Lathe rigidity

Agreed, at most it should be a tight push fit, it should not slide on without significant effort. Also the front bearing with should be a light press fit on the front of the spindle, while the rear bearing (as you have deduced) needs to be able top be moved to adjust the preload.

Neil

Thanks for the clarification Neil.

The front bearing is likely to be a tight press fit (I'm not sure what that means, but that's my guess) requiring a lot of grunt on my press to push it in place. Thanks btw to the poster who said that you should only press on the part of the bearing which you were pushing in (e.g. the outer race and not the inner). It took me a while, but I can see ho to do that.

As you say, the rear one needs to move, but move with strong manual pressure or only move with a press?

Also, can anyone suggest the correct grit to use for the polishing? I'm on 10 micron at the moment.

Iain

What matters is what the diameter started at and what size you are aiming for. I would not reduce the diameter more than 0.012mm, or 0.0005" at most. You have the old bearing inners to test the fit, they should now be noticeably easier to press on and off the rear end of the spindle than when they were originally removed. Tighter than a push by hand fit.

Isn’t loctite ‘bearing fit’, or similar products, designed to take up too easily sliding fits?

Hi Ian,

I have tried to understand your problem, but I still do not know why you are blaming the bearing press fits for your inaccurate and not rigidly mounted collets. It seems to me that the collets are mounted via a couple of interface points, the Morse taper and the spindle Morse taper mounting resulting in a collet positioned too far from the front bearing. This cantilever will magnify any errors present and reduce rigidity, making the bearing fits easier does not seem to be, in my mind, the solution. Simply put the collet should be mounted as close to the front bearing as possible for maximum concentricity and rigidity. If I have failed to understand the problem I apologise and I am sure it will be pointed out by others, but good luck anyway.

Alan

Loctite bearing fit would certainly cure a loose bearing, but in this instance, the fit needs to be just tight enough to be easily adjusted, but not loose. Bearing fit Loctite would be too much of a good thing.

The collet question which has been somewhat neglected during the thread might be cured by using one of the collet plates which bolt directly on to the spindle flange. Like this one:

**LINK**

.

Yes ... but using it would make adjusting the pre-load a 'non-trivial' matter.

MichaelG.

.

Edit: 'old mart' beat me to it

Edited By Michael Gilligan on 09/09/2019 18:43:16

Alan and co. - I get the same deflection on a 3 jaw chuck mounted directly on the spindle.

And I have bought a spindle mounted collett chuck - one from ARC as it happens. I didn't know they existed until an earlier post introduced the idea.

Before I get to check how well that works, though I have to improve the fit of the rear bearing, re-assemble the lathe and try again.

That's not going to be today though

Iain

Are you suggesting that the outer or inner race moves within their fixed positions when adjusted. I think not - I think one is moving (by an infinitesimal amount) one race towards or away from the other. In fact, unless moving from no preload to some preload, the bearing races will not move at all.

Problem I see with using any form of retainer on the inner rear race is that by the time you have assembles it and then run the lathe to bring it upto temp the retainer will have gone off and you won't be able to make fine adjustments.

Same applies if you need to adjust it at a later date, you would need to break the loctite joint before you could adjust.

.

Well that's three of us that understand

MichaelG.

In that case you had better explain it in a little more detail - ‘cos I don’t.the loctite will only be set when the bearings are hard against the shaft or seat. Are you suggesting that they will be non-concentric? If so, that can be eliminated by curing in position with bearing load applied. Once cured, the races should not need to move at all. A sliding fit is not a wibbly-wobbly fit, is it?

The problem is that something like 641 bearing fit sets in about 30mins at 20deg C and time reduces to about 10mins at 40deg C though there are a few other variables that affect the time. Time starts ticking as soon as the bearing is placed onto the shaft and the air is excluded from the retainer, contact with metal is the catalyst.

This gives a problem as you need to apply the Loctite, assemble all the parts, run the machine to bring it upto working temperature and then make fine adjustments to set the preload. It is very unlikely that you could do this in the amount of time available.

Also quite possible that someone who has not done it before may need to have more than one go at getting the preload right which will be difficult once the Loctite has gone off.

Lastly there is the issue that if you ever need to take things apart again you need to break the bond which would need the headstock, spindle etc bringing upto 250deg C not that easy to do while trying to pull the asembly apart.

The need is for a close fitting bearing inner ring that can still move under the load applied by the adjusting nut/nuts. You can't do that with the ring bonded to the spindle.

Edited By JasonB on 10/09/2019 13:15:27

Not much point in needing to use the press as you won't be able to use that on the assembled parts to set the preload.

As Michael says do several test fits and then aim for the bearing to need the C spanner on the nut to move it without excessive force but not so slack that you can turn the nut by hand. You will need to slip the gear and pulley onto the spindle to test this unless you can find a suitable spacer.

not used diamond grit so can't help but again as Michaels says you are really just trying to remove high spots left from turning the spindle or coarse grinding.

.

Exactly that ^^^

MichaelG.

.

Infinitesimal, perhaps ... but definitely real : Any change of preload requires movement.

MichaelG.

Not true. If a force is applied, an acceleration should occur (Newton’s Second Law of Linear Motion) - unless that force is nullified by an equal force in the opposite direction. That is all that occurs when a preload is changed. No force, no preload.

Every preload force is opposed by an equal and opposite force - unless something breaks (and preload will reduce to zero). Bearings don’t move, only the force applied changes. You can’t actually have zero preload as the bearing races could be together or miles apart! If anything moves, it will be the housing, not the bearing, per se.

 ... One last attempt :

https://www.nationalprecision.com/library/articles/bearing-preload.php

Edited By Michael Gilligan on 10/09/2019 15:48:29

As I see it, the front bearing is fitted to the headstock, pressing on the outer race and the assembly is then pressed onto the spindle, pressing on the inner race. Then the rear bearing is pressed onto the spindle until the outer race is up against the rear shoulder. It is important that both races are pressed as a unit. The fit of the rear bearing on the spindle should be a very light press fit which will ensure that the front bearing is not overstressed. I would support the headstock while the rear bearing is fitted. If thought is not applied to the assembly process, then brinelled bearings will result. Taper rollers would be easier to fit than the deep groove sealed bearings used on this model lathe.

So NDIy, can you explain this.

When my X3 mill arrived there was movement in the spindle of about 20thou sideways and up/down . I adjusted the preload by turning the nut and locknut and got that down to an unmeasureable amount. Please tell my how the gap between bearings that allowed that movement was eliminated if one of the bearing races did not move along the spindle to take it up?

Edited By JasonB on 10/09/2019 15:49:52