Jump to content
UltravioletPhotography

Spectroscopy fun with some lens tests


dabateman

Recommended Posts

I was able to recently obtain a JAZ spectrometer with two modules to look at some UV and IR. I have been playing around with it to learn its many oddities.

Maybe obvious things but things I have learned that affect the signal when trying to determine the transmission through a lens.

 

1. Most critical is the alignment of the fiber optics. This is obvious. As you slightly shift the optical fibers you will affect the light detected.

 

2. The path length seems to matter. As in, a longer tube will affect the transmission. I was surprised by this not being negligible. But makes sense, more air between the fibers increases the scattering of the light, which decreases your signal.

 

3. Orientation of the lens matters. If the light enters the back of the lens and the detector probe is at the front of the lens you will get lower transmission than if the light is coming in through the front and the detector probe is located at the rear of the lens. This might make sense as the optics have changed, flipping a lens it becomes macro.

 

4. The focus of the lens matters. If the lens is set to infinity or to its most macro position will affect the transmission of light. This doesn't just seem additive based on increasing the tube length of the lens as you focus closer.

 

So for these plots you can ignore the maximum transmission number. I am not sure I have controlled everything properly. Just wanted to share the cut off points for these screen captures. For this set my 39mm custom lens was set as 100% transmission. Which is a single element fused silica lens. The "exposure time" was 1000 ms for all lenses with average of 3 scans and boxcar setting of 3. This detector has a range of 178nm to 876nm, using grating #2, an L2 lens on the sensor and slit 25um. I am using a deuterium/halogen UV visible light source. All lens were measured with light entering the front of the lens and detector located behind the lens and with the lens set at infinity focus position.

 

Ignore the settings in the screen grabs. I see i have them on the second module which was not used for transmission measurements. It just allowed me to collect raw data at the same time for each lens in the IR portion of the spectrum. It has grating #4 and appears to be exactly 3 times less sensitive to my other module at the overlapping wavelengths. Thus why its set to 3000 ms.

 

Pentax 85mm UAT lens:

post-188-0-58217200-1589095701.png

 

KSS 60mm Macro f3.5 lens:

post-188-0-52385200-1589095722.png

 

Nikkor 80mm F5.6 EL lens:

post-188-0-09544200-1589095749.png

 

iGoriginal 35mm f3.5 lens:

post-188-0-34118900-1589095766.png

 

My 39mm lens with my SvBony #D 0.5x focal reducer:

post-188-0-87305400-1589095817.png

 

Sigma 30mm f2.8 Micro four thirds AF lens:

post-188-0-36751500-1589095898.png

 

Canon 40mm f2.8 STM lens:

post-188-0-31262300-1589095914.png

 

Canon 40mm f2.8 STM lens on the Metabones Speed booster ultra:

post-188-0-65698200-1589095930.png

Link to comment

Interesting Dave, I can see how you have set this up & your standard procedure etc...you didn't mention it but I presume this is with the aperture wide open too ?

Your 'Pentax 85mm UAT lens' went off the wavelength scale but was low transmission %.....why....?

Link to comment

Yes aperture wide open for all lens.

I don't know what is going on with the UAT. If that is real, or a fiber out of alignment issue.

I will have to retest that.

 

One other thing I did notice is that wide angle lens focus the light better and result in a better transmission than telephoto lenses.

I saw a like linear step in my fused silica elements, when looking at the raw data. Bicx 25mm was best transmitting, with decreasing singal for each following focal length in linear progression. I have 25mm, 50mm, 100mm, 200mm, and all dropped transmission.

So that can be an effect as well.

Link to comment

Thanks Dave

I wish testing lenses & filters was not so complex....

 

Then, perhaps not so much "fun" ??

Link to comment

If spending a lot of money & not getting the results you want.....then it is all too hilarious.... :lol:

 

Its not about what I want, it is what it is.

The JAZ system lets me do everything, absorbance, transmission, fluorescence, reflectance and since it has its own onboard computer, I can run around outside and measure reflectance off stuff.

 

But none of that was what the instrument was purchased for. I want to measure and calculate extracellular potassium levels. Might end up using 340/380 absorption ratio or 510nm fluorescence of PBFI. My module 2 has a grating 4, lp475 filter and 25um slit. I may get a 100 or 200um slit if need more signal.

 

These are just for fun to see how it works.

Link to comment
Andy Perrin

Do you have anything to validate these curves with? I know that at least for the EL-Nikkor 80mm/5.6, Ulf has done prior tests, and also Klaus. Those might be a good start for benchmarking. The one thing I've learned from Ulf and Jonathan about spectrometers is that it's possible to make some VERY subtle errors. (One recalls the case of the peculiar R72 filters...)

 

In particular, I would like to know whether the results at very short wavelengths are believable for the KSS lens. It looks like it drops off before 200nm, but that seems like it may be questionable.

Link to comment

Just playing around still Andy.

I completely believe the KSS curve. It models what the design says for cut off and my experience with trying to image at 185nm.

I don't believe the UAT curve and a retest is needed. But I do have the same dip around 300nm for UAT that Jonathan previous reports. So may not all be wrong.

Link to comment
Andy Perrin
Assume it is wrong until proven correct (in a certain range of wavelengths). Do tests to show it's working properly. It's the same rules as with computer models — wrong until shown to be right.
Link to comment

Assume it is wrong until proven correct (in a certain range of wavelengths). Do tests to show it's working properly. It's the same rules as with computer models — wrong until shown to be right.

 

Nope computer models are always wrong.

 

Link to comment
Andy Perrin
Ummmm...no? Computer models that correctly model the physics will agree with the real world very well. If this is your sense of humor again, I’m not amused.
Link to comment

I too would be very interested in details of your setup.

I found that my DH-2000-BAL light source did not produce adequate illumination for testing lenses using 600um fibers.

My current direction is something using a Godox AD200 with uncoated quartz tubes for illumination, but life keeps getting in the way of my experiments.

Link to comment

Ummmm...no? Computer models that correctly model the physics will agree with the real world very well. If this is your sense of humor again, I’m not amused.

 

Yes sorry, I correctly modeled many NMR experiments as the physics and math are known and well defined.

I was making a joke about the current state of affairs. But should watch out for the extreme generalness of that statement.

 

Multiple methods are always best.

Link to comment

I too would be very interested in details of your setup.

I found that my DH-2000-BAL light source did not produce adequate illumination for testing lenses using 600um fibers.

My current direction is something using a Godox AD200 with uncoated quartz tubes for illumination, but life keeps getting in the way of my experiments.

 

Thats very close to my current setup. For all these initial tests I have a 600um fiber 22 inches long going from an AIS DT1000 deuterium tungsten halogen light source into a home made dome. I should take a picture. That fitted with 52mm ring to screw into the front of a lens. Then at the back used either an equal home made dome or m43rds lens cap with fiber (antenna connector, as ripped these off old web cameras). From that I connected a bifubricate optical fiber that is two 300um fibers that split into both my detectors.

Every thing is light tight, from source to detect. My biggest concern is optical fiber alignment, as any little wiggle, does effect the peak transmission.

 

Link to comment

I would start with testing transmission of known filters, to begin with to learn to work with the spectrometer.

 

There are many pitfalls in measuring lenses.

I think a collimated beam into the lens is a better setup.

Then an integrating sphere to collect the emerging light.

 

If everything is setup correctly with a centered beam, decreasing the lens's iris from fully open would not change the detected signal.

 

Some Jazz-spectrometers are not very stable in one can believe what's written on the web.

I have never used a Jazz.

 

Your setup is OK to do a check of the UV-cutoff wavelength.

I would define that where the level has dropped by 50% from it's peak value.

At low levels there is much errors due to stray light in array spectrometers.

Link to comment

Ulf,

I am glad you commented. Where have you read anything on the stability of the JAZ spectrometers? I haven't yet read anything negative, but I haven't read everything.

I probably should have asked a certain member here about them as I think I found code written by them to decifer the on board data format.

 

Lenses are definitely tricky. I think what I am seeing if a focal length effect. Where a long focal length lens will only slice out a small angle from the input light and deliver that to the detector. Thus a wide angle lens takes in more light than a telephoto lens by design, so just get more transmission that way. It seems to evenly scale with focal length in my hands. A 25mm BICX single lens element has twice the transmission of a 50mm single element, which has twice the transmission of a 100mm element, when all lens are setup with the exact same total light path. Just drop in different focal length lens elements.

 

To solve that your idea would work. I have two culminating lenses. But not sure if I can use them in the way I have it configured for lens tests. I have no plans to buy an intergrating sphere. True lens transmission is not my goal for this setup.

 

I am just having fun here. Knowing where the lenses cut off into UV, I find informative and I think this setup is valid for that. I do say to ignore the total transmission here at the top.

 

The very first thing I measured was my light sources. All where mostly as expected. My newest unknown 385nm light was only 400nm, sadly. My 25W 185nm light blew out, I had used it for only about 2 total lamp hours. But the seller gave me a full refund, as should have lasted 8000 hours. Will order from them again.

My little 3W 185nm bulbs where falsely advertised and exploded. I did get a partial refund on them.

 

Be extremely careful with 185nm or ozone lights. The 3W ones no matter what the seller says are maximum 12V, 0.3 Amps in the E17 format. NONE of them will support 120V.

 

So I don't know my lowest UV limit yet. Not sure if I can see the 185nm mercury line. That is not the goal of my setup though. I only have grating #2. If I really want deep UV, I would buy an additional module with grating #10 and a 100um slit. That might be the best configuration to get the most UV light in. Just as grating #14 might be the best for IR. But I am happy with #2 and #4 with 25um slits. I think I am getting at least the calculated resolution of 1.5 nm, and I have great sensitivity and range.

I know I can differentiate between 297 and 302nm mercury lines easily. I don't think I can differentiate the 312nm and 313nm dublet, but will double check.

 

Now I just need an excellent fluorescent potassium indicator that is commercially available or can be obtained.

 

Link to comment

Lens testing for absolute transmission is very difficult to do. My set up is a collimated 200w xenon lamp, which then goes through an adjustable pinhole. The light from this passes through the lens and into an integrating sphere, before being collected for analysis. It was harder to get this right than it was to make the system I have for measuring sensor sensitivity. Even with mine I get some non-optimal readings occasionally, especially with lenses with small apertures.

 

David, are you using an integrating sphere? If not then I'm not sure how you're going to be collecting everything thats coming out of the back of the test lens for analysis.

Link to comment

Lens testing for absolute transmission is very difficult to do. My set up is a collimated 200w xenon lamp, which then goes through an adjustable pinhole. The light from this passes through the lens and into an integrating sphere, before being collected for analysis. It was harder to get this right than it was to make the system I have for measuring sensor sensitivity. Even with mine I get some non-optimal readings occasionally, especially with lenses with small apertures.

 

David, are you using an integrating sphere? If not then I'm not sure how you're going to be collecting everything thats coming out of the back of the test lens for analysis.

 

Yes I will not be able to do absolute transmission measurements. I think I understand that. The best I could do it run a standard curve with single lens elements to determine the "focal length modification factor", then use that to determine a relative transmission for a lens set tested on that day with the calibration curve.

However, I don't think I care. I rather just know the UV and IR cut off to best select a lens for those wavelengths of light. Then under my specific light and subject with my specific filter, determine a relative T-stop value for a lens, like I did at 370nm with EL80. To best choose my lenses for the image I have in my mind.

 

I have no interest in testing other peoples stuff. Or even controlling for everything.

Once I get my cell system up and running. I may not even be able to find time to test my fun stuff.

Link to comment
Absolutely David, if you don't need absolute measures, then a much simpler setup will give you plenty of info about cutoff.
Link to comment

David,

 

I just saw a comment on a forum here:

https://www.researchgate.net/post/Field_Spectrophotometer_for_Vegetation_Reflectance

Look for Geert J J Verhoeven's comments.

 

To setup a proper absolute transmission-test you need a strong light-source with a reasonably narrow beam that can enter the lens's optical axis and not hit the iris, preferably at least a bit closed.

Then you need an integrating sphere with a good wide-band reflecting matted surface. If all light exiting the lens is integrated the same way as the unobstructed beam you can fond the losses caused by the lens by observing the differences.

The main problem here is that the sphere must be big enough to allow all light from the lens to enter. That means that the light from the wall's of the sphere will be much dimmer than the beam itself. The output from your deuterium light-source is so weak that you would need a very long detection time. During that time everything will have changed due to drift.

 

You can do this with power LEDs for short wavelength parts, but unfortunately such LEDs are only available down to 360nm.

 

I do not know any other way than with a powerful short gap continuous xenon lamp as Jonathan uses in his setup.

Such light-sources can be even stronger than 200W. I have seen described 1000W units.

 

You can build a collimator yourself with a small planoconvex fused silica lens, that is placed with it's focal point at the fiber's end.

As the material has dispersion it is a good idea to limit the tested bandwidth and make the collimator adjustable for optimal focus.

Link to comment

Well I do also have the JAZ PX pulsed xenon flash lamp. So may test that out as well.

 

I thought for an integrated sphere to work, it would need the entrance to match the lens exit. I see you are saying that.

 

So something I would have to construct, if I want to go down that route. Just need a large rubber ball and some liquid PTFE. Might be doable in the distant future.

 

For now I just want to know which lenses are horrible in the IR, or deep UV as the glass and/or coatings cut that wavelength range off.

The Canon 40mm f2.8 STM lens seems to be the best right now for both IR and UV. The Sigma Art 30mm f2.8 lens cuts off in the IR.

 

Ulf,

Thanks for the link. I only see one user having issues and it wasn't fully explained what he was doing. I am not worried. If I see saturation, I now know to see if the unit is over heating. My entrance slits are small and my fibers are 600um. I am not coupling a lens directly to an open non-slit port.

He also seemed to throw a large blanket that all ocean optics spectrometers were bad. I don't think you or Jonathan have had issues with yours. So I am not worried.

 

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...