Lathe Spindle Design (Seals, Bearings)

[barnbwt]

Greetings,
I would like to pick your brains regarding some basic aspects of lathe spindle designs. Namely, those of the bench-top variety with a servo-controlled spindle that can operate as a fourth-axis. My slant-bed design is to be a modest size, with a 5C spindle about 8" in length through the headstock, and I am seeking confirmation or advice on my approach to this most crucial portion of the machine.

I've decided to go with the tried-and-true opposed angular contact pair at the nose end, and a simple deep groove ball bearing at the drive pulley. The workpiece is generally going to be as large as ~2" x 8" long & running at a max of 4000rpm, driven by a 1.5hp servo motor and ~400oz-in steppers. My first instinct is that the added capacity of roller bearings will be unnecessary, and would mostly serve to generate waste heat turning so fast. The goal is a very compact (3ft cube) unit that is rigid enough to do entry-level-quality light CNC prototyping (400+ lbs total weight) and since I won't have lots of damping mass, reducing the vibrations generated at the spindle and toolpost takes priority. My chief concern is that the length of my system seems to be somewhat short for it's 1.25" bore diameter, and that putting both ends of the angular contact pair at the same end may be counterproductive to reducing off-axis motion (compared to the thermal expansion advantage of keeping them close together on the spindle shaft)

The other question has to do with seal systems, which it seems run from very simple (the sealed rubber skirts of my machine) to highly sophisticated (positive pressure through non-contact labyrinth). Is it typically sufficient to rely upon the seals of a bearing pack alone for a machining/coolant environment, or is a secondary contact seal preferable? Vendors suggest that 4000rpm is safe for contact seals with grease, but should I expect greatly reduced seal life and go with non-contact shields instead? If I do a small amount of fine live tool machining or finish grinding that produces dust-like particles while the spindle is stationary, will a labyrinth plate behind the chuck be sufficient, or should this change my approach to something more proactive? I would like a fairly maintenance-free solution that doesn't entail taking apart the spindle for frequent cleaning if practical.

Lastly, though I think I know the answer already, has anyone found success using the more-compact double-row angular contact bearings in lieu of an opposed pair of single-rows, with an additional angular contact or deep groove bearing at the tail end to add preload? My guess is that absent great precision, such an arrangement moots one of the double rows, resulting in only a single row of balls at each end of the shaft carrying all the load.

TCB

Similar Threads:

• Replacing spindle bearings on a D&M 5 CNC Lathe
• Spindle Design and Bearings questions
• Lathe spindle bearings and spindle material
• Lathe spindle bearings and spindle material
• Need Spindle Bearings for CMS Lathe

[mactec54]

Greetings,
I would like to pick your brains regarding some basic aspects of lathe spindle designs. Namely, those of the bench-top variety with a servo-controlled spindle that can operate as a fourth-axis. My slant-bed design is to be a modest size, with a 5C spindle about 8" in length through the headstock, and I am seeking confirmation or advice on my approach to this most crucial portion of the machine.

I've decided to go with the tried-and-true opposed angular contact pair at the nose end, and a simple deep groove ball bearing at the drive pulley. The workpiece is generally going to be as large as ~2" x 8" long & running at a max of 4000rpm, driven by a 1.5hp servo motor and ~400oz-in steppers. My first instinct is that the added capacity of roller bearings will be unnecessary, and would mostly serve to generate waste heat turning so fast. The goal is a very compact (3ft cube) unit that is rigid enough to do entry-level-quality light CNC prototyping (400+ lbs total weight) and since I won't have lots of damping mass, reducing the vibrations generated at the spindle and toolpost takes priority. My chief concern is that the length of my system seems to be somewhat short for it's 1.25" bore diameter, and that putting both ends of the angular contact pair at the same end may be counterproductive to reducing off-axis motion (compared to the thermal expansion advantage of keeping them close together on the spindle shaft)

The other question has to do with seal systems, which it seems run from very simple (the sealed rubber skirts of my machine) to highly sophisticated (positive pressure through non-contact labyrinth). Is it typically sufficient to rely upon the seals of a bearing pack alone for a machining/coolant environment, or is a secondary contact seal preferable? Vendors suggest that 4000rpm is safe for contact seals with grease, but should I expect greatly reduced seal life and go with non-contact shields instead? If I do a small amount of fine live tool machining or finish grinding that produces dust-like particles while the spindle is stationary, will a labyrinth plate behind the chuck be sufficient, or should this change my approach to something more proactive? I would like a fairly maintenance-free solution that doesn't entail taking apart the spindle for frequent cleaning if practical.

Lastly, though I think I know the answer already, has anyone found success using the more-compact double-row angular contact bearings in lieu of an opposed pair of single-rows, with an additional angular contact or deep groove bearing at the tail end to add preload? My guess is that absent great precision, such an arrangement moots one of the double rows, resulting in only a single row of balls at each end of the shaft carrying all the load.

TCB

A double labyrinth seal is best, most use a single

Most of the double row Ac Bearings have some clearance, and no way to preload them, the more expensive double row can be preloaded, so if you are looking lower cost 2 single Ac Bearings Back to Back in the front of the spindle and a deep groove at the rear preloaded would be best

[barnbwt]

That's about what I figured, as far as the two-row jobs. Shame, since it would certainly reduce the volume of the stack. Those split-race deals seem like an ideal solution with the best of both types, but as is too often the case...too dang much money, lol.

As far as labyrinths, is there any advantage to the radial vs. axial variety? Axial seems easier from an assembly standpoint, but radial won't sling lubricant nearly as much (this will be grease lubed, btw)

http://weelibrary.weebly.com/uploads...34371_orig.jpg

At this time, I'm thinking stacked 7X10 (50mm bore) AC bearings, and a single 6X10 at the tail end for the belt to pull on. Exactly which bearing series will depend on what deals come up. That 50mm spindle OD should leave a decent amount of material to surround the 1.25" 5C bore (with room for collet closer tube, of course)

TCB

[mactec54]

That's about what I figured, as far as the two-row jobs. Shame, since it would certainly reduce the volume of the stack. Those split-race deals seem like an ideal solution with the best of both types, but as is too often the case...too dang much money, lol.

As far as labyrinths, is there any advantage to the radial vs. axial variety? Axial seems easier from an assembly standpoint, but radial won't sling lubricant nearly as much (this will be grease lubed, btw)

http://weelibrary.weebly.com/uploads...34371_orig.jpg

At this time, I'm thinking stacked 7X10 (50mm bore) AC bearings, and a single 6X10 at the tail end for the belt to pull on. Exactly which bearing series will depend on what deals come up. That 50mm spindle OD should leave a decent amount of material to surround the 1.25" 5C bore (with room for collet closer tube, of course)

TCB

I mostly make them axial style, either style work well, no problem with oil or grease being thrown around, there should not be any Grease or oil in the labyrinth seal as shown in the diagram, you only have a max clearance of .002" /.05mm to .004" /.1mm between the faces, and if use a rotating slinger on the front face nothing can get in through the labyrinth

[barnbwt]

The gloop in the labyrinth churning about is what keeps dust & coolant out, right? Hence the Zerk fitting on the second diagram? Sounds like if I work in two or more zig-zags in conjunction with a couple shields on either side of the greased area to deflect spray I should be fine for low pressure coolant and greased lubrication. From my very brief research, it sounds like maintenance on these systems is basically to periodically squirt some grease into the bearing area, and very infrequent tear-down & cleanup presumably when the bearings fail? I'd like to avoid systems that require active measures like pumps & the like, so such a regime would be ideal.

Now, on the assembly front, my plan was to have a D1-4 flange permanently scabbed onto the front of the 5C bore spindle shaft* (or machined as a single piece), which means that the spindle must install from the nose-end of the housing. Is the general practice to press the thrust-bearing pair onto the shaft first, and the tail bearings after the fact, or is it preferred to have the (looser by necessity) tail-end bearing present to aid in alignment? Because of the large nose-flange, I think I'd have to press the thrust bearings into the housing first, and then put the shaft through them while supporting the inner races with a spacer tube (or even from the center race of the tail bearing against that spacer). I suppose I'll have to figure out how to get the labyrinth plate affixed across the front of the housing opening before the spindle flange is pressed up against it (perhaps I'll access the plate's fasteners through the cam-lock holes or something).

As for disassembly, it seems intuitive that I'd have to do all my installation steps in reverse, starting with removing the tail-end bearing so the AC races may be pressed out the front. I think the only way this would be possible would be to have the tail-end bearing be a fairly loose slip fit on both the shaft and housing, secured to the shaft with a jam nut, but this seems like it would invite runout (especially with such a short spindle). Any thoughts?

TCB

*not necessarily a deal breaker since 5C can also mount things like small chucks in a pinch, but the versatility of cam-lock is highly desirable

[barnbwt]

Pictures are interesting, so here's a mockup of the (very) preliminary design. There is a labyrinth on either side of the AC bearings to keep the grease in one place. The 'outer race spacer' doesn't need to be pressed tight between the two sets of outer races (whatever preload will be fixed at the spacers between the AC races), but my thought was it could be used to press the whole stack of outer races out the front for disassembly. The 'endcap' has a labyrinth feature at its inside &outside edges because the face between them (where its mounting screws are) is accessible through the large camlock holes in the nose of the spindle (when not in use they will be plugged, but I don't expect them to remain perfectly clean)

[barnbwt]

Some minor updates to the design, along with some new questions:

-inner race diameter has been decreased to 45mm, since there's apparently a massive price-inflection when you go up to +50mm (like, nearly double for P4 bearings)
-all bearings are now 45x85 size, greatly simplifying the boring & fitting operations (just need a single gage cylinder, tool setup, etc)
-since the nose-end bearings are now the same size, I will use a set screw(s) to secure the outer race spacer so the nose AC bearings may be pressed up against it (simplifies spindle removal, too)

The new questions;
-I found a good deal on a high-quality bearing for the tail-end of the spindle, but it is not sealed. Would a press-in teflon shaft seal be a good option for keeping its grease in place since it is not as exposed to debris/fluid? Adding a second labyrinth seems like it would add considerable bulk & assembly difficulty, and 65x85mm seals seem to be a common size (the inner-race/pulley spacer the seal would sit on is 65mm OD).
-Since both nose & tail bearings are now swimming in grease, is there any sense in maintaining the interior 'void' area (see above diagram) between them, or should that whole volume be filled with grease? Would a single grease baffle near each bearing to keep the grease loosely contained internally with a dedicated Zerk fitting be a better approach?