Testing a SvBony 0.5x focal reducer lens for photography and UV TriColour

[Stefano]

David has previously shown that this lens is good for UV down to UVB: https://www.ultravio...but-not-magical

And here is an image quality test: https://www.ultravio...-tests-at-313nm

Bernard Foot has also tested this lens for its intended purpose, and too reported good results: https://www.ultravio...th-svbony-et-al

 

So I tested one myself. I built a lens using the following parts:

 

 

 

The 25-52 mm adapter ring barely screws on the lens ring since the lens is thicker than the original UV filter. If you want to build a similar lens, either glue the step-up ring or try to find a deeper 25 mm filter ring. The lens should be mounted with the most curved side towards the subject and the flatter side towards the sensor. The focal length of this lens is about 45-48 mm (I used 48 mm to calculate the apertures below).

 

Image quality

Being a single lens, there's no correction for spherical and chromatic aberrations, so you will get both. You will need to refocus at different wavelengths.

 

Full-spectrum Canon EOS M, Chinese BG39 (2 mm).

 

Fully open (about f/2.4), ISO 200, 1/30 s exposure:

 

About f/5, ISO 800, 1/30 s exposure:

 

Stopping down even more will improve image quality a lot, especially the edges, but of course this will require longer exposure times.

 

UV (ZWB2 (2 mm) + Chinese BG39 (2 mm)), lens stopped down between f/3 and f/5, eyeballing:

 

ISO 25600, 1/15 s exposure:

 

Rubik's cube. All colors absorb at 310 nm and 340 nm, but the white squares are still a bit reflective at 387 nm, and they appear red here:

 

 

ISO 100, 8 s exposure:

 

ISO 25600, 1/30 s exposure:

 

UVB performance and TriColour

This lens is quite transparent to UVB, at least down to 310-313 nm (see David's test in the link at the beginning), while even lenses with a very good UV reach (such as my Soligor) do not transmit much there, just a few percent at most. This makes taking UVB images and TriColours easier, as I put the 310 nm images in the blue channel. With this lens there is a gain of at least 2-3 stops compared to a Soligor 35 mm f/3.5 lens at 310 nm.

 

The focus shift between 310 and 340 nm is not very strong, but between 340 and 387 nm it is visible in live view (I would have expected the opposite). In the images below, I had to zoom-in some channels (especially the red one) and stretch them horizontally and/or vertically to overlay them as best as possible. I did that in Paint. In the 310 nm images I only kept the green channel, as most of the signal is recorded there.

 

Channels:

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.

 

Empty glass with paper tissue and rocks. This image was probably taken with the lens either fully open or very open, as it is quite soft. The absorption of the glass at 310 nm is visible as a yellow tint.

 

Magnifying glass (top), plastic lens (bottom left), car headlight lens (bottom right). All the lenses are black at 310 nm, and the plastic lens is black at 340 nm too (hence the red color):

 

Rubik's cube. All squares absorb UV, except for the white ones, which reflect at 387 nm and appear red here:

 

2 mm thick ZWB2 filter (left), 3 mm thick ZWB1 filter (right):

 

From here, I also stretched the channels to reduce color fringing. Before, I only zoomed them.

 

CFL bulb. Both the base and the tube are white to the naked eye (athough with slightly different shades). In UV, the base becomes a neutral gray (quite unusual in my opinion) and the tube becomes orange. This color means two things: firstly, the reflectance gradually decreases at shorter wavelengths (a sudden drop would usually look red or yellow), and secondly, the glass the tube is made of is mostly transparent at 387 nm and still partially transparent at 340 nm. I have detected the 365 nm mercury I-line coming from this very bulb in the past: https://www.ultravio...-a-spectrometer

 

Polycarbonate goggles with atypical temples. I already showed them in this topic. Both temples absorb at 310 nm, but only one absorbs well in UVA:

 

Some trees with some haze:

 

Daisy:

 

One day I should try an aspheric lens, if quality is significantly improved. But this lens is still nice and cheap.

[dabateman]

Stefano the SvBony element is a positive meniscus lens.

You want the concave side to face the sensor and the convex side to face the subject.

It come backwards from a camera point of view as its ment to screw into the telescope and your eye is in the front.

So if you flip it from how it comes into the 25mm ring, then you are good to go.

 

I am not sure that you have it in the correct orientation. Try flipping it in the ring and see if its better or worse.

[Stefano]

Yes, this is what I did. Since it is a positive meniscus (converging), the concave side is flatter than the convex side. So for converging lenses, you have to put the flatter side towards the sensor.

[Andrea B.]

wow, a do-it-yourself lens!! Very nice experiment.

[Stefano]

Thanks Andrea. This is literally the simplest lens possible after a pinhole, I would like to build a better one one day. My main goal now was UVB, next will be image quality.

[diant]

Thanks Andrea. This is literally the simplest lens possible after a pinhole, I would like to build a better one one day. My main goal now was UVB, next will be image quality.

Stefano, if UVB is your main goal now, you can buy similar positive quartz lens on AliExpress and have guaranteed 90% UVC-UVB-UVA transmission. Something like this or this...

[Stefano]

UV-grade fused silica lenses are good to ~180 nm and still transmit something at 160-165 nm, so they would be perfect for UVA, B and C. That's another thing I should get sooner or later, I have a long list...

 

An aspheric fused silica lens, like the one Jonathan tested, should hopefully leave me with chromatic aberration only (it should correct for at least most of the spherical aberrations). For TriColour or UVB imaging, the bands are narrow and so there is not much chromatic aberration within them, and the CA between them can be fixed in post-processing. A lens corrected for both chromatic and spherical aberrations would of course be better, but I think that unless I am lucky I won't find one for cheap. They just aren't cheap.

[Fandyus]

Might be a bit late but you should try this with a fluorescent UV bulb as well. I wonder what that would look like in a UV trichrome.

[Stefano]

Hi Fandyus, sorry if I am late too. I don't have a fluorescent UV bulb, just a normal one (shown in one of the images). It too emits some UV, I could try that.

[Fandyus]

On 11/6/2021 at 4:29 PM, Stefano said:

Hi Fandyus, sorry if I am late too. I don't have a fluorescent UV bulb, just a normal one (shown in one of the images). It too emits some UV, I could try that.

Oh no problem, you should definitely get one though, they're quite cheap on eBay and emit a lot of UV, the "CRI" is quite low but you can still take false color UV pictures with them. I got a 40w one, it's quite powerful. https://www.ebay.com/itm/E27-15W-40W-UV-Ultraviolet-Fluorescent-Blacklight-Light-Bulb-Lamp-CFL-AC-220V-/254540196779?mkcid=16&mkevt=1&_trksid=p2349624.m46890.l6249&mkrid=711-127632-2357-0

[Stefano]

Thanks for the advice. They have quite a broad spectrum too judging from images posted here.