The basic principles of incandescent lamp design predict that the
rate of evaporation from the filament determines lifetime. Temperature
accelerates the evaporation and thus shortens the life while higher gas
pressures slow evaporation and prolong the life. If the evaporation were
to occur uniformly over the entire filament, the resistance of the wire
would increase as the wire thinned, resulting in a corresponding drop in
temperature. Under these ideal conditions, the lamp would last forever.
Instead, the wire burns through at a specific spot, while the remainder
shows little or no reduction in diameter. These spots are caused by some
localized defect and have become known as hot spots. Becker in 1925
was the first to suggest a hot spot model of filament failure.9 In his
model, the lamp filament would fail when a portion (hot spot), which
was initially hotter than any other section, reaches the melting point of
the wire. The time to failure depended on the initial temperature of the
wire and the temperature difference of the hot spot. Horster et al. were
among the first to try to put a quantitative interpretation to the effect hot
spots have on the lifetime of incandescent filaments.10 They were able to
demonstrate a relationship between the smoothness, or lack of hot spots,
along a tungsten wire and its weight loss at failure. The smoother or
more uniform the wire, the more uniform the evaporation will be, and
the greater its weight loss at failure.
The ’70s brought on considerable efforts, particularly at General
Electric and Westinghouse, to further understand the hot spot phenomena.
At Westinghouse, filament failure research was being conducted by
a number of scientists, including Fran Harvey, George Comenetz,
Herman Johansen, Al Pebler, Heinz Sell and Chet Dawson. While Dr.
Harvey was developing a theoretical model for hot spot growth,11 Dr.
Comenetz was showing that the necking down or thinning of the filament
wire effected the life of coiled filaments as well as straight wires.12
Because it is thin, the wire’s cross-section at that point is of higher resistance
than at other points along the wire. Rather than have an overall
drop in current, which would occur if the entire wire thinned, the current
is hardly affected by localized thinning. More energy is dissipated
at this point then elsewhere, causing the wire to burn hotter or form a
hot spot. Dr. Harvey’s and Dr. Comenetz’s model, like those before
them9,10 would predict that because of the higher temperature, thinning
would accelerate at the hot spot until the lamp eventually failed at this
Effect of Coil Pitch
When lamp manufacturers wind tungsten wire into either a single
coil or that coil into a second coil to form respectively a single coil or
coiled coil filament, the distance between each turn of the coil is very
uniform; when mounted into a lamp, however, that coil is stretched,
threaded through and pulled by supports and finally flashed (lighted
briefly) to evaporate the getter and set the metallurgical structure of the
wire. While the manufacturer tries to use processes to minimize changes
in coil spacing (called coil pitch), variations do occur. Al Pebler used x-
ray radiographs to correlate this variation in coil pitch with the failure
point in both coiled coil and single coil filaments. Filaments burn the
hottest where the coil pitch has been decreased (turns become closer together)
most during processing, and this is the point at which they will
often burn out.
Influence of Internal Flaws
Thinning of wire and variations in coil pitch are two ways that hot
spots form, but what if the wire’s cross section diminishes by the development
of internal discontinuities? In another section of this book, Sell
and Moon discuss how lamp-grade tungsten wire is produced from
tungsten powder doped with potassium silicate and aluminum chloride.
The potassium dopant is the source of submicroscopic bubbles that form
when the filament is heated. Figure 1-4 shows metastable, elongated
pores in drawn, annealed tungsten wire, that was previously ion implanted
with potassium and aluminum. The pores are seen beginning to
spheroidize into the more stable rows of bubbles commonly seen in
lamp grade annealed tungsten wire. Chet Dawson showed that thermal
gradients, such as hot spots, act as a driving force for bubble growth.13
Bubble migration up the temperature gradient is the most likely of several
possible mechanisms that may be responsible for this growth. When
two bubbles come in contact with one another they coalesce and form a
larger bubble. The result of this phenomena would be thinning of the
wire from within as well as from without with a shorter resulting lifetime
than would be expected from the hot spot evaporation model
Experimental Proof That Hot Spots Influence Lamp Life
Other than chemical attack, most incandescent lamps fail by the
development of hot spots. Dawson et al., using a current pulse scanning
device (see Figure 1-5), conclusively followed the development of hot
spots in T10 single coil filament lamps from the beginning to the end of
life.14 By pulsing the filament, they traced the temperature profile of the
filament when such temperature differences were greatest. This may
also explain why lamps fail so often when they are switched on.
When a lamp is turned on, the temperature at a hot spot rapidly
overshoots the steady state burning temperature of the filament. This
occurs because thermal conductivity does not have time to level the
temperature closer to the steady state burning temperature of the rest of
the filament. Since the peak temperature a hot spot reaches on switching
may be several hundred degrees higher than its steady state temperature,
it will often fail when this temperature overshoot approaches the
melting point of tungsten. This did occur in the work of Dawson et al.
where the temperature overshoot at switching increases during lamp life
until the lamp filament eventually fails at the fastest developing hot
"Revolution in lamps: a chronicle of 50 years of progress" - книжка 2001 г., куча вкуснейших и неочевидных подробностей о, казалось бы, простейшем устройстве - причём порой таких, которых я нигда в литературе не встречал. Вероятно, многие результаты разных внутренних исследований Филипса и т.п. ранее не публиковались вообще, как это нередко бывает с отраслевой прикладухой, тем более имеющей коммерческую ценность.
Просто обалденно, уровень "Истории свечи" Фарадея, если не выше. Аж перевести хочется. Ну, с доработками. Глубокими. До "Истории лампочки накаливания". Но - ...