I have been reading more and more of people making exposure boxes with LEDs instead of the conventional florescent tube construction. UV LEDs have become more readily available and I managed to pick up 200 390nm LEDs for $16 USD on ebay. I am interested in building one but have a few questions…
I need to verify some theory and math for figuring out how many LEDs I need. This is what I have so far.
theta: viewing angle of the LED
d: grid spacing between adjacent LEDs
h: distance from LED to PCB
Now the way I understand viewing angle is that it describes a cone of half power. In my mind that says that if I fix the distance from the the viewing point to the LED then the intensity at theta/2 is half the intensity of what it would be if I was viewing directly perpendicular to the LED.
From this I worked out that:
d = 2htan(theta/2)
Can someone verify this will give even light intensity over a surface?
The question has been burning in my head ever since I posted this… I decided to do some simulations.
If the simulation is accurate then the effect of properly spacing the leds just makes one BIG led. I am going to try and add an optimization stage to my simulation and see if by varying output power based on location on the leds I can smooth out the intensity.
I also attached a simulation for when the leds are too far apart.
I modeled the an led as a point source with radiation pattern X. The radiation pattern I used was from a particular led I am planning on using but the simulation could technically use any pattern since it is entered as a table of relative intensity levels as a function of angle with respect to maximum intensity. I then scanned through each pixel of the image and added up the contribution of light intensity from each point source at that particular location. I used the formula in my first post to determine how close the led’s needed to be in order to match up half-power intensities with each other.
I actually expected to see nodes and anti-nodes in the simulation but was very pleased to see how smooth the radiation was. I don’t think it will be too difficult to optimize this.
I would be happy to help you out and run the simulation for your led’s, just post your radiation pattern and led configuration and I will see what I can come up with.
Well, I meant what sort of software did you plug the numbers into? I’d like to try out a whole slew of LEDs, RGB versus separate Red/Green/Blue grids, and find a cost balance between large theta and just throwing more LEDs at it.
I’m not as strong mathematically as I was a decade ago so I’m more inclined to just slide values around. :mrgreen:
The distance between the LEDs and the glass makes a big difference. If they are too close together, there will be hot and cold spots. I used a scanner that is about 4 inches deep and it works really well.