CHAPTER FIVE

LEADING-IN  WIRE  DEVELOPMENTS

The electric current which heats the filament inside the bulb to incandescence is carried to the filament by two wires which pass through the glass chamber. These wires must make an air-tight joint with the glass in order to preserve the vacuum and, in a gas-filled lamp, to prevent either the air entering or the gas leaving the bulb.

In the beginning, platinum was the only material known which would answer the purpose and it was used for many years. It was comparatively cheap in the early days, about six dollars an ounce, but the cost gradually increased. As the cost increased the amount used was reduced.

Substitutes for platinum have been sought almost from the beginning, and some lamp manufacturers quite early used nickel-steel wire with fair success. This nickel-steel alloy can be made to have the same expansion as glass, but it does not stick to the glass and is never one hundred per cent efficient. The seal between the leading-in wires and the glass is made at high temperature when the glass is soft and, as it cools down, the wire and glass must stick together as they contract to the lower temperature. This is a greater range in temperature than that between the lighted and unlighted lamp, in which the wire and glass must also stick together to make an air tight seal.

The leading-in wires present other problems beside making a tight joint with the glass. They must be good conductors of electricity, which nickel-steel is not. So in many lamps, copper wires were and are now welded to the wire imbedded in the glass. At the present time pieces of copper wire are, in all lamps, welded to these short wires imbedded in the glass, passing outward to connect with the terminals of the base to which they are soldered. In vacuum lamps, copper wires are used going inward to connect with the filament, it having been found that copper is the best material for filament connection. It is always oxidized in making the lamp, because the oxide acts as a beneficial "getter" in vacuum lamps. In gas-filled lamps, nickel wire is much better than copper for filament connections. Copper oxide has a bad effect, while clean nickel is the best material known for the purpose.

Originally the seal between the leading-in wires and the glass was made by fusing a piece of small glass tubing around each wire, two such wires with their glass "petticoats" being inserted in a stem tube. The end of the stem tube which went inside the bulb was closed by fusing it around the glass petticoats on the wires. In the latter part of 1880 the glass work of the stem seal was greatly simplified. The petticoats of glass were omitted, the end of the stem tube being flattened together about the two wires. This method has been used ever since.

Clamps

The connections between the leading-in wires and the filament have been a problem from the beginning. At first a little screw clamp was used which held the enlarged end of the carbon filament in its jaws, the other end of the clamp being fastened to the platinum leading-in wire. At first these clamps were made of platinum and later of nickel. These screw clamps were used until early in 1881, when the copperplated connection came into use. In this arrangement a piece of copper wire was welded to the platinum wire, the latter being sealed in the glass. The other end of the copper wire was flattened quite thin, bent double, folded about the enlarged end of the filament and the connection made good by copperplating.

EDISON LAMP, 1889

The length of the seal was reduced so that less platinum was necessary for the leading-in wires.

As long as this connection was used the filaments were made with enlarged ends so that the part of the filament which was in contact with the copper would not become hot enough to melt the copper or vaporize it.

In 1886, carbon paste joints were introduced. At first the carbon paste was made of india ink and fine graphite, but soon a better paste was made of two kinds of graphite, one of which contained a considerable amount of clay. This graphite mixture was mixed with a binder composed of a solution of sugar and gum arabic. These paste joints were baked in an oven to about 400 deg. F. to partly carbonize the binder, otherwise in damp weather some of the joints would absorb moisture and become loose before they were sealed in the bulb.

For large sized filaments a special paste was used, consisting of coarse graphite, soft coal and coal tar pitch with sugar and gum arabic binder. After these joints had been baked, each was painted with a little red phosphorus and was heated red hot on a fine gas jet. The heating decomposed the hydrocarbons, drove out a lot of gas and smoke, and left a hard piece of coke for the joint which gave out very little gas during exhaustion.

EDISON LAMP, 1890

The amount of platinum wire in the seal was further reduced by imbedding the welds between the copper and platinum wires in the seal.

At the present time a material called "aquadag" is used for the paste joint in the few carbon lamps made. Aquadag is an extremely fine graphite powder mixed with water. When dry it becomes pure carbon and yields practically no gas in exhaustion. This enables the exhaustion of the present carbon filament lamps without lighting up, for most of the gas which appeared in the previous carbon filament lamps during exhaustion came from the paste joints.

When the pressed tungsten filament came into use, the connections between the filament and the leading-in wires were made by fusing the two together with an electric arc. This was done in a reducing gas atmosphere to prevent burning the filament. This practice was continued as long as pressed filaments were used.

At first the connections for drawn wire filaments were made by forming short tubes in the ends of the leading-in wires, inserting the ends of the filaments in the tubes and flattening and crimping the tubes on the filament ends. This made a good connection, but the construction was expensive. It was simplified by what is called the hook connection. The end of the leading-in wire was flattened and folded over on itself, forming a hook. The end of the filament was placed inside the hook and the hook pressed hard on the filament, which imbedded the hard filament wire in the softer leading-in wire. This also made a very good connection and is in general use today. Some large size filaments are electric spot welded, and the very largest sizes are electric arc welded, to the leading-in wires.

Substitutes for Platinum Leading-in Wires

The first substitute wire commercially used on a large scale was that invented by Byron E. Eldred, which is covered by a patent applied for in October, 1911, and granted in December, 1913. This wire consisted of a nickel-iron alloy core which was dipped in an acid copper-sulphate bath to give it a slight coating of copper, then silver plated and further covered by a platinum sheath. This composite wire was so proportioned in its parts that it was designed to have a slightly lesser coefficient of expansion than glass, so that in cooling down from the high temperature at which the seal is made to the temperature at which this part of the lamp operates, a pinch effect of the glass on the wire was obtained. It was commercially used from 1911 until the early part of 1913.

The use of the non-oxidizable platinum outer sheath was deemed necessary, as glass would not "wet", that is, make a hermetic seal with, or stick to, a bare wire of any base metal or of nickel-iron or other alloy if the wire were made large enough to be used as a leading-in wire. The intermediate copper and silver was for the purpose of making a tight union between the nickel-iron core and outside platinum sheath which could not be directly made.

Dr. Colin G. Fink, of the Research Laboratories of the General Electric Company, invented an improved wire which was put into commercial use in 1913, superseding Eldred's wire.

ASSEMBLY OF MATERIALS OF DUMET WIRE

This leading-in wire, which took the place of platinum in 1913, consists of a nickel iron core with a copper sheath. After brazing the two together and drawing to the proper diameter, the wire is coated with borax.

Fink's wire consisted simply of a nickel-iron core, dipped in acid copper-sulphate to give it the thin copper coating, and inserted in a brass sheath, in order that the outer copper sheath could be readily brazed to the nickel-iron core. This wire has an expansion coefficient that was practically the same as that of glass. The sheath is about 20 per cent by volume of the wire, the proportions of the core being about 45 per cent nickel and 55 per cent iron. This wire is even better than platinum itself and its use has resulted in a much smaller percentage of leaky lamps. While copper oxidizes readily, it was found that if the pinched seal is heated somewhat longer than formerly, the glass absorbs the oxide and makes a very tight union. This wire is called "dumet" wire. Dr. Fink applied for a patent in June, 1912, which was granted in June, 1924.

The sealing in of dumet wire was improved by W. L. Van Keuren, of the General Electric Company, by coating the wire with borax. Van Keuren applied for a patent on this in December, 1913, which was granted in June, 1918.

COATING DUMET WIRE WITH BORAX

The dumet wire is heated to slightly oxidize it, and is then dipped in a solution of borax which in drying and heating forms a copper borate with the oxide and makes a ready seal with the glass. Under the conditions which exist in sealing the wire in the very hot glass, the copper borate is largely absorbed in the glass and the union between wire and glass becomes very tight. The wire makes a tighter joint with glass than platinum, and is a better material for the purpose. It is also relatively inexpensive to make compared with platinum, which has risen steadily in cost and is now well over one hundred dollars an ounce.

About twenty years ago, Geist invented a leading-in wire composed entirely of copper. He used a copper wire about sixteen thousandths of an inch in diameter and flattened it at the point at which it was sealed in the glass, so that it was very thin. He also made a round hole in the center of this flat part. For some reason he could not make these seals consistently effective, but he did succeed with a large majority of them. Recently this invention has been further developed and when the flattened parts are made much thinner, about one and one half thousandths of an inch, cross section, the wires make perfect seals. An automatic machine has been developed and many thousands of trial lamps have been manufactured using all copper leading-in wires of this type. The copper unites so firmly with the glass that even though it shrinks more than glass, the shrinkage of these very thin parts does not pull them away from the glass. No hole is now made in the thin section of copper.

In cases where the requirements to be met by the lamps necessitate the use of a specially hard glass in the seal, which will stand high temperatures without softening, large tungsten wires are used for the leading-in wires. Since the temperature expansion coefficient of such hard glass is about the same as that of tungsten, the combination of the two results in a tight seal.