How Do Car Lightbulbs Manage to Last Despite Harsh Conditions?

[Rothiemurchus]

Why do car lightbulbs last so long when they get knocked around so much and the filaments are so thin and get so hot?
Has it got anything to do with the frequency of the shockwaves passing through the filament? Or does the car's suspension do a better job than I think it does?

 

[Tide]

Do headlamps really have a long lifespan?

 

[HallsofIvy]

Well, I haven't had any problem with my headlights in 5 years. In the meantime, I've changed bulb in my home many times.

 

[Tide]

Yes, but I'd guess that you run your headlamps for fewer hours each day than you do your household lightbulbs.

 

[pervect]

I gather most cars nowadays use halogen lights.

http://auto.howstuffworks.com/question387.htm
"most cars currently use halogen lights in their headlights"

http://home.howstuffworks.com/question151.htm
describes halogen lights, which chemically re-deposit evaporating tungsten back on the filament.

There are also some articles on how the newer HID automoblile lights work at the end of some of the URL's above.

There may be some information at
http://faq.auto.light.tripod.com/

though it seems to be mostly about the pricey HID headlights. I haven't so far found any MTBF figures for headlights.

You can use halogen bulbs at home in specially designed fixtures - and of course they use the halogen bulbs - I'm not sure how long they last (they probably last a lot longer when used with a dimmer that most fixtures incorporate), but they are pricey, and not as energy efficient as fluorescent bulbs (which also last a long time). Halogen's run hot too, they may be a fire hazzard.

There were some other links I ran into that pointed out that electromigration lowers the length of bulbs running off of DC significantly (half the life), this life reduction should apply to automotive bulbs.

 

[russ_watters]

If 20% of your driving was at night and you averaged 30mph, over the typical 150,000mi life of a car, that's 1,000 hours for your headlights. Not very long at all.

 

[Integral]

If 20% of your driving was at night and you averaged 30mph, over the typical 150,000mi life of a car, that's 1,000 hours for your headlights. Not very long at all.

Daylight running lights are becoming more and more common, required, I believe, in some countries. We have a '98 Pontiac with DRL and 130K miles, same head lights. Though they do run at a lower power in the day time, they are on and they do not fail.

If we just consider a simple filament bulb the resistance of the 12 V filament must lower then that of the 120V filament by a factor of 100 for the same power. Now assume the same material and fix the length, this means that the radius of the 12V filament must be larger then the 120V filament by the same factor. Looks to me like a 12V filament bulb simply can have a stronger filament. It could be made both shorter and fatter thus reducing its susceptibility to breaking.

The Gas Discharge tubes would be a whole different story, but since there are few if any gas discharge tubes in household use we really cannot compare them.

My personal experience with household halogen bulbs is that they are fragile and not, on the average particularly long lived.

 

[pervect]

I'd expect lightbulbs to follow the typical Arrhenius law

http://www.weibull.com/AccelTestWeb/arrhenius_relationship_introduction.htm

This law is used in both chemistry and reliability engineering, for the later case it is written as

Life = C*exp(-Ea/kT)

where T is the temperature, in this case the temperature of the filament.

I would expect that if we operate a non-halogen lightbulb at the same temperature as a halogen light bulb, we would clearly see the advantage that the pricier halogen lightbulb has.

Because of the exponential relationship, a small difference in temperature can mean a huge difference in lifetime. Unfortuantely, the same is true for power output. Using the black body law as a reasonable approximation, the total light energy output goes up as something like the fourth power of the temperature.

To a certain extent I'm guessing here, since I don't have any data on the lifetime of bulbs. But I feel strongly that it's important to normalize the temperature of the filament to the same value when comparing the lifetime of different incandescent bulbs.

 

[Rothiemurchus]

If a tungsten filament gets up to 2500 celsius wouldn't the tungsten flow a bit like a liquid and get thinner at the top?
Is there any evidence that bulbs last for less time in colder climates where
the contraction/expansion ratio of filament and glass/quartz would be greater?
Or do warmer,wetter climates shorten bulb lifetime?

 

[Mk]

evaporating tungsten back on the filament.

How does a metal evaporate? Same way water does? My dad's had a pontiac for 12 years, lights never went out. I know some cadilacs have fully LED lights. How does and LED create light?

 

[pervect]

If an atom or molecule on the surface gets too much energy, the attractive forces binding it to the surface are not strong enough to hold it, and it escapes.

The amount of energy needed will be the lowest for a molecule at the surface, because a moleucle on the surface is not bonded on all sides.

This is one way of justifying the use of Arrhenius's law, because the probability of a molecule having an energy far out on the "tail" goes up exponentially with temperature. Thus the probability of escape goes up likewise, and the lifetime goes down.

 

[Integral]

There has to be other factors then simply temperature. By increasing the volume there are more electrons available so more can be lost before failure occurs. Increasing the volume will have to be accompanied by an increase in surface area. A larger surface area will mean that MORE electrons will be lost per unit time, but since the volume increases by the cube of the radius, while area increases as the square there should be a net gain in life for a larger filament operating at the same temperature of a smaller filament.

I would guess that many failures are due to thermal shock and or simple mechanical fatigue (due to continual vibration) . (It is my observation that most household bulb failures occur at turn on, which is why I feel it would be more accurate to specify the life of a bulb in on/off events then hours of operation... that's a separate issue) Would not a shorter fatter filament be better able to with stand the vibrations inherent in automobile operation?

 

[Rothiemurchus]

If the vibrations have a long wavelength then they won't make the filament resonate - and this would be help it last longer.

 

[Sirus]

I know some cadilacs have fully LED lights. How does and LED create light?

LEDs are not used in headlights, only in some signal, brake light, and interior applications.

 

[pervect]

I wouldn't say that temperature was the only factor that was important. But I would say that temperature was an important factor, (I'd go so far as to say that it was probably the most important factor), and that I'd expect the lifetime to vary with temperature according to the exponential law I mentioned over a fairly wide range of temperature.

The law is a bit empirical, but it's fairly widely used for semiconductors (not just lightbulbs).

One thing I didn't mention is that there are often several potential failure mechanisms, each with their own rate constant and activation energy.
I should add I'd be rather skeptical about pushing the law _too_ far. While it has some theoretical basis, it should still be seen as an empirical law, a sort of engineering "rule of thumb".

I also stumbled across another website about lightbulbs

http://invsee.asu.edu/nmodules/lightbulbmod/burnout.html

which was quite interesting, and included some scanning electron microscope pictures. It mentioned a few other phenomenon- one is that when a particular region of the filament gets thinner, it tends to get hotter, which tends to cause a runaway feedback effect.

One of the most interesting general quotes about lightbulb failure is below:

Even when the light bulb is perfectly sealed from air it still “burns out,” but not as fast. Tungsten directly vaporizes from the filament surface while at the extremely high temperatures tungsten filaments operate at. The pressure of tungsten vapor over a filament is 10-4 Torr at 2,757oC.

W(solid) + heat ------------> W(gas)

As the trace of tungsten vapor leaves the surface by sublimation, the cool argon/nitrogen gas around the filament causes the vapor to solidify as smoke, which slowly settles on the glass bulb giving it that irregular gray tint as it ages. Filaments fail by brittle fracture when they become too weak from thinning.

.

 

[pervect]

Found another website:

Halogen bulbs cost more, but may have a lifetime of up to triple a normal light bulb of the same wattage, and at the same time be anything up to a fifth more efficient at producing light.

http://www.naturalhandyman.com/iip/infelectrical/howlightbulbswork.shtm