a slant on pre-loading bearings

[darryl]

Gentlemen. Having cleaned up quite much in the shop, I have uncovered an old project which I should have finished some time ago- the homemade tablesaw. This is the one with two cable-operated fences, one normally parallel to the blade, and one perpendicular to it. The table is built and tested, all is well. I'm back on the project again.

I'm now at the point where I need to machine the bores for the arbor bearings. I'm using a pair of 6203 sealed bearings, and what I plan to do is epoxy one of them into its bore, then have the other one a very close sliding fit into its bore. The arbor (spindle) will be pinched between them, in other words pressure on the outer race of the movable bearing will take up the play. The drive pulley will be mounted between them, and a flat micro v belt used.

In taking out play in the arbor, I thought to use the equivalent of a threaded ring that would bear against the outer race. Further thought suggested that I might put a certain measured force against the side of the bearing, then turn the ring until it just seats and no further. A sealant of some kind would fix the ring in this position, and would also be used to prevent the outer race from spinning in its bore. Design constraints place the bearings just under 3 inches apart.

In reality, instead of a threaded ring, I'll have a retaining washer that will press against the bearing, and that washer will be held with four bolts. The procedure would be the same- apply a certain amount of force to the washer, then lightly run each bolt down. The end result will be the same- a pre-load based on a measured applied force.

I realize that this type of bearing is not meant for large axial forces, but in applying a small force meant to take up play and nothing else, they should be fine. I haven't been able to find suitable information so far as to how much force to apply- I'm thinking of something like two pounds or so.

I could just use a normal method and apply the force, then run it for 10 minutes or so, checking for heat build-up. Back off if it runs too hot, increase if it stays cold. But all I want is to reduce looseness to zero, not end up with a continuous axial force at a greater level than really required. I'm thinking that if I use the temperature sensing method, I'll end up with more pre-load than I want.

So that's it- in short I'm asking how much side force should I put on the bearing?

Going a little deeper- the front bearing, the one that's next to the saw blade and the one that gets epoxied in place, is also initially floating in its bore. This will allow the arbor to control the tilt of the bearing to some degree, and also give me a way to make a fine adjustment of the axial positioning of the arbor. When the correct position is found and the arbor has aligned the bearings as it wants to, the assembly is left while the epoxy cures. A retaining ring is then placed over the front bearing and the bolts loctited. This is just a mechanical insurance that the bearing cannot move out of it's bore, if perchance the epoxy let's go of it. If I ever have to replace the bearing, it will still be possible. Once the front bearing is done, I'll move to the rear bearing and apply my pre-load method.

A bit more about the design for those who are interested- the blade tilt mechanism in this design will allow for tilt in one direction to about 31 degrees, and in the other to a bit over 45 degrees. At zero height, the surface of the table, the blade will stay in position through the entire tilt range. This is something I fine-tune with the axial positioning of the arbor and its bearings. With the blade at 90 degrees, I'll have over 3 1/2 inches of blade height capability (with a 10 inch blade), and that will not change with tilt in the normal direction, but will lessen in the 31 degrees direction. Blade height is achieved by sliding the entire arbor housing/motor assembly along four guide posts, and will be motor driven via a pair of leadscrews linked with a chain drive. Setting the blade angle will be done by loosening a clamp, then manually swinging the drive assembly from the side of the saw. A change from 90 degrees to any predetermined angle setting will take about five seconds or less- a few seconds more for odd angles- no hand cranking involved. Index points will be placed at normal spots- 22.5, 30 degrees, 45, etc. You won't even have to see a dial to know when you've hit it right on.

So much for now- I'll have pictures when it's together.

[dp]

A project without a picture is like a song without a melody; a kiss without a squeeze; soup without a spoon; spring without a robin; a string of homilies without a point...

[darryl]

I will post a few pics of the arbor assembly once I have the bores done and a trial fit-up is possible.

[philbur]

What will temperature rise in the bearings/shaft do to your carefully planned "cold" preload. Setting preload by temperature seems to be ideal for this application. Just check the temperature by hand the first few times you run it and adjust until the bearings are warm to the touch when at stable conditions. You don't care what the preload is as long as you have a little at normal running conditions.

Phil

PS: Temperature will give you the average preload. Actual loading will only give you a point preload which, depending on bearing and mounting precision, could be different at other points in the rotation. If you set the preload by measured load and it runs hot what will you do, probably back it off until it doesn't run so hot!!!!!

I think spring induced preload might be good for controlling preload for not so precise bearing/mount setups, its also self adjusting for none uniform thermal expansion.

Just a few thoughts.

[darryl]

I've considered the effects of expansion- for one thing, because the housing is aluminum, and there's lots of it, it will take awhile for the mass of it to heat to any great degree. Also because the distance between bearings is less than three inches, the expansion of the arbor should be minimal.

At this point, I'm more concerned with getting the bores inline and square to the housing. I'll have to check my mill to see if the downfeed tracks perpendicular to the table in both directions. I plan to mount the housing on parallels, then bore both sides through without moving anything. If the blade ends up not being parallel to the saw table slots, I have a small range of adjustment I can do there, but not much. I'll need it very close from the start.

Spring loaded pre-load- one thing I don't want to have happen is for side load on the saw blade to overcome the spring and let things run loose at that point. I know there shouldn't be any side loading on a saw blade, but in practice it is going to happen. For a high enough spring tension to overcome that, I think the pre-load will be too high.

Lots of things to consider. Thanks for your thoughts.

[philbur]

Minimal and acceptable are not the same thing.

It seems your mind is already made up as to what your intend to do, however I still think that setting preload by temperature is the best/simplest method for this application.

Phil

PS: Don't forget that these are common or garden, ordinary deep groove ball bearings.

the expansion of the arbor should be minimal.

[GerryR]

Use 7203 angular contact bearings which are made for this type of spindle application. They can take a great deal of axial loading (compared to a deep groove bearing), and can be preloaded using your described method or with Belville-style washers. Just a thought.

[JoeLee]

I cant see where the use of precision angular contact bearings are needed for a table saw spindle!!! I've never seen a table saw that had anything but standard conical ball bearings. It sounds like way overkill to me. What precision those bearings will give you in your soindle will all be lost in the blade, flex wobble heat expansion, blade mounting, movment of wood, etc. It's not a surface grinder.

JL...............

[Mcgyver]

I cant see where the use of precision angular contact bearings are needed for a table saw spindle!!! I've never seen a table saw that had anything but standard conical ball bearings. It sounds like way overkill to me. What precision those bearings will give you in your soindle will all be lost in the blade, flex wobble heat expansion, blade mounting, movment of wood, etc. It's not a surface grinder.

JL...............

who said precision? and whats a conical roller bearing? like a taper roller bearing? SO....why would run of the mill angular contacts (ie not $500 matched abec 7's) be much different, other than lighter, than tapered rollers? If though, you've seen a lot of table saws and they're all tapered roller bearings (is that the case?).....that would be a hint that they'd be the thing to use; that the application was too heavy for angular contacts

[JoeLee]

I've never seen a table saw that had anything other than a standard ball bearing, and I've seen a lot of them, old ones like Oliver, Rockwell, etc.

JL..................