Re: Quality
Re: Quality
<a href=showthread.php?s=&postid=15223011#post15223011 target=_blank>Originally posted</a> by herring_fish
I would like to have less LEDs for the same light output.
You still have options other than the Luxeon V and the K2 - get Rebel stars mounted 3-up, or Cree MC-Es, which are basically 4 XR-Es on the same die. Both of these options give you low "part count" in terms of the number of stars you have to solder and mount, but way better efficiency than Luxeon V's or K2's.
Just look at the efficiency numbers above: If you theoretically need, say, 20,000 lumens of LED light, you can get that by using 100 TFFC K2's. That would burn 365 watts, running at spec'd current and voltage (1A, 3.65v).
Or, you could get that 20,000 lumens by running 200 Rebels at spec (350mA, 3.15v) and only burn 220 watts. You could get those 200 Rebels mounted 3-up on Endor stars, and have 67 stars to work with. So it's fewer total parts than the K2 option, AND like 40% less power, for the same output!
67 3-up Endor Rebels is going to cost about $1000. 100 K2's will be like $800. BUT, this example assumes you are driving the Rebels at their design spec, while you can drive them higher, with the tradeoff of less efficiency. At 700mA, they produce 180 lumens per emitter! At that current, you would need 112 of them. Mounted 3-up, that's 38 stars. Now, the upfront cost is only $600! And, you'll be drawing only 266 watts. So, lower upfront cost, lower power draw, and lower parts count than the K2.
So, no matter what your metric is (best efficiency at spec'd current, lowest parts count, lowest initial cost, etc.) the Rebel beats the Luxeon TFFC K2 hands-down every time.
You could do a similar comparison of the Rebel to the Luxeon V, but it would be even more pronounced, since the Luxeon V is even worse than the K2 in terms of efficiency. IMHO, unless you are getting the LEDs for free or have some unique requirement, it doesn't make sense to use anything other than CREE XR-E (or similar, i.e. XP-E, MC-E, etc.) or Luxeon Rebels.