PDA

View Full Version : bearing pre-load



tony ennis
11-19-2007, 09:58 PM
I'm replacing a thrust ball bearing that's in my babbitt headstock. While I'm at it, I'd like to add a bit of pre-load to the bearing. I can do this by adding a wave or spring washer to the retaining collar on the far end of the spindle.

How much of a pre-load is needed? Typical washers exert up to 20 lbs.

Does the mighty HSM have any opinions on types of pre-load springs?

chief
11-20-2007, 12:38 AM
The bearing manufacturer should able to supply the pre-load info based on the bearing dimensions, all bearings are not created equal. Measure the thrust before install any preload device and than go from there the minium amount of preload is best, I would think 10-15ftlb is more than enough.
Here is link, I haven't read it so I can't vouch for it. Of course there is also the old tried and true method of tightening her up till she strips and then back off a 1/4 turn.
http://productsearch.machinedesign.com/research/Mechanical_Components/Springs/Preload_Springs_Spacers

SGW
11-20-2007, 08:55 AM
I don't think you need much. My South Bend 10K doesn't have any preload; just minimal clearance set by the take-up nut. Note that if you have preload, you'll need some kind of bearing on the take-up nut side of things too. If you can arrange for that, I think it would be a good addition. A couple pounds of preload ought to be enough.

tony ennis
11-20-2007, 09:32 AM
I've over-engineering through ignorance. I'll going to put a thrust bearing on each side of the left babbitt. My mad modifications will (get this) transfer force from the spindle nose, through a bronze flange bearing, and to the thrust bearing.

Previously, the lathe had been modified to transmit force from the spindle nose to the bull gear via its set screw, then to the cone pulley (which is spinning at a different rate from the spindle if the back gears are engaged), through the back gear, and into a thrust bearing. I was like WTF. The entire spindle was supported axially by a set screw. :eek:

Apparently one of the previous owners put in a bearing that was far too thick and in the wake of the mods discarded all those bothersome extra parts.

-=-=-
I also found that the bull gear and cone pulley had been rubbing together (of course, the thrust went right through them!) and ground their hubs down to the point the outer rims of the two components have touched and scored. I'll turn the hubs smooth and insert an oilite washer between them.

Carld
11-20-2007, 10:04 AM
Tony, I like your sig. line. Have you found a pictorial diagram for your lathe spindle? I kind of have the feeling you are, as you said overengeneering it.

I believe all the babbet lathes used the right hand bearing in the headstock behind the chuck to thrust from. I had a lathe that was made about 1900 and it thrusted off the bearing behind the chuck with no thrust bearing. It was a 18"x48" lathe.

tony ennis
11-20-2007, 10:38 AM
No, no diagram of a vintage headstock yet. I have a diagram from 1966. Perhaps it is representative of a 1939 headstock, perhaps not. I think I have a line on a better diagram, however.

The back end of the spindle has, from left to right, a threaded collar-with-setscrew, the leadscrew takeoff gear, and a plain collar. The latter is against the back babbitt. The only extra engineering I'm doing is replacing the plain collar thrust-bearing and pre-load spring. I could get by with just the collar but I do want the pre-load.

-=-=-
I stole the sig from a posting I saw here yesterday ;)

Carld
11-20-2007, 02:58 PM
Maybe you could put a thrust bearing on each side of the left hand bearing instead of a spring. You could adjust the end play as needed. I think you would be better off with the spindle floating rather than constant pressure on the thrust face. A little end play won't hurt and will let oil into the area.

The only lathes that I know of that have the spindle under tension are gear box headstocks and they are cone bearing/ball bearing combinations with flood/splash lube. On the other hand I have not torn into all brands of lathes yet so the door is open.

tony ennis
11-20-2007, 06:29 PM
Whenever the lathe is being used, there's probably a load on the thrust bearing. I don't see that keep 10 pounds on it at all times would hurt anything, and it would hopefully keep the spindle more stable while I'm turning something.

It's all board gabbing - there are probably 10 ways what all work about the same.

Carld
11-20-2007, 07:02 PM
Yeah Tony, it's probably not a big issue, but don't put to much load on it. If you use a wavy spring just put a little pressure, enough to take out the movement. You will be able to tell if it is to much by the wear or heat produced in the bearing. It's not rocket science.

rantbot
11-21-2007, 04:02 AM
Spindle design isn't something done by guess and by gosh. The original, before being dicked-with, almost certainly had a much stiffer arrangement than a thrust bearing with a few pounds of preload in one direction. A decent general-purpose lathe should be able to cut both ways, toward the headstock or toward the tailstock. A spring won't keep the spindle in place very well unless it's one of those belleville springs used to suspend railroad cars.

tony ennis
11-21-2007, 10:10 AM
Spindle design isn't something done by guess and by gosh.

Well, DUH, clearly it IS. :D


Unfortunately the Belleville springs had an OD that was too large. The original arrangement was a bearing on one side of the rear babbitt and a threaded collar on the other. Some people say there was no ball bearing, that the thrust bearing was just a brass flange. Other references say there is a ball thrust bearing in there. If the spring compresses when I cut away from the headstock then I'll remove it I suppose.

SGW
11-21-2007, 10:44 AM
Spindle design isn't something done by guess and by gosh. The original, before being dicked-with, almost certainly had a much stiffer arrangement than a thrust bearing with a few pounds of preload in one direction. A decent general-purpose lathe should be able to cut both ways, toward the headstock or toward the tailstock. A spring won't keep the spindle in place very well unless it's one of those belleville springs used to suspend railroad cars.

While that may well be true of higher-end lathes, I doubt it applies to the Atlas. It doesn't apply to my South Bend 10K. There is a thrust bearing for toward-the-headstock cutting only. When cutting left to right, the SB relies only on the spindle take-up nut running on a hardened washer against the left-hand end of the headstock. Given the infrequency of left-to-right cutting, it seems to be sufficient.

You do have a good point though, that any spring arrangement may be a bad idea, because a spring of any rational stiffness would allow movement -- fairly uncontrolled movement -- when cutting left-to-right. It may be better to do as South Bend does: reduce running clearance to a minimum by adjustment and let it go at that.

tattoomike68
11-21-2007, 10:56 AM
I'v tore down lots of spindles and not one ever had a spring. That sounds very cheesy to me. A thrust bearing and jam nuts would be just fine.

Carld
11-21-2007, 11:15 AM
It's clear to me that Tony has made his mind up to install a spring. When he takes a cut toward the tailstock he will find out the spring may be a bad idea. A thrust washer on each side of the bearing with end play removed would work. But that's not what he wants.

Everything we can say will not change his mind. Lathes were made for well over 100 years with babbit and thrusting on the babbit. They worked very well if kept oiled and clearances adjusted. If Tony feels that is inadequite then he won't be happy unless he changes it.

In the event it don't work as he expects he can revert to the original way. It's no big problem. If it works he will be happy and others will not.

tony ennis
11-21-2007, 11:56 AM
It's clear to me that Tony has made his mind up to install a spring

Uhh, well, I've made up my mind to consider a spring, lol


Everything we can say will not change his mind.

I try to be malleable. I'll drop an idea/approach in a heartbeat if it turns out to be a loser. My ego died years ago. I'm not the sharpest card in the lighthouse, but I know when I don't know.

There's nothing about rebuilding an Atlas spindle that's particular interesting for anyone but me - I post these things is to get opinions and see where the conversation goes.

Now rantbot says that the the setup may not be suitable for cutting left-to-right. So I have to wonder why. If I correctly follow his line of thinking, if the cutting force isn't 100% consistent, then the spindle will undulate (good word, that) axially as the cutting forces change. That's at best undesirable and almost certainly dangerous.

While it is obvious to me now, I never would have considered the adverse effect of the spring. But that's why I throw these newb questions out on the HSM.

Unless the spring is supposed to be installed completely torqued down I won't be installing it. Of course I ordered the parts last night. I have my first 'spare parts' now! :D

Carld
11-21-2007, 12:20 PM
I've never tried a spring as you are considering but my thoughts are on a heavy cut you'll not see a problem. On light cuts you may encounter load and release giving a thread like texture to the work. This is only what I think not a fact and it may not happen as I describe. I do think it will affect the quality of the cut in some way. To me it's like sitting around with a beer, whiskey, coffee, tea, etc. or all of them at once :eek: :D . you discuss what you want to do and get replies from all in attendance and decide the course to take. You won't know untill it's tried. Speculation is rampant, application is the test.

I hope you don't think I am being smart a**ed making the posts I have. I try to express what I see as problems or solutions and like you have long ago given up the notion I know everything. At 66 I no longer have anything to prove and everything to learn. If someone comes along with all the right answers all the time I excuse myself from the discussion.

tony ennis
11-21-2007, 12:23 PM
I hope you don't think I am being smart a**ed making the posts I have.

Not at all. I appreciate the input.

BadDog
11-21-2007, 12:35 PM
I don't know if this helps, but later Boyar Shultz grinders used a spring peload on the rear bearing assembly. Basically a spacer/collar with a spring-loaded ring. This ring was actually 2 rings with some pretty darned heavy springs (maybe 6 of them?) around the circumference and oriented to force the 2 rings apart. But this is a grinder, not a lathe. You can find a PDF download various places on the web that shows a cross section of the spindle and thrust collar.

rantbot
11-21-2007, 03:28 PM
Now rantbot says that the the setup may not be suitable for cutting left-to-right. So I have to wonder why. If I correctly follow his line of thinking, if the cutting force isn't 100% consistent, then the spindle will undulate (good word, that) axially as the cutting forces change. That's at best undesirable and almost certainly dangerous.
What you're doing is building in chatter. You have a mass - the spindle and workpiece - and when you add a spring, voila! A tuning fork is born.

Taking out all the free play, with a nut (a la South Bend) or a preloaded duplex ball bearing (a la Logan), or with babbit thrust washers with minimal clearances, is the way to go on a lathe spindle. Now to an engineer, even a socked-down nut is still a spring. But it's a spring with orders of magnitude higher stiffness than a small wave spring or belleville washer.

The thing to keep in mind about the Atlas lathes is that they were designed to be very cheap - but they were well designed to be very cheap. Atlas didn't simply take a fancy lathe spindle and leave expensive parts off until it became cheap enough to sell to Sears. They designed it from the ground up to use cheap components to do an adequate job. (Which is what engineering is all about.) Just swapping around simple components isn't like to give you any great improvment; more likely the opposite. Of course if the thing has already been bubba'd, that's another story.