sitelights
07-09-2001, 06:42 PM
Lamp manufacturers rate the longevity of their products using a test which is termed "average lamp life". A number of the lamps (obviously all identical and one hopes from several manufacturing batches) are energized simultaneously at the specified input voltage. If the test includes 100 identical lamps, the point at which the 50th lamp fails (say 4000 hours) is the "average lamp life". Since, in this example, 50 lamps have died and 50 have survived 4000 hours of operation the odds of a individual lamp lasting more than 4000 hours has only a .5 probability factor. It is important to note that the test is made supplying 100% of the rated voltage to the test lamps. Lower than rated input voltage results in longer average life; higher than rated voltage results in a shorter average life. No lamp manufacturer that I have contacted is willing to divulge the rate of failure (some lamps may have failed at 10 or 100 or 1000 hours) and simply refer to a "bell shaped curve".
It is important to note that halogen lamps depend on sufficient current to heat the quartz glass envelope to insure the "halogen cycle" i.e. that the tungsten from the filament is redeposited on the filament in a finely tuned gaseous environment within the bulb. Early halogen lamps were much more susceptible to undervolting than what is now available; the result being blackening of the inside surface of the bulb (tungsten accumulation) and consequent shorter life. The result of tungsten migration triggers a little understood phenomenon termed "filament notching" which could be likened to the deterioration of a link in a stressed chain until its eventual catatstropic failure . This is why a lamp can sometimes be "fixed" by tapping it while energized; the small gap in the filament will weld but fail again soon. The filament of a hot halogen lamp is much more delicate than in a cold lamp and this is why the temporary fix works but since the filament is wound in a very tight coil (like a spring) cooling will break the weld during the next cycle. "In-rush" current also has an effect on lamp life; this is the sudden heating of the filament to incandescence when the lamp is switched on at full operating voltage.
Low voltage lamps may benefit from the buffering of in-rush current through the transformer coil ("soft start"); line voltage lamps may also benefit from dimming switches since the in-rush current is bufffered by the dimming switch coil. I'd be pleased to have these theories either confirmed or denied.
My experience with premature low voltage lamp "failure" is that it is more often the socket/lamp pin contact that fails. We have been warranting halogen lamp life as part of our original three-year unconditional warranty and during subsequent three-year service agreements offered to our clients. I estimate that at any given moment we are responsible for the operation of at least 10,000 lamps scattered throughout our 75 mile service radius. This calculated risk gets our undivided attention at all times and we have kept very complete records for over 15 years. It often occurs to me that this service burden (which we bear stoicly) could be markedly reduced if the lighting equipment manufacturers would spend an extra dime on the quality of the lamp sockets used in their luminaires.
The lamp manufacturers are doing their part (witness the increase in average lamp life over the past 10 years) but the fixture makers are still too cost-driven to address the problem. Our records indicate that well over 75% of all lamp "failures" are really attributable to failed contacts within the sockets. What is needed is a mechanical- rather than a friction-fit to the lamp pins. This could be achieved with set screws to positively clamp the socket/pin connection together.
When I think of all the service calls we perform to merely re-seat the lamp in a "dark" fixture it is clear to me that we'd pay a dollar, not just a dime, premium to have a solid mechanical connection. Bi-pin sockets of this type are used in some European -sourced cable-hung systems.
It is important to note that halogen lamps depend on sufficient current to heat the quartz glass envelope to insure the "halogen cycle" i.e. that the tungsten from the filament is redeposited on the filament in a finely tuned gaseous environment within the bulb. Early halogen lamps were much more susceptible to undervolting than what is now available; the result being blackening of the inside surface of the bulb (tungsten accumulation) and consequent shorter life. The result of tungsten migration triggers a little understood phenomenon termed "filament notching" which could be likened to the deterioration of a link in a stressed chain until its eventual catatstropic failure . This is why a lamp can sometimes be "fixed" by tapping it while energized; the small gap in the filament will weld but fail again soon. The filament of a hot halogen lamp is much more delicate than in a cold lamp and this is why the temporary fix works but since the filament is wound in a very tight coil (like a spring) cooling will break the weld during the next cycle. "In-rush" current also has an effect on lamp life; this is the sudden heating of the filament to incandescence when the lamp is switched on at full operating voltage.
Low voltage lamps may benefit from the buffering of in-rush current through the transformer coil ("soft start"); line voltage lamps may also benefit from dimming switches since the in-rush current is bufffered by the dimming switch coil. I'd be pleased to have these theories either confirmed or denied.
My experience with premature low voltage lamp "failure" is that it is more often the socket/lamp pin contact that fails. We have been warranting halogen lamp life as part of our original three-year unconditional warranty and during subsequent three-year service agreements offered to our clients. I estimate that at any given moment we are responsible for the operation of at least 10,000 lamps scattered throughout our 75 mile service radius. This calculated risk gets our undivided attention at all times and we have kept very complete records for over 15 years. It often occurs to me that this service burden (which we bear stoicly) could be markedly reduced if the lighting equipment manufacturers would spend an extra dime on the quality of the lamp sockets used in their luminaires.
The lamp manufacturers are doing their part (witness the increase in average lamp life over the past 10 years) but the fixture makers are still too cost-driven to address the problem. Our records indicate that well over 75% of all lamp "failures" are really attributable to failed contacts within the sockets. What is needed is a mechanical- rather than a friction-fit to the lamp pins. This could be achieved with set screws to positively clamp the socket/pin connection together.
When I think of all the service calls we perform to merely re-seat the lamp in a "dark" fixture it is clear to me that we'd pay a dollar, not just a dime, premium to have a solid mechanical connection. Bi-pin sockets of this type are used in some European -sourced cable-hung systems.