Nichia 033A UV LED: Shiny Metal Test for Visible Output

[Andrea B.]

Yes, the RGB values can be read in any editor and most converters.

 

By the time the photograph has been "translated" from the camera to the screen it has changed from NEF to JPG and also from camera RGB to jpeg sRGB to editor ProPhoto RGB to jpeg sRGB (again) - and it has been crunched by resizing which does affect colour sometimes - and then it is seen on a monitor via a browser (Safari, Firefox, etc.) So I have no idea how true to the original colours the photograph is here on the monitor !!! :D

 

But however it measures as an RGB value, I can tell you that it is close to what my good eye sees. And what you see here is close to how the raw looks in the converter.

 

***

 

I don't know why my spoon is purple. I think that I really should photograph another spoon or find some ball bearings. What if that is a weird spoon somehow which I photographed???

[DaveO]

I think we should call the colour of your spoon "Andrea Violet" as you are probably the only person in the whole world who would see exactly that colour, which is how you adjusted your spoon photo.

 

Remember we had a previous post about what colour UV looked like http://www.ultravioletphotography.com/content/index.php/topic/717-uv-sensitivity-among-mammals/

 

It certainly all seems consistent with what others who have had cataract surgery have reported.

 

Dave

[nfoto]

Maybe giving the RGB values [or value range] together with the colour space used would make comparisons easier. We perceive colours quite differently and the human eye is not very precisely calibrated and easily led astray.

 

One should also routinely check the used monitors for any colour inaccuracy. Reading RGB of profiled colour patches is a good starting point. At the very least, neutral greys of say a Color Passport should read out with pretty similar R,G,B numbers for each patch.

[DaveO]

Yes, I agree with that. I came to digital photography from B&W printing in my darkroom so I spent an inordinate amount of time doing Zone tests and such, none of which made my prints any better but it helped me to see where I was going. So... when I "went digital" I was always aiming at making prints, so my workflow is profiled as best as I can manage. I'm looking at this on an NEC P221W monitor (22 inches not one of the giant ones) which I keep profiled with an X-rite i1 Display Pro and finishing (if I choose to print) by printing on an Epson 3800 with a custom printer profile for each paper I use. I know that's all overkill just for web images but that's the way I work.

Dave

[Andrea B.]

Those are excellent workflow practices, Dave, regardless of whether you print or not or whether you use digital cameras or scan film.

 

If any images are requested from your work for publication, you are good-to-go!!

[DaveO]

I was fortunate that I came to digital imaging after most of the really hard stuff about colour management had been sorted out and made available to us mere mortal amateurs.

 

We have a wonderful resource here in Melbourne - Image Science which is one of the best sites I've found on all topics to do with putting images onto paper without any nasty surprises. If you want to learn about setting up monitors and printers spend a few hours in their knowledge base.

 

http://www.imagescience.com.au/

 

Cheers,

 

Dave

[Andrea B.]

Excellent link, thanks! I've been reading the Monitor section - about gamuts and bit depth. Good stuff.

[Andrea B.]

2020 UPDATE of Post #23

 

I wanted to look at the raw output from files shown in this old 2015 test (back in Post #23) of the leakage from the Nichia 365nm UV-LED to determine whether the raw histograms of the files would more rigorously show that the visible violet output from this torch was supressed (although not entirely?) by the Baader UV/IR-cut filter. Indeed, it is the case that the raw histograms demonstrate this well. (I did not have Raw Digger in 2015.)

 

Of course, *now* in 2020, as Cadmium has shown us over the years, we typically use U-340 filtration on our UV-LED torches to better cut the visible violet "leak". This would also show up in the raw histograms of files made with this filter.

 

I used the second set of Daylight White Balance photos to make the raw histograms. The selection of data for raw histogram was made in the center of the (brightest) white Spectralon target. That selection excluded the whitish contaminated areas on the edge of the standard. (Yes, the standards have been cleaned up since then. Many times in fact. )

 

Here is a repeat of the gear used.

• Camera: Nikon D600-broadband
  
• Lens: Zeiss 60/4.0 UV-Planar on Helicoid
  
• Lens Filter: Baader UV/IR-Cut
  
• Light Source: Nichia 365nm 033A UV-Led Torch
  
• Light Source Filter: Baader-U UV-Pass
  
• All photos f/8 at ISO-400.
  The UV-planar is not chipped so cannot communicate aperture to be recorded in EXIF.
  

Compare the shift in the blue histogram between the unfiltered torch photo and the filtered torch photo.

 

Raw Composite and Histogram for the unfiltered Nichia torch.

 

Raw Composite and Histogram for the filtered Nichia torch.

Definitely less reflected violet here. There's about a 1-stop shift the red and green and a 2-stop shift in the blue.

(Look for EV 0 on the charts.)

[Andy Perrin]

Andrea, did you mean "*now* in 2020" here?

Of course, *now* in 2015, as Cadmium has shown us over the years, we typically use U-340 filtration on our UV-LED torches to better cut the visible violet "leak". This would also show up in the raw histograms of files made with this filter.

[Cadmium]

Nichia doesn't make a torch, they make LED's. Do you say what torch you are using? Convoy? MTE? Or s0me other brand?

Lets look at your LED with a magnifying glass.

Which one of these LED's is in your torch?

If you are using U-340 2mm on any of these Nichia LED's then you will have no 400nm+ light coming from the torch.

Baader UV/IR-Cut doesn't cut below 400nm. Baader U should. U-340 2mm does.

What you are seeing in those photos is not a result of any visual light leaking out of your torch through a U-340 2mm filter.

 

 

[Cadmium]

Nichia LED's used in Convoy S2+ 365nm Nichia LED, old and newer version (A and C).

 

Older A version

 

Newer C version

 

U-340 2mm

[Cadmium]

Andrea, PS, if your torch is one of the two you got from me years back, then the LED in those is the Nichia NCSU276A, so you don't need to remove the U-340 or take the front housing apart to examine the LED for those.

The only torch I have seen with a Nichia 033 LED in it is the MTE 301 (uses Nichia NCSU033B), however there may be other torches that have that LED, but make sure that is actually the LED you have, examine it to be sure.

[Stefano]

To see which LED you have without opening your torch:

 

- Go in a large enough room (the larger, the better);

 

- Put some paper on one side of the room, and the torch on the other side pointing at the paper;

 

- Take a magnifying glass, in this case I think smaller works better (the image will be bigger), turn the torch on, and put the lens in front of the torch at roughly one focal length apart, and adjust it until you see a sharp projection on the paper;

 

- Go and see. You should see the LED chip with the pattern and identify it.

 

That's how I did it last time. I didn't want to open the torch again, and so did it this way.

[Andrea B.]

Cadmium, I have 3 sets of UV-LED torches.

One is a variable power Nichia chipped UV-LED torch made by Dr. Klaus Schmitt.

 

Two others were made by a fellow whose name I don't remember from the flashlight website that I don't remember.

(I have the links somewhere.)

They are 365nm and 380nm Nichia chipped torches in titanium cases.

 

I call the 3 mentioned Nichia-chipped torches smply "Nichia torches" because they are made by request and have no brand name. Like, they are tailored, made-to-order torches and not off the rack. [update in next post about the brand names.]

 

At least one of these 3 is chipped with an 033A. It's probably an obsolete chip by now. Still works good though.

 

I also have the set of U-340 filtered Convoy torches from Uviroptics on Ebay.

 

In the photos above I have clearly labeled the filtration on the Nichia torch as Baader UV-pass. That is what I used in 2015. As noted, the unfiltered Nichia torch passes a small amount of violet which is suppressed quite a lot, but not entirely, by the Baader UV-pass which what I used in 2015.

---

 

Andy, we certainly do hope that "now" is 2020 because we definitely do not want to repeat this annus horribilus. :grin: :grin:

I have repaired the data typo. Thanks for the catch!

[Andrea B.]

Added: oh this is cool -- it appears that my titanium torch cases became 'branded'. They are McGizmo Haiku torches made by Don McLeish. One was especially fitted with a Nichia 033A UV-LED chip when I bought it years ago.

 

https://www.everyday...og/mcgizmohaiku

See also: https://www.candlepo...gs-(NEW-5-12-17)&p=3139908#post3139908

 

Those are really old posts in the links. I do not know whether Haiku cases are still being made.

 

 

 

[Edit: removed broken link]

[Cadmium]

Well, I am more confused than ever now, but of all those torches, you don't have one that blocks the violet?

It may be better for me to ask, do any of your UV torches block the 'leak' you think you are seeing? Or do they all do the same thing to one degree or another?

 

Incidentally, not important, but I think what you are calling the 033A is the Nichia LED that was in the first generation of the MTE torch (MTE 301).

All of the Nichia 365nm UV LED's have very similar narrow emission graphs, all peaking close to 360/365nm.

And none of those transmit much of anything above 400nm, but enough to show some visible violet/blue if not filtered, as the photo below illustrates.

 

With visual blocking filters (U-340 2mm) you will see no visual emitted from the torch, although you WILL see fluorescence, just like the photo comparison shows in the photo below, which I have somehow posted here three times in the last week or so,

but it is quite explanatory, and people ask me this all the time, it is quite common for people to get fluorescence and visual violet mixed up.

"Why does my UV torch still transmit 'blue' with a U-340 2mm filter on the front?"

The answer is, "It doesn't". If you filter the torch with U-340 2mm, the 'blue' is not from the torch, it is fluorescence from the excited target material.

 

Let me go back to my question, do any of your torches not show this purple, or do they all basically do the same thing?

With a U-340 2mm filter on the front, nothing is emitted from the torch that is above 400nm, only -400nm UV, absolutely no visible light.

 

Brown cardboard:

 

White paper:

[Stefano]

I have to say that if you shine a Convoy S2+ + 2 mm thick U-340 into your eyes (to me it happened several times briefly, I never do that on purpose and I usually wear goggles), you CAN see UV. It appears whitish violet. It is almost invisible, but shiny surfaces will reflect it and you may see some violet. UV becomes really 100% invisible around 300 nm, don't expect 365 nm UV to be completely invisible, even if in most situations it basically is.

 

So, it is normal to see some violet with a correcly filtered 365 nm UV torch with your naked eyes. If you wear UV goggles (and you should), you won't see anything at all, even directly shining the torch into your eyes.

[Cadmium]

It is an LED, they can drill through the eye of an OD needle, not a good test.

That is not what we are doing That is not what she is seeing with her photos.

She is not photographing a filtered torch directly, she is photographing an illuminated subject. The direct viewed spot is not the situation here.

If you do LED tests that way you will end up with very misleading conclusions.

[Cadmium]

Andrea, Just to sing your camel to bed, try putting another 2mnm of U-340 in front of your already U-340 2mm filtered torch. Total of U-340 4mm.

See if there is any difference, because 4mm of U-340 will block even a direct head on shot of the the camel (UV torch/LED).

Sure, putting the Baader U in front of the already U-340 2mm filtered torch should block the spot, but it doesn't, so try the total of U-340 4mm.

I think you have a 52mm x 2mm thick U-340 around there, don't you?

Try that with your same subject also.

[Andy Perrin]

This is going to be one dead camel with all these shots to the head.

[Cadmium]

Andy ;-) Yeah, sorry, no more camel talk.

People need to make solid tests, and document things with photos.

The U-340 2mm removes visible violet/blue.

[Cadmium]

Stefano, I did a bunch of tests.

The best way to remove the center red spot from a direct head-on photo of the U-340 2mm filtered Convoy S2+ 365nm UV Nichia LED torch Is to put S8612 in front of it.

I used S8612 1.75mm thick, just because it was handy, and the red spot was gone, black (it might not require that much thickness, but it certainly would not need any more thickness).

I used a Zeiss T* on the front of the camera lens to block oncoming UV from the torch.

That combination removes everything from view by the camera (except visual fluorescence).

The S8612 could be stacked with the T* on the camera lens instead (instead of the BG38 or BG40 which I usually use for visual range photos on a full spectrum camera).

 

I tried other stacks, some using the Baader UV/IR-Cut in place of the Zeiss T* on the camera lens, and those still show some of the red spot.

I don't think the red spot influences the photos when using only the U-340 2mm on the torch. I think it is way too dim in comparison to any of the other light, UV, Violet (removed by the U-340 2mm), and fluorescence.

The red spot seems to be coming from the U-340 2mm 700nm range transmission, even though the Nichia LED shows a very narrow range in graphs that should not be showing up at all in the 700nm range.

When I stacked RG850 the red spot was gone also. So I think that red spot is in the 700nm range, but is removed when filtering your camera lens to visual range.

 

AND, good news, the red spot is removed when stacking BG40 2mm on the front of the T* also, and not quite totally with either BG38 or Baader UV/IR-Cut stacked on the front of the T*,

but darn hard to see, close enough with those. So it makes sense to stack BG40 2mm + Zeiss T* on the front of your lens to delineate the incoming visual range.

[Stefano]

Steve, thanks, that's what I expected. The Nichia 365 nm LEDs are known to emit little visible light, and Convoy torches can be used as they are, without the U-340 filter, because of how clean the output is, but for weak fluorescence you want a UV-pass filter in order to have really pure UV. Those LEDs probably emit some red and infrared, if you have a diffraction grating you can check yourself with your eyes (and UV goggles), you probably would see a faint rainbow from an unfiltered torch. U-340 does a nice job at reducing the red, it strongly but not completely attenuates it, and I don't think the dim red output can spoil your UVIVF or UVIIRF images, unless your fluorescence is exceptionally weak, at this point you might want to use S8612 on the Convoy as you said (if it doesn't fluoresce too much, otherwise you may even worsen the situation), but you will need it in very few cases, if any.

 

Your RG850 test confirms it is from the red leak of U-340. It isn't the LED piercing through it, those filters block UV extremely well, but self-induced fluorescence from the LED.

[Cadmium]

There is a fairly strong amount of visible violet/blue coming out of the Convoy S2+ 365nm UV Nichia torch when they are not filtered.

That is why we filter them to remove that violet/blue so it doesn't mix with the fluorescence.

The red spot is so insignificant that it doesn't need any filtering, the camera lens visual range filter would remove that anyway.

If you want to add red end filtration to a torch, then you can, but you don't need to do that.

For a flash you should have red end filtration, because that is very strong, and you will see that, like the example Pylon posted.

 

[Andrea B.]

This topic seems to have veered off into something else. Which is OK, but perhaps everyone seems to be missing the point of what I posted?

 

I revisited this old topic to illustrate -- mostly to myself -- that the violet suppression on my "Nichia torch" under a BaaderU would show up in a comparison of raw histograms from unfiltered and filtered "Nichia torch" photos of my white Spectralon target which reflects everything but does not fluoresce.

 

Yes, I could have used better filtration on the "Nichia torch" to better suppress the violet "leak", but I didn't and I don't remember why. It was **5** years ago !!

 

That is it. End of discussion.

 

---

 

BTW, there is no blue leak from "Nichia torches". It is mos def violet light.