Spindle runout acceptable?

I have fitted new taper roller bearing to a Hobson H7 lathe headstock spindle. The lathe has yet to run under power but,there is 0.0015 " total indicated runout on the spindle nose register,diameter 1.5". The preload on the bearings has been set as per Boxfords, ie 2lb ins force to rotate the spindle from rest. My question is the runout acceptable ? or have I made an error in fitting the bearings,this a hobby lathe so no working to tenths.Also should the bearings be allowed to run in before final adjustment and runout measurement taken? Thanks for any advice or reassurance given.

regards to all Paul

Hobson h7 lathe

Is yours the 3.5 version?

Yes it is the 3.5" over the bed smaller version, I should ad the bearings were Timken, not ultra precision but not budget bearings.

regards Paul

Hi Paul

I would expect (and want) my Myford lathe spindle runout to be much better than that, but I guess whether that's acceptable or not depends on what you want to do with it. A chuck will run out more than that, so it would still be usable.

But, I doubt the error is in the bearings. Much more likely the original spindle or it's fit to the bearings.

Perhaps check the runout on the morse taper as a double check ? It's always possible someone did something funny to the spindle nose at some point and as a last resort you could always skim it in situ and machine a non-standard chuck fitting for any new backplate.

Cheers - Will

Putting too much grease in the bearings can mess up the Boxford settings. It should just be enough to lightly coat the parts of the bearing.

When I've tried it the initial reading due to spinning the parts up even as slow as that is needs to be ignored and the readings taken when things have settled down hence 4ft of rope or what ever. I also think the machine has to be run to distribute the grease but there is a catch in that. The bearings will warm up and tighten so need to be left to cool down again. That will take some time.

This subject crops up on the yahoo boxford site at times and many including me have found that the pre load is too low. Easy to tell because if a decent sized cut is taken on a rigid piece of steel it will take another cut without any change of setting - more than what might be due to the work bending. The problem can be down to wear in bearings as they are bound to wear oval over time as the forces on the outer shell are always the same when the lathe is cutting material.

What I did in the end was to use bearing heat by feeling the area under the rear bearing via a finger in the end of the spindle. I'd run the lathe for at least 15min at 500 rpm or so and then check for definite warmth - not hot but very warm. It would be best to then do the same thing with it running flat out as a check.

It's worked for me and I just have a tiny recut that I think is down to bearing wear but I haven't changed the bearings yet. I had the spindle out but the front bearing inner is way to tight for me to remove and replace at home. It really needs a bearing press and probably a heater as well.

When I bought the lathe initially I found that the bearings loosened each time I drilled a 1in hole or did some heavier turning. This must have been because they were not seated correctly. Not that uncommon a problem.

John

-

The lathe in question is a Hobson H7 which is an oil lubricated gearhead lathe, grease should not be involved.

In addition 2lb in force strikes me as very low.

A possibility is that one or more of the bearing races haven't seated fully yet. The minimal preload that is needed is not enough to do this.

Neil

Boxford did change the bearing types used at some point so the settings might not be appropriate any more. There is some comments on lathes co uk about that. From memory they did use taper rollers that had a lower cone angle and then switched to the standard cone angle ones which are also much cheaper. As far as I know they don't use super precision bearings in the older lathes but still achieve much better run out figures than that. They test to a max error of 0.0004" on the spindle and reports I have seen show 0.0002" measured. Taper on work max of 0.001" per foot, measured 0.0007". Some come out better. The taper can be corrected by straining the bed but that needs a substantial base for the lathe to sit on. Mine shows figures that are much better than this so I'd guess that the bearings do run in a bit.

I assume that the torque was measured without the spindle driving any gears ?

Bearing heat should still be an indication on a gear head lathe really. There will be some friction when they are running correctly. My reason for doing this goes back to training. Much bigger lathes but it was pointed out that lathe bearings ( gear heads) run at over 100C. Easily noticeable by touching the head stock around the bearing housing. They get very warm. A number of the machines will have had 2 angular contact or roller bearings fairly close together at the front so more sources of heat in that area. Anyway after going through the exercise of adjusting them I can now tell when the lathe has warmed up. Very very light cuts show marks on the finish which are caused by a wobble on a counter shaft pulley. The repeat pattern it generates can be seen on this

Only a few micron deep but irritating. The shot is larger than life and the finish feels glass smooth.

Machine the spindle - no way. If needed make new chuck back plates as machining the register on them will have the same effect. In fact they may have already had that done. It is possible to true up morse tapers with a reamer but DO make sure the tail stock one is correct first. That can still be trued up in the same way even if there is a slight error in the spindle. This should only be done after it's been set as accurately as possible via it's adjustments and shimming if needed and is only intended to remove a few thou not correct gross errors. The only sensible way of checking between centre alignment is via turning tests or the use of a test mandrel and dti and even then it's possible to miss interpret the results. Parallel between centre test mandrels don't seem to be available and the ones with a morse taper on the end usually have a hole in the end to take a draw bar, not much use really.

The usual way of truing up a tail stock morse taper is to fit the reamer in the 3 jaw and push the tail stock up against it with the quill part wound out. It can only be done if morse fitting still have some taper sticking out when fitted. This assumes a well fitted 3 jaw chuck.

I think the OP would best do some heavy turning with the machine and then check things again. There are also some simple turning tests that can be done to check the alignment of the spindle to the bed but bed wear can mess these up a bit. Excessive recutting with a tool at the same setting is a good indication of loose bearings. If very bad the same setting will take several noticeable cuts.

One thing most important on lathes - don't jump in. Make sure of what the problems really are before attempting any corrections otherwise things will probably get worse.

John

-

Edited By John W1 on 05/10/2015 17:10:46

Many thanks to all of you, for the trouble you have taken to enlighten me.I will run the lathe as is and take new measurements after everything has "bedded in" On the question of bearing preload, searching the net just added to my confusion as apparently authorative sources gave conflicting advice. The boxford protocol seemed the least damaging if inappropriate for this lathe. The lathe appears to be a Myford clone with "improvements" viz geared headstock, heavier spindle with larger internal bore, No 3 morse taper in the spindle nose. Are there any suggestions for initial preload sttingsfor the taper roller bearings or should I just wait and see if they get too hot?They are now lubricated by nice new full synthetic SAE 75 albeit only by splash, which seems a bit haphazard,my previous experience is with nice plain bearing with high pressure wedge lubrication at much higher revs and loadings. Thanks again for all the assistance afforded to me.

regards Paul

It is always interesting to listen to the arguments about spindle bearings "tightening up" due to heat build-up.

On a boxford (or any other machine fitted with opposed taper rollers, any heat generated in the bearing will dissipate into the supporting structures and surrounding atmosphere. As such you have to consider the relative expansion due to temperature for all the parts. In the case of a Boxford, you have a relatively long spindle (about 220mm between bearing faces) and a small rolling element in the gap between the race ways – perhaps 6mm diameter (mean diameter of tapered roller). This means that for the bearing to reduce its clearance the rollers have to expand enough, and quickly enough, to over compensate for the extension of the spindle relative to the (comparatively large thermal mass) head casting due to the heat input from the bearing. It also assumes that the spindle will expand the same or less than the head casting – extremely unlikely given the thermal mass of the two parts and their design/location (the spindle is going to be hotter than the head casting therefore it will expand more and result in an increase in bearing clearance). It is also important to remember that heat transfer is in part related to temperature difference, the greater the thermal gradient, the faster the rate of transfer. This makes it very hard to generate a heating cycle that will ultimately result in a bearing failure due to thermal expansion alone. If it is possible to seize a spindle with heat, I suspect it would be down to lubrication failure rather than simple heating due to loading especially when you consider that the bearings will carry loads measured in tonnes rather than a few hundred kg. I have not taken the time to do the calculations but there are a number of investigations that have been published on the subject of heat transfer/dissipation in rolling bearings and given the time frames involved on a lath it would require extreme conditions to generate the temperatures required in the time frames required to generate a seizure.

To complicate matters more, the rollers are tapered and moving so they would apply more axial load rather than simply expand radially as they run on the corresponding tapered race (due to the wedge action of their form). Taking the Boxford spindle again as an example, it would be quite feasible for it to extend due to elastic deformation under the working loads for the bearings so you have at least two mechanisms trying to reilive any thermal issues in the bearing.

I don’t exclude the fact that a spindle appears to get “tight” when it gets hot but I am speculating that it is not due to thermal expansion but more likely due to lubrication or some other cause?

Mark

^^^^ This.