scudinc
Enlightened
Even brighter LEDs. Of course, I won't get my hopes up until NewBie weighs in on this, but it seems like this could make every recent improvement in LEDs seem like a joke.
Just wondering, is that 4x brighter for a part that performs like typical Luxeons (~35 lm/W) or will these enhancements work for all parts? The reason I'm asking is that if you take Cree's XT-27 blue die you already have 42% efficiency so getting twice the light out the LED die brings us to 84%. Nice but I wonder if that would even be possible. Likewise, the best white LEDs tested to date are ~100 lm/W and maybe 30% efficient at converting power to light. It would be impossible to increase their brightness by a factor of 4 since that would be over 100% efficient. I'll hazard a guess then that a factor of 4 brightness increase wouldn't apply across the board.NewBie said:One is a surface structure change, and another is a shape change, each of which are supposedly able to get twice the amount of light out of the LED die. I don't know if the two techniques can be combined.
Then there are at least two phosphors that emit twice the light as the ones in use now.
If folks can get around the IP/Patent issues, or liscense them, we could be looking at parts that are 4x brighter than they are now.
jtr1962 said:Just wondering, is that 4x brighter for a part that performs like typical Luxeons (~35 lm/W) or will these enhancements work for all parts? The reason I'm asking is that if you take Cree's XT-27 blue die you already have 42% efficiency so getting twice the light out the LED die brings us to 84%. Nice but I wonder if that would even be possible. Likewise, the best white LEDs tested to date are ~100 lm/W and maybe 30% efficient at converting power to light. It would be impossible to increase their brightness by a factor of 4 since that would be over 100% efficient. I'll hazard a guess then that a factor of 4 brightness increase wouldn't apply across the board.
Regardless, if this brings even low end LEDs to at least 150 lm/W then it's welcome news. I'd personally love to see 90%+ efficient LEDs sooner rather later as it makes the heat issue practically moot.
Actually, the 27mW is at 20mA. With a typical Vf of 3.2V that gives the 42% efficiency figure I mentioned.NewBie said:Ah, but you are looking at those little die there, for little tiny SMT LEDs and 5mm LEDs which only produce 27mW from 30mA.
OK, so getting a potential 2X more light out of the die with the new process gets us to somewhere between 180 mW and 330 mW. With Vf=3.3V typical at 350 mA according to the data sheet, that would mean efficiencies in the 15.6% to 28.6% range. Even the improved efficiencies are nowhere near the present efficiencies of the 5mm dies.It is a different ball of wax, once you step up to the plate with a real die that folks would use in high power LED parts.
CxxxXB900-Sx000-A
http://www.cree.com/products/pdf/CPR3CM.pdf
On the lowest end you start out at 90mW output and head north of 165mW on the high end, with 350mA.
(side note, some have been asking about the Vf of the CREE, here is a good place to look)
What I am trying to say is that things are a world apart in the high power LEDs, as far as losses. A fella needs to think of them quite differently.
jtr1962 said:The reason I'm asking is that if you take Cree's XT-27 blue die you already have 42% efficiency so getting twice the light out the LED die brings us to 84%. Nice but I wonder if that would even be possible.
jtr1962 said:Actually, the 27mW is at 20mA. With a typical Vf of 3.2V that gives the 42% efficiency figure I mentioned.
OK, so getting a potential 2X more light out of the die with the new process gets us to somewhere between 180 mW and 330 mW. With Vf=3.3V typical at 350 mA according to the data sheet, that would mean efficiencies in the 15.6% to 28.6% range. Even the improved efficiencies are nowhere near the present efficiencies of the 5mm dies.
Anyway, I guess that answers my question. White power LEDs can easily get an overal 4 times efficiency boost and still not be anywhere near their limits. It would be nice to see some 150 lm/W parts, at least in the lab, with these improvements.
BTW, any idea if that 42% efficiency of the XT-27 die is starting to bump up against any real limits, or can we expect 5mm dies with efficiencies of 50%, 60%, or more in the near future?
HarryN said:I hate to break this guys bubble, but I had a discussion about using Q dots and nano wires as a replacement for phos with LL almost 18 months ago. At that time they were already well into a testing program comparing various materials to each other and standard phos materials.
It is one thing to demo a glow, it is entirely another to prove reliability and consistent color mixing. Keep in mind, that the Q dots and nano wires emit in VERY narrow bands, so great care would be needed to avoid having your LED emit in "spikes".