US1992797A - Method of and apparatus for treating filaments - Google Patents

Description

Feb. 26, 1935. P ABEL 1,992,797

METHOD OF AND APPARATUS FOR TREATING FILAMENTS Filed Dec. 31, 1932 [N VENT 0R. WILLIAM F ZABEL BY MMVZM tend to twist and warp out of shape.

machine.

Patented Feb. 26, 1935 PATENT OFFICE I METHOD OF AND APPARATUS 'FOR TREAT- ING FILAMENTS William P. Zabel, Cleveland Heights, Ohio, assignor to General ElectricCompany, a corporation of New York Application December-31, 1932, Serial No. 649,705

5 Claims.

My invention relates to filaments for electric incandescent lamps and similar devices "and to methods of and apparatus for their manufacture. More particularly, my invention relates to coiled metal filaments, especially to those filaments .which are operated athigh temperatures and ject of my invention is to eliminate these difliculties.

The coiling operations set up strains in both primary and secondary coils of a coiled coil filament, part of which are released immediately upon removal of the coils from the winding Further distortion occurs when the mandrels are removed from the coils. New strains are set up in bending the coils to form them into desired shapes. These various strains are still further released during heat treatment of the coils by ordinary methods, flashing of the mounts and lighting of the filaments. in the lamps.

Heretofore, molybdenum mandrels have been generally used for primary'and secondary coils. Such coils on mandrels are heat treated at 1200- 1600 C. in electric furnaces in an atmosphere of hydrogen. The mandrels are then dissolved in a mixture of acids, which is a slow and tedious process having some corrosive and, therefore, damaging efiect on the tungsten filaments. The above temperatures are too lowto effectively set the tungsten filament and, being applied after the coils have been cut to lengths, a certain amount of distortion occurs in said coils. Another method of applying a higher temperature treatment consists in heating the coils on their molybdenum mandrels in an atmosphere of hydrogen by passing current through the secondary mandrel. However, since molybdenum melts at approximately 2500 C., temperatures cannot be used which even approximate the operating temperature of the filaments when they are burned in lamps. Furthermore, it is dangerous to treat the coils on molybdenum mandrels near the melting point of the molybdenum because of contamination of the coils by diffused molybdenum which results in blackening of the lamps. Tungsten coils which are treated by passing current through inserted tungsten mandrels in the presence of hydrogen permit of using still higher temperatures close to the melting point of tungsten (approximately 3300 C.) without danger of contamination from'the mandrel. Heretofore no way was known of dissolving out or-otherwise removing the primary mandrel when used The obin a coiled coil filament without also destroying or attacking the filament.

According to my invention the primary coil is wound on a mandrel which may be readily dissolved out, preferably a steel mandrel. The secondary coil is then wound on a mandrel consisting of a refractory metal having a melting point substantially as high as the metal comprising the filament. With a standard tungsten filament I prefer to use a tungsten mandrel for said secondary coil. Said secondary coil generally comprises a series of spaced segments with an extra long space between consecutive groups of segments, each group constituting a filament. The primary coil is clamped or 'welded to the secondary mandrel thus holding the coiled coil on the mandrel just as it was coiled. The coil is then heated to approximately 1100 C. in an atmosphere of hydrogen to remove a considerable part of the coiling strains in the primary coil. The next step is to remove the primary mandrel, hydrochloric acid being used preferably, to dissolve it out if it is of steel, The coiled filaments, mounted on their secondary mandrels are placed in electric terminals and current passed through the assembly or through the mandrel alone in a treating bottle, in the presence of hydrogen, the filaments being heated according to a predetermined schedule to approximately 2800 C. The clamps or welded bonds are now removed from the ends of the coils, and the secondary mandrel is withdrawn, leaving the coiled coil filaments free from all mandrel, ready to bemounted as they are, or bent to other shapes.

In the drawing, Fig. 1 is-an elevation of a filament on a primary mandrel; Fig. 2 is a partial elevation of the coil of Fig. 1 being wound around a secondary mandrel; Fig. 3 is an elevation. of series of segments of primary coils wound around a, secondary mandrel and afiixed thereto; Fig. 4 is a side elevation in section of heat treating apparatus; Fig. 5 is a section along the line 5-5 of Fig. 4; and Fig. 6. is a front elevation of a stationary clip of the apparatus.

Referring to Fig. 1, a, tungsten filament wire 10 is coiled about a mandrel 11, preferably 'of steel wire. This coil is then wound around a secondary mandrel 12 (Fig. 2), preferably a tungsten wire, leaving a suitable spacing betweensegments 13 (Fig. 3) and an extra long space between the groupof segments 13 and an adjacent group of segments 14. Saidgroup of segments 13 constitutes one filament. Before 55 removing the assembly from the coiling machine, the starting and finishing ends of the primary coil are clamped to the secondary mandrel 12 by means of small nuts and bolts or other suitable devices. v

The next operation is to clamp, preferably by welding, the primary coil to the secondary mandrel at two points 15-16 (Fig. 3) about a half inch apartin the space between groups of segments 13-14. Since tungsten cannot be readily welded to tungsten, the welds 15,-16 are made of nickel, although other bonding metal could be used which does not dissolve in hydrochloric acid or strong alkalies. The bolts can now be removed from the extreme ends of the group of filaments and the coiled coil on mandrels is se-' curely located just as it was coiled. If the tungsten mandrel is small and quite ductile, this 0011 after having been welded can then be continuously heated by passing it through an electric furnace in the presence of hydrogen at a suitable temperature of nearly a white heat (about 1100 C.) toremove a considerable part of the coiling strains in the primary coil. The individual groups of segments 13-14 can then be cut apart between the clamped points 15-16. If the tungsten mandrel 12 is heavy and inherently brittle, it' has-been found more convenient to first separate the individual groups 13-14 and then heat them in an electric furnace at approximately 1100 C. in an atmosphere of hydrogen.

The next step is to remove the steel mandrel 11 from the primary coils. This is done by inserting the coiled coil assemblyin a hydrochloric acid solution which dissolves out the said steel mandrel 11, leaving the secondary mandrel 12 intact and the coils still attached by .the welds 15-16 in their proper locations. There now remain a number of segments of coil 13 wrapped about the tungsten mandrel 12 and attached to it by the nickel bonds 15-16 at each end with a free length of primary coil 10 and tungsten mandrel extending outside said bonds, as shown by full lines in Fig. 3. The ends of the coil and mandrel assembly are now clipped into electrical terminals of the apparatus shown in Fig. 4 which is fully described hereinafter, and current is passed therethrough in a treating bottle in the presence of hydrogen. The current is regulated according to a predetermined schedule to arrive at any desired temperature for the assembly. It is desirable to make the schedules as simple as possible, yet obtain the advantages of germinative grain growth and removaLof strains without appreciable loss of tungsten. The grain growth occurs very rapidly (less than thirty seconds) in straight drawn tungsten wires of small diameter between temperatures of approximately 2200 to 2800 C. I haveobtained the desired result by heating filaments of two to five mil diameter in accordance with the following schedule: The

- current is regulated through the mandrel, or the mandrel and coil, so as to raise the temperature from a red heat to 225 0 C. in approximately fifteen seconds, then to- 2800 C. in thirty seconds, and down to a red heat again in thirty seconds. The time schedule may be extended for larger filaments. This treatment is sufficient to reduce strains of coiling and other deformations so that they will not later cause distortion of the filament in the lamp.[

' The heat treating apparatus comprises a stationary clip 17 (Figs. 4 and 6) having a pair of jaws 18-19 and a movable clip 20 (Fig. 5) hav-' clip 17, while the jaw 19, also preferably made of molybdenum, is mounted on a slide 23 which is slidably mounted in said clip 17. A cover plate 24 is mounted on the front of said clip 17 to retain the slide 23 therein. The slide 23 is moved up and down by a lever 25 which is connected to said slide by a pin 26 and is pivotally mounted on a pin 2'7 in the cover plate 24. The jaws 18-19 are normally kept closed'by a spring 28 mounted on theend of the lever 25 and on a pin 29 in the clip 17. Said clip 17 is mounted on a tube 30 which is mounted in a vertical support plate'3l and insulated therefrom by an insulating tube 32 surrounding said clip tube 30. Said clip tube 30 is cooled by a flow of water entering through a tube 33 mounted thereon and passing along the annular space between said tube 30 and tained therein by a cover plate 36. Said slide 35 is moved up and down by a lever 37 which is connected to said slide by a pin 38 and is pivotally mounted on a pin '39 in the cover plate 36. A spring 40, mounted on the end of said lever 37 and on a pin 41 mounted in the clip 20, serves to normally keep the jaw 22 closed against the jaw- 21. The clip 20 is mounted on a copper tube 42 which is slidably mounted on a tube 43. Said tube 43 is mounted in the support plate 31 and insulated therefrom by an insulating tube 44 surrounding said tube 43. cooled by a flow of water entering through a tube 45 mounted thereon and passing along the annular space between said tube 43 and a tube 46 mountedtherein, then out through said tube 46. A pair of bolts 47, mounted in the clip 20 and in a block 48, pass through a clamp block 49 which is mounted on the clip tube 43. Springs 50. mounted on said bolts 4'7 between the blocks 48 and 49, tend to keep the clip 20 in yielding contact with said block 49, thereby allowing said clip 20 on the tube- 42 to slide along the clip tube 43 to compensate for expansion and contraction of the mandrel 12 which is held between the pairs of jaws 21-22 and 18-19 and kept under tension so that it remains straight. By locating the clip block 49 at difierent points along theclip tube 43, diiferent lengths of mandrel may be handled by the device. Terminal clips 51-52 mounted on the'clip tubes 30-43 are connected 10 they are enclosed in a treating bottle 53 which is mounted on rollers 54 and moved along horizontal tracks 55 mounted on supports 56. The open end of said bottle 53 enters a groove 57 (Fig. 4) in the support plate 31 in which is located a rubber tube 58 which is compressed by the end of said bottle 53 to form a gas tight joint. Said tube 58 is held in said plate 31 by cover plates 59. A flow of hydrogen is maintained through said bottle 53, entering through a tube 60 .and flowing out through tubes 61-62, said tube 61 being provided with a series of holes 63 The clip tube 43 is along the bottom thereof. The bottle 53 is kept cool by a flow of water through a jacket 64 surrounding said bottle, the water enteringthrough a tube 65 and flowing out through a tube 66.

After treating the filament 10, the nickel bonds are clipped off the ends thereof and the mandrel 12 is withdrawn, leaving the coiled coil filament free of all mandrel, ready to be mounted as it is or bent to other shapes for use in lamps, particularly in projection lamps or otherhigh wattage lamps. The coils, having been heated on straight mandrels throughout the treatment, are in perfeet axial alignment whereas, minus said mandrels the coils would be bent, particularly in the portions between segments.

.It is of course entirely possible to dispense with the welds l-l6 on each individual group of filament segments and, instead, merely clamp the series of groups at each end and treat them thus, rather than singly. Filaments-made according to my method have been found to be substantially free from warping and twisting throughout the life of the lamps. The heat treating schedule disclosed herein may of course be used with equally good results on standard single coil filaments supported on mandrels, as well as on the coiled coil construction. If, as mentioned above, a series of groups 13-14 are heated at a time, sleeves or coils of a refractory substance such'as tungsten may he slipped over the end portions of the filaments between groups 13 and 14, thereby dissipating the heat therein so that said ends do not reach the high temperatures of the coils themselves, thereby leaving said end portions relatively ductile for mounting on the lead-in wires.

What I claim as new and desire to secure by Letters Patent of, the United States, is:

11 The method of making metal filaments which comprises winding a filament wire around a primary mandrel which is considerably less refractory than said filament, winding the coilv thus formed with its mandrel around a refractory secondary mandrel, heating the coiled coil structure thus formed to nearly a white heat to remove a considerable part of the strains in the primary coil, removing the said primary mandrel, heating the coiled coil on the secondary mandrel to approximately 2800- C., and then removing said secondary mandrel.

2. The method of making metal filaments which comprises winding a filament wire around a' primary mandrel which is readily soluble compared with saidfilament, winding the coil thus formed with its mandrel around a refractory secondary mandrel, heating the coiled coilstructure thus formed to nearlv a white heat to remove a considerable part of the strains in the primary coil, dissolving out the said primary mandrel, heat ing the coiled coil on the secondary mandrel to approximately 2800' C., and then removing said secondary mandrel. v

3. The method of making metal filaments which comprises winding a filament wire around a steel wire primary mandrel, winding the coil thus formed with. its mandrel around a refractory secondary mandrel, heating the coiled coil -structure thus formed to nearly a white heat toremove a considerable part of the strains in the primary coil, dissolving out the said primary mandrel, heating the coiled coil on the secondary mandrel to approximately 2800 C., and then removing said secondary mandrel.

4. The method of treating coiled tungsten filaments of small diameter wire which comprises heating said filaments in successive steps to approximately 2200 C., then to. approximately 2800 C. in approximately thirty seconds to allow germinative grain growth to take place.

5. The method of heat treatingtungsten filaments of small diameter wire which comprises heating said filaments in successive steps to a red heat, then to 2250 C. in approximately fifteen seconds, then to 2800 C. in approximately thirty seconds.

WILLIAM P. ZABEL.

Images