3
\$\begingroup\$

I recently learned that modern computer mice take tiny pictures many times a second and compares them to understand in what direction it is heading. And that the little light at the bottom is to create shadows for uniform colored, textured surfaces to get a decent image.

I was wondering, is it possible to extend how far away the image sensor can pickup on a surface? Right now, if I hover my mouse over about 0.1" above the surface, I still get movement of my cursor, but beyond it fails to detect. If I were to put a strong magnification lens right after the sensor, would I be able to hover the mouse higher?

\$\endgroup\$
4
  • \$\begingroup\$ Not easily. I do not think you can do it with retroactive modifications or while maintaining eye safety. \$\endgroup\$
    – DKNguyen
    Nov 21, 2019 at 3:24
  • \$\begingroup\$ Eye safety? I thought it was just a red LED at low power. That could still damage the eyes? \$\endgroup\$
    – joe
    Nov 21, 2019 at 3:25
  • \$\begingroup\$ If you magnify it or make it real bright with a different LED. But the main reason for the distance is the emitter shines at an angle to cast shadow and the receiver points straight down. It forms a triangle and move too far then the surface illuminated isnt in front of the sensor anymore. \$\endgroup\$
    – DKNguyen
    Nov 21, 2019 at 3:27
  • \$\begingroup\$ There are known examples of using the motion correlation IC from optical mice in simple DIY robotic applications. \$\endgroup\$ Aug 16, 2020 at 7:59

1 Answer 1

5
\$\begingroup\$

Yes, what you suggest is possible. (Congratulations on your cogent thought process).

The mouse navigation compares "features" from one image frame to the next, and calculates how much X-movement and Y-movement has transpired. If no image features can be found, then no movement can be calculated. An out-of-focus image contains no features: a mouse's lens is extremely near-sighted.

A lens of a different focal length can focus further away. For example, a fixed focal-length camera usually extends focus to infinity. Doing this converts a mouse to a camera (See my profile photo; it was done with this modified mouse):
mouse camera

Having a sharp focus very far away likely allows features to be found, and XY movements to be calculated. Still, pointing it at a featureless sky would confuse the internal calculator, and yield no motion-detection.
However, the mouse is now useless as a mouse on a mouse-pad because features very close are out-of-focus.


Notice that the mouse's LED light source has been removed. Daylight is sufficient to create an image. The internal chip can accommodate a fairly wide range of light levels (but not as wide as your eye).
For a mouse on a pad, the LED is needed for illumination, since the mouse itself shields room light from the small space between mouse and pad.


The plastic lens included with this mouse could not be moved close enough to the internal chip to allow infinity-focus. It was replaced with a 4.5mm focal length plano-convex glass lens. The lens quality need not be high since pixel-count is so small.
Lens is mounted in a machined-brass housing that press-fit over the ADNS-2610 chip's cover. The tiny cover hole (to allow light inside) was drilled out to about 3mm dia. The cover has a 5.6mm diameter step over which the brass housing press-fit. The cover can be pried off the chip with care.
Interface to this 8-pin chip is serial I2C. This is an older chip that requires +5V Vdd. Avago data sheet is very well-done.
ADNS-2610 with cover removed Have tested its XY scanning ability with infinity focus, but not used it as a mouse this way: you may find that much gesture movement yields small cursor motion on your screen.
More modern chips provide more pixels and pixel download rates are fast enough to allow video frame rates when used as a camera. Getting data sheets for them and getting small quantities for experimenting seems to be a problem.

\$\endgroup\$
5
  • \$\begingroup\$ The fact that your picture was taken by a mouse camera is awesome! Where did you procure your lens? I would like to experiment with this as well. \$\endgroup\$
    – joe
    Nov 21, 2019 at 17:33
  • \$\begingroup\$ @joe have edited with modified lens mount. Enjoy experimenting...this is a very interesting chip with many applications beyond mousing. \$\endgroup\$
    – glen_geek
    Nov 21, 2019 at 18:43
  • \$\begingroup\$ Thank you! I think once I get lens, I'll think of something I can do with this. \$\endgroup\$
    – joe
    Nov 21, 2019 at 22:53
  • \$\begingroup\$ By the way, where did you get your lens? I am going on sites like Edmund Optics and they seem to be expensive! \$\endgroup\$
    – joe
    Nov 22, 2019 at 1:23
  • \$\begingroup\$ Apologies, the lens in that mount was likely 4.5mm focal length. Plano-convex is OK too. Another lens tried came from a $2 laser pen. Many of these have a screw mount for focussing - very useful. A larger 6mm F.L. lens had a "tele-photo" effect on an image. This chip has an internal 8-bit register SQUAL providing a count of "features detected"...useful to adjust lens focus. \$\endgroup\$
    – glen_geek
    Nov 22, 2019 at 14:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.