spindle bearing fitment and retention

[PeterT]

Thanks. So dumb question. If we squeeze the bearing set via red arrows on outer race until they contact, they become set to their stated preload. Why do we need the ledge near green arrow at all? For example if if there was a 0.005" gap there, where else could the bearings migrate laterally? Aren't they fully constrained by their 1) their OD & ID faces inline along spindle & housing 2) the AC races & 3) the red arrow axial squeeze on outer races locking them axially?
What suppliers or catalogs are you looking up matched ACB sets for this application?

 

[TorontoBuilder]

Thanks. So dumb question. If we squeeze the bearing set via red arrows on outer race until they contact, they become set to their stated preload. Why do we need the ledge near green arrow at all? For example if if there was a 0.005" gap there, where else could the bearings migrate laterally? Aren't they fully constrained by their 1) their OD & ID faces inline along spindle & housing 2) the AC races & 3) the red arrow axial squeeze on outer races locking them axially?
What suppliers or catalogs are you looking up matched ACB sets for this application?

to equalize the load across the face of the bearing.

I'll use a torque wrench to tighten both ends so that in theory the same force will be on both the inner and outer races.

 

[RobinHood]

Are you using a sleeve to constrain the inner races (blue to green arrow) of the ACBs?

Putting pressure on only the inner race of the first ACB @ green arrow (with your nut on the pulley end) will not properly constrain the set-up.

You could use the nut on the pulley end to push the inner race of the deep groove ball bearing, spacer sleeve, inner races of the ACBs to the shoulder (at green arrow). This arrangement requires the outer race of the deep groove ball bearing to be floating.

 

[TorontoBuilder]

View attachment 56662

Are you using a sleeve to constrain the inner races (blue to green arrow) of the ACBs?

Putting pressure on only the inner race of the first ACB @ green arrow (with your nut on the pulley end) will not properly constrain the set-up.

You could use the nut on the pulley end to push the inner race of the deep groove ball bearing, spacer sleeve, inner races of the ACBs to the shoulder (at green arrow). This arrangement requires the outer race of the deep groove ball bearing to be floating.

damn I completely overlooked that. Thanks

 

[TorontoBuilder]

I'll copy the aciera spindle mcgyver posted... use a nut to retain that side as well. Prefit the bearings and nut then insert into spindle housing....

 

[RobinHood]

That will work; this is how the spindle on my INGAR surface grinder is set-up.

 

[PeterT]

So it seems to me we have come full circle? If we replace the single deep groove bearing of the magneto spindle design with a matched ACB pair set (2 blue bearings), isn't it the same principle?

Tell me if I have my (matched ACB) recipe checklist correct.

- 4 'constraint' features, 2 on outer race & 2 on inner race once pre-adjusted (red arrows). They can be 'fixed' constraints like a lip on the shaft or housing. Or semi-fixed dimensionally like a ring, shim or spacer tube acting on either inner or outer race.

- At least one constraint needs to be adjustable, for example the threaded nut that squeezes the matched set together.

- the constraints must be perfectly in vertical alignment with one another, otherwise there will be a gap somewhere resulting in un-equal face to face loading to the bearing set. For example we don't have to draw in the purple face to face arrows between the bearings as long as the 4 red outer constraints are in alignment. But if they were not in alignment, all bets are off.

- the bearing ID & OD surfaces are sized such that they have sufficient interference on the surfaces at operating temp to stay in position radially, but they can be slid or displaced via the horizontal adjustment mechanism

Now if the ACB's are NOT a factory MATCHED SET set, we have to do or check other things dimensionally to ensure the equivalent to ^above^ because we can no longer count on (my understanding) what matched ACB's conveniently provide us: once squeezed together they deliver the specified preload or axial lay for the intended purpose.

Comment or correct away as needed.

 

[wmetfab]

Im amazed at the level of talent here znd such interesting discussions and pushing the envelope of prescision and inspiration for us home hobbyist types

Im happy (and lucky) if I can do machining and get .010 accuracy
You guys: "Oh, lets just make a new high precision spindle for the milling machine!"
Me: stick welder cutting torch and grinder... "me cave man"

Best of the season to all of you on here.
May the dreams of your next machine come true in 2025

 

[TorontoBuilder]

So it seems to me we have come full circle? If we replace the single deep groove bearing of the magneto spindle design with a matched ACB pair set (2 blue bearings), isn't it the same principle?

Tell me if I have my (matched ACB) recipe checklist correct.

- 4 'constraint' features, 2 on outer race & 2 on inner race once pre-adjusted (red arrows). They can be 'fixed' constraints like a lip on the shaft or housing. Or semi-fixed dimensionally like a ring, shim or spacer tube acting on either inner or outer race.

- At least one constraint needs to be adjustable, for example the threaded nut that squeezes the matched set together.

- the constraints must be perfectly in vertical alignment with one another, otherwise there will be a gap somewhere resulting in un-equal face to face loading to the bearing set. For example we don't have to draw in the purple face to face arrows between the bearings as long as the 4 red outer constraints are in alignment. But if they were not in alignment, all bets are off.

- the bearing ID & OD surfaces are sized such that they have sufficient interference on the surfaces at operating temp to stay in position radially, but they can be slid or displaced via the horizontal adjustment mechanism

Now if the ACB's are NOT a factory MATCHED SET set, we have to do or check other things dimensionally to ensure the equivalent to ^above^ because we can no longer count on (my understanding) what matched ACB's conveniently provide us: once squeezed together they deliver the specified preload or axial lay for the intended purpose.

Comment or correct away as needed.

View attachment 56675View attachment 56678

Huh, I thought you were instructing me?

I'm trusting that when skg says I can buy universally matching bearings that are sold individually rather than matching, and put them together either face to face, back to back or even in tandem they will work for axial clearance and meet the claimed preload

 

[PeterT]

That's why I was wondering how far down the actual catalog/purchase path you got. If you have some references, fire away.

I can see matching meaning 1) they are ground to some predetermined spec so it doesn't matter if you buy 2 or 10, you could put a pair together from any in the batch & they will yield the preload tolerance = maybe what you are referring to 2) you must buy them as a pair only & they are intended to work together to deliver that spec. If they need to be replaced, they both need to be replaced. Now here I am talking single row ACB's & specifically the style that is meant to be squeezed together, faces in contact. I have also seen double row or other variations as a cartridge but I think they are spendy & specialized, outside our scope. I had this semi-figured out some time ago but have apparently lost my notes. I had a really good bearing guide that explained this a bit better but its MIA. I get the impression that different bearing manufacturers (or is it YouTubers) use the word 'match' kind of loose & free.

 

[TorontoBuilder]

That's why I was wondering how far down the actual catalog/purchase path you got. If you have some references, fire away.

I can see matching meaning 1) they are ground to some predetermined spec so it doesn't matter if you buy 2 or 10, you could put a pair together from any in the batch & they will yield the preload tolerance = maybe what you are referring to 2) you must buy them as a pair only & they are intended to work together to deliver that spec. If they need to be replaced, they both need to be replaced. Now here I am talking single row ACB's & specifically the style that is meant to be squeezed together, faces in contact. I have also seen double row or other variations as a cartridge but I think they are spendy & specialized, outside our scope. I had this semi-figured out some time ago but have apparently lost my notes. I had a really good bearing guide that explained this a bit better but its MIA. I get the impression that different bearing manufacturers (or is it YouTubers) use the word 'match' kind of loose & free.

Yes, when I referred to SKF's universally matchable single row AC bearings I precisely meant that the bearings are ground to a specification with a P4 class tolerance and it doesn't matter if you buy 1 or 100, you can put together any pair and get the claimed performance.

This is just what SKF means when they refer to their universally matchable single row AC bearings. Only a fraction of their catalog of super precision bearings are of the universally matchable type. They are designed and manufactured to a level of precision that they can be selected and used together as if manufactured as a matched set.

To confuse the matter further SKF makes super precision single row AC bearings to the same P4 tolerance that are NOT universally matchable, and they sell matched sets of these non-universally matchable single row bearings. I believe that the non-universally matchable bearings are culls that didn't make the cut because all their bearings are precisely measured and labeled as to deviation.

Now, having said that any two single row universally matchable bearings can be used together SKF still recommends that when purchasing single universally matchable angular contact ball bearings to make a set from existing stock, the purchaser or retailer should consult the bearing deviation data on the package exterior to select bearings with similar deviations and contact angles to make up a set.

The same applies to replacement. You can replace any single row within a set without having to replace the whole set

 

[TorontoBuilder]

Yes, when I referred to SKF's universally matchable single row AC bearings I precisely meant that the bearings are ground to a specification with a P4 class tolerance and it doesn't matter if you buy 1 or 100, you can put together any pair and get the claimed performance.

This is just what SKF means when they refer to their universally matchable single row AC bearings. Only a fraction of their catalog of super precision bearings are of the universally matchable type. They are designed and manufactured to a level of precision that they can be selected and used together as if manufactured as a matched set.

To confuse the matter further SKF makes super precision single row AC bearings to the same P4 tolerance that are NOT universally matchable, and they sell matched sets of these non-universally matchable single row bearings. I believe that the non-universally matchable bearings are culls that didn't make the cut because all their bearings are precisely measured and labeled as to deviation.

Now, having said that any two single row universally matchable bearings can be used together SKF still recommends that when purchasing single universally matchable angular contact ball bearings to make a set from existing stock, the purchaser or retailer should consult the bearing deviation data on the package exterior to select bearings with similar deviations and contact angles to make up a set.

The same applies to replacement. You can replace any single row within a set without having to replace the whole set

View attachment 56695

OH and you can also buy matched sets of universally matchable bearings too...

 

[PeterT]

A) I precisely meant that the bearings are ground to a specification with a P4 class tolerance and it doesn't matter if you buy 1 or 100, you can put together any pair and get the claimed performance.

B) To confuse the matter further SKF makes super precision single row AC bearings to the same P4 tolerance that are NOT universally matchable, and they sell matched sets of these non-universally matchable single row bearings.

Thanks. This explains my confusion. So have you worked your way to the $ aspect? If a person requires 2 matched in a specific orientation, and presumably would replace them as a pair if worn or suspect, is the cost much different between 2*A vs 1*B?

Can you give me a specific PN/link of something your are considering so I know where in the catalog to be focusing? Another issue I found is I (think I) see them listed in their catalog but cant find them in the real world. Maybe its a special order thing?

 

[TorontoBuilder]

Thanks. This explains my confusion. So have you worked your way to the $ aspect? If a person requires 2 matched in a specific orientation, and presumably would replace them as a pair if worn or suspect, is the cost much different between 2*A vs 1*B?

Can you give me a specific PN/link of something your are considering so I know where in the catalog to be focusing? Another issue I found is I (think I) see them listed in their catalog but cant find them in the real world. Maybe its a special order thing?

I'm using the SKF AC bearing selector tool to sort thru product options.

Selecting inner bore, then outer diameter, then matching to get a list of viable bearings. In 17mm x 26mm there are just 7 products that may work. 6 single row options and one double row option.

I'm thinking of 71803 ACDGB/P4 but it is a special order item thru all skf distributors. I expect them to cost more than anticipated and around $700CAD for a pair.

 

[Susquatch]

Hey @TorontoBuilder - I ran across this spindle assembly by accident and thought of you and your needs. I doubt this would meet the needs of this particular project, but it might be useful on some less demanding project. So I thought I should pass it on. I would never have expected to find a spindle assembly on Amazon.

Amazon.ca

 

[PeterT]

I'm using the SKF AC bearing selector tool to sort thru product options.

That's a good link thanks. Further down the page they discuss their bearing terminology, orientations, machine examples etc.

I found some of my prior links. In the Renzetti Tips on Precision Bearing Rebuild video, he mentions this (Timkin Fafnir) document as his recommended go-to reference. Turns out I had already downloaded the PDF but apparently filed it under a different folder in plain sight. That's what was percolating in the back of my brain. I never got to the point of price quotes. At the time I was interested to know beforehand what I might be looking at to 1) replace bearings on an otherwise OK surface grinder 2) replace/improve a Chinese motorized CNC spindle adapted to light tool post grinding duty 3) various (smaller) homebrew grinding spindles. IOW the applications were all over the map. No credit cards were harmed in this process to date.

https://www.timken.com/wp-content/uploads/2016/12/5918_Machine-Tool-Catalog.pdf

 

[PeterT]

BTW - this time I displayed & saved the transcript in a MS-Word file for future reference. Going to do that more in the future when I encounter a good video. (You can turn time stamp on/off in transcript panel). Its no tthe prettiest format but info is there

~4:00 he discusses the DB style where the inner race gap is factory ground/tuned
~6:00 shows the bearing set installed (clamped together flush)
So this is what I remember. Now if most all of these are in the $$$ range then its probably a moot point for me, but anyways, that's what was sticking in my brain as 'simpler' to implement in a design