LEDs are mature

Also useful for counting stray cats as it lights up their eyes.

Reply to
John Doe
Loading thread data ...

The rate of luminance drop depends strongly on the LED current and junction temperature. The reputable manufacturers run their lamps at Imax/2 or Imax/3 and could very well achieve the claimed lifetime. OTOH, some noname manufacturers might run the LEDs at Imax and getting only a few hundred or thousand hours, before the light output drops and becomes bluish, due to loss of phosphor output.

For a street lights to last as long as the fixture (20-30 years) you still would have to drop the current to say Imax/4.

The CREE LED mentioned in this thread is rated for Imax=3 A, but characteristics are also given for Imax/2, Imax/3 and Imax/4.

Of course, running a 3 A LED at 700 mA requires 4 times the number of LEDs compared to running at Imax, thus costing a lot and it is also hard to fit that large number of LEDs into existing bulb sizes.

The measurement standards also limit how long you can extrapolate life times based on a limited (6000 hr) actual test period.

Reply to
upsidedown

You can beat 100lm/W with good CRI and power now. Lumens per current seems to be fairly constant across LED generations but Vf has dropped from 3.6 to 2.8 volts in the past few years.

The Philips web site always has nice charts:

formatting link

--
I will not see posts from astraweb, theremailer, dizum, or google 
because they host Usenet flooders.
Reply to
Kevin McMurtrie

So Philips has finally characterized their LEDs at usable Tj values of

85 C, which CREE has done for years.

For typical LEDs with CRI less than 85, IMHO the only usable for indoor lighting are the 2700 K versions. According to the referred page you get 107 lm/W at Imax/3 and 85 lm/W at Imax.

If those CRI=90 spectrum are as good as /940 or /965 full size fluorescent tubes that I use, the 5700 K LED will produce 110 lm/W at Imax/3 but only 88 lm/W at Imax.

Do not believe all the marketing hype, before checking the actual specifications.

Reply to
upsidedown

I suppose it depends on what color you're buying. The phosphor in 2700K and 5000K LED and fluorescent lamps looks awful to me so I usually buy

3000 or 3500 K.
--
I will not see posts from astraweb, theremailer, dizum, or google 
because they host Usenet flooders.
Reply to
Kevin McMurtrie

It depends on what the light is for (what room, etc.). I prefer 6500K for my workshop. I don't much care about color when I have flying sharp things near my fingers but I want to SEE.

Reply to
krw

I was wondering about that. I heard a claim that a "cooler" light makes colors easier to see. I'm not vouching for that, just wondering what the benefits are for the different frequencies. A little research would probably shed light on that.

Personally, the blue (or whatever) LED tint of old was a turnoff.

Reply to
John Doe

That claim doesn't make much sense.

High color temperature "white" LEDs have a strong deep blue spectral line and a weak continuous red/yellow continuous spectrum. A low color temperature will also have a strong continuous spectrum, thus you are going to able to detect far more colors.

Remember those low pressure sodium street lamps with only a single spectral line (pair) in the yellow, you could only see yellow and non yellow (dark) cars. With deep blue LEDs you can just separate between deep blue and non deep blue objects. With low color temperature LEDs, you can also see red, yellow and green objects separately.

Reply to
upsidedown

I'm having a little trouble following what you say here and in other replies. Do you have any citations for how well the different frequencies illuminate different things?

Reply to
John Doe

Cooler fluorescents tend to be much brighter. For this purpose, it's not the quality of the light, rather the intensity. Shadows are a killer.

Reply to
krw

Note that I'm not going to use 6500K tubes in the room I intend to use for finishing. A narrow spectrum can cause some "interesting" paint/stain colors. ;-)

Reply to
krw

There are three things to consider.

  1. The emission spectrum of the light source (in absolute units, such as mW)

  1. The reflectivity of the object, a white object reflects all col ours, a green plant absorbs blue and red light but reflects the useless green light.

  2. The spectral sensitivity of the human eye
    formatting link

For each wavelength, you have to multiply those three parameters together to get a picture of how well you are seeing something.

A narrow yellow low pressure sodium lamp will have a narrow yellow spectral line, while a high colour temperature LED has a strong blue spectral line.

If an object reflects that wavelength well, it looks light although discolored (but the eye adjusts quite well), if it reflects badly at that wavelength, it is quite dark.

Reply to
upsidedown

Look up "colour rendering index"

--
umop apisdn
Reply to
Jasen Betts

It is astonishing how good the eyes automatic white balance is. In the bad old days of colour film photography you had to use strong filters to adjust between daylight and incandescent lights and getting rid of the green casts from fluorescent strips was all but impossible.

It depends critically on what wavelengths the "things" reflect and with how much attenuation. In an ideal world you would want an emitter that approximates a 6000K black body to emulate daylight, but in practice you can get away with a strong blue LED emitter and a wideband phosphor centred on the yellow but containing some green and some red.

The typical curves of the "white" LED emitters is online at:

formatting link

Objects which only reflect light at 500nm look noticeably darker under LED "white" light (fortunately such objects are very rare).

Warm phosphors now have a boost in the bright red around 630nm.

PS If you want to convert a cool white LED into a warm one then I have found that Lee 110 Golden Amber will produce something that is very close to an incandescent lamp output in tone with some loss of light.

The harsh cold blue white isn't restful in a bedside clock for instance.

--
Regards, 
Martin Brown
Reply to
Martin Brown

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.