Ball bearing spindles

Imagine two extremes, a close fitting bearing and a loose bearing. In both cases there will be a 'wedging' effect, just as with a taper roller bearing, but less extreme.

Put axial preload on a loose bearing and the balls will bear on the sides of the race, where it is at an appreciable angle.

Take the close fitting bearing, now the bearings cannot move away significantly from the centreline of the races. The contact angle is now much less, greatly increasing the radial load.

Axial load:axial load varies with the inverse of the tangent of the contact angle.

With angular contact bearings the angle is designed to be about 45 degrees so axial and radial forces are roughly equal.

For taper rollers about 20 degrees might be typical, but the long contact surface compensates for the high radial loading.

I found this: www.astbearings.com/ball-bearings-radial-play.html

"When a ball bearing is subjected to thrust loading, a higher contact angle will result in reduced ball to raceway stresses. Greater values of radial play result in higher values of contact angle. Under these conditions, this will provide longer bearing life, lower torque, and less axial deflection. In a pure thrust situation, a 15° increase in contact angle can result in over a 70% reduction in contact stress (ball-to-raceway)."

Neil

I think Neil is suggesting that the tighter the balls are in the race, the nearer the contact line is perpendicular to the axis and any preload acts like a wedge driving the balls outwards (and inwards) rather than holding the things axially.

Edited By Chris Trice on 14/01/2019 17:23:38

Neil beat me to it.

Thanks both

Frankly, I still think that [unless deliberately built to behave as you suggest], a real-world slack/loose bearing will have balls with wider dimensional tolerance than the sooper-dooper version. The first contact will then be on the three largest balls in the set, and the preload will be applied to that tripod ...

Brinelling, anyone ?

MichaelG.

What you really want in a spindle, is a bearing where the inner and outer race and the swash of the race is as true as possible. There are tables for the precision that the outer and inner race is made to. Now sometimes you can pick up a standard tolerance bearing with its concentricity as good as a spindle grade bearing. But that is very rare.

According to the article I linked, bearing quality is NOT related to bearing spacing. You can have a high quality slack bearing.

"There is often confusion with “looseness” and precision level. Radial play is specified independently of ABEC tolerance classes for rings."

So perhaps we are worrying about the wrong things.

But an angular contact bearing will still be better than a deep groove bearing for taking end thrust.

Neil

.

Thanks, Neil

I wonder how 'what you can have' compares with 'what you typically get'

... All down to from whom you buy, I guess.

MichaelG.

Edited By Michael Gilligan on 14/01/2019 19:30:08

The way I understand ball bearings are made, they're made in batches, so if they are inaccurate in their nominal size, all the ball bearings in that batch are uniformly inaccurate and therefore the tripod scenario you suggest is not likely to be an issue within measurable tolerances. If you mixed ball bearings from different batches then there might be a tripodding (good word) issue. It's why you shouldn't be seduced into mixing new bearing parts into old bearing parts when feeling lazy repairing something.

Edited By Chris Trice on 14/01/2019 19:44:02

Thinking about it, there would have to be a radial clearance for angular bearings to work otherwise they'd just be the same as deep groove bearings but with one side built higher than the other.

.

Fair point, Chris ... but:

I can't recall where or when [and it may have been a 'top shelf' item for Metrologyphiles], but I have seen a demonstration where ball bearings were selected from a batch by the simple expedient of trapping them between two flats ... all but the three largest dropped out !!

MichaelG.

.

Edit: This may be of interest

https://www.precisionballs.com/ballgradechart.pdf

Edited By Michael Gilligan on 14/01/2019 20:13:58

Hi All

While thinking and browsing on Google looking for a spindle for my mill project I stumbled on this video on building a small spindle. as shown it uses 3C collets but that can be changed. The spindle is fitted with 3 bearings two at the front and a floating rear. Proper seals are fitted front and rear.

For fusion 360 users there is a free set of 3D plans for the spindle and the novel table top CNC lathe it is fitted to.

Link to video:
**LINK**

Link to plans:
**LINK**

Regards
John

Don't know if they are still available, but a good bearing for running a spindle is the "Magneto" type.

Ian S C

Hmmm. My understanding of magneto bearings is that they are a sort-of hybrid: a deep groove inner track and an angular contact outer track. They can therefore take some axial loading, but little, if any more than a deep groove set. I don't think they are designed for high loading. Their purpose is to allow axial location of a shaft, rather than to react substantial applied axial loads, whilst being easily separable: the housing comes apart, axially, and the shaft can be removed without having to disturb the fitting of either track. I think angular contact bearings would be better for a milling spindle.

Re Neil's earlier post. Things ain't so simple. Angular contact bearings are supposedly available in different contact angles, to be selected depending on the ratio of anticipated radial to axial forces. IIRC, 3000 and 5000 series differ in contact angle. (Sorry, too lazy to go and look. Perhaps the indefatigable finder-of-information MG will search out more...). Dunno if the availability of different geometries gets as far as our usual suppliers.

CT. I agree about the radial clearance of angular contact beraings. I would assume that the 'steeper' contact angle necessary to react substantial axial loads results in decreased radial stiffness, compared to deep groove bearings.

I'm coming round to the idea that, provided axial loads arekept within the manufacturer's allowable range, C2 bearings will provide greater radial stiffness for a given axial load, but less axial stiffness. Thus they would seem better for, say, a small lathe headstock than C3 bearings, and perhaps also better than angular contact bearings, unless a lot of drilling from the tailstock is anticipated. However, this is all reasoning from first principles, rather than working from established fact - come on, we need a bearings expert to tell us!