Checking my lenses for UV capability

[colinbm]

The other reason for using the Sunlight for these tests is that it is the light that we are photographing under mostly.

I am using this test mostly to compare my lenses against each other to see which are most suitable for UV.

[dabateman]

I don't see too much wrong with Colins method.

He can have a couple favorite lenses. Then on the day he wants to photograph something, he could scan the lens he wants to use to get an idea that day and time of its transmission. Then use the one with the focal length and UV depth he would prefer.

He will not get an absolute number. But I don't think he needs one. Also the variation in these lenses it doesn't really matter.

I think Colin just wants to know which one of his lenses can see the deepest into UV. He can scan them on multiple different days and see which one seems to have the lowest cut.

For added fun he should also scan it with the filter he wants to use that day outside in the sun to get a better idea what might be available to his camera.

[ulf]

The other reason for using the Sunlight for these tests is that it is the light that we are photographing under mostly.

I am using this test mostly to compare my lenses against each other to see which are most suitable for UV.

That makes sense, and your method is OK for that.

 

I think much of the hype for lenses reaching 320nm - 330nm is not valid if the lightsource is the sun.

Down there it is very little UV-light from the sun.

Combined with a rather low sensor sensitivity, none is normally contributing to the final image as it is overpowered by all light above 350nm.

 

Try comparing the lenses with a Baader U mounted instead. That will tell you more.

 

Not me or any relative:

[colinbm]

Thanks Dave & Ulf, yes that is my intentions.....

That would be an interesting exercise photographing the same scene with the camera & the spectroscope....;-)

[rfcurry]

Colin,

 

Klaus Schmidt laid a good foundation years ago. Maybe you can build upon his work, with proper attribution?

http://overmywaders.com/images/Lenses%20for%20UV.jpg

[colinbm]

Thanks for this Reed, that is very extensive indeed.

[Cadmium]

Does the type of fiber optic cable limit the wavelength band of the sunlight transmitted down from the roof top?

[dabateman]

Does the type of fiber optic cable limit the wavelength band of the sunlight transmitted down from the roof top?

 

Seems like a random comment but yes.

I have fibers that are capable of transmitting 190nm to 1200nm, that ate 600um in diameter.

I tested my fiber optic cable I bought for my stereo and its about 400nm up.

I bought a filter wheel and it came with 2 fibers 77525 that are 1100um in diameter. They transmit slightly differently, but only above 340nm. They are rated to transmit from 400nm to 1500nm on the spec sheet.

So the fiber makes a huge difference.

Ocean optics also has a pdf somewhere that outlines the color codes and fibers.

 

But i wouldn't expect much below 400nm for a typical stereo, or Verizon grade optical fiber.

[ulf]

Does the type of fiber optic cable limit the wavelength band of the sunlight transmitted down from the roof top?

It definitely do, as David indicated above.

 

The light conducting core must be made of a material intended to conduct UV.

Normal fiberoptic cables for visual light, like microscope illuminators do not work.

 

For spectrometers thin cores 100µm - 1000µm are mostly used.

If thicker cables are needed one can look for multicore cables used for UV-curing.

They often come as multi branch types:https://www.ebay.com...C0AAOSwgope2H-f

[ulf]

Colin,

 

Klaus Schmidt laid a good foundation years ago. Maybe you can build upon his work, with proper attribution?

http://overmywaders....%20for%20UV.jpg

 

Please observe that the used threshold for defining the UV-reach is selected at a rather odd -3 stops, giving a lower wavelength value.

These results have in some cases entered the UV-lens sticky, but without mentioning the -3 stops threshold.

That is not comparable with the normal way of defining filter cutoffs. (-1 stop)

 

Combined with the declining sensitivity of camera sensors in the UV-region the stated wavelengths are not very usable.

However the relative position in the list and the great number of lenses tested are usable for searching suitable lenses to try out.

[Andy Perrin]

Ulf, this is why I am suggesting normalizing all Colin's spectra so they all have the peak at 1 -- it's hard to tell what the "cutoff" is if the vertical scale is different on every spectrum picture.

 

Colin, if you want, I can do the analysis for you but you would have to send me the spectrum numbers in some kind of file. I can't do it from the pictures.

[dabateman]

Colin the manual says it can save files in csv format. You should be able to save your scans, then connect to a computer and download the csv files.

You will have to remember the order saved for analysis.

[colinbm]

Andy & Dave,

Yes the files are in csv & have the the lens in the file name.

I can send the csv of each of these.

[Andy Perrin]

Ok sounds good, you can contact me via PM and we can set something up.

[colinbm]

Thanks Andy, will do now.

[Andy Perrin]

Colin's numbers for relative transmission, normalized to all be 1 at exactly 620nm.

 

Same graph cut at 620nm on the x-axis.

 

And same graph cut at 320nm to 400nm.

 

The last graph makes especially obvious why looking at where they go to *zero* is a bad idea. The spectrometer is cutting off. Instead, I would judge by the 50% relative transmission.

[dabateman]

I am not too sure what more we have learned looking at them normalized.

One also has to be more careful with the data. As you normalized these data sets to each other rather than to a control source.

 

When measuring a lens the tube length, the lens orientation and the aperture will affect the measurements. Also depending on how effective the detector is a culminating the total light, you may have losses due to diameter of the rear element.

 

So you have to be careful when interpreting this normalized data.

 

Some how your normalized data makes the Sigma look much better. Before its 50% cut was around 400nm. Whereas your data places it about 378nm.

I think the original raw data Colin provided may tell a better story.

[Andy Perrin]

I don't know how you think the sigma looks better. It is clearly the worst of the lot here? Anyway, everyone has all the data, they can decide.

[dabateman]

Andy,

Try normalizing the data within its self.

100% is the maximum light through each lens or no lens. Then plot each on the same curve with no lens for comparison.

The data will look much different.

[Andy Perrin]

Dabateman, the IR results (which look wrong to me) will mess that up.

[colinbm]

Thanks again Andy for doing this conversion.

The Sigma is not the best lens, but it is the best with auto focus & aperture etc.

I can't argue about the normalizing process.

I have asked the manufacturer why the short end of the UV is truncated....?

[Andy Perrin]

Yeah the truncation thing is odd. I wonder if there is a setting somewhere.

[dabateman]

Dabateman, the IR results (which look wrong to me) will mess that up.

 

I don't think I follow.

The sun is a 480nm light bulb.

These lenses seem to shift the peak max to 550nm. Not surprising as that is what we detect most with our eyes.

The light fall off should still be ok.

You could truncate the data at 550nm, being a UV forum. But knowing which are also good in IR is useful.

Lenses focusing more IR light than no lens doesn't surprise me. Also Sigma lenses typically cut the IR down or out. So expect the Sigma 30mm to be the worst for IR. Thats due to Forveon sensor being so strong in the IR.

[Andy Perrin]

Dabateman, I think I'm done working on this. You can ask Colin for the data and analyze it however you want with his permission?

[ulf]

Nice work Andy.

 

It is good to see simplified tests that is based on sunlight.

That narrows down a bit to what is essential for sunlight-based UV-photography.

Add the low deeper UV-A sensitivity of the camera's sensor to the mix and lenses really deep UV-reach is even less important.

 

If just discussing the measurements and possible reasons to what we see, not relating to photography I think I have some answers.

These results in the deep end is not very accurate as the amount of UV-light is generally lower than the visual part.

Also from time to time the relative UV-content might vary.

The truncation thing might be due to low light in combination with some offset-drift in the instrument.

 

The odd variations in the IR-end might be explained by how the instrument is working.

It is intended to measure absolute irradiance for a wide range of wavelengths.

As the detector system has varying efficiency for different wavelengths different gain has to be applied over the range.

 

The core of the instrument is an array spectrometer with a wavelength dividing grating.

Gratings that are efficient into UV normally have low efficiency in the NIR-end, demanding more gain here.

Together with the detected signal, all offsets, drift and source errors are also gained, giving a more inaccurate measurement result in the NIR-end