I suspect the current regulator info could be handy, which covers sinks as well as sources.

His LEDs are not going to drop the same amount of voltage for each, which means one size for all resistors isn't really that appropriate.

I used information right out of the datasheet for the LEDs that our OP supplied in post #3 of this thread.
Link to datasheet: http://forum.allaboutcircuits.com/attachment.php?attachmentid=12857&d=1255789701

It gave the same VF range for red, green and blue (actually, red YELLOW and green) for a 25mA current. I simply selected the average of the Vf range to represent the typical Vf. Even if they were all at the minimum Vf of 3.2v (highly unlikely); there is so much "headroom" (10.6v) that the chance for overcurrent would be small. 14v-3.2v = 10.8v; 10.8v/430 Ohms = 25.12mA. That's less than 0.5% over specifications - for all three colors.

I've been pretty busy at work, but I also have the impression he may want to make it variable, though I could be off on that assesment. I've been meaning to redraw the standard transistor model I have using a LM317 as a voltage reference.

Well, we still don't know exactly what his controller does, although he seemed to get output from the single RGB LED that he hooked up. I don't know if it randomly cycles through colors, acts as a "color organ" to music, or anything of that sort.

Using individual 430 Ohm resistors on each RGB lead was the least expensive way that I could think of for our OP to complete the project. They might've used DIP resistors to save a little space perhaps, but that would've been more expensive.

Even if they were to try to use a common current limiting sink, they would still have to use resistors on each RGB LED.

 

No argument there, but I'm a bit suspicious of the data sheet claims for the same Vf for each LED. Historically the variation in colors is pretty large.

 

It would be a good idea for our OP to take some sample measurements for Vf @25mA; just to see if the actual components agree with the datasheet specifications.

One thing I forgot to do is to check the Rlimit power dissipation - they'll need to use resistors rated 1/2 Watt or more!