Randomly stumbled across this website claiming to sell 3535 LEDs in 97CRI. In addition to the higher minimum/typical CRI, this LED might offer two things over the 519A:
6000-6500K CCT in 97CRI. This also offers a higher-power alternative to the E21A or Optisolis emitters at this CCT.
R9 and R12 over 90. At high CCTs, even the best blue-pumped LEDs like the 519A tend to have R12 (saturated blue) hovering in the low or sub-70 range, so the claim of 90+ R12, if true, would be something new.
The site’s inquiry form requires the inquirer to be from a company, and I wonder if anyone here might be willing to give it a try.
Bummer to see such poor performance on the underlying die. Looking at the spectrum plots from the datasheet, the peak is still at 450nm like pretty much every other white LED, so the high CRI magic is probably not in the die, but in the phosphor mix. I wonder what’s preventing people from grabbing a super high efficiency emitter like the XPL2, scraping it clean, and smearing this high CRI phosphor over it, combining a high-performance blue pump with a high CRI phosphor.
Thermal resistance is the LEDs ability to dissipate heat, and all things considered, the lower the better. Higher thermal resistance means the LED gets hotter faster as it dissipates energy into heat. More heat is bad ,so we want LEDs with a low thernal resistance like the XP-L2, XP-G3/G4, XHP70.2 and 70.3 and 50.2/50.3. Higher values are not good for handheld flashlights and is indicative of an inefficient LED.
The lower the number the better something transfers heat.
So the three emitters you mentioned are close to 4 times better than the one in the OP at shedding heat.
In somewhat more technical terms:
" thermal resistance (R ) measures the opposition to the heat current (transfer) in a material or system. It is measured in units of kelvins per watt (K/W) and indicates how much difference (in kelvins) is required to transfer a unit of heat current (in watts) through the material or object."
My first thought was that some (round?) Sanan dies were used for this LEDs, but now I think these dies are maybe self-made. The specs looks like one of the first mass-produced and commercially available high-power emitters which are released fifteen years ago, at least for the electrical properties.
My only fear is that the light will be green. But seemingly they offer some sort of 3- and 5-step ellipse binning, so maybe they could be really good in terms of tint and color rendition.
Now it is more important to test these LEDs, maybe the thermal resistance is better than stated in the datasheet. If the thermal resistance is really that high, more than 2.5 - 3 Amps are not possible with this LED despite good cooling on DTP board, which is even worse than SST-12 or old XP-E(2).
@INeedMoreLumens Thank you very much for the datasheet! (Rare that this is around for Chinese emitters.)
I tried that once, as a proof of concept I used an LH351D die and attempted to separate the phosphor from a 519A. It worked as in lighting up, but removing the phosphor from the donor led was very difficult and I didn’t get it all, as corners of the glass broke and some of the phosphor was still attached.
I was originally planning on doing it to an XM-L2 (flipchip) to make something like an XM-L2 HI in high-cri, but the difficulty and expense involved in finding a good way to do it made me stop.
It is extremely cool that you tried this! The difficulty is not unexpected, and even if things did pull through the optical connection between the scraped LH351D die and the donor 519A phosphor would likely be very poor.
If we can have a bare die and raw phosphor, however, I can imagine this being much easier. Pour some phosphor powder over the die and add one drop of LED Seal or some similar sort of optical-grade clear sealant, and you have a domeless high CRI emitter.
If my understanding is correct putting phosphor over a bare die is not a super complicated, high-tech process–pretty much any LED manufacturer can do it easily, maybe even some hobbyists. It would be interesting if BLF could find a LED manufacturer (maybe through Simon?) who is willing to help us play mix-and-match between bare dies and phosphors from different third-party manufacturers, and create custom runs of LEDs made this way.
For the dies, it is not that difficult, it would be relatively easy to scrape the phosphor off a flip-chip emitter. Cree and Samsung dies are pretty tough, it just would be difficult to put the dome back on. On the LH351D for example I scraped the remaining phosphor off with a wooden toothpick, and it worked great as a 450nm blue emitter.
I didn’t think of using optical grade sealant, I was honestly thinking of mixing clear silicone with the phosphor in a known ratio, and then pouring it over the die with a mask over the non-emitting parts. That is definitely a better solution, though getting the phosphor spread evenly would be tricky. I would be very interested for example in doing that to some XHP50.2 (3v) dies, or an XM-L2 (flip chip version), so as to have good options for 5050 throw and flood.
