Are pillow block bearings accurate enough to make a lathe with?

[strantor]

I had been tossing around the idea of making a lathe with a couple of pillow blocks mounted to a casting. I found this video, and the guy did pretty much what I had been thinking of. This video shows a different type of of pillow block, or flange bearing, that I actually already have 2 of these and a shaft for them. if they were set up in this fashion, but bolted to a rigid casting instead of a welded plate, do you think that either of these 2 schemes could produce an accurate lathe? Could it do machine work comparable to commercially available lathes? I am referring only to the bearing portion/head stock.

 

[Tony Wells]

My first thoughts are no, they would not suffice. Those bearing are not designed for end loading, for one thing. If you were to add a good thrust bearing, or better yet two, and have a way to set the preload, then the main issue I see would simply be the quality of the bearing. Typical pillow blocks simply aren't precision grade bearings. You could probably make a fine wood lathe that way, but I'd still want the thrust bearing. Another issue would be the spindle itself and how it fits the ID of the bearing. There are a couple of different methods used to lock in the shaft, and neither will give you the TIR you want on a lathe spindle. Of course, you would want to finish machine the spigot once it was locked in, but if you ever needed to remove it, it's doubtful you could get it to run true again. And, the size of the through hole would be limited depending on the ID of the bearing and the needed wall thickness for rigidity and strength.

Just a few thoughts of mine, worth what you paid for them.

 

[DMS]

I have a copy of the Dave Gingery book, "The Metal Lathe", which describes building a lathe from the ground up using raw stock and aluminum castings. He uses a set of pillow blocks mounted to a temporary platform. This setup supports a boring bar that is used to cut the headstock bearing pockets in-situ using the ways as a guide (so everything lines up). Those books are gold, even if you don't plan on making your own lathe, just because the exposure to the shear ingenuity is bound to make you a smarter persone :think:.

I somewhat agree with Tony. The bearings in those are probably deep groove ball bearings (the most common type by far). They can take about 25% of their rated capacity as an axial load, which should be good enough for small work. Will it be as accurate as a manufactured lathe? Probably not. Will it work? Almost certainly. Will it do what you want? Maybe.

One thing to think about is that good quality pillow blocks are expensive (unless you have a cheap source, in which case, share . By the time you buy the stock, bearings, etc, you may have spent as much as you would have on a used 7x14 import lathe.

I went through a similar venture a couple years back when we still lived in a small apartment. I really wanted a mill, but had no room, so I decided to build a desktop cnc rounter based on one of the common designs. It worked after a fashion, but it was a bit cumbersome, and ultimately I spent quite a bit on it. I also learned a heck of a lot, and had fun. Neither of those you can really put a pricetag on.

One last thing. To get around Tony's concern about accuracy and bearing fit. Set the shaft in the blocks, and lock the set screws. Then, once the shaft is locked, turn the nose taper. This ensures that the nose taper is concentric. The only trick is that you cannot remove the shaft. If you do, the nose taper will have to be re-turned in place. I saw this technique on the Quorn mailing list in relation to producing accurate grinding spindles. In that case, the spindle is turned in it's own bearings. The result is hard to beat.

DMS

 

[Rbeckett]

There is a move afoot on another site that I frequent that is building machine tools in the third world. The designs and materials are pretty darn inovative and it seems to be a very workable design. Make magazine just did a piece on it too. Here is a link to the home page and if you look carefully you will see the engine block being re-used pretty handily, http://groups.yahoo.com/group/multimachine/ Shannon and Pat have invested a bunch of time and effort to get this project into a usable form. Have a look..
Bob

 

[Bill Gruby]

My first, second and third toughts ade NOT. These bearings just are not good (as in precision) for our use. As Tony said they are not built to take end loads. The main use of these bearings is in long shafting.

"Billy G" )

 

[Tony Wells]

Actually, DMS, I did address that concern with the same suggestion you mentioned, with the caveat that the spindle should not be removed and replaced with any expectation of the same runout.

 

[lindse34]

You may want to look for bearings that use a concentric locking mechanisms to the shaft. That would help with TIR but they are expensive in the larger diameters. MSC wants $130 ish for 2 inch units. Also you may want to consider a pair of tapered roller bearings with the outside races set so the preload against them could be adjusted. The trick though would be knowing how much preload without impacting bearing life.

As a side note when you have a troublesome installation where the shaft is continually walking out of a setscrew or eccentric locking collar, the concentric locking bearings have a tenacious hold on the shaft.

Good luck.

 

[DMS]

Actually, DMS, I did address that concern with the same suggestion you mentioned, with the caveat that the spindle should not be removed and replaced with any expectation of the same runout.

Ah yes, so you did... So much for my reading comprehension