I can confirm the graph in practice: plastic lenses almost completely blocks 365nm light and glass transmits almost all of it. I had to buy a glass lens for my Ledengin LZ1 365nm zoomie for it to work.
^
I just received the above light. Its the best UV LED I’ve come across, but it puts out much more white light than I expected. Even after adding one of those zsomething filters. Asked my wife if she also saw a bright whitish hot spot. She said she did too. Her eyes are 11 years younger than mine. Seems unlikely we both could see into the upper UV spectrum. Even with UV googles, the hot spot is pretty bright.
Doesn’t seem better than the cheaper 365 emitters I have purchased other than the output level. Unfortunately, I don’t have any other UV lights to compare to at the moment as my wife lent them to one of her co-workers. I do have a couple different emitters that I can use to build new lights with.
The emitter looks similar to the pictures above, but I haven’t closely compared it yet.
If you look at this post, you can compare how your led looks like. If the white stuff between the die-lines has disappeared, much more white light is emitted. I’m pretty sure that it is caused by the led overheating at some point, in my case the ledboard was floating because I did not tighten the pill well enough, or the solder job could have been bad, like Serp showed.
I bought a Convoy S2+ flashlight with Nichia 276A U 365 in a good deal and used the host for another build, so I had a new led spare for re-testing ( I ruined my original 276A U365).
I have improved the method a bit since the OP ( https://budgetlightforum.com/t/-/38887 ), it is a bit more standardised, and can now compare this Nicha with two other recent 365nm leds that I have recently tested with the same method. The new method is inaccurate in the sense that it assumes that the percentage of the total light output that is emitted in a 90 degrees forward beam angle is for all tested leds the same. This creates an inaccuracy but I have not yet figured out how to circumvent that. As long as the optical geometry of the led is similar across the tested leds (the usual small die with half circular silicon dome) I expect the assumption above to be reasonably valid and the leds can be compared like this. But in no way, this method has the accuracy of my method of testing visible leds.
Here's the comparison:
So the Nichia beats the l4p led but has less output and current capability as the Liteon led.
*the Royalighting 365nm led that is still available at intl-outdoor and is used in one of the Jaxman U1 flashlight versions is not in this test but from my old tests I can safely assure that its performance is under that of the l4p led.
*The Nichia has no thermal pad so can not be mounted on a DTP copper board, instead it was tested on the aluminium board it came on. Because of an observation from Serb above, I unmounted it and reflowed it again before the test to be sure it was reflowed well. Testing on the aluminium board is fair because there is no easy option to use the Nichia in a better heatsinked way.
*This method does test the amount of light around 365nm, but it does not test the fact that the Nichia emits way less visible light than the l4p led and the Liteon. This is a quality feature of the Nichia, but since the availability of the UV-pass filters that work very well (ZWB2 filter), for flashlight use it is not so much an advantage anymore.
*my standard output unit for 365nm leds has of course a relation with the output power in mW, but that relation needs a calibrated multiplier that I have not established yet. The Liteon test in the other thread suggested a multiplier of 16.5 (based on measured output and the specsheet), and the specs of the Nichia 276A (780mW at 500mA) suggest a multiplier of 25 which is in the same ballpark but srtill quite far off. More tests of specified leds are needed before I dare giving the output in milliwatt instead of my standard unit.
After Serps picture I’m quite nervous about the emitter of my own Convoy UV flashlight, have you found an indication that the emitter of your flashlight was also badly reflowed?
You tested the NCSU276A up to 1.2A, AFAIK Nichia specifies it only with 500mA, right?
Is it safe to drive it with 700mA in the UV Convoy on the aluminium board?
I remember I’ve read on the Nichia Website about mounting a UV-LED on copper but I’m not sure it was about this specific UV-LED. Would there be a benefit from a copper MCPCB even without DTP?
Hi, I bought Convoy S2+ with this LED from GB 3 months ago. I’ve used it few times, not more than 5-10min one time. The flashlight get hot after few minutes of use, so I thought heat dissapation is OK. I’ve noticed problem with my Nichia NCSU276A. There is brown spot in the centre of structure…
Brown part of structure (without white cover) emits less UV And there is a donough in the spot
Do you think I should try write to GearBest ? Where can I get a new LED in resonable price ?
No, I have not tested it, I do not have loads of these leds around. My impression that generally it is a good idea to maximise cooling of leds, but the output peak at 800-900mA is so sharp that I do not think that the decline of output over 900mA is mainly because of thermal reasons, it may just be the maximal output capability for this type of die regardless of temperature.
So my prediction is that you gain just a little bit of output when a better board is used. (I have been wrong before though )
Thanks for the update. Maybe you have the chance for a nice updated competition test between the LG LEUVA33W70RL00 and the updated Nichia 276A? ;-) The results might be interesting for building high yield UV flashlights (e.g. C8F Triple UV) without need for an additional ZWB2 filter in place.