4 ms·
Yeah. Those chips usually talk via SPI. There's interestingly a lot of options, including ways to boost quality. One sucky thing, is they use "dots per inch" wh
by kefka 10y ago
Yeah. Those chips usually talk via SPI. There's interestingly a lot of options, including ways to boost quality. One sucky thing, is they use "dots per inch" which is just completely stupid and boneheaded - seriously... Inches?!
But the ones I'm using are the ADNS-5030 for an interesting project with 3d printing. Its max poll rate is 1MHz, which is damned quick for a Atmel328p chip- it leaves only 16 instructions per SPI poll. The chip's default is 500 DPI (grr again inches), but can be changed simply by popping a "1" in a register.I'm sure different chips have much better qualities.
Also, you can pull out the dX and dY registers for relative travel since last sample. But you can also pull 256 times to get a 7bit grayscale picture of the mouse sensor! ADNS-5030 is a 16x16 7bit grayscale. And even the 'duino can pull that with relative ease and display in Processing. You can also display the fiduciary markers in which the optical analysis system determines as travel (how it gets the dX and dY).
But yeah, this article is kind worthless. Sorry. Anyways, videos can be a really bad way to pack little content in a lot of time and bandwidth. :/
- highd 10y agoYou're not trying to get absolute position from those sensors, are you? I spent far too long on that project. At least at the precision we needed, my conclusion was that you would want to acquire and process the images yourself - both to get a higher resolution/field-of-view, and to estimate the error associated with each position update (Kalman filtering or similar). You can also then "re-register" your position if your path travels over the same spot twice. Hopefully your application is less demanding!