Jump to content
UltravioletPhotography

EOS 5DS R - sensor filter transmission measurement


Recommended Posts

I recently had one of my Canon EOS 5DSR cameras converted to multispectral by MaxMax. When Dan did this I asked him to send me back the internal filters which he removed from the sensor, as I wanted to measure their transmission spectra. This post is a quick writeup of what I found. Transmission measurements were done using an Ocean Optics FX spectrometer (between 250nm and 800nm), and an Ocean Optics DH-2000-BL light source.

 

Internal filters

Two filters were returned along with the camera - a clear one and a green one. The clear one is apparently the coated ICF assembly and dust shaker, and the green on is BG38 or something similar to block IR. I noticed that the clear one started to look green if I held it an angle, and wasn't looking straight on at it. At certain angles it had a red tinge as well. Pictures below show the two filters. Firstly the green IR filter (these often crack during removal as they are stuck onto the sensor and are very thin).

post-148-0-60928200-1526644109.jpg

 

And the clear ICF filter assembly, showing the colour effects at different angles.

post-148-0-12051800-1526644125.jpg

 

post-148-0-56491900-1526644140.jpg

 

post-148-0-83974600-1526644156.jpg

 

post-148-0-47456400-1526644172.jpg

 

Transmission properties

The first set of plots shows the transmission of the two filters individually, and when combined together. This was done with the light source at 90 degree to the surface of the filters;

post-148-0-30555200-1526644093.jpg

 

As expected the green filter shows the typical IR filter profile. The ICF filter looks to be a UV/IR cut filter. Combined together they give a strong UV and IR blocking (apart from a weird little spike around 370nm), and strong red attenuation.

 

The second graph, shows the transmission through the ICF filter, when the filter is held at different angles, from flat (incoming light at 90 degrees) up to 60 degrees from flat (incoming light at 90 degrees to the filter surface);

post-148-0-56074000-1526644182.jpg

 

The graph is a little busy (I can split these out if people are interested), but shows as the filter is tilted angle of the incoming light moves away normal to the surface, the transmission characteristics of the filter vary quite significantly. With increasing angle, the transmission of the filter shifts to the left, to shorter wavelengths, with less red being transmitted. This presumably while it looks green when looked through at an angle. It also gets pretty complex, with certain wavelengths being let through and others being blocked. The red seen on the filter at certain angles is presumably reflection (I'm guessing as it isn't being transmitted, it is being reflected).

 

It needs to be tilted quite a way to get these shifts, but it does make me wonder what is happening with, for instance, extreme wide angle lenses - what is the most extreme angle the light can hit the sensor from the back of the lens? Could this lead to colour shifts towards the edges of the image? Do the effects become more extreme with mirrorless cameras where rear lens elements can get much closer to the sensor?

Link to comment

what is the most extreme angle the light can hit the sensor from the back of the lens?

 

Assuming a theoretical, totally unrealistic lens with a point shaped rear lens placed 44mm above the center of the sensor.

(44mm is the back flange distance for Canon EOS )

A full frame sensor with a diagonal of 43.3mm.

Then if i calculate correctly, I get an angle near 30° for a beam to the corner of the sensor.

A realistic lens has a diameter. The Soligor, Kuri clones has a small rear element ca 10mm in diameter.

That lens would give 22° if all of the lens diameter is used for the image.

 

I think angles of less than 20° is more realistic to expect as a worst case.

Link to comment
Ulf, in which case then the effect on a SLR would be minimal, which is good to know. It would be interesting to do the same calculation for mirrorless cameras though given the back of the lenses can get that much closer to the sensor.
Link to comment

So those internal filters must be dielectrically coated?

 

 

 

 

edit: removed erroneous remark used dichroic rather than dielectric.

Link to comment

So those internal filters must be dielectrically coated?

 

 

 

 

edit: removed erroneous remark used dichroic rather than dielectric.

 

Of the two filters there is a green one which is the IR blocker. This one didn't change colour as I changed my position to it. The other one, the one which looks clear as to look at it at 90 degrees to the surface, did change in colour as it is moved. I think this one is coated.

Link to comment

How thick are these filters, Jonathan? Knowing the thickness of the BG glass would allow us to compare it to some charts for BG glass and narrow down the "guess".

 

General question: Does a dielectric coating imply dichroic? I don't know. Or if I did I've forgotten.

Link to comment
Andy Perrin

Dielectric just means it is an electric insulator. It does not mean dichroic. Dichroic means that the filter works by reflection of unwanted light using interference effects rather than absorption (e.g. filter glass is absorbing). Sometimes a combination of an absorbing glass and a dichroic coating is used, as in the Omega 330WB80 or Baader.

 

Because dielectrics (insulators) also tend to be non-absorbing, they are frequently used in the layers of dichroic filters. It can be shown mathematically that an ideal dielectric, which does not conduct electricity at all, will not absorb light at all because the mechanism of light absorption is electrical resistance, which requires conduction.

Link to comment

Not sure Andrea. I can measure the green IR filter as that is already cracked. Not sure about the other one, will have to have a look.

 

Edit - just measured them. They are about 0.47mm each.

 

Not sure about the dielectric/dichroic terminology. Bit above my head :)

Link to comment

The filters are similar to the ones I extracted when converting my EOS 60D, quite thin too.

 

The transmission for my filters:

Blue, BG-type: 0.56mm. Green, dielectric/dichroic: 0.50mm. Magenta, both combined, 1.06mm.

post-150-0-13238500-1526759386.png

 

For my filters, I could not find any of the Schott BG types that give a transmission similar to the BG-type for 0.56mm.

It could be a Hoya glass or a special custom type made specially for Canon.

Link to comment
The green curve appears to be a 400nm longpass with magenta coating to block IR. The BG is added to give a more natural red shoulder? I'm just guessing of course. :lol:
Link to comment
  • 4 years later...

looks like the clear filter cuts off anything less then 350. so it's good it was removed. Quarts cover glass is needed for best UV I think. 
so these full spectrum conversion are really only UVA to IR capable, since most companies leave the sensor cover glass on. 

 

every camera I have taken apart has that green/magenta coating on some glass . sometimes the coating is on the green IR block glass.

 

Link to comment

found this online comparing a lot of cameras spectral range. Interesting to see the differences between manufacturers. Aparently Sony color or Canon Color is a thing perhaps due to filtration methods. 

ir uv cut.jpg

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...