Optimal Thickness of S8612: Discussion

[lukaszgryglicki]

Regarding Hoya 360 stack - I know standard here is S8612 + U-360 (both 2mm thick) - but I wonder what is the minimal thickness for this stack? Like maybe S8612 1.5mm + U-360 1.5mm is enough? Or maybe 1.75mm both? maybe one needs 2mm and another can be 1.5mm? Can somebody let me know what is the recommended minimal? Like say I want IR rejection OD4 - then I need Xmm S8612 and Ymm U-360... how about OD4.5 or maybe OD3.5? Any thoughts? Or 2mm + 2mm is generally agreed minimal stack thickness?

 

[Andy Perrin]

Lukas, S8612 1.75mm works but 1.5mm is too thin and lets IR through a tiny bit. I have both. As I was saying to doug in the other thread, it’s pointless to worry about fractions of a stop. It’s not going to affect your overall exposure enough to be worth optimizing. Getting a body that works with high ISO well makes a much more significant difference. 

[Andrea B.]

Regarding Hoya 360 stack - I know standard here is S8612 + U-360 (both 2mm thick) - but I wonder what is the minimal thickness for this stack? Like maybe S8612 1.5mm + U-360 1.5mm is enough? Or maybe 1.75mm both? maybe one needs 2mm and another can be 1.5mm? Can somebody let me know what is the recommended minimal.

 

Excellent question !! 

The answer can be calculated using the Excel based Schott filter transmission chart program. Data for Hoya can be added to this program (tedious). Our member Cadmium is best equipped to answer your question. So I'm going to refer this question to him. 

We do have some partial answers in the transmission charts found here: 

https://www.ultravioletphotography.com/content/index.php?/topic/1313-filter-transmission-charts/

and here:

https://www.ultravioletphotography.com/content/index.php?/topic/2861-filter-transmission-charts-237/

 

[lukaszgryglicki]

Thanks, let's wait then. I don't have Excel - I don't use Microsoft tools. If only I could get tabular data for S8612 and U-360 in 10nm intervals for 1mm thickness then I can calculate everything myself, say I want S8612 1.75mm and U-360 1.5mm. Lets say on 700nm S8612 1mm has 0.01  transmission and U-360 has 0.12 transmission.

Then the result is

0.01^1.75 * 0.12^1.5 and so on for every thickness and wavelength

S8612(W nm)^Nmm * U-360(W nm)^ Mmm.

W = weavelength for calc.

N = S8612 thickness in mm

M = U-360 thickness in mm.

I don't need any brand-locked program for this -0 all I need is a transmission data in 10 or 5 nm intervals, can be as a CSV file.

One more thing: how muchOD is enogh to classify as IR blocked? OD3.5? 4? 4.5?

 

 

[Andy Perrin]

At least OD4 really, but more is better. All that Schott data can be obtained from the free Schott program, which I think can spit out the CSV files. Hoya I don't think has good blocking data, hence Jonathan did a measurement here awhile ago.

[ulf]

2 hours ago, Andy Perrin said:

Lukas, S8612 1.75mm works but 1.5mm is too thin and lets IR through a tiny bit. I have both. As I was saying to doug in the other thread, it’s pointless to worry about fractions of a stop. It’s not going to affect your overall exposure enough to be worth optimizing. Getting a body that works with high ISO well makes a much more significant difference. 

I concur.

Going thinner with the S8612 gains very little exposure time, while wasting big amounts of valuable IR-suppression.

IMGO doing so is most of the time rather stupid.

 

You can naturally optimise the thickness for a certain UV-pass filter combination and get a range of different thicknesses of S8612.

 

You could also gain more by having Cadmium make custom made filter stacks that are cemented to gain ca 9% from having fewer air to glass passages.

 

The 2mm S8612 has enough IR suppression to work well with almost all stack combinations 

[lukaszgryglicki]

OK, that makes sense, I'll ask Cadmium for a cemented U-360+S8612 52mm (2+2=4mm). I currently have separate S8612 2mm and U-360 2mm (both 67mm) and I'm OK with them, I was just wondering.

 

[ulf]

9 minutes ago, lukaszgryglicki said:

Thanks, let's wait then. I don't have Excel - I don't use Microsoft tools. If only I could get tabular data for S8612 and U-360 in 10nm intervals for 1mm thickness then I can calculate everything myself, say I want S8612 1.75mm and U-360 1.5mm. Lets say on 700nm S8612 1mm has 0.01  transmission and U-360 has 0.12 transmission.

Then the result is

0.01^1.75 * 0.12^1.5 and so on for every thickness and wavelength

S8612(W nm)^Nmm * U-360(W nm)^ Mmm.

W = wavelength for calc.

N = S8612 thickness in mm

M = U-360 thickness in mm.

I don't need any brand-locked program for this -0 all I need is a transmission data in 10 or 5 nm intervals, can be as a CSV file.

One more thing: how much OD is enough to classify as IR blocked? OD3.5? 4? 4.5?

 

 

I think several people here can help you producing some results  for a few different glass thicknesses with the Schott program.

If you want to do all calculations yourself there are transmission data for some given base thickness for all Schott and Hoya filter glass.

They also publish some base refractive index that can be used to calculate the surface reflection losses.

Most transmission graphs are showing the internal transmission data without surface.

[lukaszgryglicki]

If I can only find data files I'll put them in some google sheet spreadsheet that would allow all of us to immediately see whatever stack & thickness is needed. Without installing any software on your computer - just in the browser.

 

[ulf]

Regarding the needed OD I saw some problems in extreme daylight situations. with my Baader U.

It has more than OD4:

[ulf]

6 minutes ago, lukaszgryglicki said:

If I can only find data files I'll put them in some google sheet spreadsheet that would allow all of us to immediately see whatever stack & thickness is needed. Without installing any software on your computer - just in the browser.

 

https://www.schott.com/en-in/products/optical-filter-glass-p1000266/downloads

https://hoyaoptics.com/colored-glass-filters/

[lukaszgryglicki]

Almost 0.01 around ~740 means OD2-2.1 ... but OD3.5 will be so much more block - possibly just enough for everything? Why we need OD4 (everybody says that), I think OD3.5 even maybe OD3 will be OK?

[lukaszgryglicki]

Regarding data - will take a look - at the first glance I see PDF links and visible charts or tables - not that useful - I would need to create data files manually, not impossible, but very time consuming. As a programmer - I prefer to write 1000 lines of code and create useful data with a program (then put on some user-friendly google sheet) that just copying data from one table to another (BORING and even 100 lines will be hard to do - prefer 1000 lines of code instead :p).

 

[ulf]

11 minutes ago, lukaszgryglicki said:

Almost 0.01 around ~740 means OD2-2.1 ... but OD3.5 will be so much more block - possibly just enough for everything? Why we need OD4 (everybody says that), I think OD3.5 even maybe OD3 will be OK?

Look again. The scale starts at 100 (%) OD4*4.1!

 

!00% = OD0

[lukaszgryglicki]

Wow, thanks, what a stupid mistake - my bad!

 

[Andy Perrin]

Lukas it's literally all been done by many people here (including me) many times before. But knock yourself out. These were entered by me by hand in exactly the time-consuming manner you described. And then I wrote my thousand lines of code (not included, left as an exercise).

 

U360.m

 

S8612.m

[dabateman]

I think we all tried this.

Stefano even tried playing with 3 or 4 total filters to best optimize something with very thin options. 

But ultimately it depends on what you can actually buy or what Cadmium will make and sell you.

 

I was hoping this NBP365 filter would work out.  But it leaks too much IR and still needs to be stacked with zwb1 or U340

https://www.ebay.com/itm/29-1-0mm-365nm-UV-Narrow-Bandpass-Filter-NBP365-/274773696739?mkcid=16&mkevt=1&_trksid=p2349624.m46890.l49286&mkrid=711-127632-2357-0

 

[Stefano]

Yes, I think it was this: 

 

[Andy Perrin]

Yep, and I tried making a two-axis version waaaay back here:

 

[Cadmium]

I think 2mm + 2mm is best. If those are separate filters, then you can use thew 2mm for other things much better.

I recommend 2mm.  1.75mm S8612 and 1.5mm would work too, but you are better off with 2mm.
 

[ulf]

Stefano's OD tables mentioned above confirms my finding about S8612 2mm and U-360 2mm.

I did the same measurement for that stack as I did for the Baader U I showed the transmission graph for above.

With the S8612-stack there were no low level bumps in VIS or lower NIR at all. The stack is >OD5

[lukaszgryglicki]

Thanks, good to know that it was done before.

Anyway, *.m files are very useful, especially S8612.m - thanks (U360.m could have a bit more data, but still very usable). Does anyboady have links for those *.m files? I would dowenload U340.m and possible some UG*.m files.

 

[Andy Perrin]

Lukas, I made those .m files myself from the datasheets provided by Schott and Hoya. I do have more but I don’t recall if I did U340. 

[lukaszgryglicki]

If possible, and not a problem, can you share your *.m files? I still need to convert them to pure data files, but this can be easy automated and will save a lot of manual work.

 

[Andrea B.]

Remembering that UV-pass filter stack optimization is a 2-variable process of cutting IR and raising UV transmission, one at the expense of the other, please don't everyone make themselves nutso over trying to decide between such things as S8612 x 1.75 and S8612 x 2.00.

 

It's not that I myself don't enjoy such optimization efforts, I do. But the real optimizations lie in the areas of good UV illumination, a sensor with good image quality (high dynamic range and low noise), a very UV-capable lens, and practice. Put some effort into learning to use well what you have, into refining your image making skills both in the field and in the editor.

 

OK, enough advice from Mom. I need to go make some photos!