In any serious application I would not consider driving LEDs other
than with constant current drive.
Exception might be if maximum current worst case was under LED
acceptable max and the brightness was acceptable in all cases.

With Denny's 3 LED example of 3 x 3.4V LEDs =3D 10.2 V you have  enough
headroom to run down to 11V or less using electronic constant current
circuitry.

At eg 350 mA (slightly over 1 Watt per 3.4v block*) you get about 1/3
Watt dissipation in the regulator per Volt of "headroom" with a linear
regulator. So at say 14V you get about (14-10.2)/3 =3D ~~~ 1.5 Watt
dissipation. A switching regulator will reduce this substantially for
higher power LED strings.

I say "3.4V block" as, while the system may LOOK like 3 LEDs in series
most strings of N LEDs  will effectively be 3 groups each of N LEDs in
parallel. If a simple series resistor is used for current limiting
then ideally each group of 3 physical LEDs should have its own series
resistor. Using only one resistor per system and placing N/3 LEDs in
hard parallel (as is very very very commonly done in these or similar
LED arrays, then differences in LED Vf:I charateristics can lead to
massive differences in individual LED currents. MOST Asian sourced LED
products using multiple LEDs use inferior or no LED current balancing
or control.

Unfortunately, you cannot rely on claims of "long life" without other
supporting evidence. Data sheet specs are only as good as the
reputation of the manufacturer or the independent test house who
verifies them. (A tour of a manufacturer's test lab and demonstration
of their instrumentation may be impressive - but is utterly no
guarantee that the supplied test data will be accurate - or indeed
bear any relationship to reality. (Ask me how I know :-) ).

..

             Russell
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.