I was reading the specs on a bunch of red/green (diffuse white when off)
bi-color LEDs (real party animal I am). I noticed that the standard
was for about a 4x higher 'mcd' brightness rating for the green output
vs the red. Eg 3mcd red, 12mcd green.
Does this result in even brightness as perceived by the human eye? Does
turning both on at full brightness produce a uniform yellow? Or do you have
to take the ratio into account?
Green will look brighter than red, but human vision will "compress"
range of brightness to make 12 mcd look only about 2.5-3 times as bright
as 3 mcd.
As for making yellow: 3 mcd of red and 12 mcd of
usual-yellowish-LED-green will add up to 15 mcd of yellow, rather than
only a slightly more yellowish green.
My calculations:
Typical cheap red LED that gets only 3 mcd has dominant wavelength
(color specification that is approximately but not always exactly "hue")
close enough to 650 nm. CIE chromaticity coordinates would be about
x=.725, y=.275, z=0.
Typical cheap green LED in a non-green-tinted package has dominant
wavelength in the upper 560's to close to 570 nm, and CIE chromaticity
would be about x=.438, y=.56, z=.002.
The 1931 CIE standard has "y" related to luminosity as a function of
wavelength. (x is roughly reddishness, y is roughly greenishness, z is
roughly bluishness.) The "photopic function" of wavelength had an update
in 1988 making it differ from the "Y-bar function" in the deep blue and
violet, but this usually only slightly invalidates the below calculations,
and not at all when wavelengths shorter than greenish blue are not
involved.
So, as for "X-impact", "Y-impact" and "Z-impact", I would do this:
First: Multiply chromaticity coordinates by ratio of photometric
specification to the y chromaticity coordinate.
That results in:
Red: X becomes 7.91, Y becomes 3, Z remains zero
Green: X becomes 9.39, Y becomes 12, Z becomes .043
Add these up, and you get:
X totals 17.3, Y totals 15, Z totals .043
To get a chromaticity specification from these, divide each by the sum
of all three. That yields:
x=.535, y=.464, z=.001
This is close enough to the color of about 583 nanometers, which is
yellow and not greenish, although a little less orange than usual
yellow/"amber" LEDs (which have dominant wavelength typically in the upper
580's or close to 590 nm).
If this 2-color LED has a red chip that is of an orangish shade of red
as opposed to pure red, expect the yellow to be slightly less orangish
still, but probably not greenish.
What to watch out for:
1. Photometric specifications for LEDs are often only loosely approximate.
2. Some 2-color red/green LEDs have a red chip that gets less efficient
and more orangish as current increases above a milliamp or two, while the
green chip has efficiency increasing with current until the current gets
past 10-30 mA or more. Such an LED may have a lemon-yellow color at 20
mA per chip, but a more orangish color at lower currents, possibly even
reddish orange at 1 mA per chip, and may be a slightly greenish shade of
yellow at 30 or more mA per chip (if it does not overheat). This is the
result if the red chip is a GaP red with peak wavelength around 690-700
nm, and red LEDs with such GaP chips are sometimes called "low current
red".
- Don Klipstein (
[email protected])
