UV TriColour and UVB photography with a Soligor 35 mm f/3.5 lens

[Stefano]

Jonathan sold me some 25 mm filters some time ago, including two BrightLine filters, a 340/26 and a 387/11 (2 pieces). These filters pass two distinct UV bands with no overlap, and I thought they could be useful for TriColour UV. Since they leak IR, I had to stack the 340/26 filter with 2*2 mm thick ZWB1 and the 387/11 with 2 mm thick Chinese BG39.

 

For the blue channel, I used two chinese 310 nm filters stacked to eliminate any leak (alone, they leaked some 340 nm light).

 

I putty-mounted all filters in 30-26 mm step-down rings. To avoid vignetting, I had to reverse-mount my step-down rings, so I mounted a 46-43 mm ring on the Soligor, then a 43-37 mm ring on top, then a 37-30 mm ring upside down and then the 30-26 mm rings containing the filters. I can not screw completely this assembly on the lens because something touches the metal ring with the lens name and the serial number. The only rings which allow for reverse-mounting are those with wide threads.

Below two 46-37 mm step-down rings, they are the same except for the threads.

 

46-43 mm, 43-37 mm, 37-30 mm and 30-26 mm step-down rings mounted as described:

 

*Edit: I noticed that sometimes narrow thread adapters allow reverse-mounting.

 

The filters putty-mounted on the 30-26 mm step-down rings. From left to right, 310 nm bandpass filter (stack of two), brightLine 340/26 and brightLine 387/11. The BrightLine filters are "lipstick mirrors" as Andrea would say, and that makes dust very visible (not in images luckily). The 310 nm filter has an orange side and a gray side. The orange side is UV-yellow and IR-orange.

 

310 nm filter, orange side:

 

My Soligor, according to Ulf's data only transmits 1% at 310 nm, which means I have to use very long exposure times at very high ISO and so my images are noisy at 310 nm.

 

These two are my first attempts at UV TriColour:

 

Camera: Full-spectrum Canon EOS M

Lens: Soligor 35 mm f/3.5

 

Red: BrightLine 387/11 filter + Chinese BG39 (2 mm);

Green: BrightLine 340/26 filter + ZWB1 (2*2 mm);

Blue: Chinese 310 nm bandpass filter*2.

 

You can notice a dichroic discoloration and a very noisy blue channel.

 

Magnifying glass:

 

Wooden house:

 

UVB photography

The blue channel in the photos above is in the UVB band at 310 nm. Since this wavelength is recorded basically in the green channel only, any signal in the red and blue channels is noise, and by keeping the green channel only I can cut some of the noise. This is not enough to have good images though.

 

Some UVB photos at 310 nm. Lens (Soligor) at f/3.5 or f/4. I took the green channel only.

 

ISO 25600, 60 s exposure, brightened by 50% in Windows Photo editor

 

ISO 25600, 30 s exposure

 

ISO 12800, 1/30 s exposure. This is the Sun. I don't know what causes that weird reflection effect.

 

ISO 25600, 30 s exposure. here I also tried to improve the image by adjusting the contrast and the brightness.

 

Comments

Although the Soligor makes it possible to take 310 nm photos, it is at the limit of what's possible. A better lens (such as a quartz lens or a lens transmitting significantly more than 1% at 310 nm) would help immensely.

[Andy Perrin]

Nice photos. I am thinking that ideally colorful tricolors should have overlapping filters so that when both filters are “on” you get intermediate colors, as we do with visible light. I know we all work with what we have though. I am curious to see how this does with a quartz singlet

[Stefano]

If the SvBony focal reducer David tested has good transparency at 310/313 nm (like 50%), then I might use that. Assuming it has 50% transmission, that is 50 times more than the Soligor, which is a significant difference.

 

My filter stack has a (theoretical) peak transmission of about 25%, so that has an impact too.

[diant]

If the SvBony focal reducer David tested has good transparency at 310/313 nm (like 50%), ...

Stefano, 50% is absolutely impossible.

SvBony focal reducer - usual astronomical 2-elements multi-coated unit. MC only reflect a ton of UVB. Its elements of course are made from different glass, one of which usually more opaque in UV. If you obtain 10%, you may consider yourselves a lucky one, I think.

[Stefano]

Looking at David's photos it seems like it is 2-3 stops below quartz, so maybe the transmission is around 20%, but I don't know.

[diant]

Where I can see that David's photo? May we say that they have equal f-ratio? 20% (if such case will be) - is something near a miracle. But a miracles yet there exist :)

[Stefano]

They are here: https://www.ultravioletphotography.com/content/index.php/topic/4787-homemade-lens-tests-at-313nm

[Andrea B.]

Well I do wish I had those filters so I could trial them for you on a UV-Nikkor or UAT to see how things come out.

 

Good write-up on your tricolor experiment. Thanks for posting this.

[Andrea B.]

Stefano, out of my own curiosity, can you make an indoors UV photo of one of your sunflowers using that 310 nm stack and your 310 torch?

[Stefano]

I tried with the UV LED but I barely get an image. I may try it later.

[diant]

They are here: https://www.ultravio...-tests-at-313nm

Ok, I've looked there. But I don't know how good is these "U330WB80 Improved filter and 313bp15 filter".

I can only note (from my astronomical experience) that interf. BP filters very seldom have OD>4 in cutting regions (more usual around 3). But in this situation every bit of leakage in UVA region (where sensitivity every CMOS much higher than at 313nm) may contribute its share in forming of integral image, instead of true UVB. And I don't know how to separate them...

[Andy Perrin]

330WB80 is a dichroic filter by Omega Optical and the OD varies, but probably 4-5 (or better) out through 1000nm. The stack of the two should be very well blocked since the 313bp15 will also contribute. The skepticism would be justified except that we have a lot of experience on the board with that 330 filter so we can be pretty confident it’s okay.

 

ETA: Here is the OD of the current version. It's OD 6 even, out to IR where there is a weak spot of OD4. But we can tell from the images that no IR is getting through. The vegetation is not white or anything...

https://www.omegafil...om/product/4399

[diant]

Andy, I say mainly not about 330WB80 but about 313bp15.

330WB80 may be as good as it can in region longer 380 nm. But in region 380nm and shorter it is opened. And if we speak about UVB (313nm) imaging, it has equal clear window from 313 to 380. And only filter that remains for separate UVB 313nm from UVA is 313bp15. But can we be sure that it have no sensible leakage in UVA (330-380) where CMOS chip is much more sensitive?

No skepticism, I really don't know two points: 313bp15 OD in UVA and Olympus EM1mk1' CMOS real UVB/UVA sensitivity. Here I'm only reasoning to clear this points and escape possible confusion in what we see - real UVB 313nm image or mainly UVA one.

 

Exoterras (yet they exist in different spectrum) also often have a large peak in UVA (as we see in spectrum below).

[Andy Perrin]

Hm, a UVA leak in other words. Well, it is a little weird that he didn't get anything with the 310nm LED.

 

That camera I think has in the past been shown to record UVB in dabateman's tests? I think the bigger issue is that the LENS (as Stefano noted) is not transmitting much at that wavelength, so if anything is making the leaks bad, I would think that would be the number one issue.

 

dabateman has a crime scene imager with an image intensifier tube that works out in UVC even, and a quartz lens, so I think his tests can taken as comfirmed examples.

[Stefano]

I don't want to interrupt your discussion, but I want to post a test for Andrea.

 

Rubik's cube, room lights, no filters.

f/4, ISO 100, 4 s exposure:

 

310 nm LED ~4 cm away from the cube, double 310 nm bandpass filter on the lens, f/4, ISO 25600, 30 s exposure:

 

Well, there's not much to see. I can't take sunflower photos.

[Stefano]

If I point my camera straight at the LED, I can easily see it. But also it's output is rated at 10-15 mW of output power, so not much.

[Andy Perrin]

Intensity is what matters though, not power alone. Would be interesting to know the intensity vs intensity of sunlight at that same wavelength for the given distance from the cube.

[diant]

Stefano says about exactly radiant power - 10-15mW UVB light at 1W electrical power. I have also experimenting with these leds a little and was disappointed a bit with their long-long parasitic tails. Intensity of sunlight at 313nm has huge dependence on many factors. I think you know it well. For example: if imaging makes in room with glass window - 313nm will be near zero etc.

 

10-15mW UVB 310 nm Led can easy beat sunlight 310nm if distance be some first cms. Total sun light power in summer noon day is not large 100mW/cm2. The rest is obvious.

 

By the by. Leds 310 show some wonderful features. Among them - possibility of exciting luminescence of some non-luminescent (in usual UVA) substance. I discover that UVB force to glow with gentle bluish-white light the glasses containing Lead (flint glasses). And if remember that flint Pb-glass was obligatory element every lens until near the end of XX century (when Pb-free glasses exclude Pb-flints in Schott, Hoya etc.) one easily conclude what milk glow light should pour over CMOS diffusely, if we use a lens with even if one old flint at tens sec exposures (if such lens of course show any UVB transparency).

[Stefano]

If you want to do some calculations, the LED has a 60° beam. It can also be that with the LED my camera sees 310 nm and in sunlight it sees more between 310 and 315 nm.

 

[dabateman]

 

Stefano, 50% is absolutely impossible.

SvBony focal reducer - usual astronomical 2-elements multi-coated unit. MC only reflect a ton of UVB. Its elements of course are made from different glass, one of which usually more opaque in UV. If you obtain 10%, you may consider yourselves a lucky one, I think.

 

Nope,

The SvBony 0.5x focal reducer is a single positive meniscus element lens about 21mm diameter mounted in a 1.25 inch holder.

You can easily remove that single coated element from the holder and mount it into a 25mm filter ring.

Its K9 glass equivalent to Schott B270 glass. Its very high transmission at 310nm, atleast 80%. But I can't properly do absolute transmission curves, so best not show one.

[Andy Perrin]

Intensity of sunlight at 313nm has huge dependence on many factors. I think you know it well. For example: if imaging makes in room with glass window - 313nm will be near zero etc.

Of course not through a window. And not under clouds, etc.

 

But the point is exactly "10-15mW UVB 310 nm Led can easy beat sunlight 310nm if distance be some first cms."

 

Stefano said it's not much, but it certainly is, if we are comparing the LED to the sun in his photos above.

 

If you want to do some calculations, the LED has a 60° beam. It can also be that with the LED my camera sees 310 nm and in sunlight it sees more between 310 and 315 nm.

Usually LEDs aren't that monochromatic. There's usually 20nm spread or so, isn't there?

[dabateman]

Andy, I say mainly not about 330WB80 but about 313bp15.

330WB80 may be as good as it can in region longer 380 nm. But in region 380nm and shorter it is opened. And if we speak about UVB (313nm) imaging, it has equal clear window from 313 to 380. And only filter that remains for separate UVB 313nm from UVA is 313bp15. But can we be sure that it have no sensible leakage in UVA (330-380) where CMOS chip is much more sensitive?

No skepticism, I really don't know two points: 313bp15 OD in UVA and Olympus EM1mk1' CMOS real UVB/UVA sensitivity. Here I'm only reasoning to clear this points and escape possible confusion in what we see - real UVB 313nm image or mainly UVA one.

 

Exoterras (yet they exist in different spectrum) also often have a large peak in UVA (as we see in spectrum below).

 

Its good to be skeptical. I have tested this many times and don't place all the history in each post.

 

The Exoterra UVB lamps have a strong 313nm peak and the UVA and visible Mercury lines.

 

The EM1MK1 can see down to 280nm, but not lower.

 

My copy of the 313bp25 filter is very good, but does have a slight IR leak. So I stack it with the U330WB80 improved filter to filter that IR leak out. There is no UVA at all going through the two filters.

 

[diant]

> The EM1MK1 can see down to 280nm, but not lower.

Wow! It is very interesting how you test it?

Good "UV-tail" of CMOS sensitivity)) What sensor it has, I really wonder...

 

> I stack it with the U330WB80 improved filter to filter that IR leak out. There is no UVA at all going through the two filters.

Very well, ok, and this point we can closed.

 

I have pair Exoterra UVB lamps, but they have not 313 peak at all. But yours may be others, yea.

 

> SvBony 0.5x focal reducer with one K9 glass...

Very strange. I've read its description of sites. Two elements declares. Please, look carefully - is it not glued patch (2/1)? (third weak light glare should be in such case). Were it be one element 0.5x reducer, it would be of fearful one really))

 

Chenese K9 may be something different from Schott K9. I mean especially its UVB transmission.

[OlDoinyo]

i assume the 35mm in question is the Kyoei version rather than the Komine, Kawanon, Tamron, or Hanimex versions?

[Stefano]

This is my lens: https://www.ultravioletphotography.com/content/index.php/topic/4412-new-camera-first-tests-and-a-long-story-to-tell/page__view__findpost__p__43319