Engineering principle behind this type connection

[Maplehead]

I agree that, in general, identifying degrees of freedom is important. But you need to do more. Here, you want a single degree of freedom (rotation about the vertical axis). Now you need to identify loads. You have, as I understand it in the example, primarily axial loads. So different 1-DOF rotational joints will handle loads differently. A deep-groove ball bearing will handle some axial load but that's not it's strong point. A thrust bearing will handle axial loads, but not radial loads, and most will not handle radial alignment (i.e., they are not suitable for restraining all but 1-DOF).

A tapered roller bearing is great at a mixed axial and radial load, but they require preload (a force smashing the two halves of the bearing together). Generally, they are used in opposing pairs, so that they preload each other essentially. If, as here, you have unidirectional axial loading, you could use a single one as long as the preload is sufficient. What does that mean? It depends on the applied load. If you have a pure axial load, then really no preload necessary. But as soon as the bearing needs to resist a moment from radial loading away from the plane of the bearing, then you need preload. Better than that would be a second bearing.

What is difficult about what you're asking is that is is too simple. You need much more than what the joint is called. And there are many ways that such a joint could be constructed, depending on constraints. Bushings (plain bearings) are good for low speed, high load, but will have more friction than rolling-element bearings.

It seems like physical size is important to you. So you may think about a solution that can nest the two structures (one for radial/moment loads, one for thrust). How about a cylindrical bushing to keep everything centered? That would restrains 4 degrees of freedom, leaving 2 unconstrained (axial movement and rotation about the primary axis are unconstrained, while movement along and rotation about the transverse axes are constrained). Then add a needle thrust bearing on bottom to handle the axial load (restrains axial movement). That seems like a potentially good approach.

You could do as Dave suggests - a single bearing. But not many bearings are well suited to operate in that configuration with your proposed loading. I've been considering the ball turner but you mention another jig that you are interested in. It's entirely possible that the jig would work really well with a single deep-groove ball bearing. Even the ball turner might if we can estimate the loads a little and check available bearings.

There's a lot for me to digest here but thanks. In terms of loads my primary use for this ball turning jig is for turning a 3/4" ball from 3/4" mild steel rod. I also would have another hole for the cutter to do the same for a 3/8" ball made of mild steel.
The Sehr version I made does work but it's kinda sloppy. (My fault.) I have to be pushing down on the disk opposite the cutter otherwise the disk has wobble and thus does not cut cleaner and perfectly at the center, (3/8" up the 3/4" rod).
Lastly, searching on DOF came up with a lot of cool related stuff.

 

[Maplehead]

There's a lot for me to digest here but thanks. In terms of loads my primary use for this ball turning jig is for turning a 3/4" ball from 3/4" mild steel rod. I also would have another hole for the cutter to do the same for a 3/8" ball made of mild steel.
The Sehr version I made does work but it's kinda sloppy. (My fault.) I have to be pushing down on the disk opposite the cutter otherwise the disk has wobble and thus does not cut cleaner and perfectly at the center, (3/8" up the 3/4" rod).

I agree that, in general, identifying degrees of freedom is important. But you need to do more. Here, you want a single degree of freedom (rotation about the vertical axis). Now you need to identify loads. You have, as I understand it in the example, primarily axial loads. So different 1-DOF rotational joints will handle loads differently. A deep-groove ball bearing will handle some axial load but that's not it's strong point. A thrust bearing will handle axial loads, but not radial loads, and most will not handle radial alignment (i.e., they are not suitable for restraining all but 1-DOF).

A tapered roller bearing is great at a mixed axial and radial load, but they require preload (a force smashing the two halves of the bearing together). Generally, they are used in opposing pairs, so that they preload each other essentially. If, as here, you have unidirectional axial loading, you could use a single one as long as the preload is sufficient. What does that mean? It depends on the applied load. If you have a pure axial load, then really no preload necessary. But as soon as the bearing needs to resist a moment from radial loading away from the plane of the bearing, then you need preload. Better than that would be a second bearing.

What is difficult about what you're asking is that is is too simple. You need much more than what the joint is called. And there are many ways that such a joint could be constructed, depending on constraints. Bushings (plain bearings) are good for low speed, high load, but will have more friction than rolling-element bearings.

It seems like physical size is important to you. So you may think about a solution that can nest the two structures (one for radial/moment loads, one for thrust). How about a cylindrical bushing to keep everything centered? That would restrains 4 degrees of freedom, leaving 2 unconstrained (axial movement and rotation about the primary axis are unconstrained, while movement along and rotation about the transverse axes are constrained). Then add a needle thrust bearing on bottom to handle the axial load (restrains axial movement). That seems like a potentially good approach.

You could do as Dave suggests - a single bearing. But not many bearings are well suited to operate in that configuration with your proposed loading. I've been considering the ball turner but you mention another jig that you are interested in. It's entirely possible that the jig would work really well with a single deep-groove ball bearing. Even the ball turner might if we can estimate the loads a little and check available bearings.

Isn't my primary load radial and not axial?

 

[jwmelvin]

There's a lot for me to digest here but thanks. In terms of loads my primary use for this ball turning jig is for turning a 3/4" ball from 3/4" mild steel rod. I also would have another hole for the cutter to do the same for a 3/8" ball made of mild steel.
The Sehr version I made does work but it's kinda sloppy. I have to be pushing down on the disk opposite the cutter otherwise the disk has wobble and thus foes not cut cleaner and perfectly at the center, (3/8" up the 3/4" rod).

I'm not really sure but it seems like axial loads there wouldn't be more than several hundred pounds, and likely quite a bit less.

Take an example bearing: SKF 6002. It has a 641 lb basic static load rating. The information about loads says that axial load can be 50% of that, but for bearings with bore less than 12mm, only 25% of the basic static load. So that puts us about in the ballpark for axial load (but not much margin).

Your load isn't pure axial, depending on where it's applied relative to the bearing diameter. You likely want a larger bearing. They do get more expensive, so you start having to balance factors. But it seems like Dave is right - a single ball bearing may work for you. And these are not expensive bearings. Even a name brand 6005 is not much. According to SKF, that size has a basic static load rating of 1472 lb. So you're starting to get to something that is more likely to survive the application.

If you have room for almost 3.5" OD, then use something like a 6209.

 

[jwmelvin]

Isn't my primary load radial and not axial?

I thought that in this style tool, the cutting load is pushing down on your rotating assembly, along its axis. There will be some load radially away from the spindle, which combines to a radial and moment load on your assembly. It's the moment load that will be difficult to handle with a single bearing. A large one, like the 6209, would handle that better (wider base can resist the moment much better). But using an axially spaced radial bearing is better. Just depends on what you need and how much space you have. I'm thinking you are most constrained by axial space, so using a wide thrust bearing or a large deep-groove will be best.

 

[rabler]

The other option that has not been brought up is to use some sort of anti-friction bushing. You are not looking at a lot of motion, but at some high pressures, so a bushing approach would work. Bronze or delrin ...

 

[Maplehead]

Isn't my primary load radial and not axial?

Here's a pic of my jig now along with a video showing the wobble. I can take out the wobble by tightening up the axle screw but then I cannot rotate the disk. And again, isn't the jig taking on radial load and not axial?

 

[Maplehead]

I thought that in this style tool, the cutting load is pushing down on your rotating assembly, along its axis. There will be some load radially away from the spindle, which combines to a radial and moment load on your assembly. It's the moment load that will be difficult to handle with a single bearing. A large one, like the 6209, would handle that better (wider base can resist the moment much better). But using an axially spaced radial bearing is better. Just depends on what you need and how much space you have. I'm thinking you are most constrained by axial space, so using a wide thrust bearing or a large deep-groove will be best.

Yeah, if axial space is up and down then yes, I don't have much space, as you can see in my latest pic.

 

[Maplehead]

I thought that in this style tool, the cutting load is pushing down on your rotating assembly, along its axis. There will be some load radially away from the spindle, which combines to a radial and moment load on your assembly. It's the moment load that will be difficult to handle with a single bearing. A large one, like the 6209, would handle that better (wider base can resist the moment much better). But using an axially spaced radial bearing is better. Just depends on what you need and how much space you have. I'm thinking you are most constrained by axial space, so using a wide thrust bearing or a large deep-groove will be best.

To me, and I'm probably wrong, the load is radial, as in the red arrow, and not axial, as in the green arrow.
To me the load is a "pushing across" load, not a pushing down load.

 

[rabler]

You need to back up a bit. Your working two problems, tolerances, and friction.
The up and down motion at the handle is an issue of tolerances. You are working with a very short axle length.
Imagine a shaft held at two ends, with .001 inch tolerance in the ends. If the shaft is a foot long it will not move as much as a shaft that is only an inch long. More angle of play in a short shaft.

You might start by using a sharpie to blue up the contact faces and see where the disc is touching the cutout. If the contact point is near the center bolt, you have the same effect as a short shaft. If you hollow out the center a bit so the contact point is near the edge, then you have a much more stable base.

 

[jwmelvin]

To me, and I'm probably wrong, the load is radial, as in the red arrow, and not axial, as in the green arrow.
To me the load is a "pushing across" load, not a pushing down load.

I’m sure there is a mix of loads. The ratio will depend on the tool geometry, depth of cut, and feed rate. Let’s say for now that both loads apply.

That means your tool needs to resist not just axial load but also moment load. Plenty of ways to do it.

A large deep-groove like Dave proposed and I discuss above is pretty straightforward and if sized for your largest permissible diameter, should handle the loads.

A plain bearing such as a bronze flange bushing would likely work but I’d make the flange pretty large diameter. And use a preload bolt and spring.

Think about a needle thrust bearing. Would be great for pure axial load but as moments apply, far side of the bearing is compressed (that’s what it’s made to handle, so no problem) and the near side is pulled apart. If that pulling apart overcomes the preload, you now have a wobbly thing. The nice thing about the ball bearing is that is handles the arbitrary loads quite well.