Tech/science applications of dual-band blue + NIR filter?

[enricosavazzi]

I chanced upon a very small number of prototypes of a Schneider lens model that never made it into production and is undocumented. Testing is difficult because of odd lens mount and filter threads, and short registration distance (no-no on DSLRs, difficult but possible on mirrorless). These lenses have 26 mm FL, no geometric distortion, transmit VIS, NIR and at least some UVA, are built and weigh like high-end machine-vision lenses, cover APS-C but not completely full-frame, but I know nothing more. I need custom adapters made before I can really test them.

 

All have a dual-band dielectric filter mounted at their rear, that transmits blue/cyan VIS and 750-900 nm NIR. I am aware that this type of filter is used for automated measurement of soil leaf cover. Are there other technical or scientific applications for imaging with this type of dual-band filter? It might also be used for night-time surveillance without artificial light sources, but I am not positive.

[colinbm]

I am using a Sigma fp that is full frame & Leica-L mount  & the registration is 20mm.
If they could be mounted on a helicoid & used for macro.
What focal length are they ?

 

[enricosavazzi]

It is a 26 mm f/5.6, so in practice a wideangle on full-frame. On full-frame it gives black corners when focused at infinity and mounted in normal orientation (i.e. not reversed). The dark corners are quite small when focused at 30 cm from the subject.

 

So far, my tests can only be preliminary and improvised. One test shows a flat field of focus with aperture fully open, at a distance of a little less than one meter from the subject (a wallpaper). So this lens seems designed to work optimally at some point between the close-up range and infinity. I can see in the hand-held test images dust specks down to a diameter of about 2 pixels on a 24 Mpixel full-spectrum Sony A7R II without any filter. The room illumination is a mixture of white LED and daylight with plenty of NIR, so the amount of focus shift between VIS and NIR seems too low to cause any blurring (as expected from a lens originally equipped with a dual-bandpass blue and NIR filter).

 

When reversed on zero extension it gives approximately 1.2x and strong curvature of field. I see no potential at this magnification, too far from the optimum focus range of this lens. There might be some potential when reversed at higher magnification, if the curvature of field becomes lower (as I expect). The optical scheme is strongly asymmetric and the pupil ratio 2.54. This means that, reversed, the lens speed becomes a nominal f/2.2, therefore diffraction when reversed at high magnification will be low enough not to be a problem. The working distance should be between 15 and 20 mm, which is short but not impossible. The rear filter mount works like a lens shade when the lens is reversed, and is integral with the lens barrel.

 

In normal orientation the lens must sit quite close to the sensor to focus at infinity. The registration distance from the mounting flange seems to be very roughly 25 mm and the lens protrudes 10 mm behind the flange, so only around 15 mm of clearance to the sensor.

 

The lens attachment is an odd M42 x0.5, which is neither the standard M42 nor T. It might be meant for focusing by rotating the lens within a custom adapter. This thread is quite short (5 mm), so if this is true the lens is designed for a restricted focus range. Other machine vision lenses in my possession have much longer focusing threads, so this is another odd feature of this lens model.

[colinbm]

Thanks Enrico, a bit weird.