If you ever visited some industrial surplus shops, very often you would see some cameras and lenses used in industrial automation. But you probably do not know that these lenses can have very high optical performance and features we want: high resolving power, large image circle, low distortion, and often very long working distance compared to some of other lenses we use.
There are other lenses, but the best lenses are the so called line scan lenses. These line scan lenses often have very high resolving power that it can match or beat a microscope objective. Here is a lens that I ran into and bought it after checking scratches, molds, etc. I was told that it is a 12K line scan lens. After some googling, I found its specs here. So, it looks like it has 7um resolving power, over 118mm working distance, and image circle of 86mm.
With spec sheet like that, though at around 1.25x, it can beat some of the best objectives like this Mitutoyo 1x objective (though Mitutoyo 1x is APO) that has resolving power of 11um, working distance of 11mm, etc.
Here are some of images taken with this lens.
The setup shown above consists of a subject platform with 3 degrees of freedom (roll, pitch, yaw) and a platform with one degree of freedom (forward and backward). All of these 4 degrees of freedom is to allow fine alignments to take an image.
This is how the lens is setup — lens, M44x1-to-M52x0.75, M42-to-M52 filter adapter, M42-to-Nikon AI, 3 segements of 28mm AI extension tubes, and finally, NEX-to-AI adapter so that a Sony AIII can be mounted.
Image shot with the setup above. Note the vignetting pointed by arrow (all 4 corners), this is caused by tubing, not by the lens. The lens can produce image circle of 86mm
The image above looks very sharp (though I wish the printing on it can be sharper). If you look at the large square grid, there is very little distortion. The 1 deg and 2 deg pattern at the left and right corner are sharp, too, indicating good corner to corner coverage.
Making It Telecentric
One of my goal is to make it telecentric and after playing with it, I finally got it very close to telecentric by adding some spacer and an aperture assembly. The spacer is about 32mm long between the back of the ProOptics lens and the aperture assembly. Important note, depending on your aperture assembly and how you adapted your lens, this distance might be different. Final result is very encouraging, yielding about 0.026% scaling factor between each image. The best part of this configuration is that you can remove or add additional extension tubes behind the aperture assembly without affecting telecentricity!
Making it telecentric by adding some spacer and an aperture assembly. Correction: it is M44x1 to M52x0.75 in the picture. It is custom made.
After making it telecentric, this is one of the result. This was stacked from 151 images for a 12mm deep stack. Scale from first to the last is about 1.045, so the scaling factor for between adjacent images is about 0.03%, a bit more than 0.02% for ideal measurement, but good enough. Without this telecentric modification, it is 0.052%.
100% crop of another image
This is a small part of peacock feather, the bottom part usually looks “fluffy” but under this lens, each fiber looks pretty stiff.
100% crop of above image
Conclusion: next time you visit an industrial surplus shop, if you see this kind of lens, particularly line scan lens and it is “cheap”, grab it!