Angular contact bearing quick question - Practical Machinist
Aug. 18, 2025
Angular contact bearing quick question - Practical Machinist
Hi, i have a quick question concerning preloading angular contact bearings.
say you had a lathe spindle fitted with a roller bearing at the pulley end and a pair of AC bearings at the spindle end arranged in DB/ back to back fashion, with a preload nut threaded on the spindle to push the inner rings of the ACs together to add preload.
pressumably as the spindle is rotating, they should be fitted with an interferance fit to the spindle, now my question is will the inner rings be able to move together when the nut is tightened, given the tight fit? second question - could the same preload be acheived by having a fixed collar infront of the front AC in place of the nut, and have the preload nut on the read end of the spindle behind the roller, pulling the spindle back, and the collar against the ACs ?
any info or advice would be greatly appreciated.
cheers,
josh. I had a similar situation with my lathe, but only 2 tapered roller bearings. Both inner races were tight on the shaft making the preload adjustment kind of impossible. I ended up making a sliding fit under the rear bearing inner race. So far I do not see any negative effects.
As to your second question, you can put the preload nut at the end of the spindle behind the roller bearing. In this case you are not pulling the spindle back, just putting an axial force on the roller bearing and rear AC bearing inner races. You will have to use a tubular spacer between roller and rear AC bearings. I do not like this method if there is any significant distance between roller and AC bearings due to unpredictable thermal expansion of the spindle and gear casing. If I'm understanding correctly
Two cone bearings back to back you want the inner or rear race to be against a shoulder, and you only need to move the outer race to preload both cones which are fixed to the shaft. Any bearing at the end of the tube can be used but you don't want to add thrust there unless it is a thrust or angular contact bearing, ball or roller. Hi thanks for your reply, surely if i put the nut behind the roller bearing, tightening would put pressure on the roller and the front of the angular contact inner race (nearest the chuck) via the collar infront of the forward most AC?
would it be preferable to have the rear nut tightening against the inner ring of the roller bearing or against the housing so as to leave the roller free ?
there isnt too far between the angulars and the roller, its a 25mm OD spindle with 70mm between roller and rear AC.
many thanks,
josh. I see. Could be achieved using a radial contact ball bearing at the rear or a thrust bearing at the rear and take up device beyond that, can't apply thrust load on a needle roller type bearing.
Depending on what work your doing the chuck bearings will be loaded once working with a drill, or live center but not with turning or parting, both tapered bearings need preload to survive and remain accurate.
You could swap out the rear roller and add in one that can take thrust either a thrust rated ball bearing, cone race setup or even thrust bearing with washers each side. Some lathes are made this way already. So It is a small lathe spindle. I understand the desire to adjust AC bearings preload at the rear end - there is space limitation and accessibility issue. You can do it with the rear nut pushing on the rear bearing inner ring, then spacer and finally the rear AC bearing inner race.
But 25 mm diameter under AC bearings is rather small unless it is only for very small parts using collets directly in the spindle bore (no 3-jaw or 4-jaw chucks). Even small chucks have significant length, which would make a 25 mm diameter spindle non-rigid.
When the two bearings are placed back-to-back, with both inner and outer races touching, the preload is correct. It is built in by the bearing manufacturer.
On a typical Bridgeport-type milling machine spindle, the two bearings are back-to-back, but spaced apart by two identical length spacers. One tubular spacer for the inner races one tubular spacer for the outer races, both identical length. It's the same as putting the bearings back-to-back without spacers.
You warm the bearings in oil, quickly fit them to the spindle and tighten the spindle nut fully. This assembly then slides into the housing, not a tight fit in this case. The cap nut contacts the outer race and holds the whole assembly in place. The preload is automatically set. No adjustment, just do it up tight.
Are angular contact bearings suitable for a lathe? A small one perhaps? I think lathes use tapered roller bearings at the chuck end, albeit high precision type. For the original poster:
Your proposed configuration is how hardinge often sets up their lathe spindles. An angular contact pair in front, and a single redial bearing in the left hand side.
The AC contact pair have their inner races constrained to be constrained in contact between the front spindle face and a nut that threads in from the rear. The outers are constrained in contact between an outer clamp ring that bolts to the front of the headsctock casting, and a bore inside the front of that casting.
The installation is such that the inners can be brought into contact by tightening the ring nut even if the fit is quite tight on the spindle as no force is applied thhrough any bearing balls at all.
The left bearing you describe as a roller bearing, in hardinges cas is simply a radial ball bearing. The fine point is a steel spindle and cast iron headstock casting have different thermal expansions. Hardinge accounts for this by allowing the outer race of the left bearing to be constrained radially in a cylindrcal bore in the casting, with no axial constraint. So allowed to float.
In the HLVH all of the three barings have their inners clamped up via one single nut, with a long precision "preload cylinder" to space the left bearing at the left side of the headstock. This makes the asssembly of the spindle as a single unit into the headstock a bit tricky. Other hardinge lathe headstocks use simpler assemblies but the AC pair on the right and a single radial on the left is a common design. In my 90's belt-drive Mazak cnc lathe spindles, there is a large tapered roller bearing at the rear, and the front has at least a pair (sometimes 3) of angular contact bearings along with another large tapered roller bearing.
Why the setup on the chuck side of the spindle? The TRB's preload and installation sets only the radial play in the spindle, whereas the AC bearings set only the axial play.
What's interesting is the spindle is designed so that the outside diameter of the AC bearings actually have clearance in the housing bore---they want the AC bearings to only carry axial thrust loads. This clearance also gives the AC bearings room to expand from heat. The adjacent front TRB carries all axial loads.
Newer Mazak spindles are integral, and thus I believe they don't use a roller bearing at the rear anymore. Older spindles had the roller at the rear, designed to carry the side loading of the belts.
And as said, this rear bearing doesn't constrain the spindle shaft from growing in length from heat.
ToolCat In my 12" lathe the spindle bearing design is very simple, but effective. It has only 2 precision tapered roller bearings. The outer races are constrained in the headstock casting. The inner race of the front bearing is a tight fit on the spindle. The inner race of the rear bearing is axially adjustable by a nut behind this bearing. This is how you can adjust the preload.
Tapered roller bearings are quite rigid by design and can carry both radial and axial loads. The distance between both bearings is only around 8-9", so thermal expansion does not play a big role. For a little difference in thermal expansion of the spindle and the headstock casting the casting walls can cave in or out a little to compensate for it. This is a crude and cheap design, but it works in the short spindles. Need a ton more info to give any specific advice. But to answer some of your questions
Yes, tightening a front shaft nut will preload the front pair overcoming the interference--which is small maybe 3-6 micron depending on bearing size.
No, the same force isnt applied in your second example, unless they are in tandem and you are pulling against something that works in an axial direction.. so not a cylindrical roller like you are proposing. Unless you mean a cup/cone style roller which are not really used in spindles as the tolerancing isnt usually good enough. There are exceptions, but they are very pricey. Plus its not a good application since all you are trying to do is control forces of belt tension, a CRB is better, cheaper, and tighter.
What is the size of the shaft/housing you are trying to make?
What loads will be applied to the spindle?
Speed you want to run?
Lubrication (i assume grease?)
Budget?
Drive? ex belt, coupling, motorized, gear? Dismantling and rebuilding a hardinge spindle. The original angular contact bearing was so worn it had negative preload - aka radial motion. The original bearings are in place in the first photo, the single radial rear bearing has been removed.
Second photo shows headstock casting, and one of the outer race clamp rings. Note the rim that's sized to contact the outer race. Note the spot in the center of the cylindrical bore, where the rear bearing sits. The oil gasket thickness is critical here.
New bearings installed on spindle, old ones in background. These were about $300 or so, 20 years ago. Note the match marks lining up on the outer races.
Hardinge went back and forth beween oil and grease lubrication. These are oil lubed.
say you had a lathe spindle fitted with a roller bearing at the pulley end and a pair of AC bearings at the spindle end arranged in DB/ back to back fashion, with a preload nut threaded on the spindle to push the inner rings of the ACs together to add preload.
pressumably as the spindle is rotating, they should be fitted with an interferance fit to the spindle, now my question is will the inner rings be able to move together when the nut is tightened, given the tight fit? second question - could the same preload be acheived by having a fixed collar infront of the front AC in place of the nut, and have the preload nut on the read end of the spindle behind the roller, pulling the spindle back, and the collar against the ACs ?
any info or advice would be greatly appreciated.
cheers,
josh. I had a similar situation with my lathe, but only 2 tapered roller bearings. Both inner races were tight on the shaft making the preload adjustment kind of impossible. I ended up making a sliding fit under the rear bearing inner race. So far I do not see any negative effects.
As to your second question, you can put the preload nut at the end of the spindle behind the roller bearing. In this case you are not pulling the spindle back, just putting an axial force on the roller bearing and rear AC bearing inner races. You will have to use a tubular spacer between roller and rear AC bearings. I do not like this method if there is any significant distance between roller and AC bearings due to unpredictable thermal expansion of the spindle and gear casing. If I'm understanding correctly
Two cone bearings back to back you want the inner or rear race to be against a shoulder, and you only need to move the outer race to preload both cones which are fixed to the shaft. Any bearing at the end of the tube can be used but you don't want to add thrust there unless it is a thrust or angular contact bearing, ball or roller. Hi thanks for your reply, surely if i put the nut behind the roller bearing, tightening would put pressure on the roller and the front of the angular contact inner race (nearest the chuck) via the collar infront of the forward most AC?
would it be preferable to have the rear nut tightening against the inner ring of the roller bearing or against the housing so as to leave the roller free ?
there isnt too far between the angulars and the roller, its a 25mm OD spindle with 70mm between roller and rear AC.
many thanks,
josh. I see. Could be achieved using a radial contact ball bearing at the rear or a thrust bearing at the rear and take up device beyond that, can't apply thrust load on a needle roller type bearing.
Depending on what work your doing the chuck bearings will be loaded once working with a drill, or live center but not with turning or parting, both tapered bearings need preload to survive and remain accurate.
You could swap out the rear roller and add in one that can take thrust either a thrust rated ball bearing, cone race setup or even thrust bearing with washers each side. Some lathes are made this way already. So It is a small lathe spindle. I understand the desire to adjust AC bearings preload at the rear end - there is space limitation and accessibility issue. You can do it with the rear nut pushing on the rear bearing inner ring, then spacer and finally the rear AC bearing inner race.
But 25 mm diameter under AC bearings is rather small unless it is only for very small parts using collets directly in the spindle bore (no 3-jaw or 4-jaw chucks). Even small chucks have significant length, which would make a 25 mm diameter spindle non-rigid.
im toying with the idea of building a small lathe, and the rear bearing furthest from the chuck is just a standard roller bearing, with 2 angular contact bearings back to back at the chuck end.The best bet is to put a spacer between the ac's. You can then set the preload off the machine(by grinding the spacer) then assemble them as a preloaded pair and locking them in the housing . For expansion a straight roller at the rear. I am no expert - but the angular contact bearings I have used have no adjustable preload.
When the two bearings are placed back-to-back, with both inner and outer races touching, the preload is correct. It is built in by the bearing manufacturer.
On a typical Bridgeport-type milling machine spindle, the two bearings are back-to-back, but spaced apart by two identical length spacers. One tubular spacer for the inner races one tubular spacer for the outer races, both identical length. It's the same as putting the bearings back-to-back without spacers.
You warm the bearings in oil, quickly fit them to the spindle and tighten the spindle nut fully. This assembly then slides into the housing, not a tight fit in this case. The cap nut contacts the outer race and holds the whole assembly in place. The preload is automatically set. No adjustment, just do it up tight.
Are angular contact bearings suitable for a lathe? A small one perhaps? I think lathes use tapered roller bearings at the chuck end, albeit high precision type. For the original poster:
Your proposed configuration is how hardinge often sets up their lathe spindles. An angular contact pair in front, and a single redial bearing in the left hand side.
The AC contact pair have their inner races constrained to be constrained in contact between the front spindle face and a nut that threads in from the rear. The outers are constrained in contact between an outer clamp ring that bolts to the front of the headsctock casting, and a bore inside the front of that casting.
The installation is such that the inners can be brought into contact by tightening the ring nut even if the fit is quite tight on the spindle as no force is applied thhrough any bearing balls at all.
The left bearing you describe as a roller bearing, in hardinges cas is simply a radial ball bearing. The fine point is a steel spindle and cast iron headstock casting have different thermal expansions. Hardinge accounts for this by allowing the outer race of the left bearing to be constrained radially in a cylindrcal bore in the casting, with no axial constraint. So allowed to float.
In the HLVH all of the three barings have their inners clamped up via one single nut, with a long precision "preload cylinder" to space the left bearing at the left side of the headstock. This makes the asssembly of the spindle as a single unit into the headstock a bit tricky. Other hardinge lathe headstocks use simpler assemblies but the AC pair on the right and a single radial on the left is a common design. In my 90's belt-drive Mazak cnc lathe spindles, there is a large tapered roller bearing at the rear, and the front has at least a pair (sometimes 3) of angular contact bearings along with another large tapered roller bearing.
Why the setup on the chuck side of the spindle? The TRB's preload and installation sets only the radial play in the spindle, whereas the AC bearings set only the axial play.
What's interesting is the spindle is designed so that the outside diameter of the AC bearings actually have clearance in the housing bore---they want the AC bearings to only carry axial thrust loads. This clearance also gives the AC bearings room to expand from heat. The adjacent front TRB carries all axial loads.
Newer Mazak spindles are integral, and thus I believe they don't use a roller bearing at the rear anymore. Older spindles had the roller at the rear, designed to carry the side loading of the belts.
And as said, this rear bearing doesn't constrain the spindle shaft from growing in length from heat.
ToolCat In my 12" lathe the spindle bearing design is very simple, but effective. It has only 2 precision tapered roller bearings. The outer races are constrained in the headstock casting. The inner race of the front bearing is a tight fit on the spindle. The inner race of the rear bearing is axially adjustable by a nut behind this bearing. This is how you can adjust the preload.
Tapered roller bearings are quite rigid by design and can carry both radial and axial loads. The distance between both bearings is only around 8-9", so thermal expansion does not play a big role. For a little difference in thermal expansion of the spindle and the headstock casting the casting walls can cave in or out a little to compensate for it. This is a crude and cheap design, but it works in the short spindles. Need a ton more info to give any specific advice. But to answer some of your questions
Yes, tightening a front shaft nut will preload the front pair overcoming the interference--which is small maybe 3-6 micron depending on bearing size.
No, the same force isnt applied in your second example, unless they are in tandem and you are pulling against something that works in an axial direction.. so not a cylindrical roller like you are proposing. Unless you mean a cup/cone style roller which are not really used in spindles as the tolerancing isnt usually good enough. There are exceptions, but they are very pricey. Plus its not a good application since all you are trying to do is control forces of belt tension, a CRB is better, cheaper, and tighter.
What is the size of the shaft/housing you are trying to make?
What loads will be applied to the spindle?
Speed you want to run?
Lubrication (i assume grease?)
Budget?
Drive? ex belt, coupling, motorized, gear? Dismantling and rebuilding a hardinge spindle. The original angular contact bearing was so worn it had negative preload - aka radial motion. The original bearings are in place in the first photo, the single radial rear bearing has been removed.
Second photo shows headstock casting, and one of the outer race clamp rings. Note the rim that's sized to contact the outer race. Note the spot in the center of the cylindrical bore, where the rear bearing sits. The oil gasket thickness is critical here.
New bearings installed on spindle, old ones in background. These were about $300 or so, 20 years ago. Note the match marks lining up on the outer races.
Hardinge went back and forth beween oil and grease lubrication. These are oil lubed.
Dismantling and rebuilding a hardinge spindle. The original angular contact bearing was so worn it had negative preload - aka radial motion. The original bearings are in place in the first photo, the single radial rear bearing has been removed.Your pictures do not show up. Is it just me?
Second photo shows headstock casting, and one of the outer race clamp rings. Note the rim that's sized to contact the outer race. Note the spot in the center of the cylindrical bore, where the rear bearing sits. The oil gasket thickness is critical here.
New bearings installed on spindle, old ones in background. These were about $300 or so, 20 years ago. Note the match marks lining up on the outer races.
Hardinge went back and forth beween oil and grease lubrication. These are oil lubed.
Angular contact bearing query | Page 2 - The Hobby-Machinist
I will need to contact the seller and get them replaced then as I ordered angular contact bearings.
Ay up, just googled " B 2RS bearings" and all search results show angular contacts, same dims too so it looks like I'm ok, just an unusual design maybe?
Sorry you got burned. Hope he doesn't suddenly have trouble communicating when you get back to him.
McMaster Carr has what you want
https://www.mcmaster.com/#standard-ball-and-roller-bearings/=180g6l7
Id be shy of anything a lot cheaper than their price. AC bearings are pricey even for standard duty like these. The sellers were touting them as unsealed or sealed units, the pic a few posts up supplied by alittlerusty is of a sealed unit also which is why I assumed they were legit (skf going by the screen grab, thanks alittlerusty, that was a big help by the way and just wanted to mention that).
Thanks for the link to McMaster Carr, I've never seen so many types of bearing, felt all exited, like a kid in a sweet shop. I think I'm going to enjoy brousing that site lol. Some places list -B-2RS deep grove and some are listing it as Angular contact. witch is strange at the least.
Is there a greater than sign on the side > https://www.123bearing.com/bearing--B-2RS.php
I rebuilt spindles for 15 years and all the angular bearings I use had a big difference from one side to the other. But bearings where $ 40.00 High speed precision
I will need to contact the seller and get them replaced then as I ordered angular contact bearings.
Sorry you got burned. Hope he doesn't suddenly have trouble communicating when you get back to him.
McMaster Carr has what you want
https://www.mcmaster.com/#standard-ball-and-roller-bearings/=180g6l7
Id be shy of anything a lot cheaper than their price. AC bearings are pricey even for standard duty like these. The sellers were touting them as unsealed or sealed units, the pic a few posts up supplied by alittlerusty is of a sealed unit also which is why I assumed they were legit (skf going by the screen grab, thanks alittlerusty, that was a big help by the way and just wanted to mention that).
Thanks for the link to McMaster Carr, I've never seen so many types of bearing, felt all exited, like a kid in a sweet shop. I think I'm going to enjoy brousing that site lol. Some places list -B-2RS deep grove and some are listing it as Angular contact. witch is strange at the least.
Is there a greater than sign on the side > https://www.123bearing.com/bearing--B-2RS.php
I rebuilt spindles for 15 years and all the angular bearings I use had a big difference from one side to the other. But bearings where $ 40.00 High speed precision
If you want to learn more, please visit our website Angular Contact Ball Bearing.
Want more information on F607 Bearing? Feel free to contact us.
For more information, please visit Rimao.
38
0
0
Share:
Previous: None
Comments
All Comments (0)