When you say you're molding your own ball are you referring to the actual ball? If so, why?
Mainly to get the right diameter (75mm), and to have it empty inside so it's light and has relatively low inertia. Also, I have full control over surface, finish, color, etc.
If I make it, it'll mean I have full independence regarding the parameters of the trackball.
It's also very hard to get it right, and that makes it super attractive
.
I've done this once with a 3d print directly, printed from 2 parts, glued, polished. It worked quite well for a prototype, so now I want to make a real functional thing.
Have you tried the BTUs with a ball? How smooth is the movement? What model have you bought?
Used an old trackball I had on them, but they were not angled properly, etc. Still, the results are very promising. There's nothing I can see go wrong at this point that can't be countered somehow.
I've got these:
http://www.boschrexroth.com/en/us/products/product-groups/goto-products/goto-linear-motion/ball-transfer-units/index 12mm no collar.
Almost no force required to get them rolling.
They feel very smooth without load.
They feel harsher under load, like when you press them against a table and try rolling them around - they oscillate a bit, but still near zero friction. Feel of these is identical to most not-hyper-precise linear motion systems that use ball recirculation.
The only thing I'm worried about now is that the oscillations will transfer to the ball, and make it feel harsh. If that happens, and will prove unpleasant, I'll try dampening it somehow.
Rubberized ball much? Anyone likes Thinkpads? :]
My printed BTUs are not providing a satisfying movement. The ones I printed yesterday turned up ok, the nitride ball could spin freely inside, but the friction is still too high. Probably a precision housing and better materials are required.
In the mean time I bought a m570 trackball and it is smoother but the ball is way to small and light for my novice taste. I only bought it to have something to benchmark against.
Well, it's going to be pretty hard to make a ball spin in a 3d printed element. The steel ball will require some momentum to start rotating, and sliding against the surface, regardless what plastic is used to house it (even if it's super slippy, it's gonna accumulate dirt, for example. Remember ball mice?)
This is exactly why BTUs were made. They are designed to go over this particular problem by simply rolling two elements one against the other. They should be OK as long as you don't put too much sand in them, and parts of them won't start rusting. They just roll over stuff.
When they stop, you replace them.
You could try something like this, with plenty of lube:
Oh, hey, I have an idea.
LET'S CALL THIS THE IRON MAIDEN TRANSFER UNIT
Should be printable if prepared right. (assembled from 2 halves probably)
But from my POV it's best to go with a ready-made element. Those parts are complex enough by themselves.
The only thing that could actually be better than BTUs are magnetic bearings.
They would be several orders of magnitude better, and you wouldn't have to deal with all the electronics usually used to stabilize those systems, because the speeds and momentums on this would be very low, and would not require superb precision (by magnetic bearing standards. Would still be damn precise.)
Though you want to shield user limbs from long term magnetic field exposure due to health concerns.