3d printed trackball, metal bearing - 50mm ball - teensy 2.0 - pmw3360 sensor

[iso]

Being stuck inside, had to start another project, this is work in progress

Ball: On Amazon search "manual ball roller" I got the plastic version - Fairly hard to remove from its casing without scratching the ball.
Bearing:  On Amazon search "uxcell 6812ZZ"
To add - Sensor: https://www.tindie.com/products/jkicklighter/pmw3360-motion-sensor/
To add - Teensy: https://www.pjrc.com/store/teensy.html
Ball slides on 3 zirconia/ceramic bearing balls I got from a bearing I cracked open, steel will do just fine with plastic, if you add a glass/steel ball, switch the "sliders" to tungsten/ceramic.
The ball is trapped inside, wont fall if turned upside-down.
You can see 6 buttons but only 5 will be useable, left/mid/right click and navigation/dpi switcher.
I removed the bearing seals and washed off the lube with acetone, althought was very smooth and gave a nice feeling, wasnt spinning freely as I wanted so, had to be done.
Scrolling wheel got 24 holes at the bottom, where the magnets gonna be, when I edit videos I know a full turn will be 1 second if the footage is 24/fps. I plan on having multiple "setups" for the wheel, 30 holes and 60, those can be swapped on the fly.
I dont have a way to clamp down the sensor just yet, will update this when I do.
The firmware does not support flipping axis, or better said, I dont know how to doit so I will use a little program to flip x and y, this needs to be done as the sensor will be facing up, not down.
The base of the ball assembly slides to allow about 17 degree of movement, same with the hand support ball-head.
Button assembly is adjustable too, and the sensor case will allow 360 degree so I can move the buttons to the pinky if i want to or use it with my left hand.

Everything printed in medium gray PLA, I will add the OBJs soon so anyone can print their own.

I attached a PDF with hookup instructions (old but fairly easy to follow) and link to the firmware.
If anyone can help with modifying the firmware and flip axis would be greatly appreciated, pleae upload the HEX so I can test it right away.

[theKM]

nice! you're a man on a mission!

I've been looking at ways of dropping the use of Logitech for something that will let me use a billiard ball, but to do that I need to make my own bearings. I found this page of a dude working on it: http://www.madox.net/blog/2017/06/17/tribology-diy-trackball-2/

...and just followed his findings to make some from 3mm silicon nitride balls, they really do roll super nice, as smooth as logitech's stock bearings.

info for what it's worth

[iso]

Well... I`ve put all the electronics in the other day and I didnt like it.

Re-modelled 90% of it, the only thing that didnt change is the sensor case and magnet wheel/scroll

Stood too high from the desk and my elbow start to hurt after 20 min or so
All buttons were set for thumb, makes it very hard to use combinations
Magnet density was wrong, or better said because of the strength, the reed switches would close the circuit no matter where the magnets were, I`ll get the 1mm so I have better control over triggering distance.

I hope to find the time next week to add the electronics and I`ll update with the OBJs

[iso]

Got it working, all files attached.
The 3 x 5 grid will have programmed specific shortcuts, I will be using that with 3ds max and Virtual dub
Left to do:
1. Add another Teensy and have some spare Cherry ML`s as buttons
2. Combine plate, teensy case and Cherry ML`s plastics in one big piece

[nevin]

just seeing this. sent you PM

[Renanssf11]

Magnet density was wrong, or better said because of the strength, the reed switches would close the circuit no matter where the magnets were, I`ll get the 1mm so I have better control over triggering distance. text to speech

[iso]

Yes, depending on magnets strenght the density will vary. If you dont want to buy new magnets just lower the magnetism, heat them up or break them in smaller pieces

Thanks for mentioning, I forgot.