Intl-outdoor's 3535-size 365nm UV-led on 16mm Noctigon (Royalighting RY-3535P)

UV-leds have neither the power nor the wavelength to be the least effective in killing germs. UV-cfl tubes and other UV-sources have way more power and go much deeper into the UV-spectrum. Leds become only interesting if a directed beam is needed, cfl tubes can not do that.

Picked one of these up. You think ~1350mv is good

My in-law was saying that it shouldn’t shift with higher power

No, the Ledengin shows a shift of a whopping 0.3 nm when going from 400mA to 1000mA, I'm not sure if anyone is concerned by that ;-) .actually, temperature has a considerable larger effect: 3nm going from 20 to 85degC.

This chinese led surely will have other characteristics, but I would be surprised if the current->wavelength shift will be disastrous.

But apart from this, why run it at 1350mA ? Stick to 1 amp and you are almost at maximum power, you just make heat with the last 350mA (and stress the led unneccessarily).

I’m like dbc… I’ll try it at a amp then thanks djozz

If you refer to the KD drop-in that DCBstm posted here above somewhere, with the two die's side-by-side, this led will perform better already if only because it is a single die that will have a better beam, even in a small reflector. But I would think that the output is also better because it is a more modern led: the datasheet is from 2012, the two-die led from KD has been around for much longer.

The 16340 host is not the UF-602C host, but it is noticably smaller, and of better quality. It is this flaslight, unfortunately sold out at Fasttech. The 2x7135 NANJG driver came with the stock light :-)

Kaidomain P-60 drop-in

LED ENGIN LZ-1

Images taken with Canon 20D, in RAW, Manual mode, on a tripod, 1 second f/4.5 at ISO100. 10’ from bookshelf to camera sensor. No sharpening or any other adjustments applied, RAW converted to JPEG for sharing.

Base image with Auto settings and shot in JPEG

The Kaidomain drop-in is less bright with mutliple cells, the driver cuts the amperage the higher the voltage is to maintain a constant output current. I’ve tried it with as many as 6 cells (hope I’m remembering that right, I used 18500’s with multiple extension tubes on an L2P) directly comparing output from 1-6, 1 being the brightest in overall output.

Convoy has introduced an 18350/16340 version of the S2 that will be available pretty soon. Richard has some coming in. :wink:

I wonder if we got different products, then. I’ve only tried one and two cells, but it’s much much brighter on two cells. My DMM confirms it, too. Even the amperage goes up with two cells. Overall power used increased by a factor of about 4 or 5, IIRC, and it shows.

Mine does look the same as what I’ve seen posted elsewhere, two dies right next to each other under one dome, but it seems people’s results vary. In any case, I may just have to get one of these nicer LEDs and make myself a new little UV light. It’s too bad the upgraded 602C host was only made for a short time.

Awesome review. Thank you for doing this for us.

You should consider the Nichia NCSU276A

Under promises and over delivers if you want a good 3535 UV part

Spec’s look good, to bad it uses a XB pad layout, not XP like these other’s. Wonder how expensive they’ll be to the US. Luxeon makes a good UV emitter to but they’re just as much money as the LZ1 is for less output and lower driver current at the same higher vF.

yes you would need to use a 119B board

Would using Kapton tape, e.g.:

work with these emitters and an XP-G board?

Yes you just need to keep the anode and Cathode pads of the XP centre thermal pad to avoid shorts

I'd love to have a look at that Nichia one, but it is 40 dollars shipped (leds.de), and that is too much at the moment for one led, some testing fun and then a modded light in a wavelength that I have already 'covered' twice.

Perhaps later sometime :-)

EDIT: Also, I am cheap ;-) , and so is this forum, tests of cheap parts are more interesting than tests of expensive parts, especially if that reveals that you can make for cheap a flashlight with the performance of something more expensive. In case of the (affordable) tested led of the OP has the advantage of fitting on a DTP copper board so that it can be run at 400mW output. I made for 17+9=26 dollars a (estimated) 300mW OTF uv-flashlight with it with a performance of one that would cost you many times more (a quick internet scan learns that usually they are over 200 dollars). That is what BLF is about :-)

The issue is with heatsinking, because there is no center thermal pad those style LED’s are designed to heatsink threw the front, to overdrive them, or even run them at max current you need to have them heatsinked correctly, the best way it threw the reflector touching the base board (the base of the COB LED, not the MCPCB), thats what they’re intended to use.

I’m with you, why take a chance its not a good performer, or deal with the lackluster heatsinking, or even mess with it at all when the LEDENGIN LZ1 (thats currently the top performer) is cheaper to get, easier to use and a better design

If calculated cost of parts, measurement tools and time, actually not that cheap, until data is release for public. So thanks again.

If not tested how do we know?

I think we must have some kind of fund from contribution of members at their will for test purpose.
Let sb or someone else hold it.

This would NEVER work.

If you like someone’s testing / work PP them yourself, a community fund will never work, ever. Also instead of money you could buy the LED and send it to DJozz or someone for testing, not all of it has to be out of his pocket…

A community fund just will not work, to many upset people thinking their work is the “best” and that they deserve the fund’s more than someone else, its just not even feasible.

edit: How I know is the datasheet, it’s not even close to the LZ1.

CK dont say that cuz it does work and LPF has had MANY MANY MANY ‘murder funds’ no we dodnt go kill someone but instead killed diodes in the name of testing. We killed or tried to kill a $300 diode once. Obviously no one gets anything out of it other than test results but weve done it alot on LPF

I don’t mean community fund for contributors, but fund to buy parts. I think it’s different.