Long time since I read the GHT write-up, but I suspect that he saw the greater longitudinal stiffness of roller bearings as a good reason for fitting them to his Myford feed-screws, as much as the friction benefit perhaps.

Think again John. **LINK**

Murray

The needle thrust bearings are ideally suited for this application, one of the main things for them is the low profile over a ball thrust race.

Ian S C

Needle thrust bearings are often used where a lower profile or small diameter are required when compared to ball or roller thrusts. One thing to bear in mind when mounting is that the housing washer is better fitted into a sized counterbore as there is no radial location as in a ball type which has a grooved track. I would normally fit the housing washer into a recess and the shaft washer onto the shaft of all thrust bearings to prevent the washers moving, but I accept it is not important on hand rotation.

You can also use accurately machined surfaces instead of hardened thrust washers to save space. This approach is endorsed by SKF.

Neil

Apart form the conditional "excessive" (nothing is designed for excessive anything by definition), this is incorrect. One of the principle reasons for using TRBs is their ability to accept axial and mixed loads. The only conition is that a single row can only accept load in one direction.
It's the taper on the rollers that allows them to roll without skidding.

This needle thrust conundrum is something that I have often thought about. If the needles were perfectly parallel and the race surfaces dead flat then scuffing of some sort must take place so it would not be a 100% rolling bearing. Today though I found that some manufacturers of this type of bearing 'relieve' the ends of the needles to reduce stress points.

Maybe this relief effectively make the needles microscopically barrel shaped so that at light loads its nearer to point (rather than line) contact. In conventional ball races, I thought that the inner and outer race race ball grooves would always be exactly a true part of a circle, however I now know that a lot of technology is used to shape the groove to give the best performance and that parabolic and other profiles are used.

If the whole point of needle thrusts is to minimise the overall axial length of a thrust bearing then another approach might be to use ball thrust bearings with very small balls. It seems to me that in the same radial distance that a needle occupies there could be several concentric ball thrust races.

Ian P

Preload is required to minimise end play – as close to zero as possible. Wheel bearings should be tightened until they drag and then released by a fraction (or number) of flats of the securing nut (dependent on thread pitch and nunber of flats, of course). Running tight would overheat the bearing.

I have engines with steel crank and cast iron crankcase and others with sreel crank and aluminium crankcase. Obviously each require a different preload, to take account of thermal expansion at running temperature, to achieve the permitted end-float (always positive, never zero – which might be ideal – because the bearings might be axially loaded).

Consequently, I am not sure how much end thrust most of our spindle bearings would take continuously. My lathe has ball thrusts either side of the drive pulleys and they have survived for about 50 years of hobby use. Never yet seen any that have failed (given sensible treatment – lube and setting up an cleanliness).

I'll link this again, it's a good read:

http://www.model-engineer.co.uk/news/article/lathe-bearings/19136

Neil

Links work OK for me.

It hows how influential ME was in 1947 that British Timken felt the need to respond to one of LBSC's lobby chats in such detail!