Real UV flashlight

[Smeerworst]

The 365nm 300mW flashlight is in my country

But customs like's to play with it for a while :tired:

PDF file for those intrested watching..

http://www.lohenstein.de/bilder/UV-Led.pdf

 

[Atomicrox]

Let us know if it leaks any visible light. Mine does emit UV but leaks heavily...

 

[Cyparagon]

LEDs are not monochromatic. UV LEDs are no exception. There will be some visible light. 365nm is visible anyway.

 

[Smeerworst]

Dark room
Left 365nm
Right >400nm ?

Light room
Left >400nm ?
Right 365nm

>400nm above minerals

365nm above minerals

UV reactive old mainboard
>400nm

UV reactive old mainboard
365nm

 

[DashApple]

Very nice , I plan to get a Ledengine 11 Watt 365nm led , with a output of 1200 mW

 

[Atomicrox]

LEDs are not monochromatic. UV LEDs are no exception. There will be some visible light. 365nm is visible anyway.

Are you sure 365nm is visible? Maybe we are just seeing very weak fluorescence..


Nice pics, SW!

 

[Cyparagon]

It can be seen through a spectroscope, and it is indeed further from violet. Some of it causes fluorescence in the eye, but if it was ALL fluorescence, your entire field of vision would light up instead of seeing a point as you do.

We know already that after a spectroscope there is pure UV light because it separates light based on color. Now think of a green beam going through olive oil. The orange light does not take the path of the laser, it goes in random directions. The beam is still green after it has passed through. Now imagine a ray of UV light going through the eye and hitting the retina. What color is it when it hits the retina? It's still UV.

 

[Sigurthr]

That is one nice UV torch! Gotta get a 365nm led one of these days...

 

[zyxwv99]

365nm LEDs don't cost thousands of dollars. 355nm and 375nm laser diodes do. (365nm laser diodes are not available.)

Humans can see 365nm. According to CIE1931 our photopic (color) sensitivity goes like this:

360 0.000003917000000
365 0.000006965000000
370 0.000012390000000
375 0.000022020000000
380 0.000039000000000
385 0.000064000000000
390 0.000120000000000
395 0.000217000000000
400 0.000396000000000
405 0.000640000000000
.
.
.
555 1.000000000000000
.
.
.
825 0.000000641530000

Back in the 19th century, Herschel, Stokes, Matthiessen, and Helmholtz each independently did research that involved shining uv light of various wavelengths into their own eyes (using a monochromator). Below a certain wavelength the violet changes to indigo, then blue, then grayish blue, then bluish gray, and finally around 290-300nm just gray.

Helmholtz determined (by dissecting a freshly dead human eyeball) that the retina is weakly fluorescent, accounting for the color shift and the gray. Once no more blue is visible, one is no longer seeing uv, just the fluorescence of the eye. (A similar color reversal takes place in infrared, with the red taking on an orange hue, but for entirely different reasons.)

Below 380nm the cornea begins to interfere with brightness and visual acuity. Below 350nm the aqueous humor does the same except worse. At 320nm all you see are shapeless blobs.

 

[Atomicrox]

Thanks guys, that's great info!

zyxwv99, can you post a link to the full table? the one I have here only goes from 396nm to 708nm (which is why I assumed it wasn't visible at all).

About the IR, at which wavelength does it change to orange? 808nm looks like a very deep red to me and people here say anything above 900nm is completely invisible.

 

[zyxwv99]

Since I'm new here I can't post links, but here goes.
(Just use search-and-replace on [colon] and [dot].)

The chart:

(from Light-Emitting-Diodes-org)
sample chapter (ch 16) from the book Light Emitting Diodes by E. F. Schubert (Cambridge University Press, 2006)
http[colon]//www[dot]ecse[dot]rpi[dot]edu/~schubert/Light-Emitting-Diodes-dot-org/Sample-Chapter[dot]pdf
appendix 16.1 chromaticity table for photopic (color) vision from CIE1931 page 290 of original text (page 16 of PDF document)
appendix 16.2 scotopic (night) vision CIE1951 (next page)

Color reversal:

Color Vision: From Genes to Perception
edited by Karl R. Gegenfurtner, Lindsay T. Sharpe (Cambridge University Press , 1999)
page 93
Perceived "yellowing" at long wavelengths (figure 3.3 on page 93 and figure 3.4 on page 94)
The text says it starts at 700nm
http[colon]//books[dot]google[dot]com/books?id=4zQMQLLVkFYC&pg=PA93#v=onepage&q&f=false
This phenomenon is also known as "infrared color reversal" referenced to "Brindley (1955)"

More on uv, including color reversal:
"The color of ultraviolet light"
http[colon]//www[dot]jstor[dot]org/discover/10.2307/1418730?uid=3739256&uid=2129&uid=2&uid=70&uid=4&sid=21102358641961

Random claims (they can't both be right):

"The generally accepted IR point of invisibility to the Human eye is 945nm. "
http[colon]//www[dot]nightvisionforums[dot]com/phpBB3/viewtopic.php?f=1&t=5366

"The human eye has nonlinear response to light from around 700nm up to around 1350nm and can see different wavelengths within that range at different intensities, however across that entire range, will see the light if it's bright enough.

Above that, there's no response from the eye at all."
http[colon]//www[dot]ar15[dot]com/archive/topic[dot]html?b=6&f=18&t=378252

 

[Sigurthr]

848nm is still deep red to me, though honestly now that I think about it 808nm looks redder than 848nm does (848nm looks like a very weak 680nm to me). So perhaps it is just a perceived shift towards and not a perception as orange.

I've definitely noticed the graying of UV as the wavelength gets shorter. 313nm is pretty much gray with a tinge of violet.

 

[Atomicrox]

Thanks for all the info, guys!

I wonder if anyone on these forums has ever managed to see 980nm...

 

[zyxwv99]

A few days ago I started a thread entitled What's the longest wavelength of laser light you have ever seen? Not too many replies. 808nm obviously. Someone else said they couldn't see the dot on a 1W 980nm laser. That's when I thought about how damaging it must be to look at the dot on something like that, even if you can't see it. That's why I think such experiments should be done with diffuse light. That's also why I ordered a 5 W 945nm flashlight. Eventually I will need increasingly expensive filters for it since unfiltered I'll probably be seeing the tail end at 845.

I'll be doing more research on this online.

 

[Sigurthr]

I just got mine in today! There is a fair amount of white spill over, but not enough to cause any issue for targets beyond a few feet. I'll have to see if I can source any wood's glass rounds in the right diameter. For close inspection it does wash out the 365nm a bit (but not any fluorescence from the 365nm).

Oddly enough I'm starting to question if the driver is different on this one than Cyparagon's. At 8.4V (2x 16340) the tailcap current is 565mA and it gets HOT with very short run time (<2min). The LED drop-in module was over 140F after about 90sec run time. This makes sense as the power draw is 4.74W and that is a lot of dissipation for such a small heatsink with little to no host conductivity. At 4.2V (1x 18650) the tapcap current is only 410mA and it appears to be equally as bright as with two 16340's, but the power draw is only 1.72W. So far it has been running for about 10minutes and the drop-in module is only 103F. I'm letting it run until cell voltage drops below 4V and I'll check the tailcap current again. Overheat issues aside, if the Po is equal between the two battery setups the 2x 16340 arrangement would only yield ~70min of run time (2.8WH cells at 2.3W/cell draw = 1.17Hr), where as the 18650 would yield 5hr20min of run time (assuming the driver is boost or buckboost).

The module temps make sense when looking at the power inputs when you consider that if at beast we're getting a 58% efficiency (1.72W/1W) you're only wasting 720mW as heat with an 18650 but 3.74W with 2x 16340s.

Update: 350mA draw at 4.05V. I don't have access to my lab variable supply at the moment, which would make this a lot simpler.
Update 2: 295mA @ 3.93V. Definitely looks to be not fully satisfying the diode's Vf. This confirms the need for 2x 16340s, but I can't say I like the high Pdiss involved. I worry for the life of the LED.

 

[zyxwv99]

I'm curious about the white spillover. It's probably fluorescence, but where is it coming from? On a cheap uv keychain light I can see the plastic/resin LED bulb glowing white. Glass will do that a little bit due to microscopic cracks and scratches. However, a major source of fluorescence is the human retina, more so at lower wavelengths. Is there any way you could put that through a spectrometer?