>> A major problem is that you are "sailing too close to the wind" for
>> the simple circuit that you are trying to use. You are trying to
>> approximate a constant current source by using a constant voltage
>> source and a series resistor.

Russell said it better than I did.  The use of a series resistor
and constant voltage supply to approximate a constant current
source is only going to work "well", if the voltage drop across
the resistor is "large" compared to the variation in voltage drop
across the other components with changing conditions and parts.
If you were driving three luxeons (at, say, Vf = 3 to 4 V), you'd
have about R=15 ohms and your current would vary from 1A to 0.8A,
or about 20%.  Driving 5 series luxeons you'd use a 9 ohm resistor
and your current would vary from 1A to .44A, or close to 60% (and
it'd be worse with 6 luxeons.)

You don't say which parts of your design are fixed, or how cost
sensitive things are.  The current you're trying to get (1A) is
close to the limits of "simple" LM317-based linear current regulators
unless you go to a fair amount of trouble in heat management.
That would hold true for any linear regulation scheme,  of course,
even resistors.  (With 3 series LEDs and 15 ohm resistor, you
dissipate 15W and probably want at least a 20W resistor!)

Switching regulator design has always seemed complex to me, but
it's getting to the point where you can buy simple-to-use chips
designed for driving power LEDs and use their suggested circuits.
Those tend to be relatively expensive (but on the third hand, you
may be able to replace the 24V regulated supply you're using now
with something cheaper...)

BillW
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist