Pros & cons regarding high transmission UV pass filters: Baader U (3 versions) & Chroma U-Bessel

[LarsHP]

I consider buying a high transmission UV pass filter; at least 70% peak transmission, with a broad bandwidth.

Subject: Biological UV in low light condition.

 

It appears to me, that the Baader U-Venus filter (#2458291) is the most popular model, disregarding stacks. However, Baader makes two other UV only filters, the UBVRI Bessel U-filter (#2961753U) and the SLOAN/SDSS U-filter (#2961703U) which has a similar bandwidth as the popular model. I asked Baader Planetarium customer service for a comparison graph and received the one you see attached below.

 

The UBVRI version has lower transmission than the two others, so I won't consider that, but the SLOAN version has noticeably more transmission in not only the shorter wavelengths, but also at the peak of the U-Venus filter. On top of that, the SLOAN model is cheaper. In mounted, two inch filter size, the SLOAN is €213.45 while the U-Venus is €280.67 (excl. VAT). So, it should be a no-brainer to get the SLOAN over the U-Venus, right?

 

After checking the graph in the longest wavelengths, I saw something that made me hesitate: The SLOAN graph picks up in the 1000 to 1200nm range. The optical density log file (attached below) shows that the optical density is around 3.5 or less from about 990nm to 1200nm, down to about OD 2.66 from 1089 to 1093nm. From my understanding, the CMOS sensor itself isn't very sensitive in that end of the spectrum, being close to zero at 1100nm. In addition, solar radiation is not nearly as strong in the 1000-1200nm range, and also less than in the 700 to 900nm range, as it is in the visible spectrum. To me, all this seems to be fine, but I have no experience or expertise that allows me to say this with confidence.

 

To the experts here, my question here is whether these data is something to worry about?

 

Another option, is the very expensive Chroma U-Bessell (#27050) filter, which is appears to be a purely dichroic one. From about 330nm to about 390nm, it has more than 90% transmission. Optical density outside the UV range is around 4 or better up to about 1170nm and beyond, where it leaks heavily. That filter seems ideal to me, since you can use it as a broadband UV filter alone, or use it in a stack as a red and IR suppression filter without the usual cut in the short end of the UV range. The only downside is the price tag: $700 for a mounted two-inch filter! Unmounted 36mm size may be a solution, since that is less than half price. My question here is then:

 

Is the inclusion of the 380-390nm range considered a "fault" or otherwise problematic?

 

For instance, is there some "irrelevant" UV light going on there in biology?

 

From what I have read, the UV range is not very well-defined, since some use conservative values and consider the visible spectrum to end at about 420nm, while in laboratory tests, some people have been able to see all the way to 310nm(!)

 

BTW: Yes, I am new here. Thanks for letting me in!

baader_sloan_sdss_u_filter_transmission_od_log_t.xlsx

[ulf]

Hi Lars, 

Welcome to the forum!

 

You have found an interesting set of filters.

 

First let's discuss the importance of aiming for a really high peak transmission.

Naturally transmission affects the exposure time, but how big is the actual gain?

 

A change in transmission from 65% to 75% will change the speed by 1/5 of a stop!

That is hardly noticeable in exposure or exposure time in most situations.

 

The transmitted spectral range will affect the images much more.

 

The UBVRI-U filter have the lowest peak transmission and a transmission of 15% at 400nm. That would be considered as a grave violet leakage from many UV-purists, but will still show acceptably black UV-signatures in flowers when present.

In most cases it is actually very hard to distinguish between 395nm UV and 405nm violet.

 

As you mention sunlight it is worth noting that the extended range into UV of the SLOAN filter will not improve the detected image much as the camera sensors are much less sensitive at 330-345nm than at 370nm. That will be even more true if you will not use a dedicated lens with really deep UV-reach like the UV-Nikkor.

Not until you block the upper UV-A the 330-345 nm will contribute to the images.

 

IR leakage is quite important, especially where the sensors have high sensitivity, like at the typical U-filter hals materials secondary peak around 700nm. Beyond 900nm that becomes less important but I do not have any experience with 

 

Normally if your light source is a wide-band UV-source like the sun or a converted flash, then due to the fast declining sensor sensitivity to shorter wavelengths of UV, normally the filter transmission closest to visual are dominating.

 

The falling slope around 365-375nm for all three filters are quite close.

This means that you will gain much less exposure speed with the Sloan filter than you would expect.

 

Also Dichroic filers are more sensitive to the incidence angle of the incoming light beams.

They are normally designed for normal incidence, depend more on lens shades and do not work well at all with wide angle lenses.

70mm FL or longer is recommended.

[Andy Perrin]

Ulf's summary is good. The natural conclusion (assuming you plan to use it in sunlight primarily and not with UV-B light sources) is to pick the cheapest option, possibly disregarding the UBVRI-U with the violet leak. By the way, keep in mind that dichroic filters have some big tradeoffs - you may find in the end (like many of us...) that you prefer a good stack of absorption glass!

 

Quote

After checking the graph in the longest wavelengths, I saw something that made me hesitate: The SLOAN graph picks up in the 1000 to 1200nm range. The optical density log file (attached below) shows that the optical density is around 3.5 or less from about 990nm to 1200nm, down to about OD 2.66 from 1089 to 1093nm. From my understanding, the CMOS sensor itself isn't very sensitive in that end of the spectrum, being close to zero at 1100nm. In addition, solar radiation is not nearly as strong in the 1000-1200nm range, and also less than in the 700 to 900nm range, as it is in the visible spectrum. To me, all this seems to be fine, but I have no experience or expertise that allows me to say this with confidence.

Past 1100nm, the sensitivity drops dramatically. From 1000-1100nm, there is still significant sensitivity, but I think the UV end may outweigh that still - I do not know of any tests of this filter, though, so it's hard to be 100% sure.

 

Quote

Is the inclusion of the 380-390nm range considered a "fault" or otherwise problematic?

Nothing problematic about it. If it's the only thing you can see, that is a bit of a loss if you want false color in your images, but you won't have that problem if you choose a suitable lens. 

[dabateman]

If you buy the Sloan I would love to see some images.

I can't find any reports on cloudy nights forum with problems like the massive IR leakage from the Optolong filter. 

I have been meaning to replace my cracked Baader venus u filter and this one at 50mm might be better.  Its 3mm thick though and my Baader venus u is 1mm thick. So may nothave the same punchy contrast I love.

 

[dabateman]

We know the Baader venus u filter is 1 mm UG11 glass with coatings to block IR.

From the 1996 paper referenced in this thread:

https://www.ultravioletphotography.com/content/index.php?/topic/6069-use-of-sloan-sdds-photometric-filters-for-distinguishing-two-nir-bands/

 

The Sloan filter seems to be 1mm UG11, 1mm BG38 and quartz glass with coatings to block IR. This seems like an interesting combination and might be better blocked than the Venus u filter. 

 

[LarsHP]

Thanks to everyone for your kind responses.

 

To Ulf: I have the Coastal Optical 60mm f/4 UV-VIS-IR Macro lens, which should keep transmission up all the way to 320nm, and both a full spectrum converted Sony a7 and a monochrome converted Nikon D600. I also have other "UV capable" lenses, but with the Coastal Optical lens, the extension from the SLOAN filter would be welcome. 

 

To Andy Perrin: It appears that your considerations regarding the SLOAN filter leak plus sensor sensitivity in the 1000 to 1100nm range is right. See the reply from Baader Planetarium below.

 

To dabateman: Thanks, I didn't realize the difference in filter thickness. Also note what Baader says regarding the SLOAN version in the quote below.

 

 

Regarding the higher transmission of the SLOAN filter at the shortest wavelengths, I consider that the strong point of that filter, precisely because digital sensors have low sensitivity there and many "UV capable" lenses drops off there as well. I have realized that, when looking at UV transmission graphs, we don't see what the resulting spectral transmission is. For completeness’ sake, the graph from lens and sensor should be added, but who knows the graph of their (full spectrum converted) camera?

 

 

This morning, I got the answer from Baader Planetarium that I thought they had forgotten to answer. In short, they unfortunately advise against the SLOAN version because they don't guarantee full NIR blocking.

 

Here's their e-mail response in full:

 

"Thanks for getting back to us. Let us explain the differences between the three mentioned filters and it usecases in detail:

Sloan-U-Filter – purely intended for photometry:  will result in some leakage in the NIR – as clearly shown in the graph. Reason being: the Sloan Filter is not meant for daylight imaging but aims at maximum T in UV – as modern counterpart to the old UBVIR-filter sheme.

U-Venus-Filter: from the beginning was designed to not just work in the astronomical field but provide the best off-band blocking in the industry - out to 1200 nm – way above the sensitivity range of VIS-grade CCD or CMOS-chips (not meant to be used w. NIR-enhanced CMOS-chips). This at the same time is the reason why transmission must be lower for this very filter. If first class off-band blocking is the prime consideration then other parameters must suffer because it is impossible to accumulate such a huge amount of coating layers without making such a coating stack become porous or even prone to cracking under a wide temperature change. Actually this U-Venus-filter does provide best results under daylight conditions – very many imagers of nature, industry, insurances and forensic investigators use it for just a UV-imaging purpose – to mention just a few: to show an insect view of nature, to show the effect of protective sun-creams on human skin, to show hidden and repeated repair on car paint surfaces,  to reveil bite-marks or blows on human skin.

UBVIR-U-Filter – purely intended for photometry: does have a similar off-band blocking characteristics as the U-Venus-filter but follows a classical transmission requirement for this family of variable star photometry, whereas the peak transmission in UV is required by tradition to be lower even – in order to mimic older U-filters made from color glass.

We hope this helped for your decision and answered your questions."

[Andy Perrin]

11 hours ago, LarsHP said:

To Ulf: I have the Coastal Optical 60mm f/4 UV-VIS-IR Macro lens, which should keep transmission up all the way to 320nm, and both a full spectrum converted Sony a7 and a monochrome converted Nikon D600. I also have other "UV capable" lenses, but with the Coastal Optical lens, the extension from the SLOAN filter would be welcome. 

You won't see any difference to your images even with the Coastal Optics lens because the longest wavelengths always dominate. The most common beginner mistake in UV imaging[*] is to not fully grasp just how rapidly the sensor response decreases from the longer end of UV-A to the shorter end of UV-A and on into UV-B. Exponential decay is a heckuva thing! And on top of that the solar output is going down rapidly as well in the same range.

 

The end result is that the light you are seeing in your images are primarily the longest wavelengths UV that the filter allows (assuming no leaks elsewhere in the spectrum). You can expect plenty of 400nm-370nm, some 370-350nm, and not much else unless you add additional filtration to block some of the former.

 

[*] After not blocking the IR leakage, I guess, although in recent years that problem has received enough discussion that even beginners usually know about it.

[ulf]

Thanks Andy!

That is what I have tried to convey so many times. You phrase it so well.

 

To summarise the need of UV reach of a lens you need enough reach to get a reasonable recordable spectrum band forming your image.

Then you get reasonably good false colours after WB.

For filters like Baader U, SLOAN, or filter stacks that means lenses like the Focotar, EL-nikkor 80mm and also the Coastal 60mm.

 

Lenses with substantially more marginal UV-reach can still give UV-images, with longer exposure times, but the false colours are much less vivid.

 

The only time a deeper UV-pass for a filter is important, is when the filter/filters also properly block the upper UV-A wavelengths.

 

That is true for example with 8mm thick filters made of U-340 glass or 8mm of ZWB1 glass, when  combined with suitable lenses.

(Due to their thickness there is no need for a BG-IR-blocker at the OD at 700nm is good enough alone)

 

The reach of the Focotar-2 50mm or EL-Nikkor 80mm is not deep enough.

I do not have a Coastal Optical 60mm lens and do not know if the reach of that lens is deep enough  with those types of filters.

 

I noticed that those filters work well on my FS-camera with my UV-Nikkor.

[LarsHP]

Thanks, Andy and Ulf. The point both of you share regarding the shorter UV wavelengths on digital is noted.

 

Regarding the spectral transmission of the Coastal Optics 60mm, Klaus Schmitt has measured it and compared it to the UV-Nikkor. See the attached graph. As you can see, the CO lens reaches 320nm nicely, but drops sharply below that. I also plan testing UV on film, so the possible extra reach in the filter I get, will be of use there, if not on digital.

 

I have sent another e-mail to Baader Planetarium (annoying customer!), because it struck me that the graph and the OD values in the 1000-1200nm range apparently doesn't show the same thing. The graph shows about 1-3% leak there, while the OD values in the same range are from 2.66 to 3.6.

 

These OD values should translate to less than 0.1%, if I am not mistaken? If that's correct, then the graph is wrong...

 

What I like about the Sloan version, apart from the higher transmission around 321-350nm and lower price, is:

1) It is available in unmounted 50.4mm size, meaning I can put it in a regular 52mm filter ring,

2) VIS + IR suppression is quite high: from OD 5.4 and above. (I have seen sunlit images here on UVP, showing leak from the Baader U Venus filter.)

 

[ulf]

20 minutes ago, LarsHP said:

Thanks, Andy and Ulf. The point both of you share regarding the shorter UV wavelengths on digital is noted.

 

Regarding the spectral transmission of the Coastal Optics 60mm, Klaus Schmitt has measured it and compared it to the UV-Nikkor. See the attached graph. As you can see, the CO lens reaches 320nm nicely, but drops sharply below that. I also plan testing UV on film, so the possible extra reach in the filter I get, will be of use there, if not on digital.

 

The graph shows about 1-3% leak there, while the OD values in the same range are from ≈

 

These OD values should translate to less than 0.1%, if I am not mistaken? If that's correct, then the graph is wrong...

 

What I like about the Sloan version, apart from the higher transmission around 321-350nm and lower price, is:

1) It is available in unmounted 50.4mm size, meaning I can put it in a regular 52mm filter ring,

2) VIS + IR suppression is quite high: from OD 5.4 and above. (I have seen sunlit images here on UVP, showing leak from the Baader U Venus filter.)

 

 

Besides the sensor sensitivity the UV content in sunlight also drops drastically, below 350nm.

That affect imaging on film too, still making the upper UV-A dominant.

 

Yes the graph seams incorrect. 

OD 2,66 - 3,6 is 0,25% - 0,025% transmission.

 

Not all Baader U filter versions have leakage problems.

Very often OD just above 4.0 is enough if the leakage is in the range VIS... 800nm where sensor have high sensitivity.

[Andy Perrin]

Yeah, there have been many Baader U versions over the years, and if you look at tests on this board from long ago, they may not be representative of the current version of the Baader U. 

[LarsHP]

Ulf: Thanks for confirming the graph can't be right, even though I didn't get the percentage right (below 0.25%, not below 0.1%).

 

Andy: I can't find the thread where I found the comparison image (so attached below instead), but the file says it's from October 2021. I consider that recent, but that doesn't say anything about when the filter was made...

 

I think I will order two 2mm QB21 for my ZWB2 2mm filter and see how that works out. The TSN575 cuts too much in the shorter end of UV for my taste.

[ulf]

The QB21, 2mm do not cut enough IR at the secondary transmission peak for ZWB2 as you can see in the second image.

The QB21, 2mm is an excellent filter to restore the camera for normal VIS images.
 

You would likely be surprised how little difference there would be with two QB21 compared to a TSN575, as an IR-blocking filter

I think a better alternative would be a QB39, but it must be 2mm thick!

 

It looks like the lens used above was rather wide, from the green corners in the first image with the Baader U alone. 

 

Are you aware of that beside the attenuation of the filter material each air to glass passage gives losses of ca 4.5%. You can calculate those losses alone. The 4.5% is an estimation and depend of the refractive index of the material.

4.5% from six surfaces give 0.95^6 = 0.95 x 0.95 x 0.95 x 0.95 x 0.95 x 0.95 = 0.735 a loss of 26.5%

A total of 4mm QB21 will also cut more in the shorter end of the UV.

 

If you live in a hot humid place it is a good idea to store the filters in a cool dry place when not used.

Several optical filter glass materials corrode due to humidity. How much varies by the glass type.

 

Colin  (colinbm) lives in such an environment.

Some years ago he experimented with a local product similar to the windshield coating product RainX and I think it worked well for him.

If he reads this he might tell us more.

[LarsHP]

Ulf: Thanks again.

 

I already have two QB21 filters, but unfortunately only in 1.5mm thickness. They are too thin for "normal" pictures.

 

The idea is to buy two QB21 in 2mm thickness and try that with the ZWB2 in 2mm. Hopefully that works for UV as seen in the test image above, then I have that option. If it doesn't work out, then I can use them for what they are intended, as you point out. I guess I will benefit from these either way.

 

If I find out that 2x2mm QB21 + 2mm ZWB2 works really well, I will try to put at least two of them together in the same filter ring in order to minimize the light loss you point out. Will that help, or will the loss still be there even when stacked directly on top of each other?

 

The QB21 filters are anti-reflection coated. I thought that also has a protecting effect?

 

Anyway, I still aim for either a Baader U filter (hopefully Sloan version) or getting the Chroma U-Bessel, but then probably only in 36mm, unmounted.

 

- - -

 

For readers of this thread who may consider buying the Chroma U-Bessel #27050 filter unmounted, I was told by customer service that the thickness is 3mm.

[ulf]

Yes, AR-coating normally protect the filter glass-material from corrosion and the surface transition losses will be reduced, especially if the AR-coating is designed for the correct wavelengths.

The coating has to be tailored to fit the correct wavelength band.

[dabateman]

2 hours ago, LarsHP said:

Ulf: Thanks again.

 

I already have two QB21 filters, but unfortunately only in 1.5mm thickness. They are too thin for "normal" pictures.

 

The idea is to buy two QB21 in 2mm thickness and try that with the ZWB2 in 2mm. Hopefully that works for UV as seen in the test image above, then I have that option. If it doesn't work out, then I can use them for what they are intended, as you point out. I guess I will benefit from these either way.

 

If I find out that 2x2mm QB21 + 2mm ZWB2 works really well, I will try to put at least two of them together in the same filter ring in order to minimize the light loss you point out. Will that help, or will the loss still be there even when stacked directly on top of each other?

 

The QB21 filters are anti-reflection coated. I thought that also has a protecting effect?

 

Anyway, I still aim for either a Baader U filter (hopefully Sloan version) or getting the Chroma U-Bessel, but then probably only in 36mm, unmounted.

 

- - -

 

For readers of this thread who may consider buying the Chroma U-Bessel #27050 filter unmounted, I was told by customer service that the thickness is 3mm.

The losses Ulf indicates are the transition gaps. Stacking them on top of each other will not solve this. It will actually create more problems with Newton rings. To avoid it you would need to glue them together,  which is tricky and maybe not worth it. 

[LarsHP]

Ulf: Thanks for confirming. Hopefully the coatings work in both UV and VIS.

 

dabateman: Thanks! Then I won't make the extra effort and just stack the filters normally.

 

I understand it so that the issue is the air-to-glass transition, and that stacking without having a 100% airtight connection between the glasses won't cancel the light loss. Semi-full connection (not glued) will then have areas with 100% contact while in other areas, there will be a tiny gap => Newton rings.

[photoni]

17 hours ago, LarsHP said:

The TSN575 cuts too much in the shorter end of UV for my taste

.

Yes, but it's the only one that cuts IR well with ZWB1 or ZWB2
Tangsinuo only sells 1.5mm thick QB39~BG39, but this is not enough to cut IR like TSN
I have two QB39s if I put them together + ZWB2 the result is like the TSN

[Alaun]

got some pictures taken with

 

D810 full spectrum, UV-Nikkor,  f11, ISO 200

 

Row 1: Baader U      &  S8612         , Time 0.8 s   

Row 2: Baader U                             , Time 0.5 s   

Row 3: Baader SLOAN  &  S8612   , Time 1.0 s

Row 4: Baader SLOAN                   , Time 0.8 s

 

 

Column 1: WB on 1st picture

Column 2: WB on 2nd picture

Column 3: WB on 3rd Picture

Column 4: WB on 4th Picture

 

Edited because I had mixed up rows and columns

[LarsHP]

photoni: Thanks. The odd thing is that in the test I posted above (found here on UVP), the stack appears to work just as well as the Baader U + 2mm QB21. It even looks a bit brighter.

 

Alaun: Great to see that we have someone with both the Baader U Venus and the new Sloan U-filter! Yes!!!

 

What is your impression of the Sloan version regarding out of band leak in sunshine versus the U Venus?

 

Regarding your comparison, I don't really know what to gather from it. Please explain what you found out, except a small difference in exposure?

 

What does the color difference seen in the upper right-hand quarter of the composite image versus the color in the lower left-hand side imply?

Is that a difference in transmission of wavelengths we see here?

[Alaun]

Well, below are extracts from Raw Digger from the 4 pictures.

Left column is the Baader U, right column is the Sloan

First row is with S8612, second is without. 

(here rows and columns  are correct ;-))

 

Lighting was fulll sun light, but behind a window.

Due to the sligthly different exposure times the camera has chosen,

you have to be cautious when interpreting the results.

 

My feelings: yes, there are some slight differences with respect to wavelength.

I would have similar concerns about IR leakage for both filters.

Within the "upper part" of the UV band (full sun light behind window ;-) ), the Baader U

has a small advantage. After individual WB the differences are negligible.

[LarsHP]

Thanks again. Good to know that they block out-of-band wavelengths equally well.

 

Since this is behind windows (regular glass), I expect the lower wavelengths where the Sloan version has more transmission to be cut away mostly, if not completely, before the light hits the target.

 

I frankly don't understand why the U Venus needs about half a stop less than the U Sloan. It should be about the same, and if any difference, the U Sloan should need slightly less. According to the comparison graph I got from Baader (see first post), the Sloan U is about 2/3 stop more transparent around 322nm, which we obviously don't see after the light has passed the window, but for the rest of the band, the two filters follow each other quite closely.

 

The only place where the U Venus filter is more transparent in a way that could show up in images, is around 395nm, where the Sloan U cuts completely off, while the U Venus still has some transmission (around 5%). This may be the reason, given the graphs and that the sensor is more sensitive there than at shorter wavelengths.

[photoni]

 

 

@Alaun 

your RAW _D814907.nef is 200iso f:11 1/2"=0.5"
your RAW _D814909.nef is 200iso f:11 1/1" ??? what does this mean?-  1,1" ?  - 1"?

.

I'm not a technician but I think it's best to take some test photos...
with direct sun, without window glass
use familiar flowers, this season there are white daisies that turn purple and dandelion flowers that turn yellow with a black center
It would be better to only compare the images coming from the RAW balanced with a small piece of Teflon tape or sheet (better if it is present in the image)
Thank you

Toni

-

@LarsHP asked about the speed gain, it would be appropriate to insert the time and the aperture and the differences in brightness of the RAW

 

[LarsHP]

I just saw that the U Venus filter is 2mm thick according to the specs, not 1mm, like @dabateman says his U Venus filter is.

 

dabateman: Did you actually measure your U Venus filter? If it's really only 1mm thick, then the current model must be twice as thick, or Baader Planetarium has posted wrong specs there...

[dabateman]

@Alaun,

The window test is interesting as I would expect low UV and higher IR than direct sunlight.  However, the subject may not be the best.

Do you have dandelions in your area yet? I would like to see comparison images of the Sloan Baader vs the Baader venus u in direct sunlight with a dandelion.  If there is more IR, the central bullseye of the dandelion goes brown rather than dark blackish. 

@LarsHP,

I bought my Baader venus u in 2008. Its before a couple company shifts and most likely different than the current available ones. Yes mine is only 1mm thick.