Might need bearing lesson, tennis mach ver 2

[chipmaker4130]

Don't bother looking up the ABEC7 bearings. Not appropriate for a ball chucker. Ordinary deep-groove ball bearings can also be axially preloaded, best done with spring washers to avoid over tightening. Deep groove bearings (ordinary, everyday bearings) all have an axial load spec. you can use to make your design. There is a difference in same-size bearings. Some have a minimum complement of balls, some are nearly full. You'd have to ask your bearing supply house which is which.

Unless there is a compelling reason to use cantilever shafts, the ideal solution is to go back to bearings on each end.

[Rich Carlstedt]

Mike
I am not sure how your blocks are machined for the bearings.
I assume it is a straight bore ?
As such , You can preload the setup with three thin wall sleeves
make a sleeve that fits inside the block bore and will be a spacer between the Outer races of the bearing
For your bearings, a few thousanths under the 32 mm diameter will work and and maybe .050 to .100" thick.
then make two thin sleeves for the shaft , slip fit and also only .050 to .100" thick and the lengths to be as follows.
Measure from the roller to the bearing for one, and measure the bearing to the pulley and add a few thousandths.
Install them and when you bring in the pulley, push it hard against the roller and then tighten the set screw.
Your hand force axially will load the bearings before locking the position.
Getting deep groove bearings will help your operation as well

You most likely got ABEC -1 bearings..look for ABEC - 3 or 5
Rich

[Mike Amick]

Ok, here is where I am at right now. I hammered out the current bearings (they were epoxied). Cleaned up the bearing seats
Made a temp axle out of some stainless that is a better fit just to test it. So as it sits, the bearings are just sitting inside their
area nothing holding them in.

It spins real nice. So all I have to do is secure the bearings without moving them.

And now, reading Rich's post about creating a preload. The bearing block is not bored all the way through, I just drilled
a hole bigger than the axle, then i bore a ledge for the beraing to sit on. Your preload instructions are way more involved
than I was expecting but will read it a copy of more times. .


So here are the bearings just sitting there both sides of the black.running smoothly. Now I just have to glue or loctite the bearingws
into bearing block with out messing up the allignment.

[Bented]

Don't bother looking up the ABEC7 bearings. Not appropriate for a ball chucker.

You have ruined the joke, I only use ABEC 12 bearings when building CNC garden rakes.

[A.K. Boomer]

Unless there is a compelling reason to use cantilever shafts, the ideal solution is to go back to bearings on each end.

Id have to agree with this, as beefy as he designed everything I still wonder about that channel-lock strut flexing under pressure and the repetitive shock loading of the bearings trying to cock even with loctite, Im sure it would prove to be "enlightening" watching a slow-mo vid of a ball getting squeezed and then launched with this machine...

I did look at a popular brand name ( the Lobster) to try and get an idea of how they build these internally --- even went into an owners manual but they did not have any pic with the covers off of it....

[Mcgyver]

1) don't used $1.50 bearings. For standard deep groove ball bearings the good ones aren't much money, NSK, SKF and so on. The good ones will be ABEC 4 or 5.
2) the general rule, interference fit on the rotating part , and close sliding fit on the non rotating part
3) the tenth tolerances sometimes required for that makes bearing fits demanding work, must hold your tongue just so
4) Don't use an adhesive, makes it a pain to replace the bearings
5) If you screw up on the fits, ok, use an adhesive, but make it loctite low strength cylindrical compound - the one specifically for bearings. You'll have fighting chance of getting it apart
6) If you screw up, preferably make it on the bore not the shaft, you want the shaft the right press fit so it runs true
7) I can't see why a preload is needed here, looks like straight radial loads and no reason for axial accuracy
8) retaining the outer races with an internal shoulder and then a nut or snap right is SOP and works well
9) it sounds like you have two bearings per shaft which should work fine

[Mike Amick]

That's a really nice summation Mcgyver, thanks.

Between tightening up my tolerances and understanding bearings a bit more, I think I am getting it.

Looks like a DRO is in the near future for my lathe, it's really hard (for me) to sneak up on those
sizes with the dials, goes from almost there to whoops way to much to often.

[Mcgyver]

That's a really nice summation Mcgyver, thanks.

Between tightening up my tolerances and understanding bearings a bit more, I think I am getting it.

Looks like a DRO is in the near future for my lathe, it's really hard (for me) to sneak up on those
sizes with the dials, goes from almost there to whoops way to much to often.

thanks and you are welcome. It is challenging work but doable. Agreed, its about impossible with a dial. I'd don't have a DRO on the lathe, but a tenths indicator on the cross slide works well for these more demanding jobs. The worse are high precision, smallish angular contact bearings with a tolerance range of about a tenth and half. Regular deep groove bearings are more forgiving. Bearing catalogues are slightly less painful than pins in the eye, but they do have all the info there and tolerances do vary by bearing size

[chipmaker4130]

. . .7) I can't see why a preload is needed here, looks like straight radial loads and no reason for axial accuracy. . .

Its not the axial accuracy he's trying to improve. Its the slop enabled by two ordinary bearings close together in non-precision fits, and the high-speed unbalanced rotating load applied some distance from the bearing. Pre-loading axially will indeed take most bearing-enabled wobble out, provided the inner and outer fits are improved with mounting compound.