I'm getting confused! FIL wants an expensive reading lamp, but surely the only (main) difference is the LED being used compared to cheap lights.
I can find high CRI bulbs but they have low colour temperatures. How can a CRI bulb of >95 only have a colour temperature of 2700K? Isn't daylight nearer to 6500K than the warm 2700K? Or am I being confused about colour rendition and temperature? Ta.
I use these, but £15 for 4 on Amazon: Sauglae LED Corn Bulbs,
B22 Bayonet Cap, 12W, 1200Lm, 6000K Daylight White, Pack of 4
They are very bright.
To an extent you are. The bulb you describe should have good and accurate colour discrination, but colours should look like they would in tungsten incandesceent light, not as they would in daylight. This might be an advantage if the user wants restful articial light as in past days, and less good if he wants to do accurate colour matching standardised on daylight, as might be needed for some commercial or artistic purposes.
AIUI you can have a 'white' bulb that is a black body radiator that has a full spectrum, and a 'white' bulb that has three sharp R/G/B spectral lines. It is the smoothness of the spectrum that establishes CRI, while the position of the spectral lines sets the colour temperature.
I don't know how good the LED phosphors are these days, as to whether you need a special CRI bulb or just one with a regular phosphor is good enough.
Should probably find a prism to try it out...
Theo
OK, many so called white lights only have peaks to simulate the light to appear white to our easily foxed eyesight. But they can cause colours to appear different and some eye tiredness effects. If you fill in the gaps in the spectrum to create something similar to the true spectrum of sunlight, you are going to use more power in doing so, so effectively they seem less efficient. That was how it was explained to me. Others may know more. Brian
Partly perhaps... the CRI is tied to the spectrum produced by the lamp rather than its overall colour temperature.
Most "white" LEDs are actually blue LEDs coated in a fluorescent phosphor to convert the light from the LED into something approximating white light.
This tends to mean you don't get a continuos spectrum, but one with various peaks and gaps. So the more even the intensity across the spectrum, the better the CRI.
The colour temperature will also affect the appearance of colours - but if the LED is a good mimic of tungsten for example, then things should look much the same as they do under incandescent light. A good daylight lamp, should mimic daylight (if its bright enough).
(you can still get some odd effects with some fabric dyes that also include fluresscent material, since they may look different in "real" daylight since they emit light of their when stimulated by the UV in sunlight).
Nice comparison slide here:
G, H, I show daylight at various times of day. The notches and non linearities caused by atmospheric absorption and scattering.
E is a blue LED on its own D is a blue LED with a "white light" phosphor coating, actually not a bad representation, but a bit lacking in red, and too much blue, hence the familiar slight green cast.
Thanks for the link. If G is daylight, this is normally described as a colour temp of 6500K. From the spectrum of G, how could you work out the colour temp? And if you take a high CRI LED lamp, say 98, the spectrum also looks similar to G, but the colour temp is given as 3000K. How does that work? If you have a dimmable lamp, does the colour temp vary with brightness?
I am prepared for the situation where I will never understand this!
Yup - it has lots of energy well up into the blue end of the spectrum, which pushes the average colour temperature up.
(the colour temperature being just the temperature you would need to heat a perfect black body radiator to so as to get a similar light spectra).
There is a thing called Wien's "displacement" law (an approximation of what was later refined by Planck) that relates the colour temp to a peak wavelength. It's given by :
lambda = b / T
where lambda is the wavelength, and b is a constant ~ 2.9 x -10^6 m.K, and T is the colour temp in Kelvins.
So a T of 6500K gives you a peak wavelength of 2,900,000 / 6500 = 446 nm
It basically shows how the whole spectra shifts in peak energy toward the blue at ever higher black body temperatures.
(it fits in with our normal expectation of how things heat first to a dull read, then orange, then white, right through to the blue of a plasma)
There is a nice graph on the page her:
Usually there will be less short wavelength content, see the A filament spectra. However you can also "fake it" to an extent by boosting the red / yellow content of light that does also contain more blue - as is often the case with "while" LEDs that use a blue LED light source to excite a phosphor.
With a tungsten lamp, yes - very much so. With a normal dimmable LED, then no - it stays the same at all brightnesses and looks a bit odd because of it (our expectation is that higher colour temperatures are usually experienced at much higher intensities - so daylight looks natural when very bright, but starts to feel "cold" when experienced at lower intensities (hence why the light looks "colder" on an overcast day, or why a single LED daylight bulb can look very cold or blue (but lots of them together seems quite "sunny")).
Some posher dimmable LEDs have more than one chip - one having a very yellow filter, and they blend the output of both in varying amounts depending on brightness. These do dim in a way similar to an incandescent lamp, with lower levels appearing more "warm" in tone (i.e. lower colour temperatures)
Also keep in mind that the term colour temperature works in the opposite direction to the terms we normally use to classify light qualities. i.e. we say "warm white" meaning a *lower* colour temp, and might describe a very high colour temperature lamp as being very "cold"!
HomeOwnersHub 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.