This is pretty typical. Usually you'll see some that are greenish, yellowish or blueish in a particular batch. If you're using this for anything requiring precision, you may be able to ask your supplier to provide you with more consistent batches or to sort them for you (if you're using them in high volumes).
What the spec sheet usually tells you is the "typical" spectrum curve, which is usually indicative of an average of a wide distribution.
If you're sorting them yourself, lower the current through them or use the diffuser to dim the light enough to distinguish the different colours.
Another possibility might be to use RGB LEDs and adjust the color balance yourself. Probably more stable over time (both brightness and color balance) than phosphor-type white LEDs too.
Best regards, Spehro Pefhany
--
"it\'s the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
But that calibration takes effort to do well, if you are trying to make one "look like" another one next to it. I've designed both the software and (using a CIE method as a base) mathematical portions to allow a technician to "tweak" three different trim pots, one each for red, green, and blue dacs that controlled the current through each.
Before using the tool I developed, the technicians had used a standard piece of equipment (not cheap) that provided a calibrated (x,y) CIE color curve coordinate value and the technicians had a particular (x,y) they needed to "hit." But in tweaking three pots to get there, it was a rather frustrating process because adjusting the green, for example, would impact both CIE coordinates in rather complex ways. It could honestly take them a significant time (and some experience, to get better at it) to center the (x,y) where it was needed to be.
What I did was to develop instead a convolving matrix that teased out the various impacts of each of the three pots, using a single "snap shot" with a spectrophotometer to calibrate it accurately, using any initial setting of the pots (didn't matter, so long as the LEDs each put out something measurable -- rough mid-point settings was good) and that was all that was needed to exactly then provide specific slider bar displays the technicians needed to set each pot individually, without worrying about the effects each had on the (x,y) coordinate. When they got each slider centered on the 100% tick mark, whether by left or right turns, they were done with that pot. And when all three were done, the (x,y) was always exactly on target.
So it can be a little bit of fun to calibrate a good white point. Not to mention that the three dies also need to be "close" in order for them to appear to mix.
Jon
P.S. Another complicating problem, if you are being really peevish about high stability in the intensity output of any LED, can be the erratic drift some LEDs have even while holding them temperature- stable in an oven and controlling their current drive, precisely. A lot of them are poor at holding the same optical output, even under these circumstances. However, if you monitor their outputs and run a bake cycle for a few hundred hours, you can separate out those which do not appear to stabilize from those which do appear to "anneal" into a very slow drift over time. Those which are stable after a few hundred hours make for good optical standard candles, given that you stabilize their temperature and current drive, of course.
Is there any specification on the spectral spread of white LEDs?
I have a few LEDs from a batch that have slightly different colour when put through a defuser. (3mm LED). The difuser seems to filter out certain wavelengths and the colour difference can be seen.
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.