Connect with us

RGB wavelengths LED's

Discussion in 'Electronic Design' started by Ken Farrel, Apr 18, 2007.

Scroll to continue with content
  1. Ken Farrel

    Ken Farrel Guest

    Can someone tell me in nm what the optimal three wavelengths are for
    RGB color?

    I want to try simulating this on a small scale with three individual,
    high intensity LED's.

    Second, related, question. Please excuse.

    By varying the intensity of the three colors of LED's, I am looking to
    produce a seamless shift across the entire color spectrum.

    What kind of dimming circuit would best suit this application? I have
    seen current based ones, but was wondering if some kind of PWM might
    be more linear.

    Thank you kindly for any advice.

    Ken Farrel
  2. There are many meanings for "optimal" here - cost, color gamut,
    brightness, power consumption, size, lifetime, etc, and various
    coimbinations of factors.
  3. Since you are talking about LEDs, you will have to work with what
    exists or else contract to get some existing LED 'pushed' a little one
    way or another. Keep in mind, LEDs emit over a broad range, not some
    narrow line. So human perception of all of that will need to be taken
    into account for what you are planning to do.
    Some folks have already gone to some trouble to develop tri-color LED
    systems that achieve usable range of perceived color and high a fair
    degree of brightness, so that they can be used outdoors in broad
    daylight. Unless your needs are unusual, this is a market area you
    should examine to see what others have done here.
    The systems I've worked on use DACs to set an appropriate "white
    point" balance between the LEDs. This may be 100% or 25% or some
    other reference figure and the PWM is fixed during this calibration.
    But the idea is to tweak the white point to where you want it and then
    to PWM for brightness or for relative color change.

    However, you will need to be familiar with CIE color curves and how to
    compute brightness and normalized (x,y) to use against your white
    point to generate an apparent hue. You will need to invert this
    process if you intend to drive the hue and need to develop (x,y)
    values to get there. That will probably involve a calibration step
    with your tri-color LED (can be performed at most any set PWM level --
    even different PWMs for each color, if you like) to develop the matrix
    and then invert it for use.

    I've done some of what you are trying to do in a practical system that
    met my needs.
    No problem. I can recommend a book or two, if you are interested in
    human perception of colors and brightness. Lots and lots of research
    articles (some very fine ones from Edmund Land back around circa the
    last 1970s and early 1980s, if memory serves, too.)

  4. I should have added that you don't have to do it this way. You can
    avoid the DAC, I think, and just calibrate the LED and its surrounding
    design (resistor and voltage supply, for example.) Then use that
    calibration step for the rest. You'll need to do that calibration
    anyway, I think, if you are really trying to accurately sweep the
    apparent hue in some designed fashion. If you don't really care about
    any of that and can tolerate just a sweep without any necessary
    similarity in one unit from another, then you can pretty much just
    develop some function that takes in a scalar value representing your
    current color point and that develops the three PWM values to use from
    a basis function chosen for mathematical simplicity.

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day