too long. This error is rarely met with in watches made in recentyears.Itis,however,occasionallyfoundinwatchesmadeoverthirtyyearsago.Iftheerrorissmall,alonglevermaybemadetofunctionsatisfactorilybytiltingthe roller jewel toward the balance staff. This provides the necessary cornerfreedom.Ofcourse,therollerjewelshouldstandreasonablyparalleltothestaffand,iftheerrorisquiteextreme,othermethodsofcorrectionarerequired.

Theexcessively long lever is, therefore,correctedbygrinding thehornsofthefork.Theprocedureisasfollows:Placeinthelatheapieceofsoftsteelwireand turn it down to such size as just to fit the horns of the fork. ApplyingCarborundumpowderandoiltothewire,grindthehornsback,therebymakingthe levershorter.Aspecial tool forholding the levermaybepurchasedat tooland material dealers. Frequent tests should be made so as to not overdo thecorrection.

Levertooshort.Theshortlevershowstoomuchcornerandguardfreedomand is corrected by advancing the roller jewel. If the lever ismuch too short,stretchingisrequired.Tostretchthelever,placeasmallflatstumpinthestakingtoolandlaytheleverthereon.Usingapunchwiththesidesflattened,lightlytap

thelever.Iftheleverbecomesbent,turnitoveronthestumpandgiveitanotherlighttap.Frequenttestsshouldbemadeasthestretchingtakesplace.

Insufficient pallet frame clearance. One of the conditions that is oftenpuzzling to beginners is the stopping of awatchwhich, from all appearances,seemstobecausedbyafaultinthetrain.However,uponremovingthelever,thetrainrunsdown.Figure34showsthecause.Atoothengagesthepalletframeatthepointnearthepalletarbor.Thecorrectionconsistsoffilingthepalletframesothattheteethofthewheelwillpasswithoutcatching.Thedifficultymayalsobecausedbydefectivejewelsorwornholeswhichsupporttheescapewheelandlever.

SlideUp to this point in our discussion, the escapement has been banked to the

drop.There isoneadditionaladjustmentcalledslide.Slideconsistsofopeningthebankingpinsbeyondthatofdroplock,providingaslightadditionalmotionof the leverbefore thebankingpinsare reached.Slide isnecessarybecauseofmechanical imperfections such as variations in the length of the teeth of theescapewheel and the side shake in thehole jewels.Slide,however, shouldbekept small. Excessive slide increases the angular motion of the lever and itsconnection with the balance, resulting in an increased unlocking resistance, ashorter arc ofmotion of the balance, andpoor timekeeping.The banking pinsshouldbeturnedsoastostandasfarawayaspossiblefromthepalletcenter,soastolessenstrainontheleverpivotsiftheescapementoverbanks.

Figure34.

Slide is the last adjustment — the finishing touches, so to speak, inescapementadjusting.

Faulty ticking of the escapement. Watches that run erratically or stopoccasionally will often be found to have a knocking sound in one or morepositions.Beginnersmayfinditdifficulttodistinguishbetweenatickthathasaknockpresentandonethatdoesnot,butcan,withexperience,detectit.

Thisknockingindicatesafaultyadjustmentoftheescapement.Trytheguardsafety test and the corner safety test. A small corner freedom will produce aknockandwillshowupinthatpositionwherethebalancedropstowardtheforkslot.If,uponexamination,thecornerfreedomappearssufficientandequaltotheguardfreedom,removethebalanceandseeifthereisanyslidepresent.Wantofsufficientslidewillalsoproduceaknock.

If theescapement isotherwise ingoodadjustment, theaboveerrorscanbecorrectedbyopeningthebankingpinsslightly.Whenthisisdone,theknockwilldisappearandthewatchwillusuallyrunsatisfactorily.

RepairingtheDialTrainTightening the cannon pinion. The cannon pinion must fit, with some

friction,overthepostofthecenterwheel.Wantofsufficientfrictionwillresultinthecannonpinionslipping,andconsequentlythehandswillfailtomove.

Most staking tool sets include the equipment necessary to tighten cannonpinions.FitaV-shapedstumpinthedieofthestakingtoolframe.InsertaroundbroachinthecannonpinionandlaythepinionontheV-shapedstump.Selectachisel-shaped punch and give the cannon pinion a light tap with the brasshammer,thusclosingtheholeinthecannonpinion.Fitthecannonpiniononthecenterpostandfeelthefrictionbypushingthepinionaroundwiththetweezers.

Manymodernwatcheshaveablindcannonpinion;thatis,apinionwiththeend closed.We cannot run a broach through this type of pinion. In this case,shortenabroachuntilitfitsthecannonpinionwithsomefriction,andsaveforfutureuse.Doingthesameonotherjobs,onewillsoonhaveanumberofsuchbroaches that will make a set. These could be fitted to a block or box andnumberedaccordingtosize.Asetofsixbroachesshouldsuffice.

Afterfittingthecannonpiniontothewatch,putthehourwheelinplaceonthecannonpinion.Makecertainthatthewheelfitsfreely.Ifitbinds,broachoutslightly.Placethedialinpositionandnotetheshakeofthehourwheel.Always

usedialwashers,ifthespacepermits.Flattenthewasher,asthetensionshouldnotbeverygreat.Dialwashersaidinkeepingtheminuteandhourhandsapart,and thehandsare, therefore, less likely tocatchoneachother.Note,also, thattheholeinthedialiscenteredtothepipeofthehourwheel.Oftentheholeinthedialisnotcenteredandthehourwheel,withhourhandattached,willbindinthehole.Removethedialandfiletheholewithasmallroundfilesoastocentertheholetothehourwheel.Removetheburrwiththewheelcountersinksandreplacethedial.

FITTINGHANDSBeginnersaretobewarnedaboutthefittingofhandstowatches,andsome

horologistsoflongexperiencegetcomebacksfornootherreasonthanthefaultyfittingof thehands.Extremecaution, therefore,cannotbeoveremphasized,fortheaverageownerofawatchisquiteignorantofanypartofthemechanism,andno matter how carefully the rest of the work has been done, the job will bejudgedasunsatisfactory.

Seethat thehandshavesufficientclearance.Besurealsothat thehandslieparallelformostoftheirlength.Theouterendoftheminutehandshouldbebentdowntowardthedialslightly.

Restoringrustyhands.Hands that are rustyandotherwiseunattractive, ifthe conventionalblack in color, canbemade to look likenewby touchingupwithIndiadrawingink.Sharpenapieceofpegwoodtoapointandapplytheinkfreelytothehands.Sometimesit isdesirablefirst tosmooththehandswithanArkansasslip.

Sometimesthecustomerdesiresblackhandsinplaceofgoldorsilverhandssoastoreadthetimebetter.These,too,canbemadeblackwithIndiadrawingink.

CLEANINGDIALSThe silver dials found in modern watches often get tarnished and soiled

badly. If the figures are baked on, the dials can be cleaned by dipping into asolutionofcyanideofpotassium.Thisispreparedbydissolvingone-halfounceof cyanide in aquartofhotwater andadding twoounces of strong ammonia.Rinsethedialinwarmwateranddryinwarmsawdustordryer.

Painteddials,however,willnotstandsuchtreatment,asthefiguresarelikelyto be removed.These dials canbe cleanedbyusingbaking soda.Moisten thefirstfingerinwateranddipintothebakingsodaandrubthediallightly.Wash

carefullywithsoap,rinseinwarmwater,anddryinwarmsawdustordryer.

MagnetismMagnetism will cause a watch to gain time, to lose time, or, if heavily

magnetized,tostopaltogether.Totestforthiscondition,removethebackofthecase and place a compass over the balancewheel andwindingwheels. If thewatchismagnetizedtheneedlewillbedeflected.

To removemagnetism,place thewatch in thedemagnetizer and,while thecurrent ismaintained,slowlydrawout thewatch toadistanceofabout2 feet.Test thewatchagain. Ifmagnetism is still present, repeat theaboveprocedureuntilthecompassneedleliescompletelyatrest.

Never attempt to demagnetize the balance spring when it is out of themovement.

PARTIChapterFour

REPAIRANDADJUSTMENTPROBLEMSTAKENFROMPRACTICE

THISCHAPTERconsistsofacompilationofanumberofrepairandadjustmentexamples experienced by the writer over a considerable period of years. Itspurpose is to show the reader the actual procedure in specific cases,with thehopethatperhapssomeoftheseexamplesmaybetheanswertothedifficultiesyou may be having with some particular job. Consider first a few generalstatementsrelativetopointssometimesoverlookedbybeginners.

PointstoRememberinPracticalRepairingManymodernwristwatcheshaveasweepsecondhandwhichnecessitatesa

carefuladjustment.Oftenthecrystalisnothighenoughtoprovidethenecessaryclearance.One canbe fooled regarding the requiredheight of the crystal.Thewatchmayrunontherackbutnotfortheownerwhenwornonthewrist.Thereasonseemstobethattheextremityofthesecondhandwillvibratewhenwornand touch thecrystalorminutehand, causing thewatch to stopand later startagain.

Other causes for stopping that are sometimes overlooked are: (1) tightminutewheelorhourwheel;(2)rustypinions;(3)rustyescapewheelteeth;(4)setmainspring;(5) looseroller jewel; (6)roller jewel too long, touchingguardfinger;(7)out-of-flatcenterwheeltouchingbalancewheel;(8)crackedjewelsorwornholesintheplatesandbridges.

Some of the causes for erratic rate are: (1) balance out of poise; (2)magnetism;(3)looserollerjewel,sometimessolittleastoescapedetection;(4)regulatorpinspinchingbalancespring;(5)regulatorpinsspreadtoofarapart;(6)loosecannonpinion;(7)balancemotiontoolong;(8)balancemotiontooshort.

TherearemanycausesforErrorNumber8butwhateverthecauseanerraticpositionratewillinevitablytakeplace*.Assumingthatthewatchiscleanandin

goodmechanicalcondition,ashortbalancemotionwouldmostlikelybeduetoasetmainspringora springof incorrect strength.Thecorrectbalancemotion is1½turns,thatis,¾ofaturninonedirectionand¾turnonthereturnvibration.Successinwatchrepairingdependsonsecuringafullbalancemotiononeveryjobturnedout.

*The theoryandpracticeofpositionadjusting is fully analyzed inAPracticalCourse inHorology,HaroldC.Kelly,Chas.A.BennettCompany,Peoria,Illinois.

If, after a thorough examination of that problem watch, no errors arediscovered, donot conclude that the owner is at fault. Insteadwear thewatchyourself. The horologist will sometimes be embarrassed to learn that thecustomerisright.

PracticalProblemsinPositionAdjustingAdjustingwatcheswith thewatchraterecorder.Wristwatches,because

of their small size, cannot be adjusted to the close position rating generallyattainedinpocketwatches.Allowancesvaryaccordingtosize.Pocketwatchesareratedtowithin10secondsbetweenfivepositions.The10½ligneandlargercanusuallyberatedtowithin30secondsinthreepositionsbut thesmallestofwatchespresentmanydifficulties.Inthesearatingwithin90secondsforthreepositionsshouldbeconsideredsatisfactory.

In comparing the balance springs of large pocketwatches and smallwristwatches, we find that the collets and the innermost coils are largerproportionately in the latter. This is required because of constructionaldifficulties.Asaresulterrorsinherentinthebalancespringsareincreasedanditfollowsthatperfectionoftheinnerterminalisofvitalconcern.*

*Forananalysisofthepositionerrorinherentinthebalancespringseepage274.

Counterpoising, sometimes practiced in pocket watches, is not advised insmallwristwatches,asthebalancemotionismorevariablethaninlargepocketwatches.

Another observation is the relative increase of the escapement error in thesmallestofwatches.Theverticalpositionsrateslowerthanthehorizontalandaneffectivecorrectionconsistsinfittingastrongermainspring.

The use of the watch rate recorder has made the practice of positionadjusting entirely practical. A study of the following examples should proveinteresting.

(NOTE. Wrist watches are adjusted preferably to dial up, dial down, andpendant down. For greater precision, twelve down may be added. Pocketwatchesareadjustedtodialup,dialdown,andpendantup.Railroadgradesareadjustedtodialup,dialdown,pendantup,pendantright,andpendantleft.)

Adjustment1.Watch—10½lignes,15jewelsRepairs—cleaned,demagnetized,mainspringfitted

Thefirsttestafterrepairingshowedanexcessivelyslowpendantdownrate:

Because a poise error wassuspected, the balance springwas removed and the balancewas tested on thepoisingtool.Asmallpoiseerrorwasfoundandaccordinglycorrected.Withthespring replaced, the balance was placed in the brass truing calipers and thespring was examined for truth at the collet. A slight eccentricity at the innerterminal was found. After careful truing of the spring and reassembling thewatch,thenexttestshowedthefollowingrate:

Adjustment2.Watch—5½lignes,17jewels

Repairs—stafffitted,balancepoised,cleanedTheratesshowninthisexamplearesimilartotheprecedingone.

The usual routineexamination was made and the following faults were noted: (1) an eccentricmotionofthespringatthecollet,(2)balancemotionslightlylessthan1½turns,and(3)regulatorpinsopen.

Aftercorrectingtheaboveconditionsandre-regulating,thefinaltestshowedthefollowingrate:

Adjustment3.Watch—5½lignes,15jewelsRepairs—balancepoised,cleaned

Thefirsttestshowedaconsiderablelossinthependantdownposition.

Examination showed aperfectly true balance spring and correctly adjusted regulator pins. Since thebalancehadbeenpoised, there remainedonlyonepracticalcorrectionand thishadtodowiththemotionofthebalance.Checkingshowedagoodmotionbutitcouldbeincreasedslightlywithoutvibratinginexcessof1½turns.

Actingonthesuppositionthatanincreasedbalancearcwouldacceleratethependant down position, we fitted a stronger mainspring. After regulating, thefinaltestshowedamuchimprovedrate.

Adjustment4.Watch—5½lignes,17jewelsRepairs—cleaned

Thefirsttestshowedthefollowingrate:

Inthisexamplethependantdownratewasfast,which,inwristwatches,ismostunusual.Apoiseerrorwassuspected and the watch was tested in four vertical positions. The rates werefound tobepracticallyequal inallpositions.Anyerrorofpoisewas thereforeruledout.

The regulator pins were next examined and the outer coil of the balancespringwasfoundtorestagainsttheinnerpinforallbutthelongestvibrations.The outer coil was adjusted to vibrate equally between the pins and, afterregulating,thefollowingratewasattained:

Adjustment5.Watch—10½lignes,15jewelsRepairs—cleaned

The recordbelow indicatesaconditionofbalancemotionmuch too long.The

normalrateisshownforafewsecondsbut,asthemotionincreases,overbankingtakesplace,resultinginanabnormallyfastrate.

Aweakermainspringwasfittedand,withoutdoinganythingelsetothewatch,thefollowingratewasattained:

WatchesThatStopThefollowingexampleshavetodowiththedifficultyofstopping.Wedonot

pretendtohavecoveredthevariouscausesforstopping,buthaveindicated,webelieve,someoftheunusualones.

Example1.A6-sizewatchofAmericanmakewithcompensatingbalancewascleaned

andthebalancetruedandpoised.Thewatchgavegoodserviceforaboutthreemonths, when the owner brought it back complaining that the watch lostconsiderabletimeduringthenight.

Thewatchwastestedineveryconceivableway.Itwouldrunduringthedayand keep time. In the morning the watch would be several hours slow butrunning. Further examination showed an unusual condition. The clearancebetweenthebalancescrewsandthetrainwheelbridgewassosmallthat,duringthenightwhenthetemperaturewascold, theoutwardbendingoftherimswassufficienttocausethescrewstotouchthebridgeandstopthewatch.Duringthemorninghourswhen the temperature rose, thewatchwould startof itself.The

watchwasrepairedinthelatesummermonthsandranconsistently,butwiththecomingofcoolerweatherthediameterofthebalancebecamelonger.

The correction consisted in adding a small pair of balance screws andremovingthedialwashersfromunderthehighestscrews.

Example2.A small and very narrow baguette watch was brought in by a lady who

explainedthatthewatchhadbeenrepairedmanytimesbutnoonehadeverbeenabletomakeitrunforlongerthantwoweeks.

Thewatchwasgivenacompleteoverhaul;anewmainspringwasfittedandtestedonthewatchraterecorder.Anexcellentrecordwasshownbutthewatchwouldrunperhapsadayor twoandthenstop.Each timewhenthemovementwasremovedfromthecase,itwouldberunning.Themovementwastakenapartandthewheelscarefullyexamined.Fromallappearancesitstoppedinthetrainbutnosucherrorcouldbediscovered.Finallyweobserved,accidentally,thatthecenterwheelmovedwhenthecrownwastouched.Furtherexaminationshowedthat thepivotof thewindingstemtouchedthepinionof thecenterwheel.Therepairwassimple—merelyshorteningthepivotofthestem.Thiscorrected,thewatchcausednofurthertrouble.

Example3.A very small, high-grade Swiss watch, made some thirty years ago, was

assertedbytheownertohavebeencleanedsomethreeweeksbefore,inadistantcity. The owner complained that the timekeeping was not consistent andoccasionallystopped.

Thebalancewas removedandexamination showed that thepivotswere ingoodcondition.Thebalance cap jewelswere removedandone cap jewelwasfoundtobecompletelycoveredwithoil.Theothercapjewelshowedtheproperamountofoil.Theonejewelwithtoomuchoilwascleanedandreplaced.Thebalancewascleanedandputinpositioninthewatch.Thebalancenowvibratedtoafull1½turns,whichwasnotthecasebeforethecorrection.Thewatchwastestedforfivedays.Itkeptgoodtimeanddidnotstop.

Example4.AnoldmodelSwisswristwatchwascleaned,timed,andplacedontherack

forobservation.Aftertwohourswediscoveredthatithadstopped.Examinationshowed that the hour handwaswedged tight.A slightmovement of the hourhand freed the hands and the watch started. The hands were tried in several

positionsandfoundtobefreeineveryinstance.Againsettingtotime,thewatchwas placed in the rack. On the next morning the watch was found to havestoppedduring thenight.Examinationshowed that thehourhandwaswedgedtight.

The dial was removed and the dial washer was taken out. The dial wasreplacedwithoutthewasher,andtheendshakenoted.Theendshakewasratherexcessiveanditoccurredtousthatperhapstheendshakecouldbereducedandthedialwasherdispensedwith.Thiswasaccordinglydonebybendingdownthedialnearthemiddle.Replacingthedial(thistimewithoutthewasher)thehourwheelwascheckedandfoundsatisfactoryasregardtofreedom.Thehandswerefitted,checkedinseveralpositions,andsettotime.Thewatchransatisfactorilyandcausednofurthertrouble.

Ordinarilydialwasherscanberecommended,butthewashersdonotpreventthe hourwheel from climbing if there is a tendency to do so. It is logical toassume that since thehourwheelwas free for severalhours running, itwouldremainfree if theclimbingcouldbeprevented. If thedialcannotbeshaped toreducethehourwheelshake,useathickwasherwithoutspringaction.

Example5.A 15-jewel traveling clock with alarm was cleaned and balance pivots

polished.After running threedays itstopped.Thehourwheelwasfoundtobetight.Applyingpressurewiththetweezersonthehourhand,thehourwheelwasfreedandtheclockstarted.Testingthehourwheelinseveralpositionsshowednoerror.However,wedgingofthehourwheeldoestakeplaceseeminglyinspiteof all tests.Onemay reason that slightly broachingout thehourwheelwouldcorrectthetrouble.This,however,istheveryworstpractice.Neverbroachoutthehourwheelunlessitactuallyfitstightonthecannonpinion.

Theerrorinthisinstancewascorrectedinthefollowingmanner:Thecannonpinionwasplaced in the latheand thepointswhere thehourwheel ridesweresmoothedwiththeArkansasslipandthenpolishedwiththesteelburnisher.Nowfittingaflat-facedstumptothedieofthestakingtool,thehourwheelwasplacedthereon, wheel side up, and, with a round-faced punch, the hole in the hourwheelwasclosed.Nexttheholewasbroachedouttofitthecannonpinionclosebutfree.Againthehourwheelwasplacedonthestump,thistimewiththepipeupward,andtheendofthepipewasclosedwithataperendpunch.Lastly,thepipewasbroachedouttofitthecannonpinion.Inthismannerthehourwheelismadetofitthecannonpinioncloselyandwithaminimumofsideshake.Tiltingis thus avoided and the hour wheel willmove up and downwithout binding.

Afterassembling,theclockcausednofurthertrouble.Ingoodgradealarmclocks,thealarmissetoffbyanupwardmovementof

thehourwheel.Hence,asmoothactionofthewheelisrequired.Thisisoftenasourceoftroubleandcarefulexaminationshouldberoutineprocedure.

ProblemsRelatedtoPendulumClocksExample1.

An 8-day pendulum mantel clock was cleaned and repaired by a studentapprentice. When the clock was set up and running, it ran with reasonableaccuracyfor threedays,when it started to lose,andaftersevendays theclockwas ½ hour slow. Obviously, something was definitely wrong. Examinationshoweddeepgroovescutinthepallets.

Theescapewheelwasshiftedonitsarborsothattheteethworkedonpartsofthepalletthatwerenotworn.Thenexttestshowedaratewithin2minutesaweekwithoutevenchangingthelengthofthependulum.

Example2.A popular-priced grandfather clock with hour and half-hour striking

movement,weight powered and fittedwith a seconds pendulum,was cleaned,holes closed, and pallets polished. After the movement was set up on anespeciallydesignedshelf,itranforaboutfourdays.

Thepalletshadbeenbrokenandsomeworkmanhadfiledupapieceforonepalletandsoftsolderedit totheremainderof thepalletframe.Wetouchedthepallet with a file and found it to be rather soft. The piece was removed byheatingthesoftsolderandthenheatedtoacherryredtohardenit.Aftercleaningand polishing, the piece was soft soldered to the pallet frame. After theescapementwasagainassembled,theclockransatisfactorily.

The palletsmust be glass hard. There seems to be a drag if they are soft,sufficienttocausestopping.

Example3.AfineFrenchclockwascleaned.Nootherrepairsseemednecessary,sothe

movementwassetupinthecase,started,andsettotime.Theclockkeptgoodtimewhilerunning,butitwouldstopafterthreeorfourdays,seeminglywithoutreason,forifre-starteditwouldrunanothertwoorthreedayswithoutadditionalwinding.

Pastexperienceservedusinthiscase.Anewtimespringwasfittedandthe

clockcausednofurthertrouble.French clocks have a reputation for acting in this manner. They are fine

timekeepers but good mainsprings and precision adjustments are essential inorderforthemtorunatall.

PARTII

THEMECHANICSOFHOROLOGY

PARTIIChapterOne

WHEELWORKOFPENDULUMCLOCKSTHEPURPOSEofwheelsinaclockistogovernthemovementofthehourand

minute hands while the escapement makes a required number of beats. Thiswould naturally require wheels that gear into pinions* so as to change thevelocity of the several wheels in order to perform the necessary function ofregisteringtime.

*Apinionisasmallerwheelwithteethcalledleavesworkinginconnectionwiththelargerwheel.

CalculatingtheNumberofTurnsLetusexaminetherelationshipthatexistsbetweenawheelandapinionofa

given number of teeth and leaves. Assume that a wheel of 96 teeth drives apinion of 12 leaves. It is evident that after the pinion hasmade one turn, thewheel has advanced 12 teeth. With two turns of the pinion, the wheel hasadvanced24teeth,etc.Itfollowsthatinordertoobtainthenumberofturnsofthepiniontooneturnofthewheel,wedividethenumberofteethinthewheelbythenumberofleavesinthepinion.

Theusualpracticeistomakeuseofthefollowingformula,inwhichWisthe

wheelandpisthepinion:Substitutingthenumericalvalues,wehave:96

Figure1.

NowconsideracombinationoftwowheelsandpinionsasshowninFigure1.ThewheelWdrivesthepinionp.TheaxisofthepinionpcarriesthewheelW′,which,inturn,drivesthepinionp′.

Itismoreconvenient,however,towritetheformulaasfollows:

CLOCKTRAINSAgroupofwheelsandpinionsperformingasinglefunctioniscalledatrain.

Thesimpleclockhastwotrains:themaintrainandthedialtrain.The main train changes a slow motion to a fast one with the particular

purposeofcausing thewheel thatcarries theminutehand tomakeone turn inthesametimethattheescapementmakesarequirednumberofbeats.

Thedialtrain,ontheotherhand,changesafastmotiontoaslowoneforthepurposeofgoverningthedistancethehourhandtravelstooneturnoftheminutehand.

MainTrainofWeight-drivenClocksFor the purpose of describing themain train, let us consider the standard

regulator movement. This movement is of simplest possible construction,containingonlyfourwheels,dialtrainexcepted.Itispoweredbyaweightandisusuallyfittedwithapendulumbeatingseconds.

Figure2.Regulatormovement.Showingarrangementofthetrain.

Figure2showsaregulatormovementoftheAmericantype.The first wheel, sometimes called the great wheel, has attached to it a

groovedbarrel uponwhich a cordormetal cable iswound.To the endof thecordisattachedtheweight.Thefirstwheeldrivesthesecondorcenterpiniontowhichisattachedthecenterwheel.(Notethatthesecondpinionis,inreality,thefirst pinion, but for convenience it is given the same name as the second orcenterwheelofwhichitisapart.)Thecenterwheelcarriestheminutehandand,ofcourse,makesone turn inanhour.Thecenterwheeldrives the thirdpinionandwheel,whichinturndrivetheescapepinionandwheel.

Calculating the turnsof a complete train. The formula for the regulator

maintrainisasfollows:inwhichthecapitallettersFCTindicatethefirst,centerandthirdwheelsandthesmallletterscteindicatethecenter,thirdandescapepinions.

Substitutingthenumericalvalueswehave:

Sincethecenterwheelmakes12turnsto1ofthefirstwheel,andobserving

that thecenterwheelmakesone turn inanhour, it follows that the firstwheelmakes1/12turninanhour.Thusbydividing720by12,welearnthattheescapewheelmakes60turnsinanhouror1turnaminute.

Calculating the number of beats. The escape wheel in the regulatormovementmustdeliver60impulsestothependulumwitheachturninordertoregister seconds.Toaccomplish this, theescapewheelmusthave30 teeth, foreach tooth delivers 2 impulses, first to the receiving pallet and later to thedischargingpallet.EliminatingthefirstwheelandlettingE indicatetheescape

wheel,theformulanowreads:Substitutingthenumericalvalueswehave:

Sincetheescapewheelmakesoneturninaminute,asecondhandmaybefixed to the extremity of the escape pinion. This is the usual practice in theconstructionofregulators.Alongpivotontheescapepinionextendsthroughaholeinthedialandthesecondhandisfittedbyfriction.

MainTrainOfSpring-DrivenClocksPendulumclocksbeingmadeinsuchalargevarietyoftypesandsizesfrom

small novelty clocks to huge tower clocks, it follows that the pendulummayvaryinlengthfromabout4inchesto13½feet.

Trains vary in construction accordingly, with more or less wheels andnumerous types of escapements. If the clock is to be fitted with a shortpendulum, resulting in a more rapid vibration, it is desirable to add anotherwheel tothetrain.Theseclocksare,ofcourse,drivenbyamainspringandtheformulabelowhasafourthwheeladdedtothetrain.ThecapitallettersC,T,F,andEindicatethecenter,third,fourth,andescapewheels,andthesmallletterst,f,andeindicatethethird,fourth,andescapepinions.

Shownbelowisanumericalexample:

Figure3.MovementofAmericanstrikingclockwithtopplateremovedshowingthearrangementoftimingandstrikingtrains.

The American mantel clock. Consider now the main train of the clockmovementshowninFigure3.ItisthetypegenerallyusedinAmericanhourandhalf-hourstrikingclocksandinsomepopular-pricedwallclocks.

NotethatthefirstwheelGdoesnotgearintoacenterpinion.Insteaditgearsintoasecondpinions.ThesecondwheelSdrives twowheels.Theseweshallcall the third pinion t,which stands above the secondwheelS and the centerpinionc,whichissituatedmidwaybetweenthetimingandstrikingtrains.

Thefiguringofatrainthusarrangedbecomesmorecomplicatedthantheonejust considered and involves a more elaborate mathematical calculation todetermine a train so as to function with a required number of beats. Twooperations are required. Formula No. 1 and numerical values are shown asfollows:STF

The escapewheelmakes 260 turns to one of the secondwheel.However,sincethesecondwheeldoesnotmakeoneturninanhour,and,sinceweknowthatthisisrequiredofthecenterpinion,itfollowsthatthenumberofturnstheescapewheelmakesinanhourmaybefoundbydividingtheratioofthecenterpinionto thesecondwheel into260.Nowthecenterpinioncontains26leavesand the second wheel 60 teeth. Dividing 26 into 60 we find that the ratio

between the twomobiles is24/13.Bydividing24/13 into260wedetermine thenumber of turns of the escape wheel in an hour. Therefore:

Wenowwishtodeterminethenumberofbeatstheescapementmakesinanhour—theescapewheelcontaining48teeth.

Itismoreconvenient,however,tocalculatethebeatsperhourbyusingoneadditionaloperation(FormulaNo.2)asshownbelow.

Thus,bycancellation,theproblemissolvedmorequickly:

POWERUNITOFWEIGHT-DRIVENCLOCKSReturning for further study of the weight-driven clock, we are confronted

with certainproblems.Suppose the clock is to run eight days.The sizeof thebarrel and the lengthof fall of theweight arepoints tobe reckonedwith.Forthesewemust resort to two formulas. Firstwemust determine the number of

turnsofthebarrelforeightdaysofrunning:Substituting:

Havingdeterminedtheturnsofthebarrelforeightdaysofrunning,wemay,bymeansoftheformulabelow,determine:(1)thelengthoffalloftheweightforagivendiameterofthebarrelor(2)thediameterofthebarrelforagivenlengthoffalloftheweight.

Statingtheformula:

Problem:What would be the diameter of a barrel making sixteen turns, if the case

permittedafallof6feetfortheweight?

Problem:Whatwouldbethelengthoffalloftheweightifthediameterofthebarrelis

1½inches?Thebarrelmakessixteenturns.

Figure4.

Doublecord.Inordertoreducethefalloftheweightwithoutreducingthebarrel diameter to impractical proportions, it is sometimes desirable to use adoublecordasshowninFigure4. In thiscase theweightfallshalfasfast,buttwicetheweightisrequired.

Maintainingpowermechanism.Whenaweightclockrequireswinding,theactofwindingtendstoreversethetrain,possiblyjammingtheescapementandthus interferingwith the timekeeping.Amaintaining power device overcomesthisdifficultyandtheclockiskeptgoing.

The mechanism, invented by John Harrison in 1745 and still used inregulators, isshowninFigure5.Thewinding ratchetwheelWR is attached tothebarrelandbotharefixedtothearborA.Themaintainingratchetwheel liesoutsidethewindingratchetwheelandisfreetoturnonthearbor.Lyingnextisthe firstwheeland it, too, is freeof thearborandalso freeof themaintainingratchet.

Actionofthemaintainingmechanism.ThewindingclickWCispivotedtothemaintainingratchetwheelMR,anditspointengagestheteethofthewindingratchetwheelWR.Theclockisrunningbyitsweightandthetworatchetwheelsturn together.Themaintaining springsSandS′ are soonunder tension and allpartsarenowturninginthedirectionofthearrowB.Nowsupposetheclockisbeingwound.ThewindingratchetwheelwillbeturninginthedirectionofthearrowC.Themaintainingratchetwheel isheldfromturningbackwardsby themaintainingclickMC,whichispivotedtotheclockplate.Theforcecreatedbythemaintaining springs S and S′ drive the first wheel in the direction of thearrowB, thusproviding thenecessarypower to the train as the clock is beingwound.

Figure5.Maintainingpowermechanism.

POWERUNITOFSPRING-DRIVENCLOCKSTherearetwotypesofpowerunitsinuseinspring-drivenclocks.Thelower-

priced clocks usewhat is termed theopenspring— that is, the spring is notenclosedinabarrel.Theouterendofthespringissecuredaroundapillerthatliesbetweentheplatesandtheinnerendhooksonanarborthatcarriesthefirstwheel.Aratchetwheelisattachedtothearborandtheclickissecuredtothefirstwheel. In winding the spring, the power is taken off the train, but, when thewinding ceases, the power is again delivered to the first wheel through theratchetwheelandclickandthencetothetrain.

Inthebettergradeclocks,thespringisenclosedinabarrelcalledthegoingbarrel.Thespringiswoundthroughthearbor;thearborcarriestheratchetandtheclickisattachedtotheplate.Sincethespringiswoundfromthecenterandtheouterendofthespringisattachedtothebarrel,itfollowsthatthepowerisalwaysmaintained to the trainwhen spring is under tension. Springs fitted togoingbarrelsshouldoccupyone-thirdofthespacebetweenthebarrelwallandthearbor.

Further information concerning mainsprings may be found in the partdevotedtobalance-wheeltimekeepers.

THEDIALTRAINFigure6showsthedialtrainofthetypeusedintheFrenchmantelclock.The

construction,however,ispracticallythesameinalltypesofclocks.Thecenterwheel,whichturnsonceanhour, isfittednear thecenterof the

movementandhasalongarborwhichextendsthroughaholeinthedial.Atubewithapinionatitsbase,calledthecannonpinion(C,Figure6),fitswithaslightfrictiononthecenterarborandturnswithitwhentheclockisrunning.Insettingthehandstotime,theentiredialtrainturnsindependentofthecenterarbor.Theupper end of the cannon pinion carries theminute hand and the pinion at thebase drives the minute wheelM. To the minute wheel is secured the minutepinionm,whichinturn,drivesthehourwheelH.Thislastwheelfitsfreelyonthecannonpinionandcarriesthehourhand.Thedialtrainissodesignedastocausethehourhandtomakeoneturntotwelveturnsoftheminutehand.

Figure6.Dialtrain

Theformulareadsasfollows,usingthelettersasgivenabovetoindicatethewheelsandpinions:

Substitutingthenumericalvalues:

The dial train may be formed by many different combinations, the onlyrequirementbeingthatthecombinedratiosequal12.

In the fourth example, thecannonpinionandminutewheelhave the samenumberofteeth,necessitatingaratioof12to1betweentheminutepinionandthehourwheel. This combination is often used in hall clockswhere the pins,

which release the striking or chiming mechanisms, are fitted to the minutewheel.

Anotherfunctionattachedtothedialtrainisfoundinrackandsnailstrikingclocks. The snail is generally mounted on the pipe of the hour wheel.Whenassemblingthedialtrainthehourwheelmustbefittedinsuchamannerthattherackdropstotheloweststeponthesnail.Thisisnecessarysothattheclockwillstrikecorrectly.

PARTIIChapterTwo

WHEELWORKOFBALANCEWHEELCLOCKSANDWATCHES

THE ANALYSIS of clock trains given in the previous chapter applies in ageneralwaytobalance-wheeltimepieces.Sincethevibrationsofthebalancearemorerapidthanthatofthependulum,itisnecessarytoaddanextrawheeltothetrain. This extra wheel is called the fourth wheel and pinion, which, in mostwatches, turns once a minute. Usually the pinion is provided with a pivot ofsufficientlengthtoreachthroughaholeinthedial.Tothislongpivotasecondhandisfitted.

MainTrainBelowisshowntheformulafor themain trainfromcenterwheel tofourth

wheel,inwhichthecapitallettersCandT indicate thecenterandthirdwheelsandthesmallletterstandfindicatethethirdandfourthpinions.

Addingtheescapepinionetothetrain,theformulareads:

Thecalculation(page116)showsthattheescapewheelandpinionmake600turnsanhouror10turnsaminute.

Calculating the number of beats. The escape wheel in most watchescontains15teethanddeliverstwiceasmanyimpulsestothebalance.LettingEindicate the escape wheel and e the escape pinion, the formula, from fourth

wheeltoescapewheel,readsasfollows:

Multiplying the 300 beats perminute by 60we find that thewatchmakes18,000beatsperhour,whichisalsoshownbygivingthecompleteformulafromcenterwheeltoescapewheel.

Fastandslowtrains.Notallwatchesmake18,000beatsperhour.Someoftheoldermodelsmake14,400beatsor16,200beatsperhour.Since thefourthwheelmakes 60 turns aminute, it is obvious that the trains differ only in theratiosbetweenthefourthwheelandescapepinion.Thetrainsareasfollows:

Smallwristwatchesoftenhavefasttrains,asthefollowingformulasshow:

PORTABLECLOCKSPortableclocksaregenerallydesigned to runeightdays. In these themain

train is the sameas thewatchexcept thatawheel isaddedbetween thebarrelandthecenterwheel.

DialTrainThedialtrainforbalancetimekeepersdoesnotdifferinprinciplefromthat

usedinpendulumclocks.Thenumberofteethandleavesdifferproportionately,butwelistherewiththemostusedcombinationsfoundinbalancewheelclocksandwatches.

TheBarrelandItsMainspring

Theearliestwatchesused thefuzee toequalize theforceof themainspringjustasthechronometerhasthisfeatureeventothisday.*Laterwhenthefuzeewasdiscarded,astopworkwasfitted,usuallytothebarrelcap.Itspurposewastomakeuseofthemiddleportionofthemainspringsoastoequalizethemotivepower.Inpresent-daywatchesthestopworkisunnecessary,sincethequalityofthe mainsprings has been greatly improved. Furthermore, the recoil clickreleasesthemainspringfromexcessivefrictionbetweenthecoils,whichwas,inpart, thepurposeof the stopwork.The fuzee isnowobsolete inwatches, andlikewise the stopwork.Bothwould be troublesome rather than helpful in thesuperiorengineeringofmodernwatches.

*Astudyof thechronometer and the fuzeebelong toa separateand specialized fieldofhorologicalpracticeanddoesnotcomewithinthescopeofthisbook.SeeHorologybyJ.EricHaswell,ChapmanandHall,Ltd.,London.

Goingbarrel.Theboxforthereceptionofthemainspringinmostmodernwatchesisofthetypecalledthegoingbarrel.Inthistypethearboristurnedinwinding the spring.Thebarrel is coveredwith teeth and, obviouslydrives thetrain.

Motor barrel. The motor barrel differs from the going barrel in that thebarrel, proper, is turned in winding the spring. In this case the first wheel isattachedtothebarrelarbor.Thetrainispropelled,therefore,throughthebarrelarbor.

Calculating the number of hours a watch will run. The barrel and itsmainspring are so designed as to run about 36 hours. The ratio between the

numberof teeth in thebarreland the leaves in thecenterpinion, togetherwiththelengthofthespring,determinesthis.Inagivenwatchthebarrelcontains75teethandthecenterpinion12leaves.Thecenterwheelmakes6¼turnstooneofthebarrel.Whenweobservethatthecenterwheelmakesoneturninanhour,itfollowsthatthebarrelmakesoneturnin6¼hours.Bycountingthenumberofturnsoftheratchetwheel,welearnedthatittakes6½turnstocompletelywindthe spring. Thus bymultiplying 6¼ (hours for one turn of the barrel) by 6½(turnsofthebarrel)weget40+,whichequalsthenumberofhoursthewatchwillrun.

Themainspringshouldrunthewatchnotlessthan32hours;36to40hoursisbetter—infact,someofthefinestwatcheswillrun45hoursandlonger.

Calculatingthecorrectthicknessofthemainspring.Wehavelearnedthatthebalancemusttakeamotionof1½turns,and,ifthewatchisingoodorder,thismotion is realizedbyadoptinga springof the required strength.Researchhasshownthatthestrengthofthespringisproportionaltoitswidth—thatis,aspringtwiceaswidedevelopsaforcetwiceasstrong.Astothickness,theratiois different, being proportional to the cube of its thickness. Thismeans that aspringtwiceasthickdevelopsaforceeighttimesasstrong.

Thisisshownbynotingthatthemeasurementofthicknessconformscloselyto the diameter of the barrel. For example, a reasonably accurate rule fordeterminingtheproperthicknessofthespringisasfollows:

Measure the full diameter of the barrel and divide the figure by 100. Thequotientisthethicknessofthespring.

Figure7.

Themetricsystemofmeasurementisusedintheabovecalculation.

Calculatingthecorrectlengthofthemainspring.Thelengthofthespringneed not be calculated in somany inches. Insteadwe note the space that thespring occupies in the barrel. Assuming that the correct strength has beenascertained,thecorrectlengthisrealizedwhenthespringoccupiestwo-fifthsofthespacebetweenthebarrelwallandthearbor.Figure7showsthisclearly.

PARTIIChapterThreeGEARING

THEGEARINGofaclockorwatch ismadeupofwheelsandpinionswhosecircumferencesarecoveredwithteethandleaves.Thefunctionofatoothis topressonaleafinamannerthatwillcausethepiniontoturnwithavelocitythatisconstanttothewheel.Thisrequiresthatthemobileshaveteethandleavessoshapedastoconformtocertainscientificprinciples.

PrinciplesofGearingInordertovisualizethispointmoreclearly,letusfirstassumethatthereisa

circlewithinthetoothedsectionofthewheelthathasacommonvelocitywithasimilarcircleofthepinionandthatthesecirclestoucheachother.Thiswouldbecompared to a rollerwithout teeth driving a smaller rollermerely by friction.The circumferences of both rollers would have a common velocity, providedthereisnoslippingaction.

Pitchcircles.Inagearing,thesecirclesarecalledpitchcircles,andalthoughtheirlocationisnotsoclearlydefined,theyhavearealexistence,nevertheless.Figure8showsthepitchcirclescuttingthroughthemiddleportionoftheteeth.Thatpartofa toothlyingoutside thepitchcircle iscalled theaddendum.Thatpartlyinginsideiscalledthededendum.Whenawheeldrivesapinion,itwillbeobserved that the addenda of the wheel teeth act only on the dedenda of thepinionleaves.

Figure8.

THEEPICYCLOIDANDHYPOCYCLOIDAddenda.Theshapeofthewheelteethmostusedforhorologicalpurposes

isderivedfromageometricalcurvecalledtheepicycloid.Itisformedbymeansofapointononecirclewhilerollingonanothercircle.Figure9showshowthecurve is formed. The portion of a circle A represents the pitch circle of thewheel.AsmallercircleBequalshalfthepitchdiameter*ofthepinionandjusttouchesthepitchcircleofthewheelA.LetussupposethatcircleB(calledthegeneratingcircle)wascutoutofcardboardandlikewisethepitchcircleA.Now,placingthesecirclesonasheetofdrawingpaperwiththegeneratingcirclejusttouchingthecircleA,fitapencilpointtothegeneratingcircleatthepointwherethe two circles intersect. Next roll the generating circle on circle A withoutslippinginthedirectionofthearrow.Insodoing,thepencilpointwilltraceoutthesideofthetooth,indicatedbythelineC.

*Pitchdiameteristhediameterofthepitchcircle.

Figure9.

Widthof tooth. It is observed that the analysis of the epicycloid as givenabovetakesaccountofonlyonesideofa tooth.Thequestionnowarisesas tohowtodeterminethewidthofthetooth.Thisisfoundbydividing360(degreesinanycircle)bythenumberofteethinthewheel.Thisgivesus,indegrees,thewidthofonetoothandonespace,generallyreferredtoascircularpitch.

Thewidthofthetoothisequaltohalfthecircularpitch;theotherhalfis,ofcourse,equaltothespace.

Figure10.

ofwhich2.25degreesisthewidthofthetoothand2.25degreesthewidthofthespace.

NextplacethegeneratingcircleB,Figure9,2.25degreestotherightoflineC.WiththegeneratingcirclenowcenteredonlineDandindicatedbythedottedcircleE,fitapencilpointaspreviouslyexplainedandrollthegeneratingcircletotheleft.ThepencilpointwilltraceouttheothersideofthetoothasshownbythedottedlineD.TheintersectionofthetwocurvesCandDformsthepointofthetooth.

Dedenda. The dedenda of the pinion leaves is determined by the samegeneratingcirclethatgaveustheshapeoftheaddendaforthewheelteeth.Themethoddiffers,however,forthegeneratingcircleisrolledinsideandalongthepitchcircleofthepinion,producingtheradiallineA,Figure10.

Thelineformedbyrollingacirclewithinacircleiscalledahypocycloid.

ACTIONOFCYCLOIDALGEARINGWhenawheelandpinionaremadetoconformwiththeaboveprinciplesof

design, a constant and uniform angular velocity from wheel to pinion ismaintained.InstudyingFigure8itwouldseemthatthepinionwouldbedrivenmorerapidlyduringthelatterportionoftheleadbecauseofthelengtheningofthewheelteethandtheshorteningofthepinionleaf.Thisisnottrue,however,for there isa slippingactionequal tohalf thewidthof the tooth,and it is thisactionthatequalizestheirangularvelocity.

Engagingfriction.The incoming toothof thewheelshouldbeginpressingonapinionleafasnearaspossible to the lineofcenters,allowingfor thefactthatthisisavaryingquantitydependingonthenumberofleavesinthepinion.Ifthedepthingiscorrectitwillbefoundthattenleafpinionsbeginactionon thelineofcenters;eightleafpinions,10degreesbeforethelineofcenters;andsixleafpinions,18degreesbeforethelineofcenters.Obviouslyitispreferablethatpinionscontainnotlessthaneightleavesinclockwork.

CorrectSizesofWheelsandPinionsPitch diameter. In order to determine the correct sizes of wheels and

pinions, it is firstnecessary tofind thepitchdiameters.For thisweproceedasfollows:Adjustthemalecentersofthedepthingtooltothetwoholesinaclockplateinwhichawheelandpinionaretorotate.Measurethedistanceoutsidethetwospindleswithaverniercaliperandsubtractthediameterofonespindle.This

gives us the center distance for one wheel and pinion. To arrive at the pitchdiameters,determinethediametricalpitch,whichisournextstep.Theformula

reads:Thediametricalpitchisnowmultipliedbythenumberofteethinthewheel,inorder to determine the pitch diameter of the wheel, and in like manner thediametricalpitchismultipliedbythenumberofleavesinthepiniontodeterminethepitchdiameterofthepinion.

Forexample,thecenterdistanceis17.15millimeters;thewheelhas42teeth;thepinionhas7leaves.Substitutingthenumericalvaluesfortheaboveformula,

wehave:Continuingtheproblemwefindthat:

Proof:

The full diameter. The height of the addenda is a varying quantitydepending on the ratio of the wheel to the pinion, but the production oftheoreticallycorrectgearsorevenknowingwhentheyexistisnotpossiblewiththeequipmentavailabletothepracticalhorologist.Theusualpracticeistoadd2.5diametricalpitches to thepitchdiameterof thewheeland1.25to thepitchdiameterofthepinion.*Experiencehasshownthattheabovefiguresarebestforallpracticalpurposes.

*Thereisoneexceptiontotheabovestatement.Forthedialtrainwherethepinionsdrivethewheelsandthewheelsdrivethepinions,asinthecaseofsettingthehandstotime,theaddendaisfiguredas2forbothwheelsandpinions.

Wefoundthatthewheelhasforitspitchdiameter29.4millimetersandthepinion 4.9 millimeters. The diametrical pitch multiplied by 2.5 gives us theheightoftheaddendaforthewheelteeth:.7×2.5=1.75Adding this to the pitchdiameter of thewheel,wehave: 29.4+1.75=31.15mm.fulldiameterofthewheelNow,figuringthepinionwehave:

.7X1.25=.875

4.9+.875=5.775mm.fulldiameterofthepinionWemay,however, figure the fulldiameterswitha lot lessworkbyadding

2.5or1.25(addenda)tothenumberofteethorleaves.Forexample:(42+2.5).7 = 31.15mm. full diameter of the wheel ( 7 + 1.25) .7 = 5.775mm. fulldiameter of the pinionCircular pitch. It will be noted that the definition forcircular pitch reads somewhat like the definition for diametrical pitch. Thedifference is:circularpitch is thedivisionof thecircumferenceof acircle (thepitchcircle),whereas thediametricalpitch is thedivisionof thediameter of acircle (the pitch diameter). In both cases the number of teeth or leaves is thedivisor.

Wemustknowtheactualwidthoftoothandspaceinordertoselectacuttertomakethewheel.Hereinlies theimportanceofcalculatingthecircularpitch.To attain this we make use of the following formula:

Substitutingthenumericalvalues:

Theproportionoftoothorleaftospaceisusually:forwheels:one-halfofthecircularpitch

forpinions:one-thirdofthecircularpitchNow,continuingwiththeabovewefindthat:

LanternPinionsThe lantern pinion gets its name from the fact that it looks like an old-

fashioned lantern.The leavesaremadefromroundsteelpinsandaremountedbetweenbrassdisks.Itisusedinalarmclocks,popular-pricedmantelclocks,andintowerclocks.

Thewheelthatdrivesthelanternpinionhascycloidalteeth.Thepitchcircleforthepinionpassesthroughthecentralpartofthepins.

Itisacuriousfactthattheincomingtoothofthewheeldrivesthepinsofthelantern pinion on the line of centers. There is no exception to this, even in alantern with six pins. Suppose, however, the lantern pinion was to act as the

driver. All of the action would take place before the line of centers and theengaging frictionwouldbeverygreat, in fact, there is somedoubt if the trainwouldrunatall.Itfollows,ofcourse,thatthelanternpinionisneverusedasadriver.Asafolloweritisequalandpossiblysuperiortoanyothertypeofpinion,inspiteofthefactthatitischeaplymade.

Figure11.

InvoluteGearingInvolutegearing is used in all heavymachinery. It is used in tower clocks

and occasionally in smaller clocks. The involute system is increasing inpopularityforhorologicalinstrumentsanditdoeshavesomeadvantages.

Figure11showsthemannerinwhichtheinvolutetoothisformed.Astringfastened to a cylinder atA is pulled tightly to the cylinder and reaches toB.GraspthestringatBandraiseittoC.ThelineBCformsthesideoftheinvolutetooth.Thecircumferenceofthecylinderiscalledthebasecircleandliesslightlyinsidethepitchcircle.

Theteethofthefollowerareformedinthesamemanner;infact,ifwefastenastringbetween twocylindersas shown inFigure12and turn thembackandforth, keeping the string tight, the curves for both wheels are formedsimultaneously.

Figure12.

Actionofinvolutegearing.Thereisatendencytothrustthewheelsapartinthe involute system, but this presents no problem, as the friction between theteethresiststheseparatingaction.Aslightdiscrepancybetweenthedistanceofcenters is not as serious an error as it would be in the cycloidal system. Theslopeof the teeth is altered,but this always remains tangent and, furthermore,thevelocityisnotchanged.

Involutegearingservesequallywellfordriversorfollowers.

PARTIIChapterFour

ESCAPEMENTSFORPENDULUMCLOCKSTHE ESCAPEMENT of a timepiece is a mechanism for the purpose of

maintaining the vibrations of the pendulum or balance. Conversely, thependulum or balance regulates the circular motion of the escape wheel to arequired velocity. That type of escapementwhich interferes the least with thefreemotionofthevibratingunitwill,ofcourse,keepthemostaccuratetime.

We shall discuss in this and the next chapter various types of clock andwatchescapementswithparticularconsiderationoftheirmeritsanddefects.

EscapementsUsingthePendulumProbablytheearliestescapementfordrivingthependulumwasconceivedby

GalileoGalileiin1610.Hemadeadrawingofitwhichhasbeenpreserved,butaworkingmodelwasnotmadeuntilafterhisdeath.

Galilei’sescapementwillbedescribedinthischapter,togetherwithanotherescapement,theverge,whichHuygensusedinhispendulumclocks.Followingthis,modernescapementswillbedescribedindetail.

Modernescapementsmaybedividedintofourclasses.Thefirstistherecoil,theoldesttypethatisstillinuse.Itisfoundinmantelclocks,wallclocks,andingrandfather clocks. The second is the dead-beat form generally found in thebetter-grademovements.The thirdclass is thegravityescapement. In this typetheimpulseisdeliveredtothependulumbymeansofleversimpelledbygravity.Gravity escapements are used almost exclusively in tower clocks. The fourthclass isaprecisionescapementassociatedwithwhat is termed freependulums,anditsuseisconfinedtoastronomicalandscientificworkwheregreataccuracyisrequired.

TerminologyAstudyoftheescapementinvolvestheuseofmanyspecialtermsthatmust

be defined and understood before we can proceed with an analysis ofescapement construction and action.The several parts ofmodern escapementsforclocksaredefinedasfollows:

Escapewheel: The wheel of the escapement that delivers impulse to thependulumthroughthemediumofthepalletsandcrutch.

Pallets: That part of an escapement that receives impulse from the escapewheelanddeliversimpulsetothependulum.

Receivingpallet:Thepalletoverwhichatoothoftheescapewheelslidesinordertoenterbetweenthepallets.

Dischargingpallet:Thepalletoverwhichatoothoftheescapewheelslidesinordertoleavefrombetweenthepallets.

Impulseface:Theliftingplaneofapallet.

Lockingface:Thefaceofapalletonwhichatoothlocks.Thisisafunctionthattakesplaceonlyindead-beatescapements.

Lettingoffcorner:Theextremeendoftheimpulsefaceofapalletwhereatoothoftheescapewheelletsoff.

Crutch:Aslendermetalpiece thatconnects thepalletswith thependulumrod.

Pendulumrod:Theconnectinglinkbetweenthesuspensionspringandthebob.

Suspension spring: A short flat spring that supports the pendulum andpermitsittoswingtoandfro.

Galilei’sEscapementClockswiththevergeescapementandfoliotbalanceweremadefromabout

theyear1000through1670.Duringthisperiodofover600years,practicallynoimprovement inescapementandbalance tookplace. It isverystrange, indeed,thatnoone,duringallthistime,everconceivedofthesimpledevicesuchastheanchorescapementforcausingthependulumtovibrate.

Galilei, after observing the swinging chandelier in the Cathedral of Pisa,suggested that the pendulum could be used for controlling the time in clocks.However hewas of the opinion that itwas practically impossiblemanually tokeepapenduluminmotion.Nevertheless,beforehisdeathin1642,hemadea

drawingofanescapement,themechanicalprincipleofwhichisshowninFigure18.

Construction.The teethof the escapewheel serveonly to lock thewheelthroughthelockingpieceL.Impulsetothependulumisperformedbymeansofimpulsepinssecuredtothefaceofthewheel.

Actionoftheescapement.Thependulum,swingingtotheleft,carrieswithit the impulse pallet I and also the unlocking palletU. The palletU lifts thelockingpieceL, thus releasing the toothT.Thewheel is now free to turn andimpulse pin P engages the impulse pallet I, thereby giving impulse to thependulum,driving it to the right.Simultaneously theunlockingpalletUdropsand permits the locking piece L to engage the oncoming tooth of the wheel.Whenthependulumreturnsaftercompletingitsswingtotheright,theunlockingpalletagainliftsthelockingpieceandtheactionisrepeated.

Figure13.Galilei’sescapement.

The chronometer escapement comprises the same mechanical principle asdoes thisescapementbyGalilei. It is a factworthnoting that thechronometer

escapement, perfected 140 years later, is themost perfect timekeeping devicethathaseverbeenconceived.Andyet,oddlyenough,Galilei’sescapementwasnotdivulgedtotheworld,andtheescapementsthatwereeventuallydevelopedproceededuponwhollydifferentlines.

VergeEscapementThe first mechanical clocks— that is, those consisting of toothedwheels

drivenbyaweight—datefromthe tenthcentury.Theearliestescapement thatattainedanyprominencewastheverge.Theinventorisunknown,althoughthenameofPopeSylvesterII(950-1010?)hasbeenmentioned.

Construction. The time counter for the earliest verge clocks was ahorizontalbarcalledthefoliot(Figure14).Itisasortofbalanceuponwhichtwoweightsaresuspendedfromnotchesneareachend.Theweightscanbemovedtowardorawayfromthecenterofthefoliotforthepurposeofregulating.ThearborforthefoliotissecuredtoacordatthetopandapivotatthebottomrunsinaholeintheblockA.Thelowerportionofthearborcarriestwoflattenedsteelpallets. The pallets and the arbor together are called the verge. The anglebetween the pallets is usually 90 to 100 degrees. In some cases, particularlythosemadeinFrance, theangle isevengreater,oftenasmuchas115degrees.Theseofthelattertypemusthavethepalletsplantedmuchclosertothewheel,resultinginalongervibrationbutlessrecoil.

Theescapewheelmusthaveanoddnumberofteethinordertoworkatall.Theusualnumberis13or15forwatchesand21,31,or33forlargeclocks.

Action of the escapement. In the illustration, the escape wheel is shownturning in the direction of the arrow.The tooth under palletB is in the act ofgiving impulse. Pushing pallet B to the left, the tooth eventually escapes,permittingtheescapewheeltoturnuntilanothertoothengagespalletC.Afterarecoilduetotheinertiaofthefoliot,thetoothdeliversimpulsetopalletC, thistimepushingthepallettotheright.SoonatoothescapesandactionisrepeatedwhenanothertoothengagespalletB.

Figure14.Vergeescapement.

Figure15.Huyghen’sclockfromHorologiumOscillatorium.PublishedinParisin1658.

Verge escapement with pendulum. Christian Huyghens adopted thependulum to the verge escapement (Figure15) about the year 1657. Thiswasconsidered quite an achievement in its day, since the pendulum rates moreconsistently than the vergewith foliot balance.More is said aboutHuyghens’clockonpage240,whereananalysisofthecircularerroristreatedindetail.

RecoilEscapementThe recoil escapement, shown in Figure 16, was invented by Dr. Robert

Hookein1675.Itissometimesreferredtoastheanchorescapement.

Construction.Theconstructionissimple.Theescapewheelisofbrassandthenumberofteethvarygreatly,dependingonthetypeofclockandthelengthofthependulum.Thepalletsaremadeofsteel,hardenedandhighlypolished.

Figure16.Recoilescapement

Actionoftheescapement.Asthependulumswingstotheleft,carryingwithit the pallets, a tooth of the escape wheel drops from the receiving palletA,Figure16.ThewheelisnowfreetoturnbutitsmotionisimmediatelyarrestedbecauseofcontactofatoothwiththedischargingpalletB.Afterthependulumcompletesitsoutwardswingandreturnstotheright,itreceivesafreshimpulsethrough the escape wheel C and the discharging pallet. Eventually thedischarging pallet releases the tooth and another tooth engages the receivingpallet.Thustheactioncontinuesaslongasthepropermotivepowerisheld.

The recoil escapement gets its name from the fact that the escape wheelrecoilsorreversesitsdirectionimmediatelyafteratoothdropsonapallet.Thisactiontakesplacebecausethepalletshavenolockingfaces.Henceatoothfallsdirectly on the curved impulse face and, as the pendulum continues to swingoutward,thecurvedpalletpushesthetoothbackuntilthependulumcompletesavibrationandstartsinthereversedirection.Theactualimpulsetothependulumisfoundbysubtractingtherecoilfromtheforwardmotionoftheescapewheel.

Escapement error. It is said that the recoil escapementmaintains amoreuniform pendulum arc than the dead-beat (Figure 17) because the powerdelivered to theescapewheel tends to reduce the recoil in itseffort tobeginanew impulse. Such action limits the free vibration of the pendulum and an

escapementerrorofgreatmagnitudeisproduced,resultinginalossfortheshortarcs. Very accurate time, however, may be realized in grandfather clockspowered by a weight for, in this case, the power at the escape wheel isreasonablyconstant.

Figure17.Dead-beatescapement.

Dead-beatEscapementGeorge Graham invented the dead-beat escapement in 1715. It is still

regardedasoneof thebest escapements forhigh-grade regulators,grandfatherclocks and others of the better quality. Except by comparison with the finestprecision clocks, it leaves little to be desired. The only defect lies in a slighttendencytochangetherateofrunningaftertheoilhasthickenedordeteriorated.

Construction. The original Graham escapement was so designed that thepalletsembracedfifteenteeth.Inthemoremoderndesign,Figure17,thepalletsspaneightteeth,althoughadesignspanningtenteeth,Figure18,hasbeenusedforlargeclocksandastronomicalregulators.Whenthelessernumberofteethareembraced,thelengthandangleoftheimpulsefacesarereducedandlikewisetherunonthelockingfaces.ThesefactorsbecomeclearwhencomparingFigure17andFigure18.

Escapement error. If the impulse of the Graham escapement could be

deliveredtothependulumanequaldistancebeforeandafterthedeadpoint,thetimeofswingwouldnotbealteredandtherewouldbenoescapementerror.Thiscannotberealizedinpractice,althoughtheGrahamescapementapproachesthisideal condition quite closely. There is about half a degree impulse before thedeadpointandonedegreeafter.Theanalysisonpage237 shows that impulsedeliveredafterthependulumreachesthedeadpointresultsinaloss,butactuallytheGrahamescapementshowsagainwhenthependulumarcsbecomeshorter.Itmaybereasonedthatthefrictionalongthelockingfacesaccountforthisunusualphenomena,for,withlesspoweronthetrain,thependulumvibrateswithgreaterfreedomandistosomeextentmoredetached.Sinceitisknownthattheimpulseerror is small, the disturbing factor ismore likely the circular error,which, initself,showsagainfortheshortarcs.

Figure18.

Figure19.Pin-wheelescapement.

Pin-WheelEscapementExcept for turret clocks, the pin-wheel escapement may be regarded as

obsolete.ItwasinventedbyJ.A.Lepautein1750.

Construction.Figure19 shows thepin-wheel escapement.Semicircle pinsstandoutfromthefaceoftheescapewheelandengagethepalletsononesideonly,sothatthepressureisalwaysdownward.Theimpulseisdividedbetweenthe pins and pallets, and, when one pin lets off a pallet, the bottom of theoncomingpinmustlocksecurelyontheoppositepallet.Theamountoflockisdeterminedbyspreadingorclosingthepalletarms.

Whenapinis lockedbetweenthepallets, insidedropisdetermined.Whenboth pallets stand between two pins, outside dropmay be observed. The pin-wheelescapementhastheadvantageofreducingthedroptotheminimum.Alsothe pallet arms may be made to any reasonable length without increasing ordecreasingthedrop—afeaturethatisfoundinnootherescapement.

Pin-PalletEscapementThisescapementwasinventedbyAchilleBrocot(1817-1878)ofParis.Itis

nowobsoletebutmaystillbe found in theolderAmericanandFrenchmantelclocks, particularly those enclosed in glass or marble cases. The escapement

oftenoperatesvisiblyinfrontofthedial,presentinganattractiveappearance.

Construction. The pin-pallet escapement has much in common with theGraham.Itisfoundinbothrecoilanddead-beattypes.Proportionsvary,buttheusual diameter of the pallets is a little less than the distance between twoconsecutiveteethofthewheel.Withthepalletsembracingtenteeth,theimpulseis three degrees, as shown in Figure 20. The pallets are usually carnelian,semicircularinform,andheldinplacewithshellac.

Figure20.Pin-palletescapement.

Denison’sGravityEscapementThe gravity escapement by E. B.Denisonwas originally designed for the

HouseofParliamentclockinLondonin1854,andperhapsnoothertowerclockhas earned a better reputation for accurate timekeeping.The fine performancemustbeascribedtotheDenisonescapementfor,whenthoroughlystudied,twoveryimportantpointsbecomemanifest.

(1) The impulse to the pendulum is entirely free of the train disturbancessuchasvariationsinthemotivepower.

(2)Theimpulsestartsatthetimewhenthependulumcompletesitsoutwardswingandcontinuesduringtheinwardswing,untilthelineofcentersisreached.

Thishasaquickeningeffecton thependulumand tends tocompensate for thecircularerror.

Construction. Figure 21 shows the Denison escapement. Two gravityimpulsearmsAandA′arepivotedasnearlyaspossibletothebendingpointofthependulumspring.TheescapewheelWconsistsof twothinmetalpiecesofthreelegseach,separatedbythreepins,calledliftingpins.TheimpulsearmsAandA′ liemidwaybetween the twometalpieces.The light, three-leggedpiecestands in front of the impulse arms and the shaded piece behind the arms.AttachedtotheimpulsearmsaretheblocksBandB′,uponwhichthelegsoftheescapewheellock.BlockB′issecuredonthefrontofthearmA′andblockBonthebackofarmA.ExtendingtowardtheliftingpinsLandUfromtheimpulsearmsaretheliftingpalletsPandP′.

ActionoftheDenisonescapement.AlegoftheescapewheelislockedonblockB′,overlappingitby½degree.ThepositionoftheimpulsearmA′andtheamountof lockingonB′ is controlledby the liftingpinL actingon the liftingpalletP′.

Thependulum,nowmovingtotheright,soonengagesthelowerextremityofthe impulsearmA′and,pushing it, releases the leg that is lockedonblockB′.TheescapewheelstartstorotateandtheliftingpinLseparatesfromtheliftingpalletP’,leavingtheimpulsearmA’freetopressonthependulumrod,therebybeingreadytogiveimpulsetothependulumonitsreturnvibration.Atthesameinstant, while the wheel is rotating, lifting pinU engages the lifting palletP,pushingtheimpulsearmAoutwardsothattheoncominglegoftheescapewheellocksontheblockBoverlappingit½degree.ThustheimpulsearmAispushedbackinadvanceofthependulum.

Figure21.Denison’sgravityescapement.

Thependulum,havingcompleteditsvibrationtotheright,receivesimpulsefromtheimpulsearmA’,whichfallswiththependulumtotheleftuntilbankingpin BP’ is reached. The pendulum continues to swing to the left for aperformancesimilartotheonejustcompletedontheright.

ReiferEscapementHorologists have for many years observed the detrimental effects of the

escapementontheuniformrateofpendulumvibrations.Intherecoilanddead-beat escapements, any variation in themotive power is carried directly to thependulum, resulting in noticeable variation in time. The gravity escapementshowsaconstantlyuniform impulse,but the locking isvariable,dependingonthepowertransmittedthroughthetrain.

In 1891 Dr. Siegmund Reifer of Munich produced a clock with a freependulum—thatis,anescapementwithouttheconventionalcrutch.Insteadofacrutch,theimpulseisproducedbydeflectingthependulumspring.

Construction. Figure 22 shows the principle on which the escapement isbasedbutwithoutalltheconstructionaldetails.

There are two escapewheels.One is for the impulse and theother for thelocking.Theimpulsewheelactsoncylindricalpalletsfortheimpulseand,whenthe impulse is completed, the locking wheel follows immediately to lock thepallets.Theouterendsof thepalletsareflattenedtoprovideasuitablelockingsurfaceforthelockingwheel.

Action of the escapement. The receiving pallet is locked on the lockingwheelandstartsthependulumswingingtotheleft.Afterthependulum,togetherwith thepallets,has traveledasufficientdistance, thereceivingpalletunlocks,permitting theescapewheels to rotate. Instantly thedischargingpallet receivesan impulse followed by a tooth locking — an order completely reversed ifcompared with the Graham dead-beat escapement. This action deflects thependulumspring,which,bytheresistanceitofferstobending,givesimpulsetothependulum.Observethatthishappensbeforethependulumhasyetcompleteditsswingtotheleft.Theoutwardswingofthependulumisthusrestrictedand,when thearc is completed, the tensionon thependulumspring is sufficient tokeepthependulumvibrating.

Figure22.Reiferescapement.

Escapement error. As with the gravity escapement, the impulse has anacceleratingeffect; thegreater thearc thegreater is theacceleration.However,the escapement error is directlyopposite to the circular error andone tends tocompensate for the other, although not perfectly. It can be shown that theescapementerrorisgreater.

ItmaybereasonedthatthegoodgoingofaReiferclockisattributedtothesmall arc of vibration of the pendulum and the uniformmotive power whichconsistsofanautomaticelectricalwindingdevice.

LeroyEscapementOneofthemorerecentprecisionescapementsisthatdevelopedbyLeroyet

Cie,ofParis,asshowninFigure23.Inprinciple,theoperationissimilartotheReifer escapement, but has the advantage of a greater detachment of thependulumfromtheescapement.

Construction. The impulse is delivered to the pendulum through themedium of two springs that are secured directly to the pendulum rod. Thesesprings, called impulse springs,meet at a pointwhere the pallets are pivoted,givingittheappearanceofaY.ThetwoendsAandBcarrysapphirepalletsthatterminate into projections called locking nibs. The inclined portions provideimpulsetothependulumandthelockingnibslocktheescapewheel.

AlsosecuredtothependulumistheframeC,whichsupportstwoadjustablestopsDandE.Thesestopslimittheinwardmovementoftheimpulsesprings.

Figure23.Leroyescapement.

Actionof theescapement. The pendulumhas just completed its swing totherightandonetoothoftheescapewheelhasalreadymovedalongtheimpulsefaceofthereceivingpalletandislockedonthelockingnib.TheaboveactionofimpulseandlockingbendstheimpulsespringG,separatingit fromthestopE.Byreasonof the tensionof thespring, thependulumisnowdriven to the left,while the frameC, since it lies above the center of suspension,moves to theright. The pendulum continues to swing, and eventually stopE overtakes theimpulse springG and, pushing it, unlocks the escapewheel. Immediately thewheelturnsandanothertoothliftsthedischargingpalletA,incidentallybendingtheimpulsespringF.Thisactionretardsthependulumnearthecompletionofitsoutward swing.When the vibration is completed, the pendulum returns to theright,sincethespringisstillundertension,torepeattheactiononthereceiving

pallet.Change of temperature would vary the elasticity of the impulse springs;

however, precision clocks are usually placed in special chambers where thetemperatureremainsconstant.SomeLeroyclocksarefittedwithelinvarspringsinplaceofsteelspringstoeliminatethetemperaturefactor.

TheLeroyclockisuniqueinthatitcontainsonlytwowheelsandindicatessecondsonly.Theclockwindselectricallyeverythirtysecondsandisoperatedbyafour-voltbatteryorequivalent.

PARTIIChapterFive

ESCAPEMENTSFORBALANCEWHEELTIMEPIECES

THEESCAPEMENTS described inChapterFour use thependulumas the timecounter. Portable clocks, chronometers, and watches require a balance andbalancespringasthetimecounterandmustofnecessityuseadifferenttypeofescapement.Weshallconsider in thischapter fourescapements thathavebeenused in balance-wheel timekeepers.These are: the lever, cylinder, duplex, andchronometer.

LeverEscapementTheleverescapementwasinventedbyThomasMudge(1715-1794)in1750.

Itdiffersfromthecylinderandduplexescapementsinthattheleverisfreeofthebalance for a greater part of the vibration. Theoretically the lever escapementapproaches the ideal; itsperformance isexcellentand it is as simpleasany toconstruct.The escapement as conceived byMudgediffersmaterially from themodern form, and because of its historical interest we shall include in thischapter abriefdescriptionof it.Twoother early typesof special interest shallalsobeconsidered.Theseare:therackleverandtheratchettoothescapement.

Themodernleverescapementisfoundintwotypes:(1)thepinpallet,usedinalarmclocks,noveltyclocks,andwatchesofextremelylowcost,and(2)theclubtooth,usedinclocksandwatchesofthebetterquality.

Figure24.LeverescapementbyThomasMudge.FromChamberlin’sIt’sAboutTimeandusedbypermissionofRichardR.Smith.

LEVERESCAPEMENTBYMUDGEItappearsthatMudgewasnotconvincedthathisescapementwasanything

very special, forheapplied it toonly twoofhiswatches.Otherwatchmakers,too,didnotrealizeitsmeritanditwassomefiftyyearsbeforeitwasgenerallyacceptedastheultimateinwatchescapements.

Figure24 shows theplanof theoriginal.Thewheel contains twenty teeth,andthepalletsembracefiveteeth.Thepallets’lockingfacesarearcsofcirclesconcentric to thepalletcenter.The teetharepointed,and this featurehasbeentraditionalwiththeEnglishtocomparativelyrecenttimes.

Figure25.Rackleverescapement.FromChamberlin’sIt’sAboutTimeandusedbypermissionofRichardR.Smith.

The forked end of the lever is different from the modern form. The twoprongs of the fork lie in different planes. Two impulse pieces on the balancestaff,inlinewiththetwoprongsofthefork,performtheunlockingandimpulseactions. The safety action is effected by a small disk of steel with a shallownotchsimilartothemoderndouble-rollerescapement.

RACKLEVERESCAPEMENTTheoldestmethodoftransferringtheimpulseofthelevertothebalanceis

the rack lever (Figure 25). The balance staff has a pinion which gears into atoothedrack.Thebalanceisneverfreeoftheescapement.Thereisnonecessityforanysafetyaction,noneedforbankingpins,andthereisverylittlefrictionorresistance in the act of unlocking. The inventor was Abbe Hautefeuille, whopublishedadescriptionofitin1722.

RATCHET-TOOTHESCAPEMENTTheEnglishratchet-toothescapementplaces theentire liftingactionon the

pallet,asshowninFigure26.Itisclaimedthatthereislessfrictionascomparedwiththeclub-toothtypebecauseofthepointedteethslidingalongthepolishedpalletjewels.Thedropisconsiderable,however,andresultsinalossofpower

whichnodoubtaccounts for the fact thatwatches fittedwitha ratchet-toothedwheelrequireastrongmainspring.

Mostratchet-toothescapementsusethesinglerollerforthesafetyactionasshownintheillustration.

PIN-PALLET—TERMINOLOGYANDCONSTRUCTIONTheescapewheelofthepin-palletvarietyismadeeitherofbrassorsteeland

usually contains 15 teeth. The teeth are so shaped as to provide a lifting orpushingeffect on thepinsR andD (Figure27). Since the teeth performmorethan one function, it is necessary to use additional terms in describing theescapement action. The slanting face onwhich the pallets lock is the lockingface;theliftingplaneistheimpulseface;theintersectionofthelockingfaceandthe impulse face is the toe; and the intersection of the impulse face and theletting-off corner is the heel. The locking face has an additional function ofkeepingthelevertothebankingpins.Theforce,calleddraw,existsbecausethelockingfaceslantstotheleft.Themetalpiecethatcarriesthepinsiscalledthepalletframe.AttachedtothepalletframeistheleverL.Attheextremityofthelever is the fork,which is divided into two parts: the slot and the horns. Theillustrationdoesnot show the fork,but it isusuallyconstructed similar to thatshowninFigure28.

Figure26.Ratchettoothescapement.

Figure27.Pinpalletescapement.

The balance carries a small steel pin, flattened on one side, called theimpulsepin.Through its engagementwith the fork slot, the balance is kept inmotion.

Aportionofthestaffismilledouttopermittheguardfingertopassinandoutduringtheimpulse.Thisfeaturepreventstheleverfromgoingacrossatthewrong time, as the guard fingerwould butt against the circular portion of thestaffshouldajartakeplace.

CLUB-TOOTH—TERMINOLOGYANDCONSTRUCTIONTheescapewheelisofbrassorsteelandusuallyhasfifteenteethwhichact

alternatelyonthepalletsRandD(Figure28).Thepalletsofrubyorsapphirearecementedinslotsinthepalletframe.Impulseisdividedbetweentoothandpalletas shown in the illustration. The pallet jewels are divided into three parts orfunctions:lockingface,impulseface,andletting-offcorner.Thelockingfacesofthe pallets are tilted to an angle of 12 degrees from the linesAB andA’B’ inorder toproduce thenecessarydraw.Thisdiffers fromthepin-pallet, inwhichcasethelockingfacesoftheteethprovidethedraw.AttachedtothepalletframeistheleverL,towhichitattachedtheforkandguardfinger.

Figure28.Leverescapement.

Incontrast to thepin-palletescapement, theclub-toothtypehasa jewelforthe impulse pin usually called roller jewel.A disk called the impulse roller isfittedtothelowerpartofthebalancestaffandthispiececarriestherollerjewel.Alsofittedtothestaffandlyingbelowtheimpulserollerisasmallerdiskcalledthesafetyroller.Acrescentorcircularnotchiscutoutoftheedgeofthesafetyroller to permit the guard finger to pass in and out during the impulse.Mostmodern lever escapements use the above type of safety action and are calleddoublerollerescapements.

ACTIONOFTHELEVERESCAPEMENTThe several variations of the lever escapement described above have the

samemechanical action. Thus the explanation to follow applies equally to allforms.

Figure 28 shows the receiving palletR locked on the tooth of the escapewheel.Thebalanceisrotatinginthedirectionofthearrow,carryingwithittheimpulserollerIR.TherollerjewelRJadvancestowardtheforkslotandmovesthelevertotheright.Thisactioncausesthereceivingpallettorecedefromthetooth until the unlocking takes place. Immediately after unlocking, the escapewheelturnsclockwiseandthetoothT,throughcontactwiththeimpulsefaceofthe receiving pallet, drives the lever to the right, and the fork slot carries theimpulse to the roller jewel. The force is sufficient to turn the balance withconsiderable velocity, unwinding the balance spring (in some caseswinding itup) until the force is exhausted. In themeantime, toothT has dropped off thereceiving pallet and another tooth has locked on the discharging pallet. Theinertiaof the balance having become exhausted, the balance spring takes overandstartsthebalanceonitsreturnvibration.Whentherollerjewelreachestheforkslot, thedischargingpalletunlocksatoothandanewimpulsebegins, thistimedrivingthebalanceinaclockwisedirection.

Beat.Theleverescapementisinbeatwhenthebalancespringissoadjustedas to permit the roller jewel to point toward the pallet center. In practice thebalancemaybeturnedanequaldistanceoneithersideofthelineofcentersandthe watch should start. If it starts on one side and sets on the other, theescapementisoutofbeat.Iftheescapementsetsonbothsides,asispossibleinsomesmallwristwatches,theescapementisconsideredinbeat.

Oilingtheescapement.Oil everyother toothof the escapewheel.This issufficient to spread theoil to thepallets after thewatch is running.The rollerjewelisnotoiled.

Escapementerror.Figure28 shows that the levermoves10degrees frombankingtobanking;5degreesoneachsideofthelineofcenters.Theunlockingaction takes up 1½ degrees, whichmust be subtracted from the 5 degrees ofangular motion to the center line. Thus the impulse which follows fromunlockingtothelineofcentersmakesgoodthelossbyonlyaslightmargin.Therealimpulseoccursslightlybeforethelineofcentersisreachedandconsistsofabout18degreesattheradiusoftherollerjewel.This,accordingtotheanalysisgivenonpage237,resultsinalossfortheshortarcs.

CylinderEscapementThomasTompion (1638-1713) invented the cylinder escapement about the

year 1695. George Graham, famous for the dead-beat clock escapement,

improveditandgaveititsmodernformsometwenty-fiveyearslater.Itappearsthat theEnglishexperiencedconstructionaldifficultieswhich theSwiss solvedbymakingboththecylinderandtheescapewheelofhardenedsteel.Theresultwas that the cylinder became popular in France and Switzerland, where menlearned tomake it cheaply.The escapement, thoughnowobsolete, is stillmetwithinrepairwork.

Figure29.Cylinderescapement.

Construction.Ahollowsteeltube(thecylinderC,Figure29)hasnearlyhalfof its circumferencecutaway topermit the teethof theescapewheel toenter.Thebalanceisattachedtothecylinder,andplugs,onwhichpivotsareformed,are forced friction tight in each endof the cylinder.The rimof the balance isfitted with a small pin which, by contacting another pin under the balancebridge,limitsthelengthofavibrationtoahalfturnonly.Thisisnecessaryinthecylinderescapement,foritcanreadilybeseenthatthebalancecannotturnmorethanahalfturninonedirectionwithoutputtingtheescapementoutofaction.Itis further evident that the escapewheel teethmustbemountedon the endsofverticalpostsprojectingupwardsfromthewheelitself.Thisisnecessarysothatthecylindermayrotatearoundatoothduringavibration.Forthesamereason,thelowerpartofthecylinderiscutawaytoagreaterextenttoprovideforafullvibration.

Action of the escapement. The balance is turning in the direction of thearrow.Thecylinder,onwhichtoothTislocked,turnsaccordingly,permittingatoothtopassinsidethecylinder.TheimpulseisgivenbytheinclinedfaceofthetoothTactinguponthecornerCofthecylinder.Impulsecompleted,thepointof

thetoothdropsontheinnersurfaceofthecylinderandremainsincontactduringtheremainderofthevibration.Uponthereturnvibration,thesametoothisfreedand another impulse takes place, this time on the corner C′ of the cylinder.Again,whentheimpulseiscompleted,anothertoothdropsontheoutsidewallof the cylinder ready to perform a new impulse as soon as the balance hascompleteditsvibration.

Beat.ThecylinderescapementisinbeatwhenthecornersCandClieatanequaldistancetotheescapewheelcenterwhenthebalancespringisatrest.Inthispositionthecylinderisalwaysreceivingimpulseandshouldstartwhenthemainspringiswoundafewturns.If,byturningthebalanceanequaldistanceoneithersideof itspointofrest,wefindthat itsetsononesideandstartsontheoppositeside,wemayassumethattheescapementisoutofbeat.

Oiling theEscapement. Oil the impulse face of every other tooth of theescapewheel.Thisisalltheoilingnecessary,asthecontactoftheteethwiththecylinderwillspreadtheoiltotheremainingteeth.

Escapementerror. The varied tension of themainspring as the timepieceruns down has less effect on the motion of the balance in the cylinderescapementascomparedwiththeleverescapement.Thismaybeexplainedifweconsiderthefrictionofthepointsoftheteethonthecylinderwall.Thisfrictionis considerablewhen themotive power is increased and just about equals theincrease in impulse. In this way any isochronal error of the balance springbecomeslessevident.

Theimpulseispracticallyequalonbothsidesofthelineofcenters,andthis,too,wouldsuggestthatthecylinderescapementshouldmaintainaconstantrate.However,inspiteofthesegoodpoints,thechieffaultofthecylinderliesinthefact that the balance is never free of the escapement and difficulties developwhenthecylinderbecomesworn.

DuplexEscapementThe duplex escapement appears to embody excellent principles of design

but,becauseofconstructionaldifficulties,hasneverachievedpopularity.ItwasinventedbyaFrenchwatchmaker,J.B.Dutertre,in1724.TheEnglishusedtheduplexmore generally than the watchmakers on the continent and some veryfinespecimensweremadewithrubyrollers.Withtheintroductionof theleverescapement,theduplexdeclinedandisnolongermade.

Construction. The escapewheel, Figure 30, consists of two sets of teeth.

Thelongerteetharecalledthelockingteethandtheshorterteeththatliewithinandabovethelockingteetharecalledthe impulseteeth.The locking teethrestontherollerduringthebalancevibrationuntilsuchtimewhentheimpulsetakesplace. Impulse ismade possiblewhen one of the locking teeth drops into thenotch in the roller.When this happens, the impulse teeth engage the impulsepalletIPtoperformtheimpulse,whichtakesplaceinonedirectiononly.Onthereturnvibration,theimpulsepalletmustswingbetweentwoimpulseteeth:Itisherethattheduplexhasaverydelicateactionandmustbeveryaccuratelymade.

Figure30.Duplexescapement.

Action of the escapement. The locking toothL is resting on the notchedrollerRandthebalanceisturninginthedirectionofthearrow.Whenthenotchreachesthelockingtooth,thetoothwilldropintoit.Althoughthefunctionofthelockingtoothistolocktheescapewheel, itdoes,incidentally,performasmallimpulsetothebalancewhenthetoothentersthenotch.Whenthelockingtoothisabouttoleavethenotch,theimpulsepalletIPhasturnedtoapositionreadytoreceive impulse from the impulse tooth. It is required that the impulse palletadvanceabout8degreesaheadoftheimpulsetoothforasafeaction.Whenthelocking tooth drops free of the notch, an impulse tooth engages the impulsepalletandimpulsetakesplace.Aftertheimpulseiscompleted,anotherlockingtoothengages the rollerand thevibrationcontinuesand finallyexhausts itself.Onthereturnvibration,thereisnoimpulse;thenotchmerelypassesthelockingtooth,causingaslightrecoiloftheescapewheel.

Beat.Theduplexescapementisinbeatwhenthenotchliesbetweenthelineofcentersandthelockingtooth.Thetoothis,ofcourse,restingontheroller.Outofbeatisafrequentcauseofstoppage.

Oiling the escapement. The roller is the only part oiled. Thickened oil isdetrimental to the duplex and the notch must be kept clean. To assure this,sharpenapieceofpegwoodandworkitupanddowninthenotch.

Escapementerror.Beingafrictionalrestescapement,thevariablepressureontherollerandthedeteriorationoftheoilcauseavariablerate.Thereisalsoanengaging friction before the impulse and a disengaging friction on the returnvibration.Theengagingfrictionisgreaterandisoftenthecauseofsetting.

ChronometerEscapementPierreLeRoy(1717-1785)created, inprinciple,anescapementuponwhich

themodernchronometerescapementisbased.ItsmodernformwastheworkofThomasEarnshaw(1749-1829).

Theescapement is sodesignedas to leave thebalancemorenearly freeofescapement disturbances than any that has yet been invented. It follows,obviously,thatitisthemostaccurateofanyintheregisteringoftime.

Figure31.Chronometerescapement.

Construction. The escapewheel (Figure 31) is of brass and the arms arereducedtohalf the thicknessof the teethfor lightness.Thelockingfaceof theteethisinclinedabout30degreestofacilitatethedraw.ThediameterofthesteelimpulserollerIRis½thatoftheescapewheelandisplantedbetweentwoteethof the escape wheel. Theoretically the roller intersects a path of 24 degreesmeasured from the escape wheel center. Actually the circular path is less,allowingforthenecessaryclearance.Accordinglythebalancearcisslightlylessthan45degrees.

ThedischargingrollerDRandpalletliebelowtheimpulserollerandengagethegoldspringofthedetent.Thediameterishalfthatoftheimpulseroller.Thesizeisimportantifthedetentistoreturnontimetolocktheoncomingtoothoftheescapewheel.

Thedetentisalengthofhardenedandtemperedsteelwhosepartsareshownintheillustrationandindicatedasfollows:F,thefoot;S,thespring;B,theblade;andH, thehorn.Ahole isbored in theblade to receive the lockingpalletLP.ScrewedtothebladeandprojectingbeyondthehornisthegoldspringGS.

Action of the escapement. The balance is rotating in the direction of thearrow. Likewise the discharging palletDP turns and pushes the detent to theright.ThisactionreleasestheescapewheeltoothfromthelockingpalletLP.Theescapewheelisnowfreetoturninthedirectionofthearrow,butitsfreemotionis retardedwhen the toothA engages the impulsepallet IP.The impulsepallethas already moved ahead of tooth A by 5 degrees, thereby securing a safeoverlapping. The escape wheel tooth now drives the impulse pallet until thedivergence of their paths separate them. The instant tooth A drops off theimpulse pallet IP, tooth B engages the locking pallet. (Return of the lockingpallet tookplacetheinstant thedischargingpalletpassedthegoldspring.)Thebalancewheelcontinuesitscircularmotion,windingupthebalancespringuntilits energy is exhausted.On the return vibration, the discharging pallet pushesaside the gold spring without disturbing the detent. After the inertia of thebalanceisexhausted,thebalancereturnsforanothervibrationpreciselylikethefirst.

Beat. The chronometer escapement is in beat when the discharging palletpoints directly toward the gold spring. In practice the balance spring is soadjustedthatthegoldspringjusttouchesthedischargingpalletonthesidefacingtheescapewheel.Inthispositionthebalancemayberotatedanequaldistanceineitherdirectionandtheescapementwillstartofitself.

Oilingtheescapement.Thechronometerescapementrequiresnooil,sincetheunlockingand impulseactionsaremore in thenatureofadirectpushwithverylittleslidingfriction.Thisisanadvantage,forthedeteriorationoftheoil,asisrequiredinotherescapements,changestherate.

Escapementerror.Boththechronometerandleverescapementsgivemostof the impulseafter the lineofcenterswhichcausetheshortarcs to lose.Thiseffect can be altered in the chronometer escapement by adjusting the relativepositionsoftheunlockingandimpulsepallets.Thischangestheamountofdropandalsothelengthoftheimpulsebeforethelineofcenters.

It will be further observed that the balance spring can be so adjusted thattherewouldbenoescapementerrorbyadjustingthespringsoastoadvancetheimpulse pallet. (This seemingly desirable feature of centering the impulse isneverpracticedforthereasonthattheescapementismoreapttoset.Insteadtheescapementisalwaysputinbeatasalreadyexplained.)

Chronometer escapement for watches. Although the chronometerescapementhasbeenusedinpocketwatches,itisnotrecommended.Suddenjarsareapt tocause theescapement toset.Theescapement ismoreexpensiveandmuchmoredelicate than the leverandeven in thechronometer,amorerobustinstrumentthanawatch,theescapementmustbehandledwithcare.

Onepointinparticularshouldbenoted.Thebalanceshouldnotberemovedwhen the mainspring is under tension. Either let the mainspring downcompletely or lock the train by wedging a piece of cork between the fourthwheelandtheplate.Thereasonforthisprecautionisthatthelockingpalletisapttobebrokenoffshouldthetrainstartwithoutthebalanceinthemovement.

PARTIIChapterSix

STRIKINGMECHANISMSTHESTRIKINGmechanismof a clock consists of a separate train andpower

unit.It isnot, inanyway,connectedwiththetimekeepingmechanismormaintrain, except that themain train sets the striking train inmotion at the propertime.

There are two types of strikingmechanisms in general use. These are thecounting-wheelandrack-and-snail.Thecounting-wheelistheolder;infact,itisasoldasthemechanicalclockitself.Itisstillusedinpopular-pricedAmericanmantel clocks and in tower clocks. The rack-and-snail type was invented byEdwardBarlowin1676andisfoundinmantelclocksofthebetterqualityandinallmodernhallclocks.

Counting-WheelMechanismTheonlydisadvantageofthecounting-wheeltypeisthecautionrequiredin

settingthehandstotime.Inadvancingtheminutehand,itisnecessarytoallowthestrikingtofinishtothehourindicatedbeforeadvancingthehandstothenexthalfhour.Inotherwords,thecounting-wheelsystemstrikesconsecutivelyandisnotself-correcting.

AMERICANMANTELCLOCKWITHCOUNTING-WHEELMECHANISMConstruction. Figure 32 shows the arrangement usually found in the

American mantel clock. The cam on the center arbor is shown in a positionreadytoraisetheliftingleverL.ThisleverisattachedtothearborA,which,inturn,carriesthewarningleverWandtheunlockingleverU.Theunlockingleverlies under and pushes the counting leverCL which is secured to the arborB.MountedonarborBaretwolevers:thecountinglever,mentionedabove,andthedropleverD.Whenthestrikingmechanismislocked,thedropleverDrestsinthe notchof the camwheelC.Also the counting leverCL rests in one of thedeeperslotsinthecountingwheelCW.

Figure32.Counting-wheelmechanismasappliedtoAmericanmantelclocks.

Action of counting-wheel mechanism. As the cam on the center arboradvances clockwise and raises the lifting leverL, four levers are brought intoaction.WarningleverWisraisedinpositiontoengagethepinPonthewarningwheel.TheunlockingleverUliftsthedropleverDandalsothecountingleverCL, since both levers are attached to arborB. This results in a simultaneousseparationofthesetwoleversfromtheirseats,theformerfromthenotchinthecamwheelandthelatterfromthedeeperslotinthecountingwheel.Thetrainisnowfreetotuibutonlyforamoment,asapinPonthewarningwhencatchesonthealreadyraisedwarningleverW.Precise!onthehour,thecampassesthelifting lever L, permitting the lifting lever to fall. At the same moment thewarningleverWalsodrops.Thisactionreleases thepinon thewarningwheelandthestrikingbegins.Asthestrikingcontinues,thecountingwheelturns,andthe counting lever drops into and raises above the shallow slots. Finally thecountingleverdropsintooneofthedeeperslotsandthedroplever,atthesamemoment,dropintothenotchonthecamwheel.Whenthishappensthestrikingstops.

Figure33.Counting-wheelstrikingmechanismasappliedtohallclocks.

HALLCLOCKWITHCOUNTING-WHEELMECHANISMThecounting-wheelsystemasusedinhallclocksdiffersindesignfromthat

found in the American clock, but the mechanical principle is practically thesame.

Construction.ReferringtoFigure33itwillbenotedthatthepinWPonthewarningwheelperformstwofunctions—(1)receivingthewarningleverpriortothe striking, and (2) receiving the drop lever to stop the striking. This differsfrom the countingwheelmechanism previously consideredwhere the strikingtrainwasstoppedwhenthedropleverdroppedandlockedintheslotinthecam.

Actionofthemechanism.ThestrikingtrainisstoppedbythepinWPonthewarningwheelengagingthelockingpointL.ThisismadepossiblebythenotchonthecamewheelbeinginpositiontoreceivethedroppingpointDandtheslotinthecountingwheelreceivingthecountingpointC.

ThedialtrainissodesignedthattheminutewheelMWturnsonceanhourinorderthattheliftingpinLP,whichsetsoffthestrikingtrain,maybeattachedtoit.The liftingpin advances in thedirectionof the arrow, and raises the liftingleverLLandlikewisethewarningleverWL.ThepinP,attachedtothewarninglever, lifts the drop lever and releases the pinWP on thewarningwheel. The

warningwheelisnowfreetoturn,makingnearlyonerevolution,whenthepinWPengagesthewarninglever.

Onthehour,theliftingpinontheminutewheelpassestheliftingleverLL,permitting it to drop. Thewarning lever drops accordingly but the drop leverdoes not, since the cam wheel has turned a sufficient distance to prevent it.Whenthecamwheelhasmadeoneturn, thecountingwheel, too,has turnedasufficientdistancetostillkeepthedropleverfromdropping.Striking,therefore,continuesuntilthecountingpointdropsintoaslotonthecountingwheel.

Rack-and-SnailMechanismThe rack-and-snail systemdiffers from the counting-wheel type in that the

hour wheel regulates the number of blows struck. This being the case, thestrikingisalwayscorrectedtothehourindicated.

HALLCLOCKRACK-AND-SNAILMECHANISMConstruction.Figure34showstherack-and-snailmechanismasappliedto

hall clocks.ThediskS, called the snail, is attached to thehourwheel andhastwelvestepswhichgovernthedepthtowhichtherackRdrops.Thetopportionoftherackhastwelveormoreteethandthenumberofteethgatheredupbythegathering pallet G is determined by the position of the snail. Suppose, forexample,theracktailTdropstotheloweststeponthesnail.Twelveteethontherackwillbeliberatedandtheclockwillstriketwelve.

Figure34.Rackstrikingmechanismasappliedtohallclocks.

Actionof therack-and-snailmechanism.ThepinPon theminutewheeladvancesinthedirectionofthearrowandraisestheliftingpieceLP.EventuallywarningleverWLliftsrackhookRH,freeingtherackwhichfallstosomestepon the snail.At the same time the stoppinSP (which is attached to the rack)recedes from the tail of the gathering pallet, permitting the train to start. Thetrain is soon stopped, however, because the pin WP on the warning wheelengagesastoppieceonthebackofthewarningleverWL.Later,onthehour,thepinPontheminutewheelpassestheendoftheliftingpieceLP,permittingthewarningleverWLtodropandfreethewarningwheel.ThestrikingtrainisnowsetinmotionandthegatheringpalletGproceedstocollecttheteethontherack.After the last tooth is collected, the stop pin SP advances in the path of thegatheringpallettailandthestrikingtrainstops.

MANTELCLOCKRACK-AND-SNAILMECHANISMConstruction. The rack system employed inmantel clocks differs slightly

from that generally used in floor clocks. The locking of the striking train inweight-drivenclocksoccursbymeansofatailonthegatheringpalletengagingthe stop pin on the left portion of the rack. This plan is not used in thosemovementsdrivenbyamainspring,asthefollowingdescriptionwillshow.

Actionofmantelclockracks.Figure35showstheFrenchclockmovementastypicalofthosefoundinmostmantelclocks.Apinonthehourwheelraisesthe lifting leverL, carryingwith it thewarning leverW.The lifting lever alsoraisestherackhookRHbymeansofthepinP,whichinturnreleasestherackR.Therackdropstothesnail,andatthesameinstantthestopleverSL(fittedtotherackhookarbor)clearsthestoppinSPonthelockingwheel.Thestrikingtrainturnsbut is soonstoppedwhen thepinWP on thewarningwheel engages thestoppieceonthewarningleverW.Eventually,onthehour,thepinontheminutewheelpassestheliftingleverLandthewarningleverdrops.Thewarningwheelisnowfreedandthestrikingbegins.ThegatheringpalletGcollectstheteethonthe rackuntil finally the rackhookdropsoff the last tooth.Thestop leverSL,accordingly,dropsandengagesthepinSPonthelockingwheelandthestrikingceases.

Figure35.RackstrikingmechanismasappliedtoFrenchclocks.

Half-hourstriking.Theusualmethodemployed inhalf-hourstriking is tomakethefirsttoothontherackshorterthantherestandtofitasecondpinontheminutewheel.Thesecondpinliesalittlenearertothecenterthanthehourpin.In thisway, the rack hook is lifted to clear the first tooth only, and the clockstrikesoneonthehalfhour.

SHIPBELLSTRIKETheshipbellstrikingworkusuallyemploystherack-and-snailsystem.The

mechanism is similar to anyhour andhalf-hour strikingclock.Themannerofstriking differs, however, as the plan adopted is for the purpose of calling thecrewtoregularworkshiftscalled“watches.”

At12:30p.m., theclock strikesone, called“onebell.”At1p.m. it strikestwo,or “twobells”; at1:30p.m., “threebells,”etc.Eachhalf-hour a “bell” isaddeduntil“eightbells”arestruckat4p.m.Theplancontinuesthroughoutthe24 hours and obviously the “eight bells” are struck at 8 p.m.; 12midnight; 4a.m.;8a.m.and12noon.After“one”bellthestrikingoccursinpairsasmaybenotedbythefollowingexample:

ChimesandChimingMechanism

Sets of bells, struck by hammers and playing fascinating tunes have beenpopularforcenturiesinclockmaking.Thereareanumberofbellcombinations,commonlycalledchimes.Mostoftheseplayatquarter-hourperiods.Someplayonechime,otherstwoorthree.Inthelattercase,thedesiredchimeisplayedbysettingahandontheclockdial.

Figure36.Westminsterchime.

Figure37.Whittingtonchime.

Figure38.Guildfordchime.

ThemorepopularchimesincludetheWestminster,Whittington,andGilford.TheseareshowninFigures36,37and38.

CHIMINGMECHANISMConstruction. Chiming clocks use three trains: (1) hour striking train; (2)

timing train;and (3)chiming train.The timing trainoccupies thecenterof themovement,thehourstrikingtrainisplantedtotheleft,andthechimingtraintothe right.The rack-and-snail system is theoneusuallyused.Occasionallyonefindsthecountingwheelusedinthechimingtrainandtherack-and-snailinthehour striking train. The mechanical principles vary somewhat in the differentmakesbutthemajorityfollowtheplanshowninFigure39.

ThechimesnailCSismountedontheminutewheelandcarriesfourpinsforraisingthechimeliftingleverCL.Alsoattachedtothechimesnail isapinforraisingthestrikeliftingleverSL.ThestrikewarningleverSWreleasesthestrikewarningwheelS when the chime gathering palletG has gathered up the lasttoothonthechimerack.

Actionof thechimingandstrikingmechanism.ThepinP, on thechimesnail,advancinginthedirectionofthearrow,raisesthechimeliftingleverCL,andthechimewarningleverCWraiseswithit.ThechimerackhookCHisalsoliftedbymeansof thedrop leverD, therebyreleasing thechimerackCR.Thetail T of the chime rack engages the largest radius of the chime snail CS,permittingthechimeracktodropadistanceofonetooth.Whentherackdrops,the tail of the chimegathering palletG is released from the rack pinCP.Thetrainstartstoturn,butissoonstoppedwhenapinonthechimewarningwheelCengagesthealreadyraisedchimewarningleverCW.Onthehour,pinPon thechimesnailpassestheendoftheliftingleverCL,permittingthewarningleverCWandthedropleverDtodrop.ThewarningwheelC isfreed;therackhookCH drops to the rack and the first quarter chime is played. The chime ceaseswhen the tail of the chime gathering palletG engages the rack pinCP. Thesecondandthirdquartersandhourchimesfunctioninlikemanner.

Figure39.ChimingandstrikingmechanismasappliedtoGrandfatherclocks.

Priortothereleaseofthechimerackatthehourchime,apinonthechimesnailhas raised thestrike lifting leverSL.Thestrike rackhookSH is likewiseraisedandstrike rackSR drops to the strike snailSS.The striking train starts,onlytobestoppedwhenthepinonthestrikewarningwheelSengagesthestoppieceonthestrikewarningleverSW.Thusitisseenthatboththechimewarningand the strikewarning takeplacebefore thehour chimeplays.A fewminutesbeforethehour,thestrikeliftingleverSLdropsoffapinonthechimesnailandthestrikerackhookSHdropstothestrikerackSR,butnofurtheraction takesplaceuntilthepinPonthechimesnailpassestheendofthechimeliftingleverCL.Whenthishappens,thechimewarningleverCW,thedropleverD,andthechimerackhookCHdropsimultaneously.ThechimewarningwheelCisfreedand the full hour chime is played.As the lastmovement of the hour chime isplaying,apinonthechimerackpullsupthestrikewarningleverSW,releasingthestrikewarningwheelS.Thusthestrikingofthehourimmediatelyfollowsthehourchime.

It will be observed that the strike warning lever is moved near thecompletionofeachquarterchime,butnoactiontakesplaceuntilsuchtimethatthestrikerackhasdropped.

PARTIIChapterSeven

AUTOMATICWATCHESAUTOMATICorself-windingwatchesarethenewestinnovationinthefieldof

timekeepingmechanisms.Theyneverneedwindingsolongastheownersgivethemnormaldailywear.

Thewindingmechanismsofautomaticwatchesarefoundinthreetypes:(1)One-direction winding with rotor motion limited by bumper springs, (2) one-direction winding with rotor motion unlimited, and (3) two-direction windingwithunlimitedrotormotion.

Theolderautomaticsusebumpersprings,andthemainspringiswoundwhentherotorturnsinonedirectiononly.Mostofthenewermodelsareconstructedinsuchamannerastowindwhentherotorturnsineitherdirection.

AutomaticWatchwithBumperSpringsFigure40showstheusualconstructionofthebumper-spring-typeautomatic.

Designsvary,butthemechanicalprinciplesarethesameinallmakes.Fittedtothecentralportionoftherotoris thewinding-upclick,WC.Thisclickengagesthe winding-up wheel, WU, and turns it only when the rotor turnscounterclockwise.ThewheelWUgears into the transmissionwheel,T, and itspinion drives the ratchet over wheel, RO. The stop click, SC, engages thetransmissionwheelandpreventsthetrainfromturningbackwardwhentherotorswingsclockwise.Ontheundersideoftheratchetoverwheelareratchetteethwhich engage the ratchet teeth of the Breguet winding pinion. The Breguetwindingpiniongearsintotheratchetwheel,RW.Thecombinedgearratiofromthewinding-upwheel to theratchetwheel issufficientlygeareddownsoas toallowtherotortodropwiththeforceofgravitywhenthepositionofthewatchisaltered.

Figure40.

When the watch is wound manually, the winding pinion turns the crownwheelwhich,inturn,turnstheratchetwheel,andthemainspringiswound.This,it is seen, is identical to any stem-windwatch.However, theBreguetwindingpinion is turned, since it gears with the ratchet wheel. No action beyond thistakesplace,sincetheratchetteethonthepinionmerelyslipbytheratchetteethontheundersideoftheratchetoverwheel.

In automatic watches the mainspring barrel contains a short, stiff spring,slightly longer than the circumference of the barrel. This spring fits securelyagainstthebarrelwallandissuppliedwithahooktoreceivethetongueendofthemainspring.Whenthemainspringbecomesfullywound,furtherwindingwillcausethesafetydevicetoslip,thuspreventingoverwinding.

Figure41.

BumperSpringAutomaticwithWinding-upLeverAnother self-winding mechanism somewhat different from the one just

described is shown inFigure41.TheOmega,VacheronandConstantin, andafew other makes use this type of construction. A toothed rack is cut at theextremityof therotor.Thisrackgearswiththewinding-uplever,WL,which isturnedtoandfroastherotoroscillates.Thewinding-upclick,WC,isattachedtothe leverWL and turns thewinding-up ratchet, WR, in a clockwise directiononly. The stop click, SC, prevents the winding-up ratchet from turningcounterclockwise.Apiniononthelowerendofthewinding-upratchetturnsthecrown wheel, C, which, in turn, drives the ratchet wheel, RW. Thus themainspringiswound.

This is possibly the simplest, and yet efficient, automatic that has beenconceiveduptonow.

Two-DirectionWindingAutomaticwithReversingLeverThe two-directionwindingmechanismsmake use of some very ingenious

devices that we shall now examine. Consider first the movement shown inFigure42.Supposetherotorturnsinaclockwisedirection.TherotorgearRG,

fixed to the rotor,engages thereversingpinion,R.Becauseof thedirectionofrotationoftherotorgear,reversingpinionRisthrustagainstthewindingwheel,W1,andturnsitinaclockwisedirection.Followingthetrainthroughasindicatedby thedot anddash center line from rotor to ratchetwheel, it is seen that theratchetwheelturnsclockwiseandthemainspringiswound.

Figure42.

Figure43.

Figure44.

Suppose, now, that the rotor gear turns counterclockwise. In this case thereversingpinionwillbethrustagainstthewindingwheelW2asshowninFigure43.Sinceanextrawheelhasbeenaddedtothetrain,thedirectionofrotationofthewheelsW1,TandRWremainsthesameandthemainspringcontinuestobewound.

AutomaticWatchwithAlternatingDiscsAnothertypeofdoublewindingautomaticmechanismisshowninFigure44.

In thismodeladevicecalledthealternatingdisc,AD, isused.On thediscaremountedtwogears,thefirstandsecondalternatinggears.Iftherotoristurnedclockwise,therotorgear,RG,drivesthefirstalternatinggearAG1directlyintothepathofthewinding-upwheelW,turningthelatterclockwise.

Figure45.

If therotor turnscounterclockwise, therotorgearremainsmeshedwith thefirst alternatinggear, as before.However,with the direction reversed, the firstalternating gear is thrust to the left while the second alternating gear AG2 ismoved to the right, resulting in the second alternating gear engaging thewinding-up wheel W. Since another gear has been inserted in the train thedirectionofrotationofthewinding-upwheelremainsthesame.ThisisseenbycomparingFigures44and45.

Inthistypeofmechanism,nodisengagingclickorratchetisrequiredwhenthewatchismanuallywound.Thefirstandsecondalternatinggearsaremerelypushedoutofaction.

AutomaticWatchwithTwoReversingClicksFigure46showsthewindingmechanismoftheTudorautomatic,aproduct

of theRolexCompany. It is designedon a verydifferent principle from thoseheretoforeconsidered.Therotorgear,RG,engages the firstwinding-upwheel,W1.Thefirstwinding-upwheelturnsthesecondwinding-upwheel,W2,andeachwheelW1andW2carriesathree-prongedclick.Theclicksarefittedtotheuppersidesofthewinding-upwheels.Fittedaboveandengagingtheextremitiesofthethree-prongedclicksarethenotcheddiscsN1andN2.Thesediscscarrypinions

thatdrivethetransmissionwheelT.Theclicksaresodesignedthatthediscsareturned onlywhen thewinding-upwheels turn counterclockwise, and since allwheelsmesh it isobvious thatoneor theotherof thewinding-upwheels turnclockwisewithoutdoinganywork—theclicksmerelyslipbythenotchesinthenotcheddiscs.

Figure46.

Figure47.

This is made clear by comparing Figures 46 and 47. If the rotor gear isturningclockwise,followthecenterlineasshowninFigure46.If,ontheotherhand,therotorgearisturningcounterclockwise,followthecenterlineasshowninFigure47.

AutomaticWatchwithOneReversingClickThisautomatic,Figure48,issimilarinprincipletotheonedescribedabove

but radicallydifferent indesign.Thereversingunitconsistsofadisc towhichthe reversingclick ismounted.Astaff is fitted through thediscand ispivotedbetweentheupperplateandathinbridge.Therearetwowinding-upwheels,W1

andW2, fitted to the staff,which turn freely.The first and smallerwinding-upwheel lies above the disc and is geared directly with the rotor gear RG. Bymeans of the click on the disc, this samewheelW1 drives the disc clockwiseonly..

Thesecondwinding-upwheelW2isfittedtothelowersideofdisc,isturnedby the reversing wheel assembly, R, and does not engage the rotor gear. Thereversing wheel assembly consists of two wheels turning as one, the upper

wheel,R1,beingslightlysmaller than the lowerwheelR2.Therotorgear turnstheupperreversingwheelR1andthelowerreversingwheelR2 turnsthe lowerwinding-upwheel,W2,which,inturn,turnsthediscthroughthefunctioningoftheclick.Thuseitherwinding-upwheelturnsthediscbutonlywhenthemotionisclockwise.

Secured to the staff that carries the disc and click is the pinion, P, whichdrives the transmission wheel, T. Through its pinion the winding function iscarriedthroughthetraintotheratchetwheelandthusthewatchiswound.Theprincipleofoperationbecomesclearbynotingthecenterlineintheillustration.When therotor turnscounterclockwiseas indicatedby thearrowA, followthecenter line A. When the rotor turns clockwise, as indicated by the arrow B,followthecenterlineB.ThearrowsB1andB2 indicatethedirectionofmotionofthewheelsR2andW2.TheheavyarrowA1indicatesthedirectionofmotionofthewinding-upwheelW1.

Figure48.

AutomaticWatchwithHeartPieceSomeofthenewertypesofautomaticmechanismsemployanheartpieceor

cam.Theonedescribedherewith,Figure49, is theproductof theInternationalWatchCompanyofSchaffhausen,Switzerland.TheheartpieceHisfittedtothe

rotor and as the assembly swings to and fro the rocking lever R is caused toswinginlikemanner.Itisinterestingtonotethatthismovementismuchlikethelever escapement where a roller jewel enters a fork slot andmoves the leverawayfromitsbanking.Twolevers,C1andC2,calledautomaticclicksarefittedto the rocking lever and are placed in such manner as to alternately turn thewinding-up wheelW. This function is at once evident when we note that therocking lever is pivoted at the point P. When the rocking lever turns in thedirectionofthearrowA,thehookoftheautomaticclickC1turnsthewinding-upwheel W clockwise. The hook on the automatic click C* idles alongcounterclockwisewithoutdoinganywork.Whenthemotionoftherockingleveris reversed, that is, turning in the direction of the arrow B, the hook of theautomatic clickC2 turns the winding-upwheel clockwise as before while thehook of the clickC2 idles along counterclockwise. Both automatic clicks areheld in position against thewinding-upwheel bymeans of theclick springS.Thus it isseen that themainspring iswoundwhen therotor is turned ineitherdirection.

Figure49.

Aunique feature of the International automatic is the ability to release thepowerofthemainspringwithouthavingtotakeofftheautomaticassembly.Andfurther, a new mainspring may be fitted without disturbing any of the partsexcepttheratchetwheelandthebarrelbridge.ThisisaccomplishedbypressingbacktheautomaticclickC1untilitentersthenotchNoftheclickC2,asshowninFigure50.Bydoing this themainspringpower is released and thebarrel istakenoutintheusualway.Afterreplacingthebarrelandmovingtherotor,thetwoclicksautomaticallyengagetheteethofthewinding-upwheel.

Figure50.

LeCoultreAutomaticWatchThiswatch, Figure 51, has nomanual windingmechanism or stem in the

usualmeaningoftheterm.Thehandsaresettotimebymeansofabuttononthebackofthecase.

Anup-and-down indicator isprovided, somethingquitenew inautomatics,andislocatedbetweenthedialcenterandnineo’clock.

Themainspringdoesnotneedaslipspringas therotor is lockedwhenthemainspringisnearlyfullywound.

Theselfwindingactionissimilartothemechanismdescribedonpage189(automaticwithalternatingdisc)andneednotbeconsideredhere.Thefunctionsthatcallforspecialanalysisare:(1)therotorlockingdeviceand(2)thewindingindicatormechanism.

Rotorlockingdevice.ReferringtoFigure53,itwillbenotedthatthebarrelarborisgivenalefthandedthread.Onthisthreadedportionisfittedadiscwhichis prevented from turning of itself by being pierced by two posts which aresecured to thebarrel cap.Thus thedisc ismovedup as thewatch runsdown,since thebarrel is turningand thedisc is turningwith thebarrel.On theotherhand, thediscmovesdownwhen thewatch isautomaticallywindingsince thebarrelarboristurninganditsthreadedportionisbeingunscrewedfromthedisc.The locking leverL has a conical head and, by reason of its shape, is pushedasideasthediscmovesdownwardwhichobviouslytakesplacewhenthewatchisautomaticallywinding.

Figure51.

Figure52.

Referring now to Figure 51, the same locking lever is seen but from adifferent point of view, that is, horizontal and the direction of motion duringwindingiscounterclockwise.This isshownin the illustrationbyfollowingthearrows. Note also that the locking lever is pivoted at the shoulder screw SS.Assumingthatthewatchisautomaticallywinding,thepinPonthelockingleverwilleventuallyengagethehookHandtherotorwillbelocked.Sincetherotorislockedthewindingwillhaveceasedforthetimebeingandthemovementwillrundown.Thelockinglever,nowturningclockwise,willeventuallyreleasetherotor.Thiswill permit thewinding to again takeplace, that is, if thewatch isbeingworn.

Winding indicator mechanism. The function of the indicator rack R isclearly shown in Figure 51. The adjusting screw AS on the locking lever Lpushesontheindicatorrackwhenthewindingtakesplace.Incidentally,theteethontherackturntheindicatorpinionIclockwiseandeventuallystopswhentherotorislocked.ComparethepositionofthelockingleverandtheindicatorrackinFigures51and52.

Figure53.

PierceAutomaticWatchAnother rather unique type of self-winding watch is the Pierce automatic

showninFigure54.Herethewinding-upweightslidesforwardandbackwardinsteel rods.Themainspring iswoundonlywhen theweight slides to the right.Theprincipleofoperationisshownintheillustrationandadetailedanalysisisnotrequired.

Figure54.

PARTIIChapterEight

STOPWATCHESANDCHRONOGRAPHSTHE PRECEDING pages have dealt with watches that register time as a

continuousphenomenon,thatis,recordingthemeanofsuntime.Therearealsotimepiecessodesignedastoindicateperiodsoftime,theirscopebeinglimitedtodetermining thedurationofsomeparticularevent,ahorseracefor instance.This involves the starting of the timepiece at the beginning of the event orphenomenon tobe timed,and thestoppingof the timepiecewhen theeventorphenomenonisterminated.Theintervaloftimeisthusdetermined.

TheStopWatchThe simplestmechanism for thispurpose is called the stopwatch. It has a

sweep secondhand encircling a largedial and aminute hand fitted to a smalldial,whichmaybestarted,stopped,andreturnedtozerobymerelypressingthecrownorbuttononthesideofthecase.

Themajorityofstopwatchesaresodesignedthatonerevolutionofthehandonthe largedial indicatesa lapseofoneminute,andthespacesaredividedtoshow fifths of seconds.The small dial shows theminutes and a total of thirtyminutestoarevolutionofthehandistheplanusuallyadopted.

Whengreaterprecisionin timingisdesired,other typesof timersaremadefor the specialized needs of science and industry. Special-purpose timers aredesignedtoshow1/10,1/20,1/30,1/50,or1/100ofasecond.Theydifferonlyinthedesignofthedial,thetrain,andthenumberofvibrationsofthebalance.

Figure55.

ANALYSISOFTHEMOVEMENTConstruction. Stop watches have the recording mechanism located under

the dial. Figure 55 shows a 19-ligne model by Ed. Heuer & Company. Thesecondsrecordingwheelisplantedinthecenterofthemovementandhasalongpivotextendingthroughaholeinthedial.Fittedtothelongpivotisthesecondsrecordingrunner,SR,akindofcannonpinion,consistingofaheart-shapedcamandapipe towhich thesweepsecondhand isattached.Theseconds runner isfrictionedtoagrooveonthepivotofthesecondswheelbymeansofaU-shapedspring,oneendofwhichpressesintothegroovethroughaslotinthepipe.Theotherendrestsontheoutsideofthepipe.Thetensionofthespringissufficientto cause the seconds runner to turn in unisonwith the secondswheel, but notenoughtowithstandtheforceofthehammer,H,whenitimpingestheheartcam.

Situatedabovethesecondswheelistheminutesrecordingwheel.It,too,hasa long pivot extending through a hole in the dial, and carries the minutesrecording runner, MR, the construction of which is identical to the secondsrunnerdescribedabove.

TheminutesrunnerandthesecondsrunnerarereturnedtozerobymeansofthehammerHandtheheartcams.Themotionofthehammerisattainedthroughthecircularmotionofthepillarwheel,P.

Themovementrunsonlyatthetimewhentherecordingtakesplace.Thisis

realizedbymeansofadelicatefrictionspring,F,which,inflippingawayfromthebalance rim,givesmotion to thebalance,or,conversely, indrivingagainstthebalancerim,stopsitsmotion.Thefrictionspringissecuredtothestart-stoplever,S,whichisalsoactuatedbythepillarwheel.

All of the above functions, it is observed, are actuated through the pillarwheel,whichisturnedbytheoperatinglever,O,throughitsconnectionwiththestemandcrown.Notealsothattheoperatinglever, thestop-startlever,andthehammerareundertensionthroughcontactwiththespringsOS,SS,andHSsoastoperformtheirrespectivefunctionsproperly.

ThespringJwhichrestsagainsttheratchetofthepillarwheeliscalledthejumper. Its function is to assist themovement of thepillarwheel to its properpositionofrest.Obviouslythespringmustbequitestrongsothatthebeakofthejumpercomestorestfirmlybetweentworatchetteeth.

ACTIONOFTHEMOVEMENTStarting function. The mechanism is at rest and hands are in the zero

position.Thecrown ispressedand theoperating leverO ismoveddownward,pushinga ratchet toothof thepillarwheeladistancecomparable to the lengthbetweentwoteeth.Asthemovementtakesplace,pillar1pushesthebeakBofthehammerHoutward, finallycoming to reston theouter surfaceofpillar1.Theheartcamsarethusreleased.

Simultaneouslywiththeabovemovement,thebeakBSofthestart-stopleverS drops off pillar 3 into the next space, thereby permitting the lever and itsfrictionspringFtomovecounterclockwise,flippingthebalanceintomotionasitgoes.Thustherecordinghandsarestarted.

Stopping function. Pressing the crown a second time the pillar wheel isturnedandbeakBSofthestart-stoplevermountspillar4.Asaresultthestart-stop lever advancesclockwise in thepathof thebalanceand themovement isstopped.During thisperiod thebeakBof thehammer remains incontactwiththeoutersurfaceofpillar1.

Handstozerofunction.PressingthecrownathirdtimethebeakBof thehammerdropsoffpillar1andtheflatfacesF1andF2 impingetheheartcams.Thustherecordinghandsarereturnedto thezeroposition.Thestart-stopleverremainsmountedonpillar4untilsuchtimeastheaboveactionisrepeated.

TheChronograph

Thechronographmaybedefinedasatimepieceinwhichthefunctionsoftheconventionalwatchandthestopwatcharecombined.Chronographsarefoundinseveral varieties. Some are so designed that the recording hands are started,stopped,andreturnedtozerobymeansofapushpieceinthecrownorasinglebuttononthesideofthecase.Thistypeiscalledthesingle-buttonchronographand themechanism is quite similar to the stopwatch. Another type equippedwith twobuttons iscalled thedoublechronograph. Inmostmodels successivepresses of one button start, stop, and again start the recording hands withoutreturningthemtozero.Thesecondbuttonisusedonlyforreturningthehandstozero. This type of chronograph is rapidly superseding the single button typebecause of its greater usefulness.A third type of chronograph, called the splitsecondtimer,isusefulinenablingtheoperatortomeasuretwocloseintervalsoftime, as for instance, the first and second winners of an auto This is madepossiblebyhavingtwosweephands,oneplanteddirectlyabovetheother.Thetwohandsmaybestarted,stopped,andreturnedtozerothesameasinanystopwatch.Howeveranextrabuttononthesideofthecasewill,whenpressed,stoponeofthehands,theotherhandcontinuestomoveasheretofore.Asecondpressofthebuttoncausesthestoppedhandtoovertakethehandinmotionandthetwohandsagainturntogether.

DoubleChronographwithPusher-typeHammerConstruction. Let us examine the mechanism of the double chronograph

with pusher-type hammer, the type usually found in wrist watches. Themechanicalprinciplesof thedifferentmakeswillbe found tovaryslightlybutthemechanismshowninFigure51maybeconsideredastypicalofthemajority.

The chronograph differs from the stop watch in that the movement runscontinuously.Becauseofthisadifferentdesignisrequiredandwefind,inplaceofthestart-stoplever,alevercalledthesecondscouplingclutch.Thislever,CC,Figure56,startsandstops therecordinghandsinamannerwhichissimilar tothestart-stopleverofthestopwatch.Thefourthwheelcarriesthedrivingwheel,D, and the seconds coupling clutch carries the intermediate wheel J. Thesewheels remain in gear and run continuously.When the pillar wheel is turnedthrough itsconnectionwith theoperating leverandbuttonnumber1, the beakBCofthecouplingclutchdropsoffapillarofthepillarwheeland,insodoing,theintermediatewheelIengagesthechronographrunnerC.Thustherecordinghandsarestarted.

Figure56.

The chronograph runner C and the minutes runner M are of similarconstruction.Oneachof therunners there isaheartcamanda longpivot thatextendsthroughthemovement.Toeachpivotahandisfittedfortherecordingofsecondsandminutesrespectively.

Thebrakelever,B,isactuatedintwoways:(1)bythepillarwheelPand(2)by thehammerH.Thebrake lever is necessary tokeep the chronographhandfrommovingatsuchtimeswhenthehandissupposedtobeatrest.

Buttonnumber2cannotbepressedwhenthehandsaremoving,asthebeakSBofthehammerwouldbuttagainstoneofthepillarsofthepillarwheel.

The fingerF on the chronograph runner pulls the intermediate wheelMIaroundonetoothatatime.ThewheelMIgearswiththeminutesrunnerMand,throughtheminutesrunner,minutesarerecordedonthedial.Howeverwhenthehammer returns the recording runners tozero, the intermediatewheelMImustbepulledasidefreeofthefingerF.ThisisaccomplishedbymeansofthebeakBBonthehammer.

ACTIONOFTHEDOUBLECHRONOGRAPHStarting function. The recording hands are at zero and the endE of the

brakeleverBrestsagainsttherimofthechronographwheel.Buttonnumber1ispressedand,throughitsconnectionwiththeoperatingleverOandthepawlOP,

pillarwheelPisturnedtherequireddistance.Thebrakeleverisraisedfromthechronograph wheel as a result of the beak LB mounting a pillar of the pillarwheel. Simultaneously the beakBC of the seconds coupling clutchCC dropsbetween two pillars of the pillar wheel, permitting the engagement of theintermediatewheelIwiththechronographrunnerC.Thustherecordingbegins.

Stoppingfunction.Pressingbuttonnumber1asecondtimecausesthebeakofthecouplingclutchtomountapillarofthepillarwheelandthehandsstop.AtthesamemomentthebeakLBofthebrakeleverBdropsbetweentwopillarsofthe pillar wheel permitting the opposite end E to rest against the rim of thechronographwheel.

Handstozerofunction.Thehandsarereturnedtozerobypressingbuttonnumber2whichactuatestheflybackleverFB.TheflybackleverinturnpressesthehammerH.Asthehammermovestowardtheheartcams,beaksBBandMBarebroughtintoaction.BeakBBpushesasidethebrakeleverthroughitscontactwith the upright pin U thereby releasing the chronograph runner. BeakMBwedgesintotheslidinggearframeSGandtheminutes-recordingmechanismiscutoutofaction.LastlytheflatfacesF1andF2impingetheheartcamsandtherecording-hands are returned to zero.When button number 2 is released, thehammerisreturnedtoitspositionofrestbythespringS.

It should be observed that the return to zero of theminutes runner is notattendedentirelybythehammerbutratherbytheactionofthejumperMJ.

Some chronographs have an hour recording wheel and dial. This addedfunction is operated by a wheel attached to the barrel, and the completemechanismisseparateandliesunderthedial.

DoubleChronographwithReleaser-typeHammerConstruction. The mechanical action of this chronograph, Figure 57, is

identical in principle to the one described above except that the hammer isreleased insteadofbeingpushed.The function isattained through thehammerbolt,HB.When button number2 is pressed for the returning of the recordinghandstozero,theflybackleverFBisactuatedanditsbeakpressesthehammerbolt,therebyreleasingthepinP.ThehammerspringSforcesthehammertotheheart cams and the recording hands are returned to zero. In this type ofchronograph,thehammerremainsagainsttheheartcamsuntilbuttonnumber1ispressedtostarttherecordinghands.

Figure57.

DoubleChronographwithReversingCamsConstruction. A new model that substitutes cams for the pillar wheel is

showninFigure58.Twocamsare riveted togetherand turnasonepiece.Themovement is first clockwise and then counterclockwise, and the oscillatingmotionisattainedbymeansofajointedoperatinglever,O.

Figure58.

ThecouplingclutchCCisdroppedandraisedbythebeakCBonthecam.ThehammerisraisedbythebeakHBon thecamandlockedbythepinP.

Pressing button number 2 pushes the pinP aside, the hammer drops and thehandsarereturnedtozero.

Thismodelhasnobrakelever.AtensionspringT,restingontheundersideofthechronographrunner,holdsthewheelfromturningaftertherecordinghandsarestopped.

AnotherpointofdifferenceistheactionoftheflexiblefingerFfortrippingtheintermediatewheelMLThewheelMIneednotclearthefingerFduringtherecording hands-to-zero operation as is necessary in the conventionalchronograph.InsteadtheflexiblefingerismerelypushedasidesincethehammerspringS isstronger.Andagain the jumperspringJ isweaker than theflexiblefinger thus permitting the minutes recording mechanism to functionsatisfactorily.

ChronographWithoutPillarWheelConstruction.Anothermodelthatdispenseswiththepillarwheelandcams

isshowninFigure59.ThehammerHistheall-importantpiece,forallfunctionsdepend upon it. The hammer rotates around the pivot CP and its several

functionsarelistedasfollows:(1)TheplanesurfaceS——SactuatestheslidinggearthroughtheeccentricscrewE.(2)TheeccentricscrewECraisesanddropsthecouplingclutchCC,and(3)theextremitiesofthehammerimpingetheheartcamsandreturntherecordinghandstozero.Whiletheabovefunctionsarebeingperformed,thehammerspringHSwillforceitsbeakHBintothenotches,1,2,and3. Thus the hammer spring assists the hammer in performing its severalfunctions.

Like thepreceding chronograph, thismodel hasnobrake lever.Thehandsare held from turning by the tension spring which is planted under thechronographwheel.

ACTIONOFTHEMOVEMENTStartingfunction.TheendEof theoperating lever0 fits intoaU-shaped

openingofthehammer.Pressureappliedtobuttonnumber1movestheendEtotheleft(arrowD)andthehammeristurnedcounterclockwise(arrowB).AstheeccentricscrewECmoves to theright, thecouplingclutch isdroppedand therecording begins. At the same instant the hammer spring HS assists themovementofthehammerandthebeakHBcomestorestinnotch1,

Figure59.

SimultaneouslywiththeabovemovementtheslidinggearSGisdroppedso

theintermediatewheelMImaybeturnedbythefingerF.

Stoppingfunction.Thischronographdiffersfromtheprecedingdesigns inthat thebuttonnumber1startstherecordinghandsbutdoesnotstop them.Tostop the hands, button number 2 is pressed, and to return the hands to zero,buttonnumber2ispressedasecondtime.

Returningnowtothefunctionofstoppingtherecordinghands,pressureonbuttonnumber2actuatestheflybackleverFB.(Incidentallythehammerliestothe extreme left and the beakHB of the hammer spring rests in notch1. Thereadermustvisualizethelocationofthepartsinordertounderstandthefunctionto be described presently.) The flyback lever, moving in the direction of thearrow, engages and pushes the operating lever0 to the right as shown by thearrowC.Thehammeris therebyrotatedclockwise(arrowA)and theeccentricscrewECraisesthecouplingclutch.Thustherecordinghandsarestopped.

SimultaneouslytheslidinggearSGisshiftedoutofaction.Coupled with the above action, the beakHB of the hammer spring drops

nearly into notch2, for the hammer cannotmove further. The reason for thismaybenotedbecausethegrooveG,underthehammer,meetsthepointPontheflyback lever. Thus the hammer is stopped dead. When button number 2 isreleased,theflybackleverreturnstoitspositionofrestandthehammerisfreed.Thehammerbeingfreed,hammerspringHStakesoverandcausesthehammertoadvanceslightlyfurther(clockwise)until thehammerspringbeakHBdropsintonotch2ofthehammer.

Hands-to-zerofunction.Theillustrationshowsthepositionofallpartsjustpriortoreturningtherecordinghandstozero.Pressingbuttonnumber2asecondtime causes thepointP on the flyback lever to press directly on the hammer,sincethepositionofthehammerdoesnotpermittheflybacklevertopassunderasheretofore.ThehammerspringassistsinthemovementandinstantlyitsbeakHB comes to rest nearly to notch 3 on the hammer. Simultaneously, theextremities of the hammer reach the heart cams, and the recording hands arereturnedtozero.

Whentherecordinghandsareagainstartedbypressingbuttonnumber1,thehammer turns counterclockwisewithout halting itsmotion at notch2. Instead,thehammerspringbeakHBcomestorestonnotch1.

Split-SecondChronographConstruction.Thesplit-secondchronographisconstructedonasimilarplan

to that of the single button chronograph. It differs only in that a specialmechanism is added together with a separate button for its operation.Constructional designs are many and varied. In some makes the split-secondwheel lies under the dial as shown in Figure55. The post of the split-secondwheel is hollow and rides freely on the pivot of the chronographwheel. Thesplit-secondhand,obviously,liesbelowthechronographhand.Inanotherdesignthesplit-secondwheelliesonthebridgesideofthemovementandhasaslenderpivot which extends through a hollow chronograph runner. In this design thesplit-second hand lies above the chronograph hand. In either design the split-secondwheelisofsimilarconstruction.

ReferringtoFigure60,weseeoneof the latestandimproveddesigns.Thesplit-secondwheelhasasmoothrimonwhichtwobrakearmsmayresttostopthewheel.Aseparatepillarwheelactuates thebrakearms.When thearmsareopen the split-second wheel turns in unison with the chronograph runner bymeansofaveryingeniousdevice.TheleverLispivotedtotherimofthesplit-secondwheel,andthespringSprovidesthenecessarytensiontoforcetherollerto the lowestpointon theheart camC.Theheart cam is secured to and turnswiththechronographrunner.Thusbothwheelsturntogetheruntilsuchtimethatthebrakeleversbuttonispressed.

Figure60.

ACTIONOFTHESPLIT-SECONDCHRONOGRAPH

Nowpressingthebrakeleversbutton,thebrakeleverscloseinonthesplit-second wheel and its motion ceases. The chronograph runner and its camcontinue to rotateand theroller is forcedoutwardfromits lowestpointon thecam, but clings to its circumference because of the tension spring. Thus therollerlevermovesoutandinuntilthebrakeleversbuttonisagainpressed.Uponthereleaseofthebrakelevers,therolleragainseeksthelowestpointontheheartcam and the split-second hand returns instantly to its proper place directlyunderneaththechronographhand.

VenusChronographConstruction. The Venus chronograph, Figure 61, is one of the newest

modelsand,inkeepingwithpresentdaytrends,isdesignedwithoutpillarwheel.The operation is on the conventional pattern, that is, two presses on buttonnumber1startandstoptherecordinghands.Buttonnumber2isusedtoreturnthehandstozero.

Figure61.

The reverser R is held to a fixed position by the reverser spring RS.Successivepressesofbuttonnumber1causesthecamC through theoperationleverO and reverserR tooscillate first counterclockwise, thenclockwise.ThecamCandthehammerHarefastenedtogetherbytwoscrewsandturntogether

asonefunctionalunit.ThehammerspringHSassistsinrestrictingthemotionofthe hammer by way of its beakHB dropping into the notches 1 and 2. Thehammercannotturnclockwisefartherthannotch2bymeansofbuttonnumber1as the downward movement of the reverser is restricted to just this distance.However,whenthebuttonnumber2ispressed,thehammerswingsclockwisetothe extent that the faces of the hammer impinge the heart pieces and therecordinghandsarereturnedtozero.Theforceofthehammerspringreturnsthehammer to the position shown in the illustration immediately upon releasingpressureonbuttonnumber2.

ACTIONOFTHEVENUSCHRONOGRAPHStartingfunction. Pressingbuttonnumber1 theoperating leverO comes

intoactionandforcesthereverserdownward.AsthereversermovesdownwarditalsomovestotheleftuntilthebeakBreachesthelowestpointLPonthecamC. As the movement of the reverser continues, the hammer H is turnedcounterclockwisethusreleasingthecouplingclutchCC.TheintermediatewheelIengagesthechronographrunnerCRandthesecondsrecordinghandisstarted.Simultaneously the motion of the cam permits the sliding gear SG to turnclockwisesoas tocause theminutes recordingmechanism tostart functioningthroughcontactoffingerFwiththeintermediatewheelML

Stopping function. Pressing button number 1 a second time causes thehammertoturnclockwisesincethebeakofthereversernowdropstotherightsideofthecam.ThehammerspringHSdropsintonotch2onthecamC.Thusthe coupling clutch is lifted and the seconds recording hand is stopped.Simultaneouslytheslidinggearisliftedandtheminutesrecordingmechanismisthrownoutofaction.

Handstozerofunction.Thehandsarereturnedtozerobypressingbuttonnumber2which,inturn,actuatestheflybackleverFB.ThepinPontheflybacklever presses on the hammer which swings to the heart pieces. Thus therecordinghandsarereturnedtozero.

PierceChronographConstruction. The principles of operation of the Pierce chronograph are

different from all other makes and for this reason some of the customaryterminologydoesnotapply.Itwillbeobserved,Figure62,thattheequivalentofa coupling clutch consists of two levers, starting leverSL, and starting springSS.Insteadofthetraditionalcenterwheel,themechanismisfittedwithahollow

centersecondswheelandisnotshownintheillustration.LyingabovethecentersecondswheelisasecondsrunnerSRwithheartpieceandfingerF.Bymeansof thestartingspringSS thesecondsrunner is forceddownward toengage thecentersecondswheel.Inthismannertherecordinghandsarestarted.Theradicaldeparturefromtheusualdesignisseeninthedirectionthatthefunctionalpartsmove. In all mechanisms previously considered the movement of the parts ishorizontalwhereasinthePiercethemotionisvertical.

However, the return of the recording hands to zero is similar to theconventional type. Even here, there is a difference in that the brakes are notactuated by the hammer H. Instead the brakes B1 and B2 friction constantlyexceptwhentherecordinghandsarerecordingtime.Thisismadepossiblesincethefrictionofthebrakesislightenoughastonothindertheproperreturningoftherecordinghandstozero.

Figure62.

ThefingerF,thatfunctionstheminutesrecordingassembly,israisedabovetheintermediatewheelIwhentherecordinghandsarestopped.Thispermitsthereturnoftherecordinghandstozerowithoutanyinterferencebetweenthefingerandtheintermediatewheel.

ThePierce chronograph iswell suppliedwith adjusting screws to properlyco-ordinate the several functions. These are shown in the illustrations by thelettersAS.

ACTIONOFTHEPIERCECHRONOGRAPHStartingfunction.Pressingbuttonnumber1actuatestheoperatingleverO

andthepillarwheelP isturned.Thismovement raises thebeakof thestarting

leverSLfromitspositionbetweentwopillarsofthepillarwheeltherebycausingthe beak to rest on the face of one of the pillars. As a result of the abovemovement of the starting lever to its new position, three springs are actuatedthrough the adjusting screws AS1, AS2, and AS3. The movement of the threespringstakeplacesimultaneouslybutweshallanalyzethefunctionsseparatelyasfollows.

AdjustingscrewAS1pressesdownwardon thestarting springSSwhich inturnpressesonthesecondsrunnerSRandthesecondsrecordinghandisstarted.The recording hand is started because the seconds runner, in its downwardmovement, engages the center seconds wheel, not shown, which runscontinuously.

AdjustingscrewAS2pressesonthesecondsrunnerbrakeB2andreleasesthesecondsrunner.

AdjustingscrewAS3pressesontheminutesrecorderbrakeB2and,throughits separation from the intermediate wheel I, releases the minutes recordingmechanism.

Stopping function. Pressing button number 1 a second time returns thestartingleverSLandthestartingspringSStotheiroriginalpositionspermittingthesecondsrunnerSR toseparate fromthecentersecondswheel. Immediatelythe recordinghand stops.Simultaneously the twobrakesB1 andB2 come intoaction and the seconds runner SR and the intermediate wheel I are held inposition.

Hands to zero function. The recording hands are returned to zero bypressingbuttonnumber2whichactuatestheflybackleverFB.Theflybackleverin turn presses on the hammer H. To give a little zip to the motion of thehammer, the springZS is fitted and the end engages the beakZB.This springresiststhemotionofthehammerwhenpressureisappliedtothebutton,andisjustenoughtocausethehammertoflyquicklyafterthebeakpassestheendofthespring.

PARTIIChapterNine

CALENDARMECHANISMSCALENDAR attachments to wrist watches and clocks have become quite

popularinrecentyears.Theideaisnotnewbutthedesignofthemechanismhasshownsteadyimprovement.Figure63showsthecalendarmechanismtypicalofthedesigngenerallyusedinmen’swristwatches.Thestarwheel,12.liesintheupper rightpart of thedial and indicates themonths. It has12points, one foreachmonth, and is corrected manually the first of eachmonth by pressing abuttononthecasewhich,inturn,actuatestheleverMC.

Thestarwheel7liesintheupperleftpartofthedialandindicatesthedaysoftheweek.Thewheelcontains7pointsandmakesoneturnin7days.Itmustbe tripped once in 24 hours and thewheel that does the trippingmust, if onetripperisprovided,makeoneturnin24hours.Thisisattainedinthefollowingmanner:

Thehourwheel,notshown,carries twowheels.The largerwheel isdrivenbytheminutepinion—infact,itisthesamewheelasfoundinthedialtrainoftheconventionalwatchwithoutthecalendarattachment.Thesmallerwheelliesabove the larger and drives the datewheel,D. The date wheel has twice thenumber of teeth as the smallerwheel on the hourwheel.Thus the datewheelmakesoneturnin24hours.ThedatewheelDgearswiththeintermediatewheelI.The intermediatewheel carries thepinPwhich trips thestarwheel,7, eachnightafter12p.m.

Figure63.

Thedatehandwheel,31,fitsfreelyoverthehourwheelandcarriesasweephand for indicating the days of the month. The wheel contains 31 teeth andmakes one turn in 31 days. The pin DP trips the date hand wheel atapproximatelythesametimeasthepinPtripsthestarwheel,7—thatis,ifthewheelsD and I are correctly timed. The timingdot,TD, on the datewheel isprovidedforthispurpose.PinPandthedotTDshouldlieonthelineofcenterswhenthetwowheelsareassembled.

Themoonphasedisc,59,isturnedbyatrainoftwowheels.ThesmallwheelonthehourwheelturnsthewheelM,which,likethedatewheel,makesoneturnin24hours.ThepinMPtripsthemoonphasedisceachdayatnoon.Themoondiscissuppliedwith59pointsandmakesoneturnin59days.Onemoonphasecontains 29½ days, hence the necessity of using a wheel of 59 points. Thisexplainswhytwomoonsarepaintedonthedisc.

All calendar records are set by small push pieces on the side of the caseexceptthedaysoftheweek,whichcanbecorrectedonlybyturningthehands.Tomake certain that the star wheelsmove the required distance, jumpers areprovided.

Figure 64 shows a calendar watch wherein the month is recordedautomatically. This feature is made possible by fitting an extra wheel Mcontaining31teeth.Onepointonthiswheeltripsthestarwheel,12,onceevery31days.Thusthemonthsarerecorded,buttheshortmonths,ofcourse,require

resetting.

Figure64.

NotesonRepairingComplicatedWatchesThroughout the pages of this book we have attempted to analyze the

essentials of clock and watch repairing; to explain the function of the moregeneral and frequently handled mechanisms of striking clocks, automaticwatches,stopwatches,andchronographs.

However,asthehorologist’sworkcontinues,theunusual,theveryold,andthe very new in timekeeping instruments occasionally come into the shop forservice and repair.Obviously, the beginner feels uneasy in tackling these jobsthatareoftenmorecomplicatedthanthegeneralrunofrepairs.Weshallgiveafewhintsbelowonhowtobesthandlethesewatches.

Routinetofollowincleaningunfamiliarandcomplicatedtimepieces. Ifthe horologist does not already have one, he should purchase that type ofmovementcoverthatcomesequippedwithatrayforwatchpartsthatisdividedintosections.Then,asthemovementinquestionisdismantled,thepartsofonecompletefunctionalassemblyareplacedinoneofthesectionsofthetray.Thismethod aids one’s memory as to the manner and order in reassembling theseparateunits.

It is a good plan, too, to start working on a complicated timepiece in the

morning so that the jobmay be finished before the day’swork is over. For abeginner, at least, the idea of starting a complicated watch of unfamiliarconstructionlateintheafternoonandfinishingthejobthenextday,resultsinafailure to remember the proper place and order of assembly of the functionalunits.

Anotherpointshouldbeobserved.Complicatedwatcheshaveagreatvarietyofscrewsofmanykinds,shapesandsizes.Whenthereisanydoubtastowhereaparticularscrewistogolateraftercleaningandrepairing,replacethescrewintheplateorbridgeimmediatelyafter thepart isremoved.Thispracticewillbefoundtosavetime.

Forbeginners,inparticular,thedismantlingofthemovementshouldbedonewithdeliberationandconcentration.Thinkwithundividedattentiononeachandeverypart removed.Usingaspecificexample, thebeginnersays tohimself,“Iam now removing the bridge that holds the chronograph wheel and minutesrunnerinplace.Removalofthehammercomesnext,andnowthechronographwheelandminutesrunnercanbeliftedout.Inreassemblingthesepartstheorderis reversed.” It is surprising howmuch thismanner of concentrating aids thememory.

PARTIII

THEORETICALHOROLOGY

PARTIIIChapterOne

ASTUDYOFTHEPENDULUMTHE PENDULUM is today and has been for three hundred years the most

important piece of clockwork. It is practically a precision instrument for themeasurementoftime.Bymeansofaweightormainspring,atrainofwheelsandanescapement,itsvibratorymotion,whichfrictionandresistanceoftheairtendtodestroy,issustained.

Thehistoryof thependulumhad its beginningwithGalileoGalilei (1564-1642),whenheobservedtheswingingchandelierinthecathedralofPisaaboutthe year 1581. Later he conceived an escapement to keep the pendulum inmotionandmadeadrawingofit.Itremainedforhisson,VincenzioGalilei,tomakeaworkingmodel,andacopyoftheoriginalinrunningconditionistobeseen at the Kensington Museum, in London. However, Christian Huyghens(1629-1695)wasprobablythefirsttoapplythependuluminapracticalwaytoclocks.Hisfirstpendulumclockwasproducedin1647.

Inordertodiscussthetheoryofthependulumwithanydegreeofclearness,it is desirable to digress for a moment and consider some of the factorsconcerning theearthuponwhichwe live, for it sohappens that therotationoftheearthuponitsaxis,thenatureofgravitation,andotherphysicalphenomenahave a direct relationship to the observed performance of the pendulum, andthesefactsmustbekeptinmind.

UniversalForcesGravitation.Allbodiesfallperpendicularlywithreferencetothesurfaceof

stillwateror,inotherwords,theyfallvertically,whichis,ofcourse,towardthecenter of the earth. The fall of a body also increases in velocity through thedistancecovered,asshowninTable1.

Theabovetableshowsthatintwosecondsabodyfallsfourtimesasfarasitfallsinthe first second. In three seconds a body falls nine times as far and in fourseconds,sixteentimesasfar.Thus,wemaysaythatthedistancewhichabodyfallsisproportionaltothesquareofthetime.

Newton’s law of universal gravitation.* Sir Isaac Newton (1642-1727)discoveredcertainfactsaboutgravityandwestateherewithNewton’sLaw:

*Newton’s systemofmechanics as explained above is valid for thependulumandother large scalephenomenaofnature.However,thetheoryofrelativityhasshownthatbeyondtheselieasystemofatomicandsubatomicprocesseswhichdonotobeyNewton’slawsatall.SeeH.HortonSheldon,Space,TimeandRelativity,TheUniversitySociety,NewYork.

All bodies in the universe attract every other body with a force which isdirectly proportional to the product of the attracting masses and inverselyproportionaltothesquareofthedistancebetweentheircenters.

The above statement of the law may be clarified by elaborating on it asfollows:

Gravityactsbetweenbodies,andtheattractionisproportionaltothemass—thatis,alargermasscarrieswithitanequallyincreasedpullofgravity.ThiswasdemonstratedbyGalilei,whenhewasprofessorofphysicsattheUniversityofPisa,bydroppinglargeandsmallironballsfromtheleaningtowerofPisa.Theyreached the ground at the same time. This demonstration astonished andbewilderedanumberofprofessorswhosawit,foritwasbelievedpriortothattimethatheavybodiesfellfasterthanlightbodies.

Later, by the invention of the air pump for producing a vacuum, thisstatementbyGalileiwasverifiedbydemonstratingthatheavyandlightbodiessuchasacoinandafeatherwouldfalltogetherinaglasstubewheretheywouldnotencounter thefrictionof theair.After theair isagainadmitted, thefeatherfluttersslowlybehindtherapidlyfallingcoin.

Newtonfurtherprovedthatgravityisthesameasthoughtheentiremassof

thebodywasconcentratedat itscenter.Hence, inconsidering theattractionoftheearthforanybodyuponitssurface, thedistanceindicatedinthelawis theearth’sradiusplusthedistancetothecenterofgravityofthebody.

The law goes on to state that the attraction of the masses is inverselyproportionaltothesquareofthedistancebetweentheircenters.Forexample,letus suppose that abodywere removed twiceas far from the earth’s center andallowedtofall. Itwouldfallonlyone-fourthasfast. If removedthree, four,orfive times as far from the earth’s center, the body would fall one-ninth, one-sixteenth,orone-twenty-fifthasfast.

Hence,anironballwouldfallfasterwhendroppedatanaltitudeofsealevelthananironballdroppedonamountain7000feetabovesealevel.Thisfactcanbe proved when counting the vibrations of a pendulum placed at differentaltitudes.

Centrifugal force.Tie a string to a stone andwhirl it around rapidly.Thestringwillstretchtighterandtighter.Increasethespeedandthereisapossibilitythatthestringwillbreak.Ifso,thestonewillflyawayatatangent.

Thisthrustisknownascentrifugalforceandtendstodistortasphererotatingonitsaxis.Infact,thisearthofoursbulgesattheequatorandisflattenedatthepoles due to the very same cause, for the earth is sufficiently pliable to be soformed.Asaresultofthisbulgingofthelandandwaterabouttheequator,theearthis26.7milesshorterbetweenthepolesthanthroughtheequator.

Centrifugalforce isgreatestat theequatoranddecreaseswitheachparalleloflatitudetowardthepoles.Atthepolesitiszeroandtheweightofallbodiesisaccordinglyaffected.

Summary.Our discussion hasmade clear three important factors: (1)Thevelocity of a falling body is the same regardless of its mass. (2) Weight isdeterminedby the locationofabodyatdifferentaltitudes,being lessathigheraltitudesandincreasingasthebodyismovedtoloweraltitudes.(3)Centrifugalforce tends to lessen the weight of bodies when the force is increased as bymovingabodyfromthesmallerparallelsoflatitudetothelarger.Atthepolesitiszero.Themaximumisreachedattheequator.

ThePendulumDefinitions.Now, returning to the study of the pendulumproper,we shall

see how the earth’s motion and gravitation affect the performance of thependulum.Itisdesirable,however,firsttoconsiderseveraldefinitionsrelativeto

thesubject.

Pendulum:Anysuspendedbodythatswingstoandfro.Bob:Themassofmetalatthebottomofthependulum.Simple pendulum: A heavy bob suspended by a slender thread.

Theoreticallyasimplependulumisaheavyparticlehavingnoappreciablesizeandsuspendedbyalinewithoutweight.

Compound pendulum: Any pendulum having a portion of its masselsewhere than in a compact bob. All clock pendulums are compoundpendulums.

Centerofsuspension:Thefixedsupportofapendulum.Center of oscillation: A point in the bob in which, if all the matter

composing the bobwere collected into it, the time of vibrationwould not beaffected.

Vibration:Onecompleteswingofthependulum.Period:Timeoccupiedinmakingavibration.Amplitude:Theanglebetweentheverticleandtheendofavibration.Angle

AOB(Figure1)istheamplitude.

Figure1.

MOTIONOPTHEPENDULUMWeknow that a simplependulum is in equilibriumwhen the threadwhich

supports the bob is vertical. However if we pull the pendulum to an inclined

positionandletitgo,itwilldescendtoregainitsoriginalpositionanditsownmomentumwillcauseittoraisenearlyasfarontheotherside.

The gravity that acts on the bobwhen inmotion is a vertical forcewhichmaybeseparatedintotwoparts:ThefirstpartactswiththeprolongationofthethreadasatD,Figure1.ThisforceisdestroyedbecauseofthethreadCwhichsupportsthebob.Thesecondactsinadirectionperpendiculartothefirst,asatE,andmanifestsallitsforcethroughtheverticalattractionofgravityonthebob,thusforcingthebobalongthelineAB.Thebobwillmoveasfarastheverticalpositionwithanacceleratedmovementbutnotconsistentlysoaswithafallingbody, for the gravitational force along the circular path is continuallydiminishing and it ceases altogether when the bob reaches the vertical. Onregainingitsverticalpositionthependulumwillraiseontheothersidebecauseofitsacquiredmomentum,althoughtheeffectofgravitywhichattractedthebobinitsdownwardmovementisrenderedvoid.Gravitywillnowactasaretardingforceandthevelocitywillbelessened.Itwill,however,raisetonearlythesameheightthat it leftontheothersideandthenreturn,executinganothervibrationlikethefirst.

Asimplependulumwillvibrateforall timeiffrictionandresistanceoftheaircouldbecompletelyeliminatedbut,ofcourse,thiscannotbeinpractice.

THEFOURLAWSOFTHEPENDULUMThefollowingarethelawsgoverningthemotionofthependulumaswould

besuggestedfromthematerialgivenintheprecedingpages.Firstlaw.Thevibrationsofthependulumarepracticallyisochronal—that

is, theyareexecuted innearly the same time forvariedamplitudes so longastheseamplitudesremainsmall.

Thependulumformula,whichwillbeconsideredalittlelater,assumesthatthependulumvibrationsareisochronal.Inpracticewedonotfinditsoandtheerrorconsistsofalossforthelongarcs.Thisisknownasthecircularerrorandamorethoroughanalysiswillbegiveninalaterportionofthischapter.

Secondlaw.Theperiodofthependulumisindependentofthemassandalsoofthematerial.

As previously stated, falling bodies of varying masses are equallyaccelerated.Inotherwordstheforceofgravityactingonthebobisproportionalto itsmass,whichmeans, simply, that altering theweightof apendulumdoesnot change the time of vibration, so long as the center of oscillation is notchanged.

However,ifweplaceasmallweightabovethependulumbob,theclockwillrunfaster,forthecenterofoscillationhasbeenraisedandthependulumhas,infact,beenshortened.

Thirdlaw.Theperiodofapendulumisproportionaltothesquarerootofitslength.

We have learned that gravity is a constant force which is required to actthroughfourtimesthedistancetoimparttwicethevelocitytoabody.Thereforea pendulum that is to vibrate twice as fast as anothermust followa path fourtimesassteep,andtorealizethiscondition,itmustbeone-fourthaslong.Statedinmathematicallanguage,wehave:

inwhichtequalsthetimeofonevibrationandlthelength.

Fourth law. The period of a pendulum is inversely proportional to thesquarerootoftheaccelerationofafallingbody.

Whenapendulumistakenfromonelocationtoanotherinwhichtheforceofgravity is different, its rate is affected in the samemanner as that of a fallingbodyandforthesamereason.Theeffect,asstatedinthelaw,wouldbesuchthatif gravity has four, nine, or sixteen times more intensity, the pendulum willvibrate two, three, or four times faster.Hence, a pendulum clock regulated tokeep correct time at the equator would run fast at the north pole. In fact thelengthof a secondspendulumforpointson theearth’s surfacewouldbeas inTable2.

THEPENDULUMFORMULAThelawsstatedaboveareexpressedinthefollowingformula:

Bymeansof this formulawemaydetermine:(1) the lengthofapendulumforagivenperiod;(2)theperiodforagivenlengthofthependulum;and(3)theaccelerationofgravityforanypointontheearth’ssurfacewhereitsvalueisnotknown.Horologistsare,ofcourse,interestedonlyinthefirsttwoprinciplesandthethirdshallnotbeconsideredfurther.

Illustrative example 1. What is the length of a seconds pendulum if theaccelerationofgravityisgivenas32.1feet(385.2inches)persecond?

Illustrativeexample2.Whatistheperiodofapendulum9.75incheslong?

Solutionofpendulumproblemsbyproportion.Thesolutionoftheaboveproblems may be simplified by noting that the third law of the pendulum,mathematicallyexpressed,isasfollows:

Squaringwehave

t2=l

Thuswemayconsidertwopendulumsofdifferentperiodsandtheirrelativelengthsbytheproportion:

Itfollowsthatthelengthorperiodofanypendulumcanbedeterminedifthelengthofasecondspendulumisknown.

Illustrative example 1. What is the length of a pendulum vibrating 1/3seconds?

Illustrativeexample2.Whatistheperiodofapendulum156incheslong?

DetrimentalForcesThereareseveralforcesthataltertheuniformswingofthependulum.Some

forcescanbequiteaccuratelycontrolled,butotherscannot.Themoreimportanterrorsarelistedasfollows:

Escapementerror CircularerrorBarometricerror Temperatureerror

Thesewillnowbediscussedintheordernamedabove.

ESCAPEMENTERROR

The pendulum of a clockmust have some sort of amechanical device tokeep it inmotion.Thismechanism iscalledanescapement. It is adevice thatperiodicallypushesthependulum,butthenatureofthispushissuchthatcertainconditionsareintroducedthataffectthetimekeeping.

Theimpulsedeliveredtothependulumthroughtheescapementshouldtakeplaceatthemomentwhenthependulumreachesthedeadpoint,thatis,whenthependulumisvertical.Thisidealconditionwouldpermitthependulumtoswingto and fro in the same time that a free pendulum performs these swings.However,themechanicalmeansatourdisposaltokeepthependulumvibratingdonotmeet theabove requirementsandweareobliged to takeaccountof thefollowinglaws:

1.Animpulsedeliveredtoapendulumbeforethedeadpointwillacceleratethevibrations.

2.An impulsedelivered to a pendulumafter thedeadpointwill retard thevibrations.

Thisprinciplecanbeeasilydemonstratedwithasimplependulum.Impulsegiven to apendulumbefore it reaches thedeadpoint causes it to arrive at thedead point more quickly than if it were acted upon by gravity alone. Givenimpulse after reaching the dead point results in driving the pendulum farther,resistingtheforceofgravityandatnoparticularlyacceleratedrate,ifany.Hencea retardation takes place, and the greater the distance the impulse takes placeafterthedeadpoint,thegreateristheretardation.

Someescapementsgiveimpulseafterthedeadpoint,othersatthebeginningofavibration,andtheratediffersaccordingly.

BAROMETRICERRORAswingingpendulumfanstheairandcreatesabreezeinitswakemuchlike

that of a rapidly moving train. But much more happens than the meredisturbance of the air, and to clarify the subject it will be necessary first toexplainsomeofthephenomenarelativetotheatmosphereproper.

Atmosphericpressure.Air,likeallgases,ismadeupofuntoldmillionsofmolecules. These molecules move about freely and at a terrific speed. Theirconstant motion produces a pressure of great magnitude and the higher thetemperaturethefasteristhemotionandthegreateristhepressure.

Atmosphericpressureisfifteenpoundstoasquareinch,oraboutonetontoasquare foot. This pressure takes place in every direction, as can be shown bypartiallyfillingaglasswithwaterandcarefullyfittingasmoothcardboardtothe

mouthoftheglass.Invertingtheglass,thecardboardwillnotfallandthewaterisseeminglysupportedby thecardboard. It isapparent that there ispresentanupwardforcethatisequaltothedownwardforceofthewater,infact,theglassmaybetiltedinahorizontalpositionandtheresultwillbethesame.Thisshowsthatthepressureofairtakesplaceineverydirection.

Weightof air. Since air is a real substance occupying space, itmust haveweight. This can be demonstrated by the simple experiment with an ordinarybalance.Determinetheweightofanopenglassflaskwithstopper.Removesomeoftheairfromtheflaskandweighagain.Theflaskwillnowshowlessweightand the actual weight of the air removed is found by finding the differencebetweenthefirstandsecondreadingsof thebalance.Experimentshaveshownthatacubicfootofairweighsabout1¼ouncesandtheairinanaveragelivingroomweighsabout500pounds.

Theaboveanalysisofatmospherewouldleavenodoubtthattheswingofthependulumwouldbeaffected.Iftheweightandpressureoftheairwereconstant,these factors would be of little concern to horologists, but since they arevariable,thependulumarcsvaryalso.Atmosphereaffectsthependulumbywayofthreeforces:

1.Inertiaoftheair

2.Frictionoftheair

3.Weightofthependulum

Inertiaoftheair.Whenapendulumstartsonitsdownwardswing,itmust,ofnecessity,pushtheair,butastheinertiaoftheairisovercome,theairtravelsalongwith thependulum. If the resistanceof the airbefore andafter thedeadpointwere equal, the timeofvibrationwouldnotbe altered. It canbe shown,however,thattheresistanceoftheairisgreaterasthependulumapproachesthedeadpointandlessafteritpassesthedeadpoint.Thisisclearwhenonerealizesthat the downward swingof the pendulum is constantly being accelerated andconsequently has to push the air all theway to the dead point. In its upwardswing, thependulumis retardedbygravityandfromthatpointon, theair justaboutkeepsupwith thependulum.Thus the inertiaof theair tends toprolongthetimeofthedownwardswingandalsoextendsthetimeoftheupwardswing.Thetotalresultisalossintime,andwitharisingtemperature(highbarometer)anditsassociatedincreaseinairpressure,thelossbecomesstillgreater.

Friction of the air. The friction of the air on the pendulum leads us to adiscussionofthebestshapeforthebob.Towerandprecisionclocksusuallyuse

thecylindricalbob.Mostothersusethelense-shapedbob.Thecylindricalbobissubjecttomoreairresistancethanthelense-shapedbob,yetthelattermeetswithmore surface friction. The more detrimental force is the air resistance,particularly if the pendulum is enclosed in a case. It is a simple matter todemonstratethisbymakingtwobobsofequalmassandmaterial,onecylindricaland one lense-shaped.By causing them to vibratewithout an escapement,wemayobserveinwhichbobthearcofvibrationfallsoffmorequickly.Thelense-shaped bob swings longer, which also implies that it would take less powerthroughtheescapementtokeepitswinging.Asidefromtakingasmallloadofftheescapement,thereseemstobenoadvantageinthelense-shapedbob.

Weightofthependulum.Wehavejustlearnedthatair,likeallsubstances,hasweight,andbecauseofthis,allmassesheavierthanairtendtobebuoyedup.Everyswimmerhasobservedthatittakeslessefforttoliftaheavystoneatthebottomofaswimmingpoolthantoliftthesamestoneaboveground.

Theeffectissimilarwiththependuluminair,forwithadenseratmosphere,theapparentweightorpullofgravityonthebobisreducedandtheclocktendstorunslower.

Thislossofgravitationalpullduetobuoyancyvarieswiththemetalusedforthe bob. It is greatest with iron and the least with lead. Brass lies midwaybetween.

Theonlypracticalsolutiontothebarometricprobleminprecisionclocksispartiallytoremovetheairfromthependulumchamberandmaintainaconstanttemperature.Fordomesticclockstheerroristoosmalltobeofanyconcern.

CIRCULARERRORChristian Huyghens, who built the first pendulum clock in 1657, made a

discovery that horologists to this day have not solved. He observed that thependulumisnotisochronousandthatthelongarcsloseandtheshortarcsgain.Also the change of rate is not uniform to the change of arc. Instead, the lossincreasesmorerapidlyasthearcsbecomeconsistentlylonger.Huyghenscalledthe error the circular error and the change of rate for different amplitudes isshowninTable3.

TABLE3

Semi-Arc Secondsperday30’ .411°00’ 1.65

1°30’ 3.72°00’ 6.62°30’ 10.33°00’ 14.8

The circular error was a real problem in Huyghens’ clock, for the vergeescapementofhisdayvibratedwithexceedinglylongarcs.Huyghensdevelopeda theoretical solution but it has never been given a practical application.Theoretically the problem is simple, for all the pendulum ball has to do is tofollowthepathofacycloidasshowninFigure2.Thedottedlineshowsthepathofthependulumball.ThesolidlineBCshowsthepathofthecycloid.Tomakethecycloidalcurve,fitapencilpointtoapointonthecircumferenceofthecircleGS (thegeneratingcircle), and roll it along the lineDE,without slipping.ThecurveBCisthusformed.ThediameterofthegeneratingcircleishalfthelengthofthependulumOA,Itisatonceevidentthatthesteepersidesofthecycloidalpath would give the pendulum added impulse that would make the long andshortvibrationsisochronous.*

*Thecauseofthecircularerrormaybebetterunderstoodifwewillvisualizeapendulum,thelengthofwhichisequaltotheradiusoftheearth.ForsuchapendulumthelinesofgravitydivergefromthecenteroftheearthasshowninFigure2B.Theearth’sgravitywouldpullthependulumbacktotheverticalpositionwithmuchgreaterforce,especiallyneartheendsofthearcs,thanispossiblewhenthegravitationalforceisdownward as is the case with the short (clock length) pendulum. This is clearly seen when comparingFigure2B,whichrepresentsapendulum3,959miles long(earth’sradius),andFigure2Cwhich showsapendulum 39 inches long. It is our opinion that a pendulum extending 3,959mileswould eliminate thecircularerrorbutthisistheoreticalfantasyandcannotbedemonstrated.

Figure2A.

Figure2b.

Figure2c.

Figure3.

HuyghensconceivedtheideaofusingcycloidalcheeksasshowninFigure3.Thependulumrodwassuspendedbytwosmallcordsandwhenthesecontactedthe cheeks, the pendulum ball was forced to follow the path of a cycloid.Huyghens’pendulumhasneverbeenusedwithmodernescapementsbecauseofmechanicaldifficulties.Othererrorswereintroducedthatweremoredetrimentalthan the one the devicewas intended to correct.Modern clocks, too, have anescapementwherebythependulumarcismateriallyreducedandthis,ofcourse,reduces thecircularerror. Ifaconstantpendulumarccouldbemaintained, thecircularerrorwouldbenoproblem;howeverthedeteriorationoftheoiltendstoshortenthependulumarcandthecircularerrorbecomesmanifest.

TEMPERATUREERRORMostmaterialsofwhichpendulumsaremadeexpandinheatandcontractin

cold.Thereforeaclockwithoutsomesortoftemperaturecompensationwillrunfast inwinter and slow in summer. In fact, a clock beating seconds and fittedwith a brass pendulum rod will lose as much as 1 minute in 24 hours if thetemperature is increased 100 degrees. We list herewith the coefficients ofexpansion of various materials generally used in clock-making. The figuresindicateachangeof180degrees,Fahrenheit—thatis,from32degreesto212degrees.

Lead..............0028 Steel..............0011Zinc...............0028 Wood..............0004Brass..............0020 Invar..............00009

It will be noted that Invar has the least expansion. Wood follows withslightlymore,whileleadandzincstandrelativelyhigh.Mercuryhasbeenfoundtobeanexcellentcompensatingmediumforpendulumswithsteelrods.

Theory of temperature compensation. The theory of compensatedpendulumsissimplythatthecenterofgravityofthebobisraisedorloweredtoanamountequal to the lengtheningorshorteningof thependulumrod.This isrealizedbyplacingtheregulatingnutatthebottomofthebob.

Suppose,forexample,apendulumrodofwoodlowerstheregulatingnutagivenamountwithagivenriseintemperature.Thebobisofleadand,sinceleadexpandseight times to thatofwood, thediameterof thebob isone-fourth thelength of the pendulum rod. The result is a fixed distance from the center ofgravityofthebobtothepointofsuspensionforalltemperatures.

Wood and lead pendulums are suitable for popular-priced regulators,grandfather clocks, and domestic clocks generally, but are not consideredsufficientlyaccurateforprecisionclocks.

Figure4.Graham’smercurypendulum.

Graham’s mercury pendulum. George Graham invented the mercurypendulum in about 1715 and it was probably the first successful attempt attemperaturecompensationwithanydegreeofprecision.Hispendulumconsistedof a steel rod and a single jar ofmercury as shown in Figure4. Two jars are

sometimes used,with the advantage that themercury responds to temperaturefluctuationsmorequickly.

Mercuryisanexcellentcompensatingmaterial,sinceitisliquidandcanberemoved from or added to the pendulum to attain perfect compensation. Forexample:iftheclocklosesinheat,thependulumissaidtobeundercompensatedandmoremercuryisadded.

Figure5.Harrison’sgridironpendulum.

Harrison’sgridironpendulum.Thispendulum,inventedbyJohnHarrisonin1720,consistsusuallyofnineparallelrods.Thedownwardexpansionofthesteel rods is compensated by the upward expansion of the brass rods. Theconstructionisasfollows:ReferringtoFigure5,thecentersteelrodextendstothelowerpartofthependulum,whereitissecuredtoahorizontalbar.Thissamebar supports twobrass rods that lead to the top.Thesebrass rods, in turn, aresecured toanotherhorizontalbar thatcarries twosteel rods to thebottom,etc.The two outermost steel rods support the bob. Thus, the five downward

expanding steel rods and four upward expanding brass rods are of suchproportionsthatthependulumwillbecompensated.

AfewclockswithHarrisongridironpendulumsmaystillbeseeninjewelrystorewindows,butthispendulumisnowregardedasobsolete.

Zinctubependulum.Thegridironpendulum,with itsnumerous steel andbrass rods, makes a rather cumbersome pendulum. A simpler construction isrealizedbysubstitutingzincforthebrass.Sincetheexpansionofzincisgreaterthanbrass,onezinctubefor theupwardexpansionisall that isnecessary.Theconstruction isasfollows:Theregulatingnutat thebottomof thecentralsteelrod supports a zinc tube that is just large enough to slip over the steel rod.Surroundingthezinctubeisasteeltubesecuredtothezinctubeatthetopofthependulum. The outer steel tube supports a lead bob at the bottom. Figure 6shows the zinc tube pendulum. Note that the steel tube reaches only to themiddleoftheleadbob.Thisisthetypegenerallyusedintowerclocks.

Invarpendulums.PendulumrodsmadeofInvar,analloyof64partssteeland 36 parts nickle, have a curious property of little or no expansion. In fact,somespecimensofthematerialhavebeenknowntoexpandincoldandcontractinheat.Invar,coinedfromthewords“elasticityinvariable,”istheresultofyearsofresearchbyDr.CharlesE.Guillaume,latedirectoroftheInternationalBureauofWeightsandMeasuresatServes.Thedifficulties in themakingofInvararesuchthatnotwobatchesareexactlyalikeandeachpendulumcallsforindividualexperimentalcompensation.

Figure6.Zinctubependulum.

ThebobforanInvarpendulumisusuallyasteelcylinder(Figure7),drilledlengthwise to admit the Invar rod. From the bottom to the center, the bore islarger soas topermit a space forabrassor steelpipecalled thecompensator,which provides the upward expansion, if necessary. The regulating nut liesbelowthecompensatorandisscrewedonathreadedportionoftheInvarrod.InmakinganInvarpendulum,thecompensatorisleftalittlelong.Testingtheclockincoldandheat,thecompensatorisshorteneduntilthecorrectlengthhasbeendetermined.

Figure7.Invarpendulum.

Invar pendulums are used today in the finest astronomical and precisionclocks.

PARTIIIChapterTwo

ASTUDYOFTHEBALANCEANDBALANCESPRING

THECIRCULARbalanceisusedinwatches,portableclocks,andchronometers.Theearliestwatchbalancewasthefoliot,similartotheclockfoliotdescribedonpage136.Themotionofthefoliotwascontrolledbyapig’sbristleinthemannershowninFigure8.

In the year 1660, Dr. Robert Hooke invented the balance spring, whichcompletelyrevolutionizedthescienceofhorology.Thisachievementpavedtheway to the development of the chronometer, which became an instrument ofprecisionasearlyas1750.

Figure8.Foliotbalancewithpig’sbristle.

Figure9.Earlyformofwatchbalance.

Followingtheinventionofthebalancespring,thethree-armedbalancewheelappeared.Thiswheel,showninFigure9,wasofsteelandmadeverythin.Latertherimwasroundedandusuallyhighlypolished.Thebalancespringwasfittedbelowthewheelinsteadofaboveit,asisdoneinmodernwatches.Aswiththeearlyclocks,thevergewasthetypeofescapementused.

Thebalancewheelinmoderntimepiecesvariesinconstructionaccordingtothequalityofthemovement.Figure10showsthetypegenerallyusedinalarmclocks and popular-priced novelty clocks. The balance is of brass, forcedfriction-tight on an axis of steel. The pivots taper to a point and run betweensteelcentersthatarescrewedintothebrassplates.

Thebalancewheelinwatchesandclocksofbetterqualityisconstructedinsuch a manner as to register time more accurately. This is realized byconsideringthetemperaturefactor—thatis,designingtherimofthebalanceinsuchamannerastocompensatefortemperaturechanges,orbyusingalloysthatremainconstant.

Figure10.Alarmclockbalance.

Figure11.Compensatingbalance.

Figure12.SolidbalanceasusedwithElinvarandNivaroxbalancesprings.

In addition, the better-quality balance is fitted with screws whichmay bealtered for the purpose of poising and timing. The pivots on the staff arecylindricalandworkinjewelsforgreaterprecision.Figures11,12,13,14,and15 show the types of balances used in watches, portable clocks, andchronometers.

TheBalanceFormulaThevibrationofthebalancewheelinmoderntimepiecesisdependentonthe

balance spring. The time of vibration is determined by a balance wheel of acertain size andweight coupledwith a balance spring of a certain length andstrength.Theformulabywhichthetimeofvibrationisdeterminedcontainsallofthesefactorsandisgivenasfollows:

Figure13.BalancebyPaulDitisheim.UsedwiththeElinvarbalancespring.

Figure14.Chronometercompensatingbalance.

Figure15.Invarbalanceforchronometer.

Figure15A.Wylershockproofbalance.

ThemassMof thebalance,asgivenabove,has todowith itsweight.Theradiusofgyrationreferstotheradiusofthebalance,which,inreality,isnotthefullradiusbutaradiusthatreachesapproximatelytothemiddleoftherim.Ifwe

couldimagineacircularringwithallofthemasscontainedtherein,androtatingaboutitscenter,themomentofinertiawouldbeequaltothemassandradiusofthebody.However,inpractice,partoftheweightofacircularbodylieswithinits circumference and the radius of gyration is likewise moved nearer to thecenter.

Themassandradiusofthebalance,takentogether,constituteitsinertia;* itfollows that a large, light balance may have the same inertia (ignoring pivotfriction)asasmallbutheavybalance.

*Inertiaisthatpropertyofanybodywhichcausesittoremainasitis—thatis,abodyatrestremainsatrest;andabodyinmotiontendstoremaininmotion.Also,itismoredifficulttogivemotiontobodiesofgreatmassand,likewise,itismoredifficulttostopthemotionofmassivebodiesthanlightbodies.

Horologistsareobligedtotakeaccountofthisfactorofinertiainthedesignofclocksandwatches.Forexample, the balance of an alarm clock is much heavier than that of a pocket watch. Alarm clocks,therefore,make14,400beats anhour;watchesgenerallymake18,000beats, and a fewvery smallwristwatchesmake19,800ormorebeatsperhour.

The balance formula is difficult to apply in practice, but the principlesgoverning timing corrections are contained in it, as may be noted in theexamplesgivenbelow.

Correctingatimingerror.Thebalanceformulaleadsustothesuppositionthat an error in the rate of a timepiece may be corrected in two ways: (1)increasingordecreasingtheweightofthebalance,or(2)alteringthelengthofthebalancespring.

Referringtothebalanceformula,itwillbenotedthat:

therefore:

T2=MorL

Thissuggeststhatproblemsrelatedtotimingmaybesolvedbyproportion.

Example1.Aclockislosing½hourin24hourswithabalanceweighing16grammes.

Whatshouldthecorrectweightbe?(The balance must be made lighter by an amount equal to the desired

increaseintime.)Therefore:

Example2.Awatchisgaining1hourin24hours.Thebalancespringis20millimeters

long.Whatshouldthecorrectlengthbe?(Thebalancespringmustbemadelongerbyanamountequaltothedesired

decreaseintime.)

TheBalanceMotionofthebalance.Thevibrationofthebalance,althoughdependenton

a balance spring instead of gravity as with a pendulum, differs from thependulumbyway of one other very important factor. It is amatter of relatedforces.Thetensionofthebalancespringisproportionaltotheangleoftwistofthebalance,whereasthepullofgravityonthependulumisnotproportionaltotheangle.

Suppose,forexample,weturnabalance¼turnandallowittoreturntoitspointofrest.Itwillreturnandcontinuetoturnnearlyasfarontheoppositeside.Nowturnthebalance½turnandallowit tovibrateasbefore.Thetensionhasbeenincreasedbytwoandthevelocityhasbeenincreasedbyanequalamount.These varied arcs show a condition throughwhich the displacement, velocity,and acceleration are equally related, and the movement is called harmonicmotion. Except for certain isolated factors, the balance spring is said to beisochronous and in accord with Hooke’s law, which states that the tension isequaltothevelocity.

However,forpracticalreasonstobeexplainedlater,thebalancemusttakea

motionwithincertainlimitstoperformthefunctionofgoodtimekeeping.This,then,isthenextpointtobeconsidered.

Theproperarcofmotion.Thebalancearcshouldnotexceed1½turns,norless than 1 turn. This can be determined readily by the following method:Supposethebalanceisatrest—thatis,thereisnotensiononthebalancespring.Nowrotatethebalance½turnandstop.Releasethebalanceandtheforceofthespringwillcauseittoreturntoitspointofrestanditsownmomentumwillcarryit½ turnfurther.Thisgives thebalance1 turn, that is,½ turn inonedirectionand½turnintheoppositedirection.Nowrotatethebalance¾turnandallowittoreturntoitspointofrestandasfarontheoppositeside.Thearcofmotionisnow1½turnsand thebalancewillcontinue tovibratebetween thesepointsaslongasthepropermotivepowerismaintained.Thearmsofthebalancebecomevisible at themoment the balance completes the arc and starts in the oppositedirection. It is, therefore, at the timewhen the balance stops that the distancecovered can be determined. Figure 16 shows the position of the arms for sixdifferentarcsofmotion.

Figure16.Thetoprowshowsthepathofhalfavibrationofthebalance.Thebottomrowshowstherelativepositionsofthearmsforafullvibration.

TheBalanceSpringThe long, fine, coiled wire that encircles the balance staff is called the

balance spring. It is commonly called the hairspring by laymen, probablybecause the first watches used a pig’s bristle to regulate the vibrations of thebalance.

The inner end of the balance spring is secured to a collet, which is fittedfrictiontightonthebalancestaff.Theouterendispinnedtoastud.Inthefinermakesofclocksandwatches,thestudfitsintoaholeinthebalancebridgeandisheldsecurebymeansofasmallscrew.Incheaperclocks,thestudisrivetedto

theupperplate.

Figure17.Flatspring.

Figure18.Breguetspring.

Thereare threeformsofbalancesprings ingeneraluse: (1) theflatspring;(2)theBreguet;and(3)thecylindrical.Theflatspring(Figure17)hasthestudfixed to the same plane as the body of the spring, with the result that thevibrations take place in an eccentric manner. The Breguet, named after itsinventor,hasaportionoftheoutercoilraisedandbentinsuchamannerastolieabove thebodyof the spring.TheBreguetoften takemany forms.Oneof themorepopularformsisshowninFigure18.Acorrectlydesignedovercoilcausesthevibrationtotakeplaceinaconcentricmanneraroundthebalancestaff.Thecylindricalspring(Figure19)coilsupwardandhastheappearanceofacylinder.Itisusedalmostexclusivelyinmarinechronometers.

Motionoftheflatspring.Ifweexamineaflatspringinmotion(Figure20),weshallobserve that thecoilshavearadialmotion—that is,amovementbywhich the outer coils tend inward and outward from the center. Furtherexaminationshowsacircularmotionwhichbeginsat thesecondcoil fromthe

regulatorpinsandincreasesaswefollowtheseveralcoilstowardthecenter.Thecircularmotionreachesitsmaximumwiththatcoilimmediatelysurroundingthecollet, and the radialmotion, so clearly evident on the outer coils, practicallydisappears. It is therefore evident that the circularmotionof the outer coils isrestrictedby the fixedpositionof the regulatorpins.The influenceof thepinscontinueswithagraduallyreducedeffectforabouthalfthetotalnumberofcoilstoward the center, until finally the inner terminal takes over and controls theinnermost coils. The coil that liesmidway between the regulator pins and thepinningatthecolletiscalledtheneutralcoil.Itissonamedbecausetheforces,originating at the outer and inner terminals, are neutralized at this point. Thevariable tensions or forces that take place on both sides of the neutral coilaccountfortheisochronalerroroftheflatspring.

Figure19.Cylindricalbalancespring.

Figure20.

DetrimentalForcesWehavelearnedthatthereareseveralforcesthataltertheuniformswingsof

thependulum.Likewise thereare several forces thatalter theuniform turnsofthe balance. The analysis of the escapement error given in the section on thependulum(page237)appliesalsotothebalance.Otherfactorsthatapplyonlytothebalanceassemblyarelistedasfollows:

(1)Isochronalerrorofthebalancespring.(2)Thepoiseofthebalance.(3)Theadjustmentoftheregulatorpins.(4)Temperatureerrorofthebalanceandspring.(5)Thepoiseerrorofthebalancespring.

ISOCHRONALERROROFTHEBALANCESPRINGThe forces present on either side of the neutral coil, as indicated in the

preceding discussion, cause a retardation or an acceleration, depending on therelative positions of the two eccentric motions as the balance vibrates. Theanisochronismthusproducedcanbevariedbyalteringthelengthofthespring.Such alterations, of course, change the relative positions of the eccentricmotions, or, in words more useful for the present analysis, they change theangular distance between the inner terminal and the regulator pins. It is thischangeofangulardistancethatdecidestheratebetweenthelongandshortarcs.Thelawsgoverningtheisochronismasconcernedwiththeabovestatementareasfollows:

(1)Whentheangulardistancebetweentheinnerterminalandtheregulatorpinsstandsatevencoils—thatis,wholecoils—-theshortarcsgain.

(2)Whentheangulardistancebetweentheinnerterminalandtheregulatorpinsstandsatevencoils,plushalfacoil,theshortarcslose.

(3)Whentheangulardistancebetweentheinnerterminalandtheregulatorpinsstandsatevencoilsplusone-fourthorthree-fourthsofacoil,thelongandshortarcsaremorenearlyisochronal.

Letus assume that the arcofmotionof agivenbalance isone turn, as anexampleofashortarc.Ifthespringispinnedatevencoils,theeccentricmotionswill stand inoppositedirections.According toLaw1, this produces againingrateascomparedwiththelongarcs.Thiscanbeexplainedbyreasonofthefact

thattheeccentricmotionoftheoutermostcoilsexertsaforce(whenwoundup)inthedirectionofthearrowA,asinFigure21,whiletheeccentricmotionoftheinnermost coils exerts a force in thedirectionof thearrowB, and, since theseforcesare inoppositedirections, there isa tendencytowardaccelerationas thearcs become shorter than 1½ turns. The maximum is reached at one turn. Inunwinding,theforcesarereversed,buttheirrelationtoeachotheristhesame.

If the spring is pinned at even coils plus half a coil, the eccentricmotionswillstandinthesamedirection—thatis,oppositetheregulatorpins.AccordingtoLaw2,thisproducesalosingrateascomparedwiththelongarcs.Sincetheforcesoftheeccentricmotionsareinthesamedirection(Figure22),thereislessresistanceordivergenceof forcesand thebalancemayvibratea little farther ;hencearetardationtakesplace.

Figure21.

Figure22.

Ifthespringispinnedatevencoilsplusone-fourthorthree-fourthsofacoil,the eccentricmotionswill stand at right angles to each other. In this case theeffectsstatedinLaws1and2willbeneutralized,andthetimepiecewillfunction

atamorenearlyisochronalrate.Ofcourse,whenthetimepieceisputtopracticalusage,themotionofthebalanceisconstantlychangingandthisfactcomplicatesresults.

Reducing the isochronal error. The superior performance of the Breguetspringintheattainmentof isochronismis thereasonfor thepassingof theflatspring. The above analysis of the flat spring would at once suggest that it ispossible to vary the isochronism bymanipulating the overcoil of the Breguetspring so as to throw the eccentric motion in some desired direction. That iscorrect reasoning; however, a spring that produces concentric vibrations willattain close enough isochronism inmost timepieces.Should extraprecisionbedesiredand themostperfectovercoildoesnotproduce thedesired results, thefollowingrulesforalteringtheovercoilmaybeused:

(1)Iftheshortarcsareslow,bringinpartofthebodyofthespringandaddittotheovercoil.

(2)Iftheshortarcsarefast,takepartoftheovercoilandmoveitbackintothebodyofthespring.

Breguetspringsarenotpracticalforthesmallestofwristwatches.Thelastcommentabovemayseeminconsistentwith thegeneralargument,

but we believe that very small wrist watches should not be fitted with theBreguet balance spring. Flat springs are still used in some of the finest Swisswatches.Theflatspringoccupieslessspace;itiseasiertohandleandadjust;andit holds its shape better. The extra precision claimed for the Breguet is lostbecauseofmechanicalinaccuraciesinthesmallestofwatches.

There is another difficulty found in very small watches fitted with theovercoil.Thebodyofthespringoftenbouncesagainstthestudwhenajartakesplace,andthusinterfereswiththetimekeeping.Sincethespaceisoftenlimited,we cannot raise the overcoil very high, so little can be done to correct thedifficulty.

Inafewwatches,theundersideofthebalancebridgecanbegroundawaytoprovidemoreroomfor theovercoil. Inourownexperience, thisprocedurehasbeenfoundpractical.

THEPOISEOFTHEBALANCEOneofthemostcommoncausesofvariationbetweenthelongandshortarcs

iswantofpoiseofthebalance,providedthebalanceliesinaverticalposition.Insomeclocks,thebalanceliesinahorizontalposition,andinthiscasethepoise

errorhasnoeffect.Ifthebalanceisvertical,thepoiseerrorworksoutaccordingtothefollowingstatements:

If theexcessofweightisonthelowersideof thebalancewhenatrest, thetimepiecewill losewhen the arc ofmotion is greater than 1¼ turns, andwillgainwhenthearcisless.

Iftheexcessofweightisonthetopsidewhenthebalanceisatrest,theresultwillbe reversedand the timepiecewillgainwhen thearcofmotion isgreaterthan1¼;turnsandwilllosewhenthearcisless.

Excessofweightbelow.Letusassumethat theexcessofweight ison thelowersideofthebalancewhenatrest.Supposethebalancevibratesatanarcofabout1turn,andinsodoingtheexcessofweightstopsatthetopofthebalance.Theforceof thespringinreturningthebalancetoitspointofrestwillreceiveaddedenergy in the formofgravityactingon theweight.Thismeans that thebalancewillreturntoitspointofrestalittlemorequicklythanwhenacteduponby the force of the spring alone. Now assume that the weight, after havingreachedthebottom,continuesthearcontheoppositeside.Theforceofgravityactingontheweightisanaddedresistancetothatofthespring;inotherwords,theresultofanaddedweightis,ineffect,thesameasifastrongerspringwereused,andthearcwillbeperformedmorequickly.

Nowsupposethatthemotionisincreasedto1½turnsand,invibratingtothisextent,theweightstartsfromitspointofrestatthebottom,turnsthree-fourthsofa circle, and stops. The force of the springwill encounter a resistance due togravity acting on the excess ofweight as it starts upward toward the top, andafterreaching the topandstartingdownward, theforceofgravityprolongs thearc.On the return vibration theweight, starting upward, is resisted by gravityuntil it reaches the top and is accelerated on its downward path. The over-alleffectisthatgravityretardsthevibrationsandthetimepiecewilllose.Figure23showsthepathoftheexcessofweightwhenthearcofmotionis1½turns.

Figure23.

Figure24.

Excessofweightabove.Nowsupposethattheexcessofweightliesonthetopsidewhenthebalanceisatrest.Invibratingtoanarcof1turn,theexcessofweight tends to prolong the vibration on its downward path, and to retard thevibration on its upward path. Thus a retardation takes placewhich is just theoppositeintheinstancewhentheexcessofweightliesbelowthebalance.

Againincreasethearcto1½turns,andagainintimetakesplace.Thiscanbe visualized in studying Figure 24, which shows the path of the excess ofweightwhenthearcofmotionis1½turns.

Neutralarcofmotion.Sinceagivenpoise error showseither againor aloss,dependingonthearcofmotion,itfollowsthatthereshouldbesomearcinwhich the poise errorwould have no effect. Such is the casewhen the arc ofmotion is slightly less than1¼ turns. Inpractice it is impossible tomaintainaconstantarc.Thelogicalsolution,therefore,istocarefullypoisethebalanceandsee that thebalance takesapropermotion—that is,amotionofnotover1½turnsorunder1¼turns.

THEADJUSTMENTOFTHEREGULATORPINSAnothercauseofisochronalerroristhecarelessadjustmentoftheregulator

pins.Infact,byslightlyopeningorclosingthepinsasthecasemayrequire,itispossibletobringthelongandshortarcsincloseagreement.

Theeffectoftheregulatorpinsmaybeexplainedinthisway:Suppose,forexample, that the regulator pins are opened slightly and the outer coil of thespring vibrates equally between the pins. We have literally made the activelengthofthespringlongerandtheclockwillgoslower.Italsochangestheratebetween the longandshortarcs.Foranarcof½turn, theratewouldbeslow,

and for 1½ turns the ratewould be fast. For arcs shorter than half a turn, theactivelengthofthespringwillcommenceverynearlyfromthestud.Now,ifthearc is increased, theactivelengthof thespringwillbeshortened,commencingmorenearlyfromthepins.Increasingthearcacceleratestherateandtheeffectwillvaryinproportiontothechangestakingplaceinthearcofmotion.

Supposenow that thepins are open, as before, but theouter coil doesnotvibrateequallybetweenthepins—thatis,thecoiltendstoleanagainstoneofthepins.Assumethatitrequiresanarcof1turntoliftthecoilawayfromthepinagainstwhichit leans.It isclearthatforarcsbelow1turntheactivelengthofthe springwill commence from the pins, and for arcs above 1 turn the activelengthwillcommencemorenearlyfromthestud.Thisconditionwillmakethelongarcsgoslower,or, inotherwords,opposite tothat intheformerinstance.Thus,itcanbeseenthattheconditionoftheregulatorpinsmaybethecauseofmanyofthedisorders intheperformanceofclocksandwatches.It isalsotruethatanintelligentmanipulationofthepinsisthequickestandsimplestmeansofcorrectingisochronalvariations.Thepracticaluseoftheregulatorpinsisstatedinthefollowingrules:

(1)Iftheregulatorpinsareclosedandthetimepiecegainsduringtheshortarcs,openthepinsslightly.

(2) If the regulator pins areopenand the timepiece loses during the shortarcs,closethepins.

(3) IfRule 2 fails to accelerate the short arcs sufficiently, bend the springnearthestudsothattheoutercoilwillstriketheinnerregulatorpinwithmoreforcethantheouterpin.

TEMPERATUREERROROFBALANCEANDSPRINGThemostdifficultproblemof thehorologistof thepastwas tocombat the

temperature error. Many ingenious devices have been tried and discarded.Perhaps the first to be reasonably successful was the one invented by JohnHarrison, who, in 1759, won the £20,000 prize offered by the Englishgovernmentforamarinetimepiece.Harrison’smethodconsistedofathinstripofsteelandasimilarstripofbrassrivetedtogetherandcarryingregulatorpinsonthefreeend.Thepiecewassecuredtotheupperplateandtheregulatorpinswerefreetoslidealongthebalancespring,makingthespringshorterinheatandlongerincold.This,itcanbeseen,providedacompensatingeffectforthevariedlengthandelasticityofthebalancespring.

Harrison’s compensating device was soon replaced by a compensating

balance said to have been invented by Pierre LeRoy about 1765 (see earlierdiscussion).Except for improvements indesign, this balancehasbeenused tocomparativelyrecenttimes.

Figure 11, page 252, shows the compensating balance for watches andportableclocks.The rim is composedofbrassand steel.Thebrass lieson theoutside and constitutes about three-fifths of the total thickness.The rim is cutnear thearmsat twodiametricallyoppositepoints.Witha rise in temperature,thebrass,whichexpandsmorethanthesteel,isresistedinitsefforttolengthenbytheresistanceofthesteel.Sincethebrassliesontheoutside,theendsoftherimmustturninward.Thisactionreducesthemeandiameterofthebalanceandlikewise its inertia, therebycompensating for the increased lengthand reducedelasticityof thebalancespring.Theopposite takesplacewhenthetemperaturefalls.ThisisclearlyshowninFigure25.Itwillbefurtherobservedthatthelooseendsremainreasonablycircularduringtemperaturechanges,buttheradiiofthecurves change, their centers being at the balance center O for the normaltemperatureandshiftingalongthearmsAA for thehighandlowtemperatures.However, the points BB, about 60 degrees from the arms, remain at a fixeddistancefromthebalancecenter,anditisatthesepointsthatalterationsforthepurposeoftimingshouldbemade.

Figure25.

Middle-temperatureerror.Itisevidentfromtheaboveanalysisthatwecanadjustthebalancescrewsinsuchamannerastocompensatefortheexpansionand contraction of the balance alone and maintain a constant mean diameter.

This, however, would not take care of the lengthening and shortening of thebalance spring nor for the changes in the elastic force.To compensate for theeffect of temperature on the spring, it is necessary to add extraweight to thelooseendsoftherim.Thisresultsinatemperatureerrorbetweentheextremesofheatandcold,knownasthemiddle-temperatureerrorbecausethebalancedoesnotcompensateequally forchanges in theelastic forceand forchanges in thelength of the spring. This is shown by insufficient compensation (weights notmoving in near enough toward the center of the balance) in the highertemperatures;andtoogreatacompensation(weightsmovingtoofarawayfromthecenter) in the lower temperatures.The result is ahigher rate in thenormaltemperature,usually2to6secondsin24hours,dependingonthegradeofthetimepiece.

Figure26.

Thisfactor isclearlyshowninFigure26.The lineAB indicates the loss intherateduetoaraisingtemperatureonthebalancespring.NotethatthelineABiscurvedwhichindicatesthatthelossinrateisnotuniformbutincreasesmorerapidlyas the temperature raises.Tooffset this ratebyacompensatingdevice,wemustproducetheoppositeeffectasindicatedbythelineCD.ThetheoreticalresultwouldbeaconstantmeanratealongthelineEF.

Figure27.

This, however, cannot be realized in practice with the ordinary brass andsteelcompensatingbalanceforreasons thatweshallnowexplain.Referring toFigure27,thelineOBindicatestheeffectofexpansiononthebrass,andOSonthesteel,whichwhencombinedasinacompensatingbalance,resultsinagain,asindicatedbythedottedradiallineOG.This isarrivedatbydeterminingthealgebraicdifferencebetweenOBandOS.Theradial lineOG indicates that theinwardmovementoftherimisindirectproportion(ornearlyso)totheincreasein temperature,whereas thevariationofelasticityof thespring isnotuniform,but,aspointedoutabove,increasesmorerapidlyasthetemperatureraises.Now,by replacing the lineOG in Figure 26, it can be seen that there would be acontinuallyaccelerated rateuntil themiddle temperature is reachedand thenalossproportionaltothegain.ActuallythiscurveisaparabolaasshowninFigure28.

Figure28.

Theaboveanalysiswouldsuggestthatthemiddletemperatureerrorcouldbereducedorperhapseliminated if the steelwere replacedwithametalof smallexpansion. Such a metal was produced by Dr. C. E. Guillaume in 1899. Hecalled themetal Invar*. Compensating balances made of brass and Invar hasreducedthemiddletemperatureerrortolessthan½secondin24hoursandhavebeenusedprincipallyinchronometers.Figure15showstheInvarbalance.

*Invarisalsousedinpendulumrods.

Alloys for balance and balance spring. Recent years have seen thedevelopmentand improvementofspecialalloysforbothbalanceandspring inwhichtheelasticityremainsconstantduringtemperaturechanges.Thesealloysaremadeofiron,nickel,chromium,manganese,tungsten,berylliumandsilicon.VariouscombinationsoftheaboveelementsareusedandmarketedundersuchtradenamesasElinvar,Corunium,Conel,Elgineum,andNivarox.

Elinvaristheoldestofthealloystobeusedinbalancesprings.ItistheresultoffurtherresearchbyDr.Guillaumeandwasproducedin1913.ThealloyhasaslightexpansionandhasbeenusedwithabalanceofasinglemetalasshowninFigure12.A compensating balancewith short bimetalic pieces has been usedwith Elinvar springs by Paul Ditisheim with excellent results. The DitisheimbalanceisshowninFigure13.

Nivarox,oneof themorerecentalloysdevelopedfor thebalancespring, issaid to be superior to Elinvar. It is the result of researches made by R.Straumann, technical director of the Fabriques demontres Thommen, S.A. Inaddition to maintaining a constant elasticity, the alloy is nonmagnetic andnonrusting.

Thus the temperature error has been solved to the great satisfaction ofhorologists.The solid balance ismuch easier to handle; it holds its shape andkeepsitspoise.ChronometermakerssinceHarrisonhaveworkedontheproblem—overaperiodofnearly200years.

Correcting the temperature error in the compensating balance. Thecompensating balance is usually adjusted to temperature between 40 and 95degrees Fahrenheit. The correction consists of moving one or more pairs ofscrewsfromonelocationtoanother.Forexample:ifthetimepiecelosesinheat,thescrewsaremovedtowardthefreeendsoftherims.Thiscarriesmoreofthemass toward the center of the balance and the inertia is reduced, thereby

effectingacompensatingactionfortheincreasedlengthandreducedelasticityofthebalancespring.If,ontheotherhand,thetimepiecegainsinheat,thescrewsaremovedtowardthearmsofthebalance.

Thechronometerbalancehasapairofmovableweightsthatmaybeshiftedalong the rims to effect temperature adjustment. Set screws secureweights torims.

THEPOISEERROROFTHEBALANCESPRINGProbably themost interestingphenomena inhorological theory is thepoise

error of the balance spring. The perfectly adjusted balance spring of a pocketwatch,placedvertically,will showpracticallynopoiseerrorwhen thebalancespring is at rest. However, at the moment the spring starts winding up, orunwinding,around thecollet apoiseerrordevelops.Thenatureof theerror isaccordingtothefollowingstatements:

Figure29.

(1)When the middle of the first half of the innermost coil of the balancespringtendsupward,thewatchwillgain.

(2)Whenthemiddleof thefirsthalfof the innermostcoil tendsdownward,thewatchwilllose.

(3)Whenthemiddleofthefirsthalfoftheinnermostcoiltendstotherightortotheleft,thewatchratesbetweentheextremesofthefastandslowpositions.

Analysisoftheclockwisemotionofthebalancespringwhenthemiddleofthefirsthalfcoilisup.InFigure29,areshownthreeillustrationsofthefirstcoilofthebalancespring.AtAthespringisshownatrestandthebalancespringfits quite closely within the dotted circle which is concentric to the balancecenter. The illustration B shows the same spring after the balance has beenturnedclockwise1/2turn,thuswindingupthespring.Itwillbenoted(atB)thattheportionofthespringoppositethepiningpointhasbeenpulledinclosetothecollet, creating a conditionwhereby the radius from the balance center to the

piningpoint is longer than thesideopposite.Gravityactson thesidewith thelongerradiusandthevibratingunitisretardedduringitsclockwisemotionandthe balance arc is shortened. Assume now, that the balance continues the arcanother 14 of a turn, as shown at C, and the pining point comes to a stop,momentarily,directlyabove thecollet. In thiscasegravityconfines toactasaretardingforceontheclockwisemotionofthepiningpoint,butatthemomentthebalanceturnscounterclockwise,gravityacceleratesthedownwardmotionofthe pining point. The over-all result is such that the time of vibration isquickenedasgiveninstatement(1).

Counterclockwisemotionofthebalancespring.Whenthepositionofrestis reached, as at D, Figure 30, the balance spring is under no tensionmomentarily.However,asthebalancecontinuestoturn,thespringisunwound.AtE thebalancehasmade½turn,counterclockwise,andatF thebalancehasturnedanadditional¼ofaturn.Notethatthespringnowhasalongerradiusonthesideoppositethepiningpointandthegravitationalpullhasbeentransferredfrom the pining point to the opposite side.The over-all result is such that thevibrationcontinuestobequickened.

Analysisoftheclockwisemotionofthebalancespringwhenthemiddleofthefirsthalfcoilisdown.Assumethatthespringisatrestwiththemiddleofthefirsthalfcoillyingbelowthecollet,asshownatG,Figure31.Windingupthespringclockwise¾ofaturn,itwillappearasshownatH.Thelongerradiusisonthesideofthepiningpoint,andasgravityactsonthisportionofthespring,itisseenthatgravityhasprolongedthearcduringthedownwardmotionoftheheavypoint.Onthereturnvibration,gravitywillresisttheforceofthebalancespring. The over-all effect is that the time of vibration is extended aswill benotedinstatement(2).

Figure30.

Figure31.

Theusual difference between the fast and slowpositions is from15 to 30seconds in24hours,dependingupon thegradeof thewatch.Adjustersof finepocket watches and railroad grades generally adapt the balance spring to thewatch in such amanner that themiddleof the first half coil around the colletstands in the direction of pendant up. The reason for this is that the verticalpositionsusuallyrunslowerthanthedialupanddialdownpositions.

BIBLIOGRAPHYBritton, F. J., The Watch and Clock Makers’ Handbook. Brooklyn, Chemical

PublishingCo.,1938.Britton,T.W.,HorologicalHintsandHelps.London,TheTechnicalPress,Ltd.,

1942.Chamberlin,PaulM.,It’sAboutTime.NewYork,RichardR.Smith,1941.Gamm,F. J.,Watches,AdjustmentandRepair.Brooklyn,Chemical Publishing

Co.deCarle, Donald, Watchmakers’ and Clockmakers’ Encyclopaedic Dictionary.

London,N.A.G.Press,1950.deCarle,Donald,PracticalWatchRepairing.London,N.A.G.Press,1946.Drepperd, CarlW.,AmericanClocks. Garden City, N. Y., Double-day&Co.,

Inc.,1947.Fried,HenryB.,TheWatchRepairer’sManual.NewYork,D.VanNostrandCo.,

1948.Fried, Henry B., Bench Practices for Watchmakers. Denver, Colo., Roberts

PublishingCo.,1954.Gazeley,W. J.Watch and Clock Making and Repairing., New York, D. Van

NostrandCo.,1953.Goodrich,Ward,TheModernClock.Chicago,NorthAmerican,1950.Gordon,G.F.C,Clockmaking,PastandPresent.London,Crosby,Lockwood&

Son,1928.Hope-Jones,Frank,ElectricalTimekeeping.London,N.A.G.Press,1940.Huyghens,Christian,HorologiumOscillatorium.Paris,1658.Kelly,HaroldC,APracticalCourseinHorology.Peoria,Ill.,Chas.A.Bennett

Co.,1944.Levin, Louis and Samuel,Practical Benchwork forHorologists. LosAngeles,

Calif.LouisLevin&Son,1944.Langman andBall,ElectricalHorology. London, Crosby Lockwood and Son,

1930.McCarthy,J.R.,AMatterofTime.NewYork,Harper&Brothers,1947.

Milham,WillisI.,TimeandTimekeepers.NewYork,TheMacmillanCo.,1942.Player,J.W.,WatchRepairing.London,CrosbyLockwood&Son,Ltd.,1945.Rawlings, A. L., The Science of Clocks and Watches. New York, Pitman

PublishingCo.,1948.Seibel,Emanuel,ComplicatedWatches.Denver,Colo.,RobertsPublishingCo,

1945.Wilkinson,T.J.,TheEscapementandTrainofAmericanWatches.Philadelphia,

Pa.,1928.Yamaguchi,Ryuji,TheClocksofJapan.Tokyo,NipponHyoron-ShaPublishing

Co.,1950.

INDEX–WATCHREPAIRbyKelly

A

Addenda,123Adjustingescapement,72isochronism,262,267position,89regulatorpins,267temperature,274

Air,itseffectonpendulum,237Alarmclock,42Alarmclockescapement,44Alloysforbalanceandspring,273Americanmantelclocks,107Anniversaryclocks,39Automaticwatches,184

B

Balanceassemblyrepairs,61Balanceformula,253Balancemotion,258BalancespringBreguet,260cylindrical,260flat,260generaldiscussion,259isochronalerror,262

poiseerror,274temperatureerror,269truingof,65

Balancestaff,62Bankingtothedrop,76Barometricerror,236Barrelcap,fittingof,55Barrelhook,fittingof,54Beatchronometerescapement,169cylinderescapement,163duplexescapement,166leverescapement,161puttingescapementin,72

Bench,26Breguetbalancespring,260Bushings,fittingof,34

C

Calendarmechanisms,221Cannonpinion,85Chimes,179Chimingmechanism,181Chronographs,204Chronometerescapement,166Circularerror,240Circularpitch,128Civiltime,14Cleaningwatches,46Clockcleaning,29,42escapementrepairs,35,44

escapements,132mainspringproblems,32trains,103

Closingholes,watches,56Compensatingbalance,269Compoundpendulum,231Cornersafetytest,78Countingwheelmechanism,171,174Curvetest,78Cycloidalcheeks,243Cylinderescapement,161Cylindricalbalancespring,260

D

Daylightsavingtime,19Dead-beatescapement,141Dedenda,125Denisonescapement,145Dialtrain,85,113watch,86

Draw,73Drop,74Duplexescapement,164

E

Earthasaclock,9Elinvar273Engagingfriction,126Epicycloid,123Equationoftime,15

Equinoxes,11Errorcircular240isochronism262,267position89,265,274temperature,269

Escapementadjusting,72chronometer,166cylinder,161dead-beat,141Denison,145duplex,164errors,81,236forkandrolleraction,76Galilei,134generaldiscussion,132,153gravity,145Leroy,150lever,153Mudge,155pin-pallet,144pin-wheel,143racklever,156ratchettooth,156recoil,139Reifer,148terminology,133verge,136wheelandpalletaction,73

Eyeloupes,24

F

Flatbalancespring,260Frictionjeweling,57Frictionofair,239Fulldiameter,127

G

Galileiescapement,134Gearing,122Goingbarrel,119Graham’spendulum,245Gravers,23Gravityescapement,145Grindingpowders,25Guardsafetytest,78

H

Hairspring,259Handsfittingof,86repairingof,37

Harrison’spendulum,246Holes,closingof,33,56Huyghens,138

I

Internationaldateline,19Inertiaofair,239Invar,273Invarpendulum,247Involutegearing,130Isochronism

adjusting,262,267balancespring,262

J

Jewels,fittingof,57

L

Lanternpinions,129Leroyescapement,150Leverescapement,153toolong,82tooshort,82

Lift,74Lockingtoodeep,81Lockingtoolight,81

M

Magnetism,87Mainspringclock,32correctlength,121correctthickness,120gauges,53watches,53,120

Maintainingpowermechanism,111Maintrainspring-drivenclocks,106watch,56

Meantime,14,15Middletemperatureerror,270Motionofpendulum,232

Motorbarrel,119Mudgeescapement,155

N

Nivarox,274

P

Palletframeclearance,83Pendulumformula,234fourlawsof,232Harrison’s,246Invar,247motionof,232terminology,230zinctube,247

Pin-palletescapement,144Pin-wheelescapement,143Pitchcircle,122diameter,126

Pivoting,34Pivots,polishingof,60Poiseerrorofbalancespring,274Poisingbalance,66Polishingpowders25

R

Rackandsnailmechanism,175,177Rackleverescapement,156Ratchettoothescapement,156

Recoilescapement,139Regulatorpins,adjustmentof,267Reiferescapement,148Removingbrokenscrews,56Repairingcomplicatedwatches,224Repairs,balanceassembly,61Rollerjewel,64

S

Screwdrivers,22Self-windingwatches,184Shipbellstrike,178Siderealday,11Simplependulum,231Slide,83Solarday,13Splitsecondchronograph,213Stem,fittingof,55Stopwatch,200Strikingmechanismsgeneraldiscussion,171repairof,36

T

Temperatureadjusting,274error,244,269

Tick,watch,84Timereckoning,8Timezones,18Timingerror,correctionof,257

Train,repairingof,33,56Truingbalance,65balancespring,67

Tweezers,23

V

Vergeescapement,136

W

Watchcleaningmachine,53escapements,153raterecorder,89

Watchescleaningof,46thatstop,96

Weight-drivenclocks,104Weightofair,239Weightofpendulum,240Wheel,fittingteethto,35Wheelworkclocks,102watches,116

Workingconditions,26Workshop,26

Z

Zinctubependulum,247

ALSOAVAILABLE

ClockRepairingasaHobbyAnIllustratedHow-toGuidefortheBeginnerbyHaroldC.Kelly

Allofthepreciseanddelicategearsandleversinaclockcanappeardizzyingto

the amateur, but this guide, with its wonderfully detailed diagrams, can clearthingsup.Itcoversthetheorybehindclockmovementanddesign,includingtheparticulars of escapements, pendulums, balance wheels, and even the sheetmusicforpopularchimes.TheAmericanstrikingclock,the400-dayclock,andthe alarm clock receive special attention. The in-depth information, includingexplanationsofclock repairing terminologyanddetailson the tools,materials,andsuppliesthatareneededforsuccess,willbenefitevenskilledenthusiasts.

$12.95Paperback•128pages

ALSOAVAILABLE

TheClockRepairer’sHandbookbyLauriePenman

LauriePenmanhaswrittenanindispensableguideforboththeabsolutebeginnerandtheexperiencedclockenthusiast.TheClockRepairer’sHandbookprovidesinformation on how to repair and maintain a clock’s delicate mechanics and

teachesthebasicsofclockrepairingthroughdetailed,easy-to-followinstructionsandmorethan300instructivediagramsandillustrations.Itincludesadviceanddirections for cleaning clock movements, pivoting and mounting, fixing trainfaults and gears, the importance of lubrication and friction, and how tomakesure thestrikeandchimesworkon thehour,everyhour.TheClockRepairer’sHandbook provides all the necessary information to troubleshoot any clock’sproblems and to make sure your clock continues to run in perfect order forgenerationstocome.

$14.95Paperback•176pages

ALSOAVAILABLE

PracticalWatchRepairingbyDonalddeCarle

Theabsorbingandeverlastingsubjectofwatchrepairinghasbeendealtwithinbooksinmanylanguagesthroughouttheyears.ButwhendeCarlefirstsetouttowritePracticalWatchRepairingin1946,itwaswiththeintentionofcreatinga

textbook“thatawatchmakercanunderstand,evenifhecan’tread.”Withover550instructionalblack-and-whiteillustrationsandanapproachthatassumesnoprior watch-repairing experience, this book achieves and surpasses that loftymission,andhasbeentoutedas“thebestillustratedbookonpracticalhorology”(HorologicalJournal)everwritten.Forthereadersinhisaudience,deCarlehasprovided well-informed discourse on every topic a watchmaker, or aspiringwatchmaker,needstoknow.WithPracticalWatchRepairing,evenalaymancanbecomeawatch-repairspecialist.

$14.95Paperback•336pages

ALSOAVAILABLE

Abbott’sAmericanWatchmakerAnEncyclopediafortheHorologist,Jeweler,GoldandSilversmithbyHenryG.Abbott

Containing hundreds of private receipts and formulas compiled from themostreliable sources, this encyclopedia is a comprehensive, masterfully illustrated

guide to traditional watchmaking and jewelry fraternity. Inside are completedirections for using all sorts of tools, attachments, and devices of the trade.Among hundreds of helpful diagrams and precise illustrations are thoroughexplanationsofadjustments;directionsformakingallkindsofalloys;areviewofalltheescapements,theiraction,construction,andproportion;anexhaustivetreatise on balances, their expansion and contraction, auxiliaries, sizes andweights, anddirections forpoising; instructions formakingand replacingnewstaffs;andmore.ThisdistinctivelyAmericanapproachtowatchmakingremainsaclassicyetusefulreferenceforbothbeginnerandskilledwatchmakers.

$16.95Paperback•386pages