• Ultraviolet Photography
  •  

Infrared block, broad UV and visible pass filter?

Infrared Filters
50 replies to this topic

#1 Avalon

    Member

  • Members
  • 16 posts
  • Location: Lithuania

Posted 30 September 2018 - 06:48

Hello. I want to build camera for ultraviolet photography and reflectography. I plan to use debayer'ed backlit CMOS camera with or without Wood's glass, ZWB3 UV filter. But I also need to find special infrared cut filter because abundant IR radiation will contaminate photos.

Problem is that there are plenty of hot mirrors which block not only infrared but also UV. After doing some tests I managed to find only one hot mirror that passes 365nm LED radiation that I scavenged from JVC gz mc100e high quality camcorder and it is small. I want to use same camera for infrared photography so putting it inside is not good option besides I'm not sure how far filter passes UV radiation. I need filter to be at least 27mm diameter.

I know there are specialized only UV pass filters such as U-Venus Baader but they are expensive, rare and not so versatile as two separate filters. I want camera to see into ultraviolet as broadly as possible although since UV-C is blocked by air alone I'll be good with filter that allows fully UV-A and preferably UV-B spectrum.

So far best IR cut filter I managed to find on Ebay was QB21 37mm filter. Filter actually blocks completely IR and allows UV-A with visible light. Only downside is that filter does not pass UV-B and UV-C, even UV-A is limited at shorter wavelenghts: https://www.ebay.com...hy/273459724074

#2 Cadmium

    Member

  • Members(+)
  • 1,748 posts

Posted 30 September 2018 - 07:20

First of all, from CHINA is a huge red flag. Filters from China are never what they say they are.
It says it is BG38, but no way is it made of actual Schott BG38 if you get it from China.
Second, BG38 and BG40 are good for visual photos with some cameras that are converted to full spectrum,
but especially BG38 is NOT good for UV stacking to suppress red/IR. Never use BG38 for UV stacking.

ZWB3 (also Chinese filter glass) is a UG5 / U-330 ~equivalent~. Stacking with 1.5mm versions of those is best done with 2mm S8612.
There is no equivalent for S8612, not Hoya, and none that is Chinese.
If you want to pass UV only, then using ZWB3 is the wrong choice, it passes UV+Blue+Green, and also red and IR. You will not be able to block the blue+green with anything.

Your camera sensor is not going to see any UVC at all, and it will not see UVB once you make a filter stack that blocks the red/IR.
In fact, UVB is really not going to be seen if you are using a filter that also sees very much UVA, the UVA will take over, and even if you blocked the UVA you would need to use very long exposures to see even a small part of the UVB range, your sensor simply isn't designed for that.
If you want UVA, and not blue and green, you will need to start with something else besides ZWB3. That is not used for UVA only or UVB, once stacked to suppress the Red/IR your UV band is reduced to UVA, but in the case of UG5/U-330, and ZWB3, you will also see blue and green, nice for 'bee vision' (UV+Blue+Green) photos, but not UVA only.

So first, start with something other than ZWB3. That is not a UV only filter, it transmits blue and green.

Edited by Cadmium, 30 September 2018 - 07:59.


#3 Andy Perrin

    Member

  • Members(+)
  • 1,594 posts
  • Location: United States

Posted 30 September 2018 - 09:44

Cadmium laid it out for you but I’ll make it shorter:

1) forget Chinese glass. That includes anything that starts with QB or ZWB. It also includes a lot of stuff marketed as by “Schott” but that isn’t really - it’s fake. I have some “BG-39” that lets huge amounts of IR through. Whatever it is, it is NOT BG-39.

2) You are never going to see UVC with a converted camera. Not only is there none from the sun, our sensors have no sensitivity to it without taking special measures like coating with a phosphor.

3) You also won’t see UVB if you also have UVA at the same time because the UVA is 20x to 40x as strong (in sun). A number of us have tried and failed to see UVB even when blocking UVA, but it can be done in principle if you have a debayered sensor and REALLY strong filtration.

4) To see UVA, go buy some real Schott S8612, 1.5mm, and some U360 2mm or so.

5) Cadmium sells filters as UVIR Optics and I have been happy with them myself. We are otherwise unaffiliated aside from this board.

Edited by Andy Perrin, 30 September 2018 - 10:54.


#4 dabateman

    Da Bateman

  • Members
  • 276 posts
  • Location: Maryland

Posted 30 September 2018 - 11:29

With out naming names uviroptics listed ONLY on ebay or etsy, sells what it actually is. Other sellers that say Real whatever, oddly sell fake stuff.

27mm is the hard part I see. Can you drop that down to 25mm or 24.5mm? If so than the 330WB80 improved, you can find on ebay, is also an excellent single glass filter that works from 290nm to 380nm and is cheapish. However as mentioned above, imaging outside you will mostly get the upper part of that range.

Read some of the forum posts here and you will see test with Baader venus outside even on a quartz lens will mostly see 380nm, even though its peak max is at 355nm. The camera sensor sensitivity and the prevalence of upper bands is too strong outside.

Your best bet maybe a U340 or U360 stacked with 1mm or 2mm S8612.
Hopefully you haven't bought any ZWB3 yet, it really is garbage.

Your two stacked glass options are really just U340 equivalent to Ug11, or U360 equivalent to UG1, stacked with BG39, BG40 or S8612.

#5 dabateman

    Da Bateman

  • Members
  • 276 posts
  • Location: Maryland

Posted 30 September 2018 - 11:34

If you want to see uvb, the only cheapish way that works is stacking the 330WB80 improved filter with the 313bp25 filter. Both can be found on ebay. But are 25mm in diameter. Then you will need a lucky herb mercury vapour lamp, to get a strong 313nm band. My 125W clear bulb has a great 313nm band.
Then you will need a quartz lens. I haven't used it but the cheapest option is the UKA 25mm f2.8 c-mount lens. Its under $600, and will work on a m43rds camera and takes 25.5mm filters. It may also cover a Sony A6000, but I am not sure. There are members here that own it.

Edited by dabateman, 30 September 2018 - 11:44.


#6 Avalon

    Member

  • Members
  • 16 posts
  • Location: Lithuania

Posted 06 October 2018 - 11:52

Probably I have to forget for now UV-C spectrum since optics also will have to be converted. I have lots of high pressure mercury quartz DRT-1000 lamps for UV-C, UV-B and UV-A spectrum. I do have also UV LED. Fluorescent tube type UV lamps are great for illumination since they already have filters but emit IR.
Want to start first with cheaper filters then move to expensive ones unless I can find optimal filter. There is ZWB1 filter which transmits even less infrared and visible light. If filter does not pass UV I can always that and return it. If UV-C can pass filter so can longer UV easily pass and UV-C produces distinct ozone smell.
Looks like I will have to spend lots of money on good quality infrared cut filter.
330WB80 spectrum looks excellent. Can use 25mm filter but I would need to design frame and zoom little bit to avoid cropping. BTW for outdoor photography will camera with 1/2" (6.4 x 4.8 mm) sensor and F3.5–5.3 maximum aperture be sufficient without needing long exposures?

#7 dabateman

    Da Bateman

  • Members
  • 276 posts
  • Location: Maryland

Posted 06 October 2018 - 13:40

First if you think you want to save money buying ZWB filters, just take the equivalent amount of money and burn it. They are a waste of time and money and you will end up buying better ones in the long run so don't waste your time.

What full spectrum camera do you own?

You will not see anything below 300nm, unless you have a special fluorescent coated sensor. A work around is to create a screen to image. I have recently done some tests with a full spectrum converted Olympus Em1, germicidal 254nm light and high quality 260bp10 filter, no ir leakage. I see nothing.

I do see just with a 300bp10 filter on the Pentax UAT 85mm lens. The lighting is tricky.

I also did a 313 filter test with my lenses. The Steinheil 50mm, igoriginal 35mm and the Nikor 80mm all just see here. The Steinheil 50mm was the best. I will add a post about this soon.

My 300bp10 filter is only 12.5mm with an open 9mm for light. I do not see vignetting on the UAT 85mm. This surprised me. With micro four thirds cameras, you can get away with much smaller filters.
Also the front element of most uv capable lenses is less than 20mm. The UAT front element is 18mm, rear element is 15mm. This is why I go with 25mm specialty filters, but my newly cheaply acquired high quality bandpass filters I am now testing and may also work out just fine for my camera.

For my 25mm filters, do a search on ebay for 25mm filters, you can get cheap uv glass ones from China with retaining rings. Remove the ring and just gently sand down the edges of the 330WB80 improved if it doesn't fit first off.
If your scared of sanding. Buy cheap 28mm uv filter and cheap 25mm to 28mm step up ring. Remove the ring from the 28mm filter and us it to hold high quality filter in the 25mm to 28mm step up ring. Then use the empty 28mm filter to expoy in a filter which has metal ring. The 330WB80 improved, does not have a metal holder.

I use the Raf camera 52mm to 25mm step adapters as they also have a female 52mm thread on the 25mm side. This allows the adapter to be used behind a lens or infront using a 52mm hood. For reverse mounting filters get a m42 to 52mm 52mm to m42 adapter set.

Edited by dabateman, 06 October 2018 - 13:44.


#8 Avalon

    Member

  • Members
  • 16 posts
  • Location: Lithuania

Posted 06 October 2018 - 15:24

I own Fujifilm Finepix F500 camera with IR/UV cut filter removed, not correct to call it full spectrum camera since CFA is not yet removed.

Well there are CMOS sensors capable of detecting UV-C as well. Here is example:
https://www.photomet...nation-img1.png

Problem is that so many materials block shorter UV wavelengths. Not just Bayer filter, sensor electrodes and microlens (in front illuminated sensors) but also silicon nitride coating which in normal CMOS camera's blocks ultraviolet and even blue light. https://www.photonic..._sensors/a47007

Besides it would not make sense to make special UV grade quartz lens for camera's if they could not detect UV-C. Optical glass does well pass UV-A and UV-B.

Edited by Avalon, 06 October 2018 - 15:32.


#9 Andy Perrin

    Member

  • Members(+)
  • 1,594 posts
  • Location: United States

Posted 06 October 2018 - 15:32

Yeah, we were not talking about sensors designed for UVC, just ordinary camera sensors that either have a Bayer or have been deBayered. I know klaus has one that can do UVC, but it is special.

I have a 330WB80 and I like it a lot.

Edited by Andy Perrin, 06 October 2018 - 15:35.


#10 dabateman

    Da Bateman

  • Members
  • 276 posts
  • Location: Maryland

Posted 06 October 2018 - 16:09

View PostAvalon, on 06 October 2018 - 15:24, said:

I own Fujifilm Finepix F500 camera with IR/UV cut filter removed, not correct to call it full spectrum camera since CFA is not yet removed.

The camera is full spectrum, but its not monochrome. When you remove the CFA, the camera becomes monochrome.

The F500 is not an ideal UV camera as it has a fixed lens. I would recommend then a SEUmk2 filter or a 390bp25 in 25mm mount filter. With that camera you will not see below 380nm. You need a camera that you can change lenses on.

#11 Avalon

    Member

  • Members
  • 16 posts
  • Location: Lithuania

Posted 06 October 2018 - 17:22

Not ideal but it's entry level camera in UV photography and don't want yet to risk destroying more expensive camera's when debayering. Later sure I'm going to get more serious camera with full frame or APS sensor with "fast" lens, ideally something like Sony A7s or other super sensitive camera's since I'm going to use camera for professional use also.
Fixed lens should be Ok for UV-A and UV-B imaging unless there is filter on lens themselves, some AR coatings do block UV. In JVC camera I mentioned before had IR cut filter that passes UV but one of lenses actually blocked 365nm UV radiation.

Edited by Avalon, 06 October 2018 - 17:23.


#12 Andy Perrin

    Member

  • Members(+)
  • 1,594 posts
  • Location: United States

Posted 06 October 2018 - 17:51

He is correct that most fixed lens cameras will not pass beyond 380nm. You are not going to get UVB. Getting to UVB is controversial usually even with quartz lens. I am starting to think every new member must rediscover this fact for themselves, though!

Edited by Andy Perrin, 06 October 2018 - 17:55.


#13 Avalon

    Member

  • Members
  • 16 posts
  • Location: Lithuania

Posted 06 October 2018 - 20:20

View PostAndy Perrin, on 06 October 2018 - 17:51, said:

He is correct that most fixed lens cameras will not pass beyond 380nm. You are not going to get UVB. Getting to UVB is controversial usually even with quartz lens. I am starting to think every new member must rediscover this fact for themselves, though!

Ok. Explain why glass lens wont pass beyond 380nm?

#14 Andy Perrin

    Member

  • Members(+)
  • 1,594 posts
  • Location: United States

Posted 06 October 2018 - 21:18

Not all glass lenses, just the ones typically built in to those kind of cameras. If you check our lens sticky, you can see there is a wide range of passbands for different lenses. It depends very much on the construction of the lens, and things like how many elements, what kind of cements are used, etc.

http://www.ultraviol...capable-lenses/

Edited by Andy Perrin, 06 October 2018 - 21:21.


#15 dabateman

    Da Bateman

  • Members
  • 276 posts
  • Location: Maryland

Posted 06 October 2018 - 21:21

Your camera has 12 elements in 10 groups. You will have coatings on those elements, and optical cement between those groups. This will eliminate a lot of uv.

Yes glass cuts out UV, about 50% per element. Thus why the best uv lens have only 3 glass elements. You can look up transmission spectra for different types of glass. You maybe surprised.

Your sensor reference are linear sensors. Which I haven't thought about modifying in a while. Its interesting as people have used scan heads on Large format cameras, but its tricky. I will have to see if this is a possible modification. The image acquisition times will be in the minutes though.

#16 UlfW

    Ulf W

  • Members
  • 338 posts
  • Location: Sweden, Malmö

Posted 07 October 2018 - 08:27

View PostAndy Perrin, on 06 October 2018 - 21:18, said:

Not all glass lenses, just the ones typically built in to those kind of cameras. If you check our lens sticky, you can see there is a wide range of passbands for different lenses. It depends very much on the construction of the lens, and things like how many elements, what kind of cements are used, etc.

http://www.ultraviol...capable-lenses/

Also cutoff wavelengths for many of the lenses here a bit optimistic.
One frequent source, KDS are using his own rather strange level to define the cutoff wavelength at -3 stop. (12,5%)
https://www.fotozone.../&hl=comprising
A more normal level would be 50% of the passband level.
The slope of the cutoff can vary between lens designs and the image sensor sensitivity also is decaying into the UV-range.
This makes the -3 stop wavelengths not very practical to show how well a lens will work for UV photography.

I believe that the results in the graph might be reasonable correct at -3 stop loss in transmission.
But it is only OK to compare UV-range of different lenses if the same cutoff level is used to define the results.

I have measured some of the lense types on the list.
A popular lens here is the Novoflex Noflexar 35/3.5. My lens of that type has a cut off at 346nm (-1 stop, 50%) and 330nm (-3 stop 12,5%).
I estimate my measurement error to be less than ±3nm
Ulf Wilhelmson
Curious and trying to see the invisible.

#17 UlfW

    Ulf W

  • Members
  • 338 posts
  • Location: Sweden, Malmö

Posted 07 October 2018 - 08:51

View Postdabateman, on 06 October 2018 - 21:21, said:

Your camera has 12 elements in 10 groups. You will have coatings on those elements, and optical cement between those groups. This will eliminate a lot of uv.

Yes glass cuts out UV, about 50% per element. Thus why the best uv lens have only 3 glass elements. You can look up transmission spectra for different types of glass. You maybe surprised.


May I add that there are many different types of optical glass.
A few of them transmit rather well down to around 300nm but many other types do not.

The transmission is dependent on glass type and thickness, cement type for cemented lens pairs..., surface coatings and the number of glass to air transitions.

Uncoated surfaces cause a loss of very roughly 4% due to surface reflections for each passed surface.
That would give an uncoated 3-element lens a max transmission of 78%

It is very difficult or impossible to create a coating that is working over a wide range of wavelengths.
That is why specialist UV-lenses like the UV-nikkor use uncoated lenses.
http://www.company7....0105f4.5uv.html

The low absolute transmission of this lens (66%) is mainly caused by surface losses.
The lens elements are made of other materials than glass better suited for UV.
Ulf Wilhelmson
Curious and trying to see the invisible.

#18 Cadmium

    Member

  • Members(+)
  • 1,748 posts

Posted 07 October 2018 - 09:34

Forget UVC, forget UVB. Narrow your total possible expectations to the 325nm to 400nm range, that is all you can expect to get, even with an optimal scenario.
!) Get some camera that you can change the lens. Not a fixed lens.
2) Don't use UG5, U-330- or ZWB3, these are NOT used for UVA pass. Get something else that is for UVA.
3) Get S8612 for Red/IR suppression (preferably, it has the deepest UV transmission reach), or BG39, or even BG40, or whatever, but not BG39 BG38 Keep in mind that there is no S8612 equivalent, only Schott makes such a filter.
4) Get a lens that transmits UV well.
Otherwise you are wasting your time. Do it right, or it will not work.

Edited by Cadmium, 07 October 2018 - 23:01.


#19 dabateman

    Da Bateman

  • Members
  • 276 posts
  • Location: Maryland

Posted 07 October 2018 - 11:18

Here is a fun article on some random single glass and ceramic filters:

https://www.lensrent...r-spectrograms/

Roger since said there spectrometer is only accurate to about 370nm, thus why some of the curves are very noisy at 350nm provided or do strange things like start to increase.

@Ulfw, isn't the minimum for uncoated reflection 4% per surface. Thus you would expect minimum of 8% loss per element for light entering (4%) plus light exiting (4%). I know even these are generalizations.

Now think how quickly this would add up for 12 elements. Don't expect to see much. However, there will be coatings that may still work at 390/380nm so thats why I use another approximation.

Cadmium has a good summary, although I think he means to not buy BG38. It doesn't have any good advantage over BG40, the cheapest option or S8612, your best option.

#20 UlfW

    Ulf W

  • Members
  • 338 posts
  • Location: Sweden, Malmö

Posted 07 October 2018 - 11:59

View Postdabateman, on 07 October 2018 - 11:18, said:


@Ulfw, isn't the minimum for uncoated reflection 4% per surface. Thus you would expect minimum of 8% loss per element for light entering (4%) plus light exiting (4%). I know even these are generalizations.

Now think how quickly this would add up for 12 elements. Don't expect to see much. However, there will be coatings that may still work at 390/380nm so thats why I use another approximation.

That was what I tried to say.

The surface reflection-losses are multiplicative. A 4% reflection will give 96% transmission, 0.96.
One lens element will have a 0.962 = 0.92 transmission.
Twelve lens elements would scale to 0,9624 = 0.38.
However there are multiple internal reflexes and angle dependencies too, making the real truth quite complex literally. :)

As an approximation the 4% loss is OK.

Coatings are very different and even the Kuribayashi 35/3.5... has coatings.
Modern lens-coatings often cut at 380/390nm, as you said.
Coatings on older lenses often transmit more UV, but many of them are still bad UV-performers due to thick lens elements and less UV-transmitting glass types.

Special optics designed to reach deep into UV and still handle VIS + NIR do not have coatings.
The fused silica collimators for my spectrometer reach 200nm. They have no coating.
Ulf Wilhelmson
Curious and trying to see the invisible.