What's new
What's new

How to measure the endplay on a precision bearing the right way ??

Joined
Jan 15, 2005
Location
The Netherlands
People ask me to measure endplay on machines I have for sale on a regular base
I always measure with a 0.001mm indicator and lift the spindle with a piece of wood
Not too much force applied
But it is soo easy to bend the shaft a couple of 0.001mm
I try to look for the moment the indicator slows down a bit while applying graduatly more force
Not perfect
Question I have now
With how much force you have to lift the shaft to measure endplay to prevent measuring bending of the shaft?
And how do you measure the force?

Peter
 
People ask me to measure endplay on machines I have for sale on a regular base
I always measure with a 0.001mm indicator and lift the spindle with a piece of wood
Not too much force applied
But it is soo easy to bend the shaft a couple of 0.001mm
I try to look for the moment the indicator slows down a bit while applying graduatly more force
Not perfect
Question I have now
With how much force you have to lift the shaft to measure endplay to prevent measuring bending of the shaft?
And how do you measure the force?

Peter
I used a 3ft long pinch bar or crowbar between the back of the spindle flange and the head stock casting. Using moderate force, Ie using one hand. This is on the usual size of shop lathe, say 15” swing by 5 or 6 ft bed length by the way. I had the DTI touching the face of the spindle flange or chuck front face and I’d expect to see some movement in the range of 0.00025 “ to 0.0005”.

Regards Tyrone
 
Seems like you could establish your own 'standard' by inserting a test rod of some set length and applying a fixed amount of force to the end (via a chain hoist and a load indicator) and establishing the deflection for known good and known bad examples. I know you restore a lot of those lathes, it's great to see someone dedicated to preserving them. If I was closer I'd get one from you, but I've been pestering my NL friend to hit you up once he has enough space for one.
 
Endplay?
This in my world would be in/out to bearing or length axis in a lathe.
Seems maybe you are talking lateral loading or run-out stiffness.
Bar bend is a easy calc. Horizontal guys way out there have to deal with this sag all the time.
A perfect bar held in the end in a perfect vee block horizontally will not say straight across the top due to this know sag.
This also pucks up people when scraping ways.
 
Endplay?
This in my world would be in/out to bearing or length axis in a lathe.
Seems maybe you are talking lateral loading or run-out stiffness.
Bar bend is a easy calc. Horizontal guys way out there have to deal with this sag all the time.
A perfect bar held in the end in a perfect vee block horizontally will not say straight across the top due to this know sag.
This also pucks up people when scraping ways.
Yes, you are correct, I was picturing deflecting them with a force similar to cutting. I wasn't thinking just the in/out axial play. Poor reading comprehension :-) Although the point would again be to apply some calibrated axial force (both ways?) and measure the difference between known good and bad setups to create some sort of baseline. Seems like that axial play would be a bit heat dependant as well, so maybe with the bearings run up to a certain temperature?

I guess the question I would have is, isn't this play also very preload dependent? I totally trust Peter, but couldn't you preload the crap out of the bearing to get a falsely tight tolerance for the sake of hitting some customer's demands?
 
I meant axial play indeed
Some bearings on lathes and milling machines do not have a preload Positive play of a couple of 0.001mm when installing a new bearing Too much preload is not good also
When I apply force to the bearing how much is lifting the shaft and when does it start to bend That is the question
 
On most lathe spindles I would pry like Tyrone toward the tailstock or wrapping a used a 4×4 to tap the spindle or wrapping a chain around the head stock and pried with a 4x4 using the chain as a fulcrum point to push the spindle toward the tailstock. Then mounted a .0001" indicator on thee chuck with indicator resting near the center. The ran the carriage up near the chuck and moved the tailstock up as far as it would go with nothing in the quill. Post a hard piece of wood and used the tailstock to push the spindle back toward the headstock.

As you have seen there are several ways to do it and they worked and didn't ruin the spindle. I agree on zero to .0.00005 movement.
 
When you talk about endplay on a lathe, then say "lift", are you referencing a vtl or are you actually measuring radial play?

Axial play or deflection is just that, movement of the shaft relative to the housing. It can be caused by a few things (not exclusively lack of preload) and depending on the design sometimes its a problem, sometimes its completely normal.

This is a pretty complex topic and we could really get into the weeds here, but since you are asking about lathes in particular thats what Ill focus on.

Virtually every lathe Ive seen has negative clearance (preload) in the bearings controlling thrust and certainly radial. Negative static preload is typically utilized for very high speed spindles as a way to accomplish the correct dynamic preload at the speeds the spindle will be used- - not lathes for the most part.

That being said, checking a lathe I would expect movement to be overcoming preload, and there is a significant difference between axial play and preload deflection. If you do have axial free play you have issues. If you push, pull etc and it always returns to the same spot, you are simply deflecting-overcoming the preload somewhat.

Now do you have the correct preload, thats a better question.

To answer your questions directly, the amount or force applied and resulting physical displacement will be different for every design. If you know exactly what bearings are in the spindle, in what quantities and orientations, you can pull a spec sheet get static load and map out your actual mounted preload to see if you are close. I use a load cell to measure the actual force being applied to the shaft, then an indicator on the shaft for the actual travel vs force, map at 4-5 places.
 
I’m not a spindle expert by any means, so please take my input with a grain of salt. As you know, I’m involved in spindle bearing replacement and adjustment of Schaublin 135 and therefore I spent some time learning the topic a bit.

To start with, I doubt there is a generic answer that is relevant to all lathes. I would assume it largely depends on the spindle design and the bearings used. So I think it’s always a good approach to first refer to the spec and manual of the specific lathe model to get some reference and understand the spindle adjustment procedure and how the measurements should be done.

With that being said, I think you might confusing different aspects:
free play is one aspect, runout is second aspect and rigidity/stiffness is a third aspect.
In addition, these aspect can be discussed separately in the radial and in the axial directions. Rigidity to tilt force might be also considered. There might be correlations between these different aspects but they can still discussed (and measured) separately.

free play to my understanding is mainly associated with the internal clearance of the spindle bearings. With modern spindle designs my understanding is that the internal clearance would be between zero and slightly negative (preload). To check the free play, light force should be used (e.g. try moving the spindle by hand with limited force) in order to avoid elastic deflection.
The rigidity of the spindle/headstock setup is another thing. In one of his videos, Dan Gelbart suggests a method of checking the rigidity. If I remember correctly he was talking about applying about 100kg equivalent force (~1KN?) using a u-shaped spring in the radial direction on the spindle and checking the displacement using dial gauge.

Dynamic runout (TIR) is correlated with other aspects, but normally should check for any machine.

To be super generic, I would say that for a used conventional lathe if most of the indicated values are less than 0.01mm one should be quite happy. one exception would be the internal clearance which should not exceed 5 microns.
 
Put your indicator somewhere else on the spindle casting and do your wood move. I bet in many cases you aren't seeing spindle play but flex of the machine itself.

As a watchmaker and former teacher of that skill, I used to put a 0.002mm Interapid on the bed of a Schaublin 102 (huge machine to most watchmakers haha) and then press here and there on the machine showing movement on the indicator, just to illustrate how elastic a big hunk of cast iron can be.
 








 
Back
Top