Elphel Development Blog

Temporary diversion

Not much progress with the 10373 board in the last 3 weeks – just cleaned up debug wiring and prepared the board for the visit from SlimLogic Ltd. The software for the 373 camera is going to be based on OpenEmbedded and I would like to set up everything correctly before getting deep in the modifying the existent codebase.

Meanwhile we tested the redesigned sensor board (will work with the 373 camera too) to make it smaller. All the current sensor boards (but the ones with a large Kodak CCD) were the same size – 31.4mm x 31.4mm – it did not change since the times of old Zoran 1.3MPix sensor. Since then the sensors became smaller and less components were needed for the power conditioning too.

The sensor board size is especially important for the panoramic video were it is desirable to keep lenses as close together as possible to reduce the effect of the parallax. For the still imagery panoramic photos it is possible to rotate a camera precisely around the entrance pupil of the camera lens, but for the video (or even some imagery) it is not a solution.

It is possible to make the parallax really small (but still not zero) by using mirrors, but that requires rather large mirrors and virtually frameless rectangular windows to protect optics from the environment. And even with the frameless mirrors and windows the entrance pupils need to be apart by at least their size – otherwise there will be vignetting near the stitching borders. It is so as looking from the point on the border between two sectors both entrance pupils of the two lenses should be apparently visible, so they can not coincide. We’ll consider making such a minimal-parallax camera with the mirrors, but it is likely to be indoors-only with unprotected mirrors.

Making a simpler multi-lens camera requires that camera lenses (and sensors boards) are put closer together. With the redesigned sensor boards that are just 15 mm wide (28 mm long), it is possible to reduce the distance between the sensors more than twice from the original 31.4mm. Unfortunately there are other elements that can prevent putting the sensor boards that close to each other, namely – the C-mount lenses (CS-mount is even worse) that we currently use with the cameras. Even when using just 8 sensors around (10 or 12 of them make problem even worse) the C-mount 1-inch thread of the adjacent lenses overlap in CAD view, so sensors have to be moved farther apart.

Alternative solution would be to either to use lenses with custom-designed mount (that I would rather avoid as we try to keep our cameras universal) or use the other, smaller mount that is called “S-mount” or “Board Lens” or just “M12 Lens” – all those names apply to lenses with M12x0.5 metric thread that usually have fixed (same as “no”) iris and the focus is adjusted simply by rotating the lens in the thread. No iris and no focus ring is fine for this application – we need fully open aperture and camera can be focused once – with small sensor size the depth of field even for the high-res sensors is usually sufficient, otherwise you need to have motorized lenses as manually refocusing 8+ lenses can be a real hassle.

There still is a major problem with such lenses – as the thread is not used to the stop (as with C/CS-mount and most larger threaded lenses) there is always some play in the the thread – significantly larger than it is acceptable for the 5MPix sensor and the F#=2 – it should be no more than 0.05..0.1 angular degree, otherwise it will not be possible to focus all the field simultaneously – objects on ones side of the frame can be in focus while ones on the other side – out of focus. For lower power lenses (as are many with M12 mount) and/or lower resolution sensors the problem is smaller and just using lens with lock nut or just thread locker is sufficient, but for these applications we need to resolve several mechanical issues:

• eliminate play in the thread during focusing
• provide precise sensor tilt adjustment to compensate for mis-alignment of the lens optical axis and the sensors.

Additional problem is just to find the good quality lens that matches the sensor resolution – many M12 lenses have rather poor quality. Unfortunately most lens manufacturers do not provide MTF data in the published specs so we need to learn how to measure it ourselves. And as we did not sign any NDA with the lens manufacturers we are free to publish our in-house lens measurement results when done.
The compact precision sensor tilt adjustment is in machine shop now, I hope to have the first unit available for testing later this week. Meanwhile we are using regular M12-to-C-mount adapter to test lenses focusing the (small) area of interest separately. So far we tried two of the 3 lenses we planned for the evaluation and the results are encouraging – we will post them as soon as we will be sure in our measurements.

Andrey