Using a locknut may help reduce fretting wear from those micro inches of movement. A disk spring or similar could also provide a similar effect with less additional cost.

What kind of bearings? If they are radial ball bearings I'm surprised you only have micro inches of movement.

What keeps the shaft from sliding out to the left - the coupling?

Hi Brian.

My information is wrong then, sorry I looked it and you are right, the play will be much more than that.
they are deep groove ball bearings.

Anyways, without some amount of preload, those sleeves and spacers will vibrate and move around on the shaft. If lubrication is poor, there will be some amount of fretting wear/corrosion. You may find this to be tolerable. If not, a nut or spring can help lock everything together.

Your assembly with retaining rings on both sides in the housing will have quite a bit of axial freedom (much more than the bearing clearances). This appears to be a chain conveyor so the axial position of the roller is not likely sensitive. A worse cases is an oscillating motion that wears out the bearing housing but this is all unrelated to your original question.

@TugboatEng

Great post! Yes, it's a chain conveyor, but I did not realize the issue with the retaining rings, thanks for pointing that out.

If I get rid of the locating rings and mirror the shoulder on bearing housings (so the bearing will be located on one side by the spacer/collar and the other side by the shoulder in the housing - outer ring), this should eliminate the assembly moving left and right because of the clearance in the retaining rings groove, correct? But I guess it makes fitting things a bit more challenging?

I wouldn't call the retaining ring clearance an issue without more details. As a chain conveyor, it should be fairly forgiving to the axial position of the roller. I only see it being a problem if the shaft oscillates side to side during operation as this can cause wear to the housing. Inserting the housing with a wear resistant bushing (some type of bronze) can solve this problem that may not even exist.

I picture a conveyor belt subjects its bearings to primarily radial loads, and the shaft stand still.
This calls for a line to line or slight interference fit between the shaft and bearing inner race, lest the inner race/ring continually creep ( "spin" ) on the shaft and wear everybody out before rotating 1 million times.
Axial inner race clamping can not be expected to prevent that creeping.
Check out page 69 of this PDF. -
https://www.skf.com/binaries/pub12/Images/0901d196...

Check your bearing catalog for guidelines and rules for proper bearing shaft and housing fits, based on the nature of the bearing loads.

The accommodation of axial thermal expansion is often set by holding one bearing outer race "fixed" axially, and giving the other "free/floating" bearing a few .001" slip fit in its housing, and plenty of axial clearance.
There are other ways, but they typically require fancier machining and fitting at assembly.
Long shot is having the free bearing mounted on a bulkhead or frame that is axially very flexible to limit the axial thrust load from thermal expansion.

Tmoose, I think the concern is with the spacers and sleeves moving. Clamping the inner race against the spacers and sleeves will prevent their motion.

https://www.schaeffler.com/remotemedien/media/_sha...

That's a really handy guide but it covers a duty/scale that is way above and beyond OP's case as it's using tapered bore bearings and labyrinth seals.

Here is another FAG design guide with a broader range of examples.

If I understand the OP's device ( kind of doubtful ) the shaft rotates, kind of like an electric motor.
As a start if the shaft rotates and the load is one directional a slip fit on the shaft on the shaft is doomed even when clamped by bearing nuts securing the inner rings.
Getting familiar with bearing catalogs and talking to a bearing OEM application engineer will confirm or deny this.
I venture the application engineer will ask many questions after receiving info in the OP's OP.

No, your link only covered tapered bore bearings in regards to conveyors.

Hi Tmoose, thanks for the link, the practical examples in that guide are pretty helpful.
As TugboatEng said, I'm using deep groove bearing in this application and overall simple arrangements.

I read about bearing settings/arrangements that I wasn't very well versed in, the past few days and now I have a much better grasp of the topic.

Thanks guys!