View Single Post
Old 09-27-2006, 12:27 AM   #9
parksgm
FLAC is for flaccid
parksgm's CarPC Specs
 
parksgm's Avatar
 
Join Date: Dec 2005
Location: Dallas, Texas
Vehicle: Mustang GT
Posts: 1,116
My Photos: ()
09.26.06 update with pics

09.26.06 update

Quick update:

LEDs have one primary disadvantage...they are point sources of light, and so require diffusion to result in even illumination of a screen placed in front of them. Unfortunately, sheet-style translucent diffusion that is sufficent to provide even screen illumination also significantly cuts light output (highly counterproductive).

So, I researched clear diffusing lenses that could be placed over each LED to optically diffuse the light near the source rather than diffusing it by obstructing it with a translucent material on the back of the screen. LEDsupply.com offers many different lenses of this sort, which can be found at http://www.ledsupply.com/20mm.php

The most promising lens for this application seemed to be one that diffuses the light from a Luxeon into a long, narrow rectangular pattern, with an 80 degree span horizontally and only 10 degrees vertically. The data sheet is: http://www.ledsupply.com/docs/L2-Optics-Flare.pdf Scroll to the second page to see the lens distribution graphic or look at the last two pics in this post. While this lens seemed to be a viable option, it is also fairly expensive, and would add $40 to the cost of the project ($3.50 each x 12 LEDs). In addition, they would require that the screen be placed at least 2.5 inches from the backlights...making the entire assembly a bit thicker than I had hoped for. So...back to the drawing board.

I reconsidered (again) the advantages of using CCFLs instead of LEDs. CCFLs provide even illumination along their length and so require much less diffusion than the point source LEDs. The reason I had abandoned pursuing their use in the past was because I wasn't sure that they could be effectively cooled. However, after attempting to do theoretical thermal calculations on both the LEDs and the CCFLs, I decided that the only way to determine if I could effectively deal with the heat that either source produced was to build a prototype backlight with each illumination method and measure their heat output. So...since I had the CCFLs on hand and the LEDs are a $120 order and 5 days shipping away, I decided to try the CCFLs first.

I built a test rig (pics below) with two 13 watt CCFL lights, turned them on, and monitored the temperature for an hour. Links to all pics can be found in my sig, and a description of the hard drive cooler I am using as a heatsink/active cooling can be found in my posts above. To my surprise, the heatsink was never more than lukewarm to the touch, and the internal temperature stabilized at 101 degrees after one hour. External air temp was approximately 80 degrees. The next steps:

1) Add two more 13 watt bulbs to the enclosure (for a total of 56 watts of CCFL light) and run the test again. This will likely be the amount of light I will need to sufficiently illuminate the LCD in sunlight.

2) Test the rig at higher ambient temperatures...similar to those found in a car.

3) More clearly describe the methods used and testing I plan to perform.
Attached Images
     
__________________

An amateur built the Ark. The Titanic was built by professionals.

Last edited by parksgm : 09-27-2006 at 12:37 AM.
parksgm is offline   Reply With Quote