>
>>>>>> >>>>> >>>>>
>>>>>> I've got a problem. I happen to carry the distinction of being the one
>>>>>> responsible for changing lights. Much as that's a problem in itself, the
>>>>>> biggest problem is.....the front porch light (surprisingly, it's not a
>>>>>> stairway light, eh?), which happens to be a hanging, upside-down,
enclosed
>>>>> fixture. And they put two screws in the thing, so it's almost impossible
>>>>>> for a single person to replace the bulb thus inserted. Really quite
>>>>>> remarkable how no one thought of this.
>>>>>>
>>>>>> Besides modifying it, which I may consider because it's just that bad, in
>>>>>> the mean time I need something that'll last. We've already tried the
"ten
>>>>> gajillion hour" CFLs, which died in all of, you know it, three months.
I'm
>>>>> guessing something high voltage (130V+?) and incandescent. Any
>>>>>> recommendations?
>>>>> They are still expensive as heck, but you can get LED lamps. Perhaps
>>>>> below US$50 each for 100W equivalent, but always over 30,000 power on
>>>>> hours life and some (more expensive) are rated at over 100,000 hours
>>>>> power on.
>>>>
>>>> Buyer beware for life expectancy of LED "lightbulbs" and the like,
>>>> especially if they are white ones.
>>>>
>>>> 100,000 hours is a widely-repeated number, to the point of
>>>> "conventional wisdom".
>>>>
>>>> White LEDs often fade significantly well before then - I dare to say
>>>> they mostly do. Colored LEDs, on the other hand, appear to me to have
>>>> most operated for 100,000 hours or more to be keeping on trucking
>>>> rather well after working for so long.
>>>>
>>>> It appears to me that better heatsinkable white LEDs take 50,000 hours
>>>> to fade by 30% at "characterizing current" ("typical current") when
>>>> heatsunk to what I call a "moderately conservative extent". Much longer
>>>> life expectancy as in 100,00-plus hours appears to me reasonable if they
>>>> are both significantly underpowered and heatsunk to extent to keep the
>>>> junction temperature well below the "old traditional" maximum recommended
>>>> LED junction operating temperature of 85 C. In general, LEDs are more
>>>> efficient when cooler and most blue, white, and non-yellowish-green ones
>>>> are more efficient when moderately or moderately severely underpowered -
>>>> even regardless of temperature.
>>>>
>>>> A main difference between white LEDs and most colored ones is that white
>>>> LEDs normally have a phosphor and over 99.9% of colored ones I have seen
>>>> don't.
>>>> The few colored LEDs that I have seen with phosphor are pastels, pink,
>>>> lavendar/"purple" (as opposed to "violet", which is nearly-UV or an
>>>> indigo-bluish-violet color with some "blacklight" effects), or a
>>>> non-amberish maybe-slightly-chartreusish yellow close to 255-255-0 on a
>>>> usual monitor.
>>>> Also beware - violet and UV LEDs with epoxy bodies tend to age fast due
>>>> to UV or nearly-UV being hard on the epoxy.
>>>>
>>>> I have seen sone reasonably credible numbers for even better 5 mm /
>>>> T1-3/4 LEDs to fade significantly (don't know whether by 30% or 50%) in
>>>> as little as 6,000 to 10,000 hours. I don't know whether this is at
>>>> "characterizing current" of 20 mA or at maximum current of 30 mA. Some of
>>>> this is news years old and some I got more recently.
>>>>
>>>> One model of a white LED nightlight that I tried had "half-life" close
>>>> to half a year. I would prefer green or blue LED nightlights over white
>>>> ones for better life expectancy as well as having a spectrum more
>>>> favorable for stimulating "scotopic vision" ("night vision").
>>>>
>>>> - Don Klipstein ( snipped-for-privacy@misty.com)
>>>Instead of using white LEDs would it not be possible to use a number of
>>>red, green, and blue LEDs create the effect of white light?
>>
>> They actually do that. It gives a wierd color rendering effect - with
>>red objects coming out "day-glo" bright and wood tones coming up very
>>reddish/pinkish. Oak has a color like that of mahogany. Skin tones come
>>up reddish.
>>
>> - Don Klipstein ( snipped-for-privacy@misty.com)
>
>So how do they make the jumbotrons?
Since jumbotrons are not illuminating devices any more than TV sets and computer monitors are, they are allowed to have a spectrum (when set to "white") that gives such odd color rendering effects.
A display screen has R, G and B at any given point set to display a desired color and brightness. Ability to do that well does not negate a combo of all 3 having an overall color of white but a spectrum that causes "distorted" color rendition of illuminated objects.
In my experience, most color rendition issues from "white" "artificial light sources" (excluding incandescent, uncored carbon arc and pure xenon arc [which excludes automotive "xenon" HID headlights that are actually metal halide once warmed up] and xenon strobes with sufficient combo of capacitance and xenon pressure) result from excess or deficiency of yellow/yellowish spectral content compared to sum of red and green spectral content.
Mercury vapor, sodium vapor, metal halide in general (sometimes only slightly), non-triphosphor fluorescent, and white LEDs have "excess yellow" and "insufficient red-and-green". That results mainly in red/reddish objects being "dulled" in color. Red objects also appear darker than "proper". Skin tones, wood and brown paper bags and brown corrugated cardboard appear less-red than "proper". The same is true with xenon strobes with less capacitance of energy-storage-capacitor and/or lower xenon pressure - especially lower value for energy storage capacitor.
RGB light sources do the opposite - with above objects appearing "more-red than proper" due to "insufficient" yellow/yellowish, and "excessive" red and green spectral content. That includes RGB LED light sources as well as (to a lesser extent) most monitors being used as a white light source.
Triphosphor fluorescents do this just a little - but they "somewhat get close to a middle ground" by having a "narrow main red band" that is a very orangish shade of red (arguably a reddish shade of orange rather than an orangish shade of red - 611 nm), and having the "green band" being a cluster of yellowish green wavelengths centered around 542-544 nm. Along with a little bit of yellow mercury wavelengths and minor yellow and yellow-orange phosphor bands.
- Don Klipstein ( snipped-for-privacy@misty.com)