>> Hmm. Just as a reality check, there is Sharps GH04P21A2G >> blue laser >> diode that emits 105mW at 5.4V, 120mA; about 16% >> efficient, while a >> royal blue luxeon 3 is rated 450mW at 3.9V, 1A or 12% >> efficient. I'd >> call that comparable and say it throws my original >> statement out the >> window... > > Pretty comparable.. But I need to operate at 850 nM Below: Comment on current LED efficiencies (20% typical, 40% known) plus a wander around related areas. _______________ As a data point, lumens (small l) as a measure of optical energy vary with wavelength as the lumen is defined in terms of a standard eye response BUT you can get a good feel for things by using lumens. For White (which the brain says is a spot somewhere near the intermediate point between complementary colours on the CIE chromaticity diagram*, 100% efficiency is about 350 lumens per Watt. The reason that daylight and cool whites tend to have higher l/W figures than warm whites is not (largely) because the former are more efficient but because they have more components closer to the eye's peak spectral response point. So ... 350 l/W is about 100% efficient, so 50 l/W is about 14%. Typical white emitters available on the market now have typical efficiencies of 50 l/W and max value of typically 40% higher, and minimum 30% lower (root 2 either way geometrically) giving 1 2:1 intensity spread around the typical values. ie Reasonable efficiency white LEDs are typically in the 10% to 20% efficiency range. However, you can buy DOTS (like COTS but dear :-) ) parts rated at 100 l/W or around 30% efficient and the best lab stuff is over 130 l/W or close to 40%. White is less efficient than the best monochrome LEDs that you can buy but not as bad as the worst as the technology used, to be able to directly make some colours at all is different than the most efficient known. White is typically arrived at either by mixing two complementaries with one coming from a LED emitter and the other from a reradiating excited phosphor (eg blue plus excited yellow phosphor) or by mixing 3 primaries directly (RGB). The blue plus phosphor approach is what made Nichia rich and famous. All phosphors are not created equal and many cheap white phosphor based LEDs show marked colour shift and/or diminution of output with time - often over a period far shorter than the 100,000 hours that everyone claims as of right for modern White LEDs. Actual lifetime tests can reveal quite a different story. Also interesting is colour shift with voltage. As LED voltage is reduced and intensity falls the colour shifts towards blue or green areas of the diagram. I assume that this is due to the phosphor having a non linear reradiation response with amplitude and being optimised for the amplitude expected at full rated current. This is especially noticeable when LEDs rated at 20 mA are run at under 1 mA. [[Yes Virginia, there can be good reasons for wanting to operate LEDs at under 1 mA average current!**]]. This colour change effect can be so pronounced that it may be worth while running LEDs in pulse modes to move the colour back towards white. Another cheap LED trap is encapsulant - anything rated to pass light for 100,000 hours needs to be resistant to colour and opacity change not just from the LED itself but from incident exterior light - especially UV. ` Russell McMahon __________________________________ * CIE Chromaticity diagram http://others.servebeer.com/misc/chromaticity.jpg Draw a straight line through the white area to touch two edges. The two corresponding colours can be summed to make the brain see white. Any number of colours can be vector summed such that, if the sum lies in the white area the brain sees white. ** But there is no Santa Claus ... http://en.wikipedia.org/wiki/Yes,_Virginia,_There_is_a_Santa_Claus -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist