test of a small xp-type UV led from the Uniquefire 602C-UV

I did a small review on the Uniquefire 602C-UV not so long ago, and yesterday I did some testing on the led that is in it, and I thought at least some of you would like to see the result. It is stated as a 395-405 nm led, so it is long wave UV, overlapping just a bit with the visible area (it has a clear purple glow). I assumed correctly in the review that the led has a xp-footprint, so for the test I could reflow the led on a 16mm Noctigon board (I have some of those spare because Intl-outdoor incorrectly sent me 4 xp-g Noctigons instead of xm-l's. Hank resent the xm-l's in the meantime, and I got the xp-g's for free for the trouble :-) ). I mounted the board on a chunk of aluminium with a reflector on top pointing at the ceiling and connected the power supply. I roughly measured current, voltage over the led, and ceiling bounce lux. It is not very useful to calculate lumens from UV-leds (the maximum output 30 lux would correspond to 62 lumens FYI, that is if the sensitivity of my lux-meter is correct at these short wavelengths), so I give the raw lux-numbers in the graph:

In the stock light the current through the led is about 1A. That gives a nice output, but as you can see in the graph, at least on copper (I even do not know what happens on the stock alu-board because I haven't tested it) you can go a fair bit higher than that, up to 2.4A with almost twice the output . However the voltage goes up quickly as well (it reminds of the Nichia 219) so with a single Li-ion you do not get there (I get 1.6A direct drive, a bit more on that later).

Concluding, if you are looking for a led in this wavelength, this led is very convenient to use in a simple flashlight mod, with its xp-footprint and a Vf similar to white leds, and as already shown in the review (although I can not compare it to many other UV-leds, I have a p60 drop-in of the same wavelength I bought two years ago, it has definitively more output than that one), the output is great! And the most convenient source for this led for me is just buying the Uniquefire 602C-UV for 11 dollars on dx (it leaves you with a nice cr123 host as well :-) ).

After the test I put the board in an old (dirty) p60 drop-in I still had somewhere, direct drive, 1.4A on a 16340 IMR, 1.6A on a 18650, and now I have a bit of a UV-thrower :

The UV does funny things with camera's, in reality the colour is all purple, instead of going from blue to purple to pinkish.

3 meters from corner:

4 meters from wall:

Now, if I want a really good UV light to find blood, pee and so on.. is this one good? Or are there better ones?

There's always better ones, but this one does have a suitable wavelength and quite some power for that. Best is to use yellow glasses to get rid of the purple glow, so you can spot the fluorescence more easily.

Big thanks for all your testing & write up! :beer: :beer:
Looks like a great led. Had to do a FT product request for the bare led. :bigsmile:

If your lux meter reads lux from a 390 nm LED, it is not a true lux meter. Lux is defined as being weighted against the photopic response curve of the human eye (basically green light). You should not be getting any significant response from a 390 nm LED (but pretty much all lux meters do).

See: Luminous efficiency function - Wikipedia


That is why I questioned my lux-meter being correct at those wavelengths. But even from a very good luxmeter with a correct wavelength-response curve I expect a certain reading because my eyes see a deep purple too (the led has some visible purple in the flanks of its output)

Even a hypothetical perfect lux meter that follows photopic response exactly would read some lux from a high power 390nm led.

The photopic response doesn’t simply end, instead it just gets very low at edges. If a source is bright enough you can see far outside where the response seems to end. 390nm is visible just less so then 405nm.
At the other end, high power 940nm leds are normally invisible so 980 would defiantly be considered ‘completely’ invisible but there are several accounts of even 980 being seen. These accounts are mostly accidents (one intentional / utter stupidity) as you will likely cause damage at the levels required for 980 to be visible.

Another issue is that leds don’t come close to emitting a single wavelength. They are labeled with their wavelength of peak power. Look for a spectral distribution curve or a spectral bandwidth number in the datasheet. Its fairly wide. 390nm leds will be emitting some 395,400,405nm. Making the led appear brighter then if it was only 390.

The keyword is “significant” response. Yes, a 390 nm LED will put out some in the blue, but it will be rather low. Also the photopic curve has some (very low) signal level there. Multiply those together and you get a very small number. If those numbers are 1%, the combined response would be .0001, so a 30 lux reading would imply 300,000 lux out of the UV led. Not likely.

Most Chinese lux meters use the same sensor chip/filter. Their filters don’t match the photopic response curve all that well (like quite a bit of response to UV and IR light). Even expensive meters have the same issues.