US2142865A - Method of manufacturing filaments - Google Patents

Description

' Jan. 3, 1939. w. p. ZABEL v 2,142,865

METHOD OF MANUFACTURING FILAMENTS Filed Marbh 24, 1957 Figl.

/ F/L AME/v7" WIRE 8r CO/LING OPERATION MANDRE'L HE) T TREA TME/VT a 2 CO/L/NG OPERA r/o/v F/LAMENT SEVERED A/VD vJECO/VDARY MANDREL W/ THDRA W/V r ///,,M,,/,,,,?, I

HEAT TREATMENT Fig. 7.

OPERATED MN:

I Inventor: LAMP William P. Zabel,

His A torney.

raw & M9 9 7 incandescentlamps and?moreparticularly to a "ments preferably of the'double (coiled coil) and.

the triple" (coiled. coiled-coil) helical 'type.-

nn z'rnon on MANUFACTURING V I. WilliamPQZabel, Cleveland Heights, flhiog-assignv a v or toGeneral Electric Company, acorporation of New York f Applicant March 24, 1937,.Serial No. 132,763

'ff'soaims. (01. 175-40) invention relates to filamentsfor electric process of forming and treating multiple coil fila- One of the-objects of mylnve'ntion is to produce double and triple helical-filaments of greater uxii- 'formityw Heretofore; the filaments varied some- 1 what in length and '-'were' often' distorted; 'In both cases the direct cause of thedefectcould not be t'ra'ced -to any particular -part-of the manufac- "turing processflargely because the'defect was likely to' appear --in 'an'yof theimanufacturing steps or even'after thefil'ament-was operated in a lampi 'The defects wer'e believed-=to-be-princi- --pally theresult ofchangescausedby the relieving of internal strains within the filament, which were produced therein by the particular coiling method used and were -not properly relieved :by

the-haflt'reatment: In some cases the coiling operation alone' caused immediate distortion of the filament coils. My process obviatesthese disadvantages by providing a more satisfactory coiling "method andalsoaheat treating method for relieving the strains produced in the wire by the saidcoiling method.

Conventional coiling methods heretofore employed have usually consisted in coiling the filament wire on a mandrel of sufficient length to be spooled, coiling the mandrel and filament coiled thereon about a second mandrel and, 'if the filament was to be of triple helical form, coiling both mandrels and twice coiled filament thereon about a third mandrel. With this method even the twice or thrice coiled filament wire, as the case may be, was coiled in a considerable length which was heat treated and then severed-with all mandrels remaining in the filament coils. The mandrels 'were finally removed by dissolving in an acid.

The conventional method of coiling is objectionable because of the length of the coiling operation since local strains are produced in the filament ;by variations in the coiling tension which gradually build up until relieved by the twisting of both the filament and the mandrel. Such strains and twists appear to be inherent in this method of operation and do not seriously affect .filament qualityvexcept in the final (second or third) coiling operations where the distortion is of greater magnitude,

Conventional heat reating consists in passing the filament and mandrel combination through a furnace at a speed which provides the desired treatment and may also consist of other treatpossible before.

ments, in which len'gths-of-said -iilament"'and mandrel combination equivalent to"several filaments (ten for instance) 'are placed in -a boat in a furnace for the desired length of -time. In

such treatmen'ts-themandrels must -be left in 5 the coils of the. filament wiretokeep-themin position andtherefore must be-considered in the heat treatment. -,The final-mandrel in multiple -..:coil filaments is of considerable sizeandi because of its mass and position within air-irregular filal0 ment; is; effective in ch'angingthe' rate and uni- -formity of heatingezand'tcooling. The final man r drel. is.a.lso of-adifierentmaterial thanthe fllament and is apt to contaminatessaid filament ment is impractical because of the larger'amount of unwinding and distortion takingpla'ce: =I'hese heat treatments take ,a relatively long time in comparison ,to my method and,- when :the' fila- '20 ments are piled one on the othenarie'objectionable because of the pressure-and restrainin 'eifect of the-filaments .on each other." w

.My process on theother hand is particularly well suited to produce more stable and uniform 25 filaments and permits manufacture of filaments having more desirably shaped end sections than The uniformity appears as an increase in consistency in the number of coils and the length of wire turn by turn (1. e. equal weight) 30 of the filaments..

According to my invention the filament wire, which is preferably of the so-called non-sag variety, is first coiled about a comparatively long mandrel preferably at a uniform pitch and, 35 if the filament is to be of the triple helical type,

- is again coiled with the mandrel about a second mandrel. Both these operations are performed in the usual manner and result in the filament wire taking a given shape and developing strains 40 and a tendency to straighten out an appreciable amount. Accordingly the filament and mandrel assembly are heat treated in-the conventional way to remove a considerable portion of the .said strains and the tendency to unwind. From this 45 mandrel and filament coil assembly individual coiled-coil or triple coiled filaments are made one after another successively. The operations in the manufacture of coiled coil filaments are, first, to lay a portion of the mandrel and filament coil 50 assembly parallel to a second temporary mandrel, to coil a succeeding portion of said assembly about said temporary mandrel at a given pitch, to draw out a still further succeeding portion of said assembly beyond the end of the temporary --.porary mandrel; -In--the.-= mamifacture. of triplecoiledfilaments. the operations are the same ex-.'

Wil'afidliiihh'the first and second manis thenjhe mandrel. The end-portions of I, --;tlie;filanientsf;iorm legs which'lie parallel to the 0' axis ,of'jthe filament and which are therefore well 1' coiled on wthetemporary mandrel which adapted to be attached to the leading in wires K of alamp. This shape of leg produced by theconventional methods 90f manufacture. The limited time of the coiling -operation .pre-. vents coiling strains from building upto a'point where they are detrimental and allows the recoilto loosen the whole filament cell on the'mandrel mandrel, to sever ,thesecond' straight portion now subiected toa heat treatment, as shown in ept'that an.asscmb1Y.-including the twice cniled' I 'pi' the original drawn mandrel and original drawn- Fi'g. 3;} designed to remove a considerable portion filament wire strains aswell as later coiling strains therein and to partially clean it. The treatment is preferably divided into two sections, the first being for ab out twelve seconds when the mandrel andthfilainenhwire assembly i2 is [passed ,throug h'la furnace l3 containing wet or humidified hydrogen**a'hd retained at 14o0 0., and the 'secondfbeingffor; about sixteen seconds when said assembly 12 is passed through a second r furnace? -'cjontaining substantially dry (dehy- -dnite cv hydrogen and retained at 1500" to 1700" so that it can be withdrawn; 'The recoil .is aiIect- .I

ed by physical characteristics of the mandrel'used in the first coil and also alterationof these characteristics in subsequent heat treatmentof both the filament coil and mandrel assembly. *The individual filaments are then heat.treated to cause them to be set in the coiled coil or triple coiled form whereupon the .other fixed mandrels are removed, preferably by dissolving in 'acid. The

filaments are then mounted on the leading-in.

C. "Ifhe'furn'aceHIS illustrated is provided with an' electricaljheating element It. In the first section of the treatment, the moisture in the hydrogen combineswlth the graphite-of the coating to partially clean the filament wire and mandrel be theoperations including thefirst heattneat- -ment, the winding of the mandrel and filament wire assemblyon the: final (temporary!- mandrel and the removal of saidfinal or temporarymantreatment)-- is done r i mandreliiisthethird mandrel when triple coiled drelbefore further work (heat on the filamentwire.

v The advantages of my process of manufacture will be more readily understood fromtthe detailed description which follows of one specific appli' cation of my process and from the drawing. The drawing illustrates diagrammatically in Figs. 1 to! inclusivethe various steps in the process of my invention. r

Patent 1,410,499 of March 21, 1922. Uncleaned wire is that still retaining the coating formed by the graphite, etc., lubricant used in the drawing process. Because of the lubricating value of'the coating incoiling the wire, and the expense of removing it, said coating is preferably left ereon.. The filament wire I0 is coiled about a wire mandrel H as shown in Fig. 2, said mandrel ll being four mils in diameter and of sufiicient length to be conveniently spooled. The filament wire 10 in thisinstanoe is coiled at a pitch of 335 turns per inch on a mandrel l I which is preferably made of molybdenum or some other material of which 7.: the filament wire and mandrel assembly are again To show specifically all features of or the uncleanedv'arietm If non-sag, the-wire. i0 is preferably made according to the Peer:-

as well as removes a portion of the strains. the secondsection of the treatment the heat re moves a considerable portion of the strains in" the filament assembly" l2. Both-sections of. the treatment are not always required in which case cnly;the second section thereof is used. This is particularly true-when the cleaned filament wire .Ilis used although for the uncleanedwire wet "hydrogen --is"*preferred. Dry hydrogen can be used :in both-heat treatments if desired.

; .-After the heat treatment :the mandrel and filame'nt wire-assembly I2 is coiled .on a-second tem- -p rary-mandrel is, as shown in Fig. 4. This filaments are-being formed. The operation is preferably performed by a machine of the type shown in Illingworth Patent No. 1,771,927 of July 29, 1930, and is an intermittent-' -oparation interrupted-atintervals while straight end portions of-the filament are produced and while the completedfilament is severed from the-remainder. As shown, one end of the mandrel and filament wire assembly i2 is first gripped between the jaws "Sand I1 whereupon the spindle l8 carrying the mandrel and filament wire-assembly i2 is rotated about the axis of the second mandrel l5 as said mandrel l5 and jaws l6 and I1 are moved at the desired rate away from the spindle iii. The mandrel, and filament wire assembly l2 passes through an off-center hole in the spindle l8 and is wound around the temporary mandrel I5 by -the rotation of said spindle l8. The movement of -tliemandrel l5 and the jaws IG'and ll away from the spiiidle It causes the windings to be advanced along the mafidrel l5. In this instance the jaws i6 and I1 hold 168 mils (4 mm.) of the mandrel and filament wire assembly l2, and the mandrel I5 is approximately 15.3 mils in diameter. The coiling 'is done at a pitch of 76 turns per inch but onlygil. turns. are placed on the mandrel l before the end of the mandrel is reached and rotation of the spindle I8 is stopped. A straight section l9 (Fig. 5) substantially the equivalent of two of the 168 mil end portions is now drawn out whereupon all motion stops.

An individual coiled coil filament in the .instance illustrated is now completely formed and, as shown in Fig. 5, the straight portion I9 is severed by the knives 20 and 2|. The knives 20 and 2| are so positioned with respect to the coiled coil portion of the filament as to leave on that end a straight leg 168mils in length, the equivv the resistancecoil 24 of'the furnace 25. The coil" ale'nt of the straight leg at the other end of said 3 0mm. The legs are both parallel to the lonkitudiual axis ofthe filament and can now be bent to som -ether'ah'g eurdesired. The temporary 1 mandrel l5 iS-theh moved to the leftto withdraw it from the filament and jaw 1'6 whereupon the jaws I6 and I! are separated, thereby freeing the filament 22 and allowing-it to drop out of posi-- tion. A new filament is started by movement of jaws l6 and I1 toward the spindle l8 which places the newend of the mandrel and filament wire assembly l2 between the said jaws l6 and I1. 'Ihe'mandrel I5 is also moved to the right and ,enters the aperture in the spindle l8 berore' the coiling operation begins. The filaments released-from themachlne are very uniform in length, have equal straight legs at both ends parallel to the" axis of the coiled coil portion and have expanded so that the latter coils now have a pitch of 69 turns per inch. The unwinding of the latter coils is not of serious consequence as it has occurred evenly from end to end. when the final, or temporary mandrel in this case, is

not removed, as in the conventional'method of operation, only'the end portions unwind and the coils are set in this position in the heat treatment. The unwindingof the filament wire on the fixed mandrel is negligible since it effects" such a very small "portion of the total amount of filament wire.

Thefilaments 22 are heat-treated again, this time to set the coiled coil or triple'colled'portion, as the case may be, and toremove further strains.

The "treatment, as shown -in Fig.-6, preferably Wakes place in' apparatus-whlchtreats said filaments individually and which isthe subject mato! myco-pending application Serial 'No. 133,854. filed'March 30,1937. In this'apparatus the filaments are placed in a tungsten boat 23 of exceedingly small mass and are inserted in "24 in this instance is heated to 2000-2 200 C. by the passage of electricity therethrough and heats the filament 22 and. boat 23 which areeremoved therefrom after 5 seconds. Tungsten-is the ma terialpreierred for'the boat since it cannot -con-.

taminate the filament and withstand the high 1 temperatures. of thetreatment. A steady ofa non-oxidizing gas, in this case substantially dry hydrogen, is introduced into the furnace 25 through the pipe 26 to prevent oxidation of the filament 22, the resistance coil 24, the boat 23 and other parts heated by the 5811110011 24/ It' presumed that a. portion of the total time of the treatment is taken to bringthe filament 22 tothe maximum temperature thereof because of the -small mass of the boat 23 and its proximity to the heating coil 24. Upon being withdrawn from the resistance coil 24,'one second is allowed for the filament 22 to-cool whereupon it is removed from the .furnace 25.

With difierent sizes and kinds of filament wire it is desirable to change the treatment accordingly and generally a treatment for between three and five seconds in a furnace at a temperature between 1900-and 2300 C. is sufilcient for most filaments of the smaller conventional sizes. The temperature of the treatment is selected from the behavior of the. filament which becomes brittle if too high a temperature is used and is not sumciently relieved'of its strains if the temperature is too'low. The optimum temperature is a compromise and is considered to be that which pro duces the most satisfactory filament both for performance and for commercial lamp production 'resistance coilf-i s betw n u If lar'ger a ran e-swe ters alifl am n r treated slightly'higher teniperatures are often much as 30 seconds? methods. The temperatures indicated were measured by an optical pyrometer sighted through a window in"the furnace wall and onto the resistance coil 24 and have been found to be about 225 C:"high(=.r than that reached by the boat 23 and the filament 22 and the truest ob- .tainable. In measuring the temperature or the resistance coil 24, the readings are corregted for the transmission factor or the window; .Diihculty is experienced in, making (absolute. 'teniperature measurements becauseipf the 'highemissivity of the tungsten coil 2'4"; tungsten boat '23 iand' -filament assembly 12 which leads "to'consider'able error in apparent temperature readings. comparison to black body conditions. Thej225 difterence in temperature between the'resistance .coil 24 and the filament 22-was foundby experipears'that m sojwattinlam assemhlyjlsonly heated to btw' a, 72* W and 2200'- c.

required and fine s? 11 ea ened na The wattage 'ofel'e rificoiwdiiiedby'tnefresistance coil is. usedto indicatef'the temperature of the treatment afterjthe -'lnitial"temperaturereadings are taken." tieatment the filament 22 contractsslightlyso that theme] coils are now at substantially 69% turns per inch.

This change in filamentlehgth hasfbeen round to vary considerablyg'in the acceptable range of v the heat treatment anditIhasjbeenIoiind possible to alter the len'gt'hlofthegfilament somewhatby changing tne"--temperature'- ot jflie' treatment.

Higher temperaturescause aigreater increase in the length of the filament.

The fixed mandrel"! r'both fixed mandrels as :the case may be, are nowaremoved whereupon the filament 22 is mounted on the inner structure of the lamp. The mandrel liispreferably removed chemically by dissolving it in acid whereupon the filament 22 is rinsed and dried in the usual way. During this operation the filament contracts until the pitch is 70 tumsper inch which shows that not all of the strains are removed in the particular instance by the heat treatments. Photomicrographs of the filament at this point show that very little if any change has'occurred in the wire during heat treatment [since it is still fibrous. The structure of the molybdenum mandrel on the other hand is no longer fibrous but indicates considerable development of grain size approaching that of single crystals.

The filaments 22 being either of coiled coil or triple coiled construction are very concentrated and, being treated according to my method, are simply mounted on the lead-in wires 21 and 28 of the lamp shown in Fig. 7 and are preferably engaged by one intermediate support 29. The straight singly coiled legs of the filament 22 are clamped in hooks at the ends of the leading-in wires 21 and 28 which are designed to hold said filament 22 when cold under slight tension to oi the lamp, but I havefo'uhd that once d i afi s t mblie t' 's ond temporary mandreh' temporary -maridrel before! further. ent and ,lmmedlately time to remove therein res ti 'g as; the second coiling-operalion and to set the filament; and removingthe v .sdrawing: the second. temporary mandrel before chiefly by the coiling operation; repeatedlyicoil- 4 facilitate uniform mounting. Aiter mou nting. the filament 22 is found to have 69 turns pefr inch in the final coils.

The lamp is nowcompletely assembled whereupon the filament 22'is permanently set and whatever getter is introduced the'reonis driven at normal and. abnor- J chiefly bythecoiling operation; forming from said mandrel-andfilament assembly a plurality oi individualcoiled filaments with straight end portions by winding theintermediate portion around a second temporarymandrel and severing theone end-portionfl from v the remainder of said off byoperating' the lamp v assembly, i x nmediately withdrawing saidsecond malvoltages. ior' very short lntervals oi time. .,temporary ma nd relbefore juriher heattreat- During. this process the filament a 'ain changesl ';ment; .heat treating the-filaments to remove a shape slightly'and in quantity production the secconsiderable amount oi the-strains produced by nd coiling is found to have apitch'of between 581 --i;h esecond coiling andto-set the filaments; and and 62 turns per inch. Th s r 'i e ofdlfierence 3 removingth first -mandr el -from;said filaments. is acceptable o m rcial lamp production 5. Inc processof malrlng multiple colltungsten since the changes in thefilaments are more genfilaments.,which consists in winding a filament eral thanbeiore. and are consequently of much wire arounda mandrelpheat treating-the manless importance. Obviously there. is. no means drel and the filament wire-assembly to remove a for calculating the amount of change which takes considerable part of: the .strains= therein producted-chiefly by the calling operation; coilingthe place in the filament 22 duringthe manufacture the -demandrel andgthe filament wire assembly. about s'ired final conditions .are reached by trial, my a second temporary mandrel and immediately process assures that these same resultsjare sewithdrawing thesecond temporary mandrel becured regularly within a practical range. I tore-further heat treatment; heating the first What I claim as new and desireiJto secure by mandrel --and.the; filament ,wire assembly to a Letters Patent of the United 'Statesis'; 3 temperature-Miapproximately 1675 to 2075'. C. --,ior iromthree to five 'secondsin a non-oxidizing ;atmosphere.=- to.=remove a considerable portion of 1. The processf oi nialsing multiple ,'cc il 1 'filameats which consists in windings. filament wire. arou en eli, heat eat se t maqd and the filam "t wire hsemhly to} siderableQ p .-6 .-;Ci-he process of making 1 mean gless chiefly by the .cblllihg' operatioh; coiling lthemanamiable-a tun sten withdrawing ,the second v 'heat' -tjreating the first mandrel andthe' filament wirea'ssembly a second a considerable .p'art of the strains .wirmaroundia zmandrelza heat treating the man zdrelfand the-filament wire assembly to remove a considerable part ofthe3strains therein produced chiefly-byrthe-coiling operation; coiling the man- 10nd temporary -mandre1:-and immediately withfirst mandrel iromthe filament coil. e 2. The process of making multiple .coil fila- .iurther heat treatment: heating the first m'anments which consists in wind ng a fil men Wire dreland the filament wire assembly .toa. temaround a ma r t e sit e ma d e perature;of approximately11775 to 1975 c. for d he filament W assembly i i fw h onabout five seconds in:a non-oxidizing atmosphere siderable partof the strains therein produced to remoye-a-considerable portion of the strains thereinresulting-irom the second coiling operation andto set .the filament; and removing the first mandrel irom the'filament coil.

- 7. The process oi making multiple coil tungsten filaments which consists inwmding a filament ing a portion of the mandrel and the. filament wire assembly about a second temporary mandrel, severing it to form individual coiledcoil filaments, and immediately withdrawing the sec- 0nd temporary mandrel before further heat treatment; then heat treating said filaments to set them and finally removing the first mandrel from the said individual filaments.

3. The process of making multiple coil filaments which consists in winding a filament wire around a mandrel; heat treating the mandrel and the filament wire assembly to remove a considerable part of the strains therein produced chiefly by the coiling operation; repeatedly coiling a portion of the mandrel and the filament wire assembly about a second temporary mandrel, severing it to form individual coiled coil filaments and immediately withdrawing the second temremove. a considerable portion of the strains porary mandrel before further heat treatment; therein produced by the second coiling operation heat treating each of the said individual filaand to set the filament; and removing the-first ments separately to remove a considerable part mandrel from the filament coil. of the strains therein resulting from the second coiling operation and to'set the filaments; and removing the first mandrel from the said individual filaments. I

4. The process of making multiplecoil filaments which consists in winding 9. fllamentwire around a mandrel; heat treating the mandrel and the filament wire assembly to remove a considerable part oi. the strains therein produced and the filament wire assembly to approximately 1500 to i700 C., for about sixteen seconds in a non-oxidizing atmosphere to remove a considerable part of the strainstherein produced chiefly and the-filament wire assembly about a second temporary mandrel and immediately withdrawing the second temporary mandrel before further heat treatment; heating the first mandrel and the filament wire assembly to a temperature of approximately 1675 to 2075 C. for from three to five seconds-in a non-oxidizing atmosphere to filaments which consists in winding a filament wireabout a mandrel; heating the mandrel and the filament 'wire' a'ssembly to approximately 1500 to 170Q'. C. .for about sixteen'seconds in a nonoxidizing atmosphere to remove a considerable part of the strains therein produced by the coiling operation; repeatedly coiling a portion of the mandrel and the filament wire assembly ab ut the strains therein resulting from the. secondfilamentsawhichziconsistsin; .winding -afilament drel and thefilamentwire' assembly about a sec wire around a mandrel; heating the mandrel by the coiling operation; coiling the mandrel 5 8. Theproeess of making multiple coil tungsten a second temporary mandrel, severing it to form individual coiled-coil filaments and immediately withdrawing the second temporary mandrel before further heat treatment; heating the individual coiled coil filaments to approximately 1775 to 1975 C. for about five seconds in a non-oxidizing atmosphere to remove a considerable portion of the strains therein resulting from the second coiling operation and to set the filaments; and removing the first mandrel from the fi1a ments.

9. The process of making multiple coil tungsten filaments which consists in winding a filament wire about a mandrel; winding the mandrel and the filament assembly about a second mandrel; heating the mandrel and the filament wire assembly to approximately 1500 to 1700 C. for

about sixteen seconds in a non-oxidizing atmosphere to remove a considerable part ofthe strains therein produced chiefly by the coilingwoperation; repeatedly coiling the mandrel and the filament wire assembly about a third temporary mandrel, severing it to form individual triple coiled filaments, and immediately withdrawing the third. temporary mandrel before further heat treat-- ment, heating the triple coiled filament assembly to approximately 1775 to 1975 C. for about five seconds in a non-oxidizing atmosphere to remove a considerable portion of the strains therein produced by the third coiling operation and to set the filament; and removing the first and second mandrels from said filaments.

WILLIAM P. ZABEL.

CERTIFICATE OF CORRECTION.

Patent No. 2, lh2,8 65.

January 3, 1959.

WILLIAM P, ZABEL. It is hereby certified that error appears in the printed specification of the above numbered patent requiring correction as follows: column, line 52, claim 1, after "mandrel" sert instead the words and immediately; line 5h, "and immediately" and insert instead asemicolon; ters Patent should be read with this correction therein that Page h, first strike out the semicolon and insame claim, strike out and that the said Letthe same may conform to the record of the case in the Patent Office.

Signed and sealed this 21st day of February, A. D. 1939.

(Seal) Henry Van Arsdale. Acting Commissioner of Patents.

Images