[Filter test] LP1100nm

[dabateman]

I purchased the Long pass 1100nm filter. It is reported to have the following spectrum:

I photographed a glass of water to see if it darkens it does seem to. But the Camera settings don't make sense to me. Does anyone think the following would be possible and what might be something better to test with.

 

Water glass, using Pentax UAT lens ISO 200, F5.6, 1/20 seconds and RG1000 filter (950nm cut off), light source was Halogen bulb 75W

 

Water glass, using Pentax UAT lens ISO 200, F5.6, 4 seconds and RG1000 and LP1100nm filter, light source was Halogen bulb 75W:

 

The water is darker, but the camera settings seem too fast. Any ideas how else to best test if I actually looking at 1100nm?

 

My RG1000 has the following spectrum that I collected years ago:

[Andy Perrin]

Those settings seem weird to me also, Halogens do put out a terrific amount of IR though.

 

Get a piece of thin silicon and see if it’s starting to go transparent?

[Cadmium]

I am confused.

When you say, "... 4 seconds and RG1000 and LP1100nm filter...", what do you mean by RG1000 and LP1100nm", does that mean you are stacking those? If so, why?

Regardless, if the RG1000 is 1/20 second exposure, and the LP1100 is 4 seconds, then that suggests something.

Like you point out, water is darker, and that is to be expected of the higher IR cutoff point, and that tendency can be seen when comparing 850nm and 1000nm even.

The exposure difference seems to make sense to me.

 

Added:

Schott longpass filters are labeled with thxe 50% Internal Transmittance (Ti) using 3mm thick filter glass.

That rule is true of all Schott longpass filters (except N-WG types, below 395nm).

So for example, RG1000 @ 3mm thick has a Ti transmission that is 50% at 1000nm, as shown on the comparison graph below.

 

Actual Transmittance (T) would be slightly different, and thickness will also change the graph transmission.

(keep in mind that typically a camera filter is closer to 2mm than 3mm)

 

The differences between T and Ti and thicknesses is more accentuated with higher NM filters such as 1000nm.

As the longpass nm becomes lower, there is less and less difference in such comparisons.

[OlDoinyo]

Dichroic spectra are only strictly accurate if the light path is perpendicular to the filter surface. Such filters will have a tendency to leak a bit when off-axis light paths are in play. I don't know how important that is in your case; this is just FYI.

1100 nm is close to the detection limit for some digital sensors. If this is an issue in your case, any out-of-band light that makes it through the filter may have a tendency to dominate the image, because the sensor is more sensitive to it.

[dabateman]

Cadmium,

Yes I am stacking them. Because I held a white led flashlight up to the Lp1100 filter and I saw faint green through it. So I assume some leakage and wanted to block all visible wavelengths.

 

Objects have a crazy glow to them with the filter, and hallogen light. The off axis effect makes sense to explain that.

 

I guess I am surprised as I thought I would be near sensor limit and need much longer exposure time. Its also the first time I have used the UAT lens for IR. Maybe if I try a regular lens it will also block IR. Sigma lenses are known to block over 850nm, I may try one of mine.

 

Thank you for the comments. I just want to make sure when I think I am seeing 1100nm, or even 1080nm, thats what I am looking at.

[Cadmium]

I may have missed something in your text above, and I am not familiar with the Lp1000 filter, so... is it a dichroic coated filter as Clark asked about?

Do you see the green from the LED through the RG1000? I guess not then, because you used it to block the green?

 

I have posted this before, but below is a comparison, and one is a stack of RG1000 2mm + S8612 1mm, which creates a small transmission starting at about 1150nm.

It requires a much longer exposure time than your filter however, and I am not exactly sure if that exposure time shown is correct, it has been a long time since I tried this test.

The main thing we see here is the darker color of the water.

 

[dabateman]

Yes the Lp1100 is dichroic filter.

I hold the flashlight to my RG1000 and I can't see anything. With my r720, I can see red. The Lp1100 I see green, not yellow or red. So I think there maybe a peak leakage in visible. A lower cut LP would be yellow like the Gg475 or orange or red. Being green I have hope that it may just have small visual peak leak.

 

I have a 2mm S8612, might be worth an overnight image to compare.

The white board my cup is on is ptfe slap so looking above, I should correct the WB to get better idea.

Thanks Cadmium.

[Andrea B.]

Who makes this LP1100? It does appear to leak. That water should be a bit darker regardless of WB setting?

[dabateman]

Andrea,

The filter was from bjomejag listed on ebay. Not sure if its an Omega overrun. I had great luck with the first two filters I purchased, the 390bp25 and the 330WB80 Improved. Both blocked and worked well. The 390bp25 is one of my favorite UV filters. But all the others I have bought since then, have been let downs. Not as described, and have had to do some work around to get what I wanted. The most recent two have had major Green leak.

 

For Cadminum,

I tried a crazy S8612 + RG1000 image. However my S8612 is 2mm not 1mm, so that probably had in impact. This test really Surprised me though. Not only do I not see the dark water. But my lenses did not perform as I had expected. The Pentax UAT was black. Could not get an image through. Not sure if that mean the limit is 1100nm on the front coating or if something else was wrong. I was able to get an okish image Using the ever amazing Sigma 30mm f2.8 Art lens. The image below is 7 minute Exposure, f2.8 at ISO 6400.

 

 

I think whatever is getting through is greater than 1100nm. The lighting on this image is two 75W Halogen bulbs positioned on the left and right about 10 inches away from the shot glass and I have the lucky herp mercury vapor lamp to image Right about 18 inches away. From the photo you can see the major light is Only coming from the mercury vapor lamp. This surprised me until I read this paper:

 

Spectral identification of lighting type and character, Sensors 10(4):3961-88 · April 2010.

 

These figures are taken from that paper.

 

The mercury Vapor lamp may have about 10x more IR output than my Halogen bulbs. My Halogens are 75W not 60W and my mercury vapor lamp is 125W not 160W, but I think the graphs would be close enough.

[Cadmium]

I am not sure what your image is seeing there, because an RG1000 2mm + S8612 2mm would suppress a lot more of the 1150nm-1200nm+ range (see graph below).

If you have BG40 2mm, that may work about the same as S8612 1mm for that range. Old rule of thumb S8612 1mm ~ BG40 2mm.

I just looked at the BG40 stack in the program, and it crosses the 1E-03 line at the same place, 1150nm/1160nm.

You could try BG38 if you have it also (same as CC1 filter, 99% sure).

Sorry the program doesn't show higher graph plots than 1200nm, but since our cameras don't see that...

If we could see up a little higher, then the water would look very black, like skin looks at that range (see link).

http://www.ultraviol...v-vis-nir-swir/

 

 

PS: In the 4 shot comparison above, I was using a simple 75W incandescent bulb.

http://www.ultravioletphotography.com/content/index.php/topic/2975-filter-test-lp1100nm/page__view__findpost__p__23682

[dabateman]

I do see something, its at 6400 ISO, f2.8, 7 minutes. But there still is something. So its possible this sensor maybe seeing above 1100nm.

I also think something is wrong with my water. I might try this again with bottle water.

[Cadmium]

Yes definitely, I think sensors see above 1100nm, for sure. How far...? I can't say, I would say 1200nm, and beyond even, given enough exposure and no other light.

Well, I doubt there is anything wrong with your water. I don't think you will see any difference between your water and distilled or RO or whatever type of water.

The water in water is what looks dark in higher IR, anything else in the water would probably not make the water look clearer.

For example, the black skin body pics, they look black because of how the water content of the body looks in higher IR, yet the body is only about 60% water, or less.

Try it with red wine... it becomes clear.

[Andy Perrin]

Cadmium no way do you get to 1200nm. Silicon becomes transparent between 1100 and 1200. If no light is absorbed in the sensor you can’t get an image. In addition the noise becomes equal to or stronger than any signal at some point. There is not just random noise which can average out, there is also fixed readout noise and amp glow. It’s better to find yourself an upconverting fluorescent screen like i was using in the old SWIR threads. That technique could be improved on. I was taking ten second exposures which certainly beats half an hour, and the wavelength range was 1480-1600nm according to the Edmund supplied spectrum.

[Andrea B.]

David, try this in strong sunlight, perhaps?

[dabateman]

@Andy,

I know, but how do you explain what we are seeing in the image?

The glass shot glass is filled with tap water on a PTFE board with PTFE board behind. Its possible the light is illuminated the PTFE and we are seeing a "negative" image but with this filter cobination only light about or above 1200nm is getting to the sensor.

[Andy Perrin]

with this filter cobination only light about or above 1200nm is getting to the sensor.

dabateman, it's the last part I don't really believe. I think you are forcing light at other wavelengths through. Which ones, I have no idea. I can tell you that at the 1480-1600nm range, water is totally inky black. Here is my photo using the upconverting fluorescent screen method:

 

[dabateman]

@Andy,

You maybe correct. My first post I have the actual absorbance spectrum for my RG1000 filter and I can see a dip start right at around 435nm. This matches with the 436nm Mercury line that is strong for my lamp. The shift in white balance to blue from the grey IR setting I have may also suggest forcing this line through, which the S8612, will easily allow. That also would explain the normal water appearance.

With an absorbance of 3.5, about 0.05% light would becoming through, which with a 7 minute exposure is possible.

All more plausible explanations than higher wavelengths.

[Andrea B.]

Which is why I'd like to see you try this shot outdoors in strong sunlight which has lots of IR. The outdoor exposure would likely be much shorter and perhaps therefore suppress some of the leakage. Then we could see if the 1100LP filter is valid.

 

But thanks for your experiments with the reptile light......

 

PTFE fluoresces some.