Updated 06-XI-2011
Incandescent
Introduction
Product Overview
Cap Nomenclature
Bulb Nomenclature
Filament Nomenclature
Technology
Operating Principle
Gas Filling Effects
Filament Coiling Effects
Vacuum vs Gas-Filled
Gas Filling Types
Getters
Burning Position
Voltage Variation Effects
Starting Characteristics
Lamp Life
End of Life & Fusing
Premature Failure
Lamp Designs
Carbon Filament
Tantalum Filament
Osmium Filament
Tungsten Filament
Advanced Filament
Infra-Red Recycling

Filament Coiling Effects

The solution to the problem of heat losses to a gas filling came in 1912 at the Schenectady Research Laboratories of the General Electric Company. Dr. Irving Langmuir discovered that although the gas is flowing very rapidly inside the lamp, there is always a film of stationary gas a few millimetres thick around the filament now known as the Langmuir Sheath. By learning to work with this sheath he made the modern high efficiency gas-filled lamp a commercial reality. GE put the first nitrogen-filled lamp on the market in 1913. It was a 300W lamp for industrial lighting

By substituting the long straight wire for a tightly coiled filament, it was found that if the spacing between adjacent windings is less than the thickness of the Langmuir sheath, then the stationary gas layer around one coil joins with the next, preventing gas flow between individual coils. As a result the surface area of the filament exposed to the moving gas is reduced, so gas losses also fall and the filament runs hotter. The relative surface area of wire exposed to the cooling effects of the gas is shown in Figure I10, which represents the end views of a straight wire, a single coil, and a coiled-coil filament surrounded by the rising gas. Note how the gas flow does not contact the filament or move between the coils, but rather flows around the entire body avoiding the sheath of stationary gas.


Figure I10 - Gas Flow around Hot Lamp Filaments - End View

Although coiled filaments do still lose some heat to the gas by conduction and convection, for the vast majority of lamps the heat loss is reduced so tremendously that the benefits of reduced filament evaporation and hotter operation outweigh the remaining energy loss. However it is important not to forget the importance of filament geometry and its desired operating temperature when designing a gas-filled lamp.

A typical single coil lamp is illustrated in Figure I11. The coil is mounted in horizontal ring, known as the Wreath or C-9 configuration, so that the cooling effect of the gas is the same over its entire length. If it was mounted axially the upper end would run hotter because hot gas rises, and the temperature gradient would cause the top of the filament burn out prematurely. This is why GLS lamps last longest when they are operated vertically cap-up or cap-down, because no part of the filament is overheated by convection currents. The introduction of the coiled filament also simplified production, because owing to its reduced length a number of support wires could be eliminated and the bulb volume reduced.
Fig.I11-Coiled Filament

Another result of coiling is to produce a black body effect due to multiple reflections of the radiation within the coil. This increases the brightness of the interior of the coil to about 175% of the exterior brightness. The massing of the filament into a smaller space is also advantageous, as for practical purposes it can be considered more of a point source allowing the fittings designers to predict performance with more accuracy.


Coiled-Coil Filaments
Since the early 1920s it was known that a coiled-coil filament would yield still further gains in efficacy, by presenting an even smaller surface area available for cooling. The first experimental lamp to be made with such a filament is illustrated in Figure I12, having been created in 1924 at the Hirst Research Laboratory of Osram-GEC in London. However there were many technical difficulties which delayed the market launch for nearly a decade, until 1933.
Fig.I12 - Experimental Coiled-Coil of 1924

The principal reason for the delay was due to the development time necessary to build fully mechanised production equipment to mount the delicate coiled coils without human intervention. Even a century ago, lamps were required in such vast quantities and made to such tight dimensional tolerances that manual assembly was out of the question indeed it is for this reason that the incandescent lamp holds the title of being the first article of any kind to be mass produced with every operation being executed by fully mechanised equipment. Additionally it was found that a perfect coiled-coil will give highly variable performance if mounted by inferior means. Any process variation that mounts the coil with differing amounts of tension will change the coil temperature, with disastrous consequences on life and efficacy. Breakdown of the gas-filling across the short filament was also troublesome, and for the first time, fuses had to be developed for incorporation within the lamp.

Once perfected, it was found that the increase in efficacy after the second coiling is considerable, but not so significant as the improvement from the primary coiling. However the coiled-coil offers additional advantages. Because it is shorter it requires fewer support wires - sometimes none at all being necessary, but certianly no more than two or three. Conversely the single coil filaments employ five to ten supports and squirrel cage types may have as many as seventeen. Each support wire causes local cooling and reduces light output by approximately 1% so the lamp is always more efficient with fewer supports.

Fig.I13 - Demonstration of Coiling Effect

The effects of coiling and heat losses to the gas filling are superbly visualised in the demonstration lamp of Figure I13, which was made up by the author at the works of GE Lighting, Leicester. The filament is an ordinary 240V 60W coiled-coil type, but part of it has been stretched out to a single coil, and the remaining portion completely unwound to a straight wire. The entire filament is made from the same piece of tungsten wire, all the same diameter. The bulb is filled with the standard 85:15 Argon:Nitrogen mixture. When switched on at quite a low voltage, the coiled-coil lights up brightly. As current is increased the single coil begins to glow, but even at 240V the straight wire produces no light at all.
In an identical lamp made up with vacuum atmosphere instead of gas-filled, all three sections of filament light up at equal temperature, because there are no gas losses. This clearly demonstrates that gas losses really are lowest for the most tightly coiled filaments.

Coiled-coiling has the greatest effect on the small diameter filaments as these have the highest surface area to volume ratio, and the most to gain. For a 230V 40W lamp the increase in efficacy is about 20% over the single-coil. It is about 10% for 100W, but only a few percent with 150W. Nevertheless coiled-coils are found in many higher power lamps because of the simpler construction with fewer supports. Many high wattage projector lamps also use coiled-coil filaments, because their compact shapes work better in the optical system.

In recent years there has been some activity to develop a triple coil incandescent filament and such a product has been commercialised in Japan by Toshiba. This offers no significant increase in luminous efficacy, but coil dimensions are reduced such that it approaches a point-source, and outstanding optical control can be attained in small reflectors. Controlling the degree of filament sag by modification of the tungsten metallurgy was a key challenge that had to be overcome, while in parallel limiting the concentration of potassium dopant in the tungsten wire to avoid problems of arcing due to the unusually high voltage gradient between adjacent coils.
Fig.I14 - Triple Coil v Coiled-Coil Filament

It should be noted that the temperature of a coiled coil filament is no greater than for the single coil type, so life is the same. All other things being equal the temperature after the second coiling would rise, with a consequent life reduction. To avoid this the wire is made slightly longer, leading to a reduction in power input for an equivalent temperature. In order to maintain the standard range of lamp wattages the power is brought back to normal by increasing wire diameter. The wire for CC filaments is therefore slightly longer and thicker than for single coils. Due to the greater mass of incandescent metal it radiates more light.