Jump to content
UltravioletPhotography

Relative Transmission of Various Lenses


bvf

Recommended Posts

In another thread somewhere whose topic I can’t remember, someone (probably Cadmium) suggested that the sensor in a full-spectrum camera might be sensitive down to 300 nm but that the (non-specialist) lens being used probably blocks out UV before you get to 300 nm.

 

This was an interesting thought because I can form images at about 320 nm, but would like to get deeper into the UV. So, time for an experiment to see whether my lenses are the limiting factor rather than the sensor - by comparing them with UV Fused Silica, which has a flat transmission spectrum across UVB, UVA, Visible, NIR.

 

Here are the results – some notes on the process are given later.

 

Firstly, looking just at UV, with Blue/470 nm taken as the starting point wavelength.

 

post-245-0-47016700-1604600738.jpg

 

A few of the results look odd, but all the measurements were re-checked at least once to either confirm or amend the results.

 

In general, the results seem to pass a sanity check:

  • The lack of UV-friendliness for the El Nik 50mm N is clearly shown.
  • The two best lenses – the Soligor 35mm enlarger lens and the Focotar 2 – are reported elsewhere in UVP as having the best UV reach.
  • The other two El Niks don’t come out too well (which I have noticed when using them), but the similar results for them would tie in with their having a similar construction.

I was disappointed that my old favourite, the Cassar S, didn’t perform better, but at least I can now justify to myself the money I spent on a Focotar 2, and can appreciate the Soligor lens that was donated by Timber.

 

The bottom line is that the test lenses are indeed limiting the sensitivity of the system at shorter wavelengths, and that the camera is capable of more than the lenses allow (as is shown by the UVFS line).

 

What this data does not tell us is whether and how much the sensor’s sensitivity extends beyond 320 nm. To check this needs a filter that will go down to 300 or 290 nm and provide strong enough blocking in visible and IR - but although the Omega Optical ebay shop has reappeared, I can’t find one (at least not in my price bracket).

 

Looking at the full wavelength range, this time using Green as the starting point wavelength:

 

post-245-0-35142600-1604600415.jpg

 

Less dramatic results in the NIR region, but:

  • Again, the El Nik 50 N performs poorest of all the lenses at the longest wavelength.
  • In fact, the El Niks as a group seem to exhibit the highest variation transmission losses in both UV and IR.
  • The Soligor (also an enlarger lens) seems to have the lowest drop in transmission and greatest consistency – I guess this could mean it’s a pretty unsophisticated lens.

Method:

 

The basis was to acquire a simple plano-convex lens made of UV-grade fused silica with of its essentially flat transmission spectrum from 300-2000 nm, and to compare UV-friendly lenses against this.

 

So at each wavelength that I had filters for, the lenses were compared by seeing which ISO setting was required to get images of equivalent brightness. The UVFS lens was taken as the baseline to compare the other lenses to. A single set of results can’t tell you very much because the lenses all had different maximum apertures and different constructions, but the series of results shows how the transmission of each lens varies with wavelength.

 

One wavelength is taken as starting point against which the other wavelengths are compared. At this wavelength, normalisation factors are calculated for each test lens using (ISO setting for UVFS / ISO setting for Test Lens): at each wavelength the equivalent factor for each lens is divided by the relevant normalisation factor to see how the transmission of the lens has changed from the starting point. By definition this gives a value of 1 for all lenses at the starting point wavelength, and the UVFS lens always has a value of 1.

 

The test lenses were:

  • UVFS – simple plano-convex lens made of UV-grade Fused Silica.
  • An Igororiginal 35mm f/3.5, which has been the subject of several threads in UVP
  • Soligor 35mm f/3.5 enlarger lens
  • Cassar S 50mm f/2.8
  • Focotar 2 50mm f/4.5 enlarger lens
  • El Nikkor 80mm f/5.6 enlarger lens (original metal version)
  • El Nikkor 105mm f/5.6 enlarger lens (original metal version)
  • El Nikkor 50mm f/2.8 N enlarger lens (later plastic version) – this lens is superb for visible macro photography, but is not a recognised UV-friendly lens. It was included to provide a sanity check that the results made sense.

The wavelengths and their respective filters were:

  • 320nm - 315BP25 (peak transmission 320nm; plus a U340 to provide additional IR blocking)
  • 345 nm - 345BP20 (plus a Baader U to provide additional IR blocking)
  • 380 nm - 380BP25 (plus a Baader U to provide additional IR blocking)
  • Blue/470 nm – Midwest Optical BP470
  • Green – Hoya G(X1) (plus S8612 and UV/IR cut filters to block IR)
  • Red/600 nm – Hoya R25A (plus S8612 filter to block IR)
  • 750 nm – Midwest Optical BP735 (plus Hoya R72 to block red)
  • 850 nm – Midwest Optical BN850
  • 1000 nm –Midwest Optical LP1000

Link to comment

In another thread somewhere whose topic I can’t remember, someone (probably Cadmium) suggested that the sensor in a full-spectrum camera might be sensitive down to 300 nm but that the (non-specialist) lens being used probably blocks out UV before you get to 300 nm.

I think you are referring to this post, right?
Link to comment

There are a number of factors... but before I talk about that, let me say your graph of several of those lenses don't match up with my results.

Focotar 2 'should be' the deepest of those compared. The El-80 'should be' fairly deep also, the Soligor should be deep (depending on the actual Soligor), however not deeper than the Focotar 2.

Soligor can be quite similar to Kuribayashi 35, depending on the actual Soligor.

That being said,

You would get below 300nm with a UV-Nikkor, and even though the UV-Nikkor has a more flat transmission than the lenses you show,

it still depends on the content of the UV illumination x the sensitivity of the sensor which also has a curve to it just like the lenses you show do.

So, yes, if you use a narrow bandpass filter, and you have no other part of the UVA/B range, then you will see the lower bandpass.

If I use a BG filter (S8612, BG40, BG38...) for visual, I don't see any UV pattern, even though those filters transmit down to ~320nm~, that is because the visual range is stronger, the sensor is more sensitive to that range, etc.,

It is the same with a usual UV-pass filter, 320nm to 400nm, it will show you the weighted part of that range, you will see no more with the UV-Nikkor than you do with the Kuribayashi,

however, if you use a narrow bandpass, lets say 300BP10, then your camera will see it, given enough exposure time, settings.

However, with a wider range, even using a deep transmitting lens like a UV-Nikkor, the lower part of your range will not show up, will not change the mix, and that can work the same way for any lower non bandpass range.

Link to comment

Well, it must be said that this test is not meant to be really precise, but indicative only. Bandpass filters have some bandwidth, ISO is not 100% reliable, etc, but you can see the difference between a good and bad lens. Maybe using Jonathan's 9 bandpass filters (those he used here) would yield better results, but nothing can beat a spectrometer in my opinion.

 

About the reach of UV lenses, I agree that after a certain point you really don't gain much more sensitivity and chromaticity, since anyway every curve (Sun's spectrum, sensor sensitivity (and Bayer filters transmission)) is downhill. UV-green, which is at ~340 nm, does never show up in UV photos, even with a quartz lens, and there aren't many UV-green objects.

 

I remember this topic: https://www.ultravioletphotography.com/content/index.php/topic/2155-uv-green-object. Unless you have a ruby or image a 340 nm bandpass filter like in a sparticle, or use a diffraction grating, you will always see blue, violet and greenish yellow.

Link to comment

I was sort of using stone-age technology for this experiment, as that is all I have available. As Stefano says, "this test is not meant to be really precise, but indicative only". And of course "ISO is not 100% reliable", plus there is variability in the flash output (although I used two flashguns so the combined variability should have been less.

 

My measurements do not answer the question of UV reach - I have no idea what happens below 320 nm. But the results did tie in with what I had observed when doing tri-colour UV photography - that the Focotar 2 showed less drop-off that the Cassar S at 320 nm compared to 345 & 380 nm, whereas the El Nik 80 and 105 showed more.

 

I understand what you (Cadmium) say about wide bandpass filters: in practice, the Baader U gives results not much different from the 380BP25, because although the Baader U passes shorter wavelengths well the sensor is much less sensitive to them.

 

I would love to get a 300BP10 filter as you suggest to see if images can still be formed. Unfortunately the only filter at around that specification that I have found (Omega Optical) also has a big spike at around 800nm, and I don't have the filters to block that to the required extent (probably about OD 8).

Link to comment

If I read Jonathan's data correctly, Hoya U-340 has about OD 4 at 800 nm at a thickness of 2.5 mm, so 5 mm should have OD 8 there, and still transmit a good amount of 300 nm light.

 

KG glass wouldn't work, as well as S8612. It isn't easy at all to make a filter that blocks UVA, VIS and NIR so well while still transmitting enough at 300 nm. OD 8 is a lot for such filters, some use 330WB80 to block IR.

Link to comment

The transmission of some of the lenses you tested Bernard are shown here - https://www.ultravioletphotography.com/content/index.php/topic/3406-lens-transmission-in-the-uv-latest-update/page__view__findpost__p__30663

 

None of them will be transmitting at 300nm. The non-zero baseline of the measurement in the link above is more an issue of stray light in the spectrometer than actual transmission.

 

The 60mm Jenoptik, the 60mm UV-Objektiv, will give some transmission at 300nm and just below. Then if you want to go really low, it's things like the UV-Nikkor, 105mm Rayfact, 105mm Coastal optics, 105mm UV Sonnar and 85mm Ultra achromatic Takumar (and various other even more rare ones). Or of course just a simple fused silica lens, if you don't mind some funky edges to your images (such as Ulf's post - https://www.ultravioletphotography.com/content/index.php/topic/3768-fused-silica-single-lens-tests-ii-150mm-plano-convex/page__hl__thorlabs__fromsearch__1).

 

Getting to 300nm and below is a big ask. The camera sensors are not very sensitive there. If you have Bayer filters then they are absorbing most of the light. If you have a sensor coverglass on the sensor, then that is probably absorbing the UV too.

 

Then there is the issue of filters. The idea of OD3 or 4 goes out of the window, and much more blocking is needed unless you have a monochrome light source (due to the extremely low sensitivity at 300nm compared to the longer wavelengths).

 

Of course this can be made easier. Some of the BSI sensor are UV enhanced such as Wihlem has (https://www.ultravioletphotography.com/content/index.php/topic/4143-a-first-test-with-my-gsense2020bsi-camera/page__view__findpost__p__39255) will of course help, but the sensor issue is still there.

 

Light sources. Sunlight is no good really. You will get some tiny amount of 300nm on a really sunny day, but it will be tiny amounts. 300nm and below is very dangerous, with burns within seconds possible. I was careless - or perhaps arrogant is a better word, thinking there wouldn't be a big problem - around a 308nm light once, and got burned for my stupidity, and that was with about 60-90s exposure total over a 24 hour period. And that wasn't all even in the direct light, but from reflections.

Link to comment

We aren't made to be exposed to light below 300 nm, and even UVB (in the wrong amounts) is harmful. And even UVA of course, but it is much less dangerous.

 

UVB is usually defined as 280-320 nm, or 280-315 nm, I prefer the first definition, but sunlight really ends at 300 nm. I once asked Jonathan to try an Invisible Vision 308 nm + thick Hoya U-340 stack, to see if he could block IR this way, and it worked much better with a monochrome sensor. You may be able to reach just below 300 nm with a color sensor, such as David did here, but you have a lot of noise, and he used a UV lamp, not sunlight.

 

A primitive lens made with a single plano-convex fused silica element works well with a small aperture or a long focal length (basically a telephoto lens in the second case), but if you want more "normal" images you will have strong artifacts at the edges. You may be able to put 2-3 elements together to make a DIY "corrected" UV lens, but you need to know precisely what you are doing.

Link to comment

The transmission of some of the lenses you tested Bernard are shown here - https://www.ultravio...dpost__p__30663

 

... Or of course just a simple fused silica lens, if you don't mind some funky edges to your images (such as Ulf's post - https://www.ultravio...__fromsearch__1).

 

... The idea of OD3 or 4 goes out of the window, and much more blocking is needed unless you have a monochrome light source (due to the extremely low sensitivity at 300nm compared to the longer wavelengths).

 

... Light sources. Sunlight is no good really. ...

 

 

It was that first post that got me lusting after the Soligor 35mm enlarger lens (in a couple of posts later on) and that resulted in Timber donating one to me.

 

What I did was very similar to Ulf - except that I got a 25mm diameter f=75mm lens (from Knight Optical). I've made some stops down to about f/17 - image quality is bottom-of-milk-bottle standard, but I will do some eperimenting with it. The point of my exercise was not to try to get 300 nm images out of those test lenses, but to judge whether the sensor might be able to get down there if I was using somehting like a UVFS lens. As you say, OD 3 or 4 (or even 6) won't be much use. Stefano suggested a 5mm U340 - which I don't have, but I do have two thinner U340s (or what claim to be U340s) so I might try using these to provide additional blocking.

 

As for light sources, I'm using Cheap-from-China flashguns, with the lenses removed. I don't think the tubes are coated. These should be able to provide illumination at 300 nm according to this

xenon flash tube emission spectrum from somewhere on the interweb:

 

post-245-0-50678700-1604787498.jpg

Link to comment

Bernard,

The flashes don't give you 300nm. I ran a test after Ulf posted an optimistic spectra and with two flash guns firing off every couple seconds, saw basically nothing at 300nm.

https://www.ultravioletphotography.com/content/index.php/topic/3242-intensity-comparison-between-canon-199a-and-godox-ad200/page__view__findpost__p__27148

 

You need a 302nm light bulb to image there.

I use 2mm thick U340 to help block my 20mm 303bp10 filter, that works ok. For added blocking I also stack both with a U330WB80 improved filter. U340 works well to help block out IR.

 

But your lenses will drop hard and fast after 320nm, then the coverglass on your sensor will also be blocking deeper UV.

 

But there is a very cheap UVC sensor that I will be posting about soon. Once I get some images. I even snagged a video of my hand under UVC, to find out that the focus is way too close with my 25mm single fused silica lens.

Link to comment
You filmed your hand in UVC? Without protection? If you imaged your skin, than that's some very rare footage, human skin in UVC. I briefly exposed my skin to in excess of 10 W/cm2 of 365 nm UVA, I looked at a 340 nm LED, but I would never try to see my skin in UVC, although the curiosity is there.
Link to comment

You filmed your hand in UVC? Without protection? If you imaged your skin, than that's some very rare footage, human skin in UVC. I briefly exposed my skin to in excess of 10 W/cm2 of 365 nm UVA, I looked at a 340 nm LED, but I would never try to see my skin in UVC, although the curiosity is there.

 

And risk destroying DNA...?

Link to comment
If I had a very sensitive UVC sensor, one of those sensors so sensitive they can count photons, then, maybe, I would give it a try at very low doses. But not otherwise. That's too much even for me.
Link to comment

Bernard,

 

But your lenses will drop hard and fast after 320nm, then the coverglass on your sensor will also be blocking deeper UV.

 

 

Thanks for the input, David. Re. the lenses, if I make an attempt to record anything below 320 nm it will be with the simple UVFS plano-convex lens, which should reach well below 300 nm - although image quality might be an optimistic term.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...