Machine spindle runout

Machine spindle run-out question?
Why is spindle run-out in a lathe, surface grinder, or mill so important to be as minimal as possible? Three jaw chucks on lathes run out about 3 thousandths of an inch anyways, and on a surface grinder, the grinding wheel is usually dressed true anyways. So why does run-out of the spindle matter so much anyways? Thanks, Brian.

In short and in my limited vocabulary...If the spindle of a mill runs out the result will be end mills that cut with one flute, reamers that act like boring bars. Oversize slots and oversize reamed holes, shortened tool life...If this is troll post from the HSM site I've just been sucked in and wasted 3 minutes of my Saturday :-)

On a lathe if the spindle has run out, your part will be out of round. Which is different than the work in a three jaw chuck running out. In that case, the work will be turned eccentric to the "chucked" diameter.

If your spindle is truly running out, then the ecentric wobble of the spindle will be reproduced in the turned diameters. Obviously this eats up your tolerance. If you are turning to a 0.001" tolerance, runout of 0.0005" would eat up half of your tolerance.

In a milling machine spindle run out causes the cutter to run out and cut unevenly. It also makes it difficult to cut features squarely.

In a surface grinder, it causes vibration and wheel imbalance resulting in poor finishes. On a centerless grinder it results in out of round workpieces. On a cylindrical grinder runout in the work spindle causes runout in the work.

Etc...

There are two sorts of runout - repeatable and non-repeatable. If you don't remount the part then the repeatable part goes away but the non-repeatable (due to bearing variation) doesn't.

With grinding it's a surface finish issue too.

So if you wan't really round - you need good bearings.

Chris P

First things first. A machine tool spindle provides an axis of rotation. It's bearings constrain the axis of rotation relative to the machine both radially and axially. The spindle is made of durable materials treated and precision finished for long life. It's proportioned so that it can withstand the cutting forces imposed on it with a minimum of deflection. On the spindle are mounting features (threads, tapers etc) for workholding or toolholding equipment.

The closer the spindle rotates to its theoretical axis the more accurate it will act on the work. The stiffer the spindle bearings (axially and radially) the better.

Plain bearng spindles run in clearanced bearings on a film of oil. While the plan bearing is by far the simplest and most economical to manufacture it has the disadvantage of being "soft;" that is, its axis displaces in rough proportion to the forces acting on it. The plain bearing depends on a supply of oil the failure of which often result in severe damage to the bearings and journals. Because the plain bearing "integrates" its inevitable manufactured-in errors, the quality of its axis of rotation is almost perfect when under a constant load RPM, and lubrication supply.

Rolling element bearings are stiffer centralizing their axis of rotation to 1/10 to 1/100 that of a comparable plain bearing. When made to the highest classes of accuracy they approach the plain bearing in the quality of the axis they generate. The RE bearing has no lower RPM limitation and it's upper RPM limit can be extended via lubricant cooling etc. Modern RE spindle bearings in VMC's run to 30,000 RPM when grease lubricated and have dynamic stiffnesses less than 0.0001" per 1000 lb load and their trace of rotation error is less than 4 micro-inches. I refer to high end VMC and turning center spindles whose costs represent a significant fraction of a $150,000 capital investment. Home shop grade equipment (unless re-cyled higher end assets) can attain but a fraction of that performance because of economics, pure and simple.

In most new machine tools, their spindle reference features are finished only after assembly in their bearings and housings, in place, as nearly the final manufacturing step. Thus the spindle nose, tapers, etc typically have TIR in the tens of millonths to the small tenths ground and finished as they are from the installed spindle bearings.

Excessive spndle nose/taper run-out is symptomatic that the RE bearing have been replaced by shade-tree wrenchers possessing tools but not the knowledge of how spindle work proceeds. Quite possibly they down-graded the spindle bearings to auto part store level thinking that there can't posibly be that much differernce between a matched $350 pair of ABEC Grade 7 DU angular contact bearings and a $16 pair of mport bearngs from eBay.

Note to shade tree guy: oh but there is a difference and the difference will bite you where you bleed worst when you next machine a bearing fit.

bcstractor said:

There are two sorts of runout - repeatable and non-repeatable. If you don't remount the part then the repeatable part goes away but the non-repeatable (due to bearing variation) doesn't.

With grinding it's a surface finish issue too.

So if you wan't really round - you need good bearings.

Chris P

Click to expand...

Ok thanks guys, this helps me understand this now, I think I understand the repeatable and non-repeatable runout. I know that with repeatable runout, if a part is not remounted in the lathe, all diameters will be turned at once and the runout will diappear. However, if it is non repeatable runout due to loose bearings wobbling, the part will not come out round. I'm thinking this is because the cutting pressure can push the spindle around, and the part will not be held steady.

As in a mill, I understand that if the cutter is running out, I know that one side of the cutter will be doing more cutting than the other. However, if the spinde is running out due to loose bearing (not bent), I would think the cutter could wobble around and not cut flat or square. If the spinde runout was from it being bent, i think this would be repeatable runout and a part could still be cut flat and square, just one side of the cutter would cut more than the other side; wearing the tool faster.

As with a surface grinder i know that wheel runout from a wheel not being properly dressed true to the spindle, will cause a poor surface finish. However, I'm thinking if the spindle was running out due to loose spindle bearings, the part would not come out flat/square because the spindle can move (non repeatable runout). If the spindle runout was from it being bent it would be repeatable run out and the part would still come out flat/square.

Let me know if I am thinking of this right.
Thanks again, Brian from PA.

Forrest, can you explain your statement please? What do you mean by "integrates" ?

"Because the plain bearing "integrates" its inevitable manufactured-in errors, "

Hi Forrest, this statement seems to go against my understanding and many statements that I find with google, for example:

"when correctly loaded, fluid bearings are typically stiffer than rolling element bearings"

reference:

http://en.wikipedia.org/wiki/Bearing_(mechanical) .... stiffness

Am I missing something.

Thanks
Phil

Forrest Addy said:

Plain bearng spindles run in clearanced bearings on a film of oil. While the plan bearing is by far the simplest and most economical to manufacture it has the disadvantage of being "soft;" that is, its axis displaces in rough proportion to the forces acting on it.

Click to expand...

Forrest Addy said:

In most new machine tools, their spindle reference features are finished only after assembly in their bearings and housings, in place, as nearly the final manufacturing step. Thus the spindle nose, tapers, etc typically have TIR in the tens of millonths to the small tenths ground and finished as they are from the installed spindle bearings.

Click to expand...

Nice post- thank you.

In regards to the above. Does this imply that any replacement of spindle bearings will by virtue of not having the reference features also reground result in degrading the accuracy of the spindle/bearing system?

And what is considered as a "new machine tool"?

I ask as I am about to embark on the rebuilding of a 40's Hardinge lathe would like to proceed in a fashion which maintains as built specifications as far as possible.

Thanks much

Machine spindle runout.

Brian,
To only add to some very good comments.
Runout will affect tool life, finish, chatter amoung other nastys.
If you had repeatable runout in a milling machine spindle, and you were running a large face mill, you could see the steps in the piece part. For example, if the runout were say .010" on a 10" face mill, you could see the steps on the milled surface. The highest (longest) tooth would take a cut .010" deeper than the lowest (shortest) tooth on the cutter. The teeth between them would each take a little more and then a little less after it passed the "highest" tooth.
This would be more apparent if the machine has a little "heal" built into the squarness of the machine spindle centerline.
I havew seen this in the bigger machines (American Hydro, York)
and (BMY, York), to mention a couple of shops local to you.
Next time you look at a finished part, look at the finish and try to determine what made the finish look like that..
Regards Walt...

Modern anti-friction, or rolling element, bearings from a good manufacturer are very high quality indeed.

But plain bearings are still used for very high cyclical loads, especially at low speeds, and super precision bearings used in certain grinders and machining centers.

For this purpose plain bearings are divided into 2 main groups:

1) hydrodynamic bearings, as typically found in automotive engines and lathe headstocks such as ancient South Bend lathes.

In these bearings an oil wedge is formed at a location almost opposite the load point by virtue of the lube oil clinging to each side of the bearing and being dragged around. Any radial load then causes an eccentricity between shaft and housing leading to formation of the oil wedge which supports the load. A drawback of this bearing is the metal to metal contact in the bearing during start-up and run-down conditions. This can be alleviated to a large extend with a hydraulic lift pump supplying oil to the bottom of the bearing prior to start-up. A running clearance is required and, depending on the size of the load, the final position of the spindle in the bearing is uncertain within the limits of the clearance. A milling or a grinding spindle could not tolerate this condition.

However, a variant of this basic design has been used for grinding spindles with success, but low-cost rolling element bearings are now preferred.

2) hydrostatic bearings, as used in turbo-generators and super precision machining. Air bearing spindles for tool & cutter grinding are another example of this type.

For this bearing oil is supplied under pressure to a number of oil pockets surrounding the shaft.

Hydrostatic bearings can be made very stiff (little deflection under load) with appropriate design and operation. As long as oil can be supplied in quantity and at requisite pressure the running clearance is not that critical.

With finish machining of the spindle locating features while running in its own bearings, spindles with this type of bearing can run true in the sub-micron range while very stiff.

In principle these bearings are not difficult to make, but their drawback is that they need a reliable hydraulic system to supply hydraulic/lubricating oil under fairly high pressure. Any failure of this oil supply while under load would wreck the bearing.

These bearings are a fascinating subject!

Arminius

On a lathe if the bearings are loose and allow the spindle to run out, the workpiece won't come out round. So on a surface grinder or mill, the workpiece won't come out flat?

Machine Spindle Runnout

Creedtown182
First, I think you are looking for some always firm fast rules.
You should learn not to use the terms always, everytime and words like them. Usually when you have runout, the quality of the finished part will be affected. In the case of milling machines, the finish may be affected. Most of the steps may or maynot be removed providing the "heel" of the spindle centerline is almost nil or even non existent. If the machine is square, the cross hatch will maybe make the finish acceptable.
In the case of a lathe, I don't think the part will necessarly be "out of round". You will end cutting on the average of the runout and the runout from the bearings will keep changing. Therefore, maybe the finish will be unacceptable but the average diameter of the part will be overall the same.
Regards Walt...
Be back in a little while, and talk about how to check for bearing runout....Walt...

Machine spindle runout.

Creedtown182
When I am checking runout (milling Machine), I use a test bar is available. If a test bar is not available, look for a longer tool that may have a ground surface out near the end of the tool holder. I havew used a longer collet holder (w/o the collet), and indicated on the inside of the holder on the collet seat area.
Rotate the spindle at a slow speed 10 to 20 rpm and indicate the collet seat surface. If the indicator goes from .0000 to .0020 and then back to .0000, you have .0020 runout. This would probably be coming from the spindle taper or maybe from the tool holder itself. However, if the indicator moves from .0000 to maybe .0030 and then back to .0005 and then back up to .0035 and then back to .0000, the indicator would suggest you have about .0030 runout inthe tool holder or the spindle and about .0005 runout in the bearing. The bearing runout will no doubt keep moving around and the indicator will not repeat.
Hope this is not too confuszing. I can't spell that word.
Regards Walt.....

Creedtown182,
The three machines you are asking about perform three different types of machining operations. I can't go into any more detail about bearings than what has already been documented. However, I would like to address your understanding of the concept of flatness. In the case of a surface grinder, flatness is entirely dependent on the quality and flatness of the linear ways. The rotating grinding wheel only determines surface finish. I.E. Wobbly wheel = bumpy rough finish. The same concept applies to milling machines. I think if you were able to locate a copy of "Foundations of Mechanical Accuracy" by the Moore Tool Company, you would have a valuable addition to your library of documents describing the nature of mechanics relating to the accuracy and precision of machine tools. WWQ

Ok thank you guys for all your responses, I really do appreciate it and you guys really did help me understand a few things. I'm a newbie with all this stuff and really have only been working with machines for the past few years. Went to vo tech night classes from 2004-2008 for precision metal machining. I'll print this thread out, that way I have it to read over from time to time. I will also look for the book you mentioned bill.
Thanks again guys, Brian.

Here's a question say you have an internal grinder that is spinning at maybe 20K rpm, if it has a few 10th's run out why does it matter so much? At 20K rpm, wouldn't one expect that it would average out to the highest spot on the grinding wheel and become almost irrelevant?

I am sure I am missing something here, but I am curious why it is? I must add a disclaimer that I have never used such a grinder either, so it could be something more obvious.

Adam

Good question Adam. I couldn't tell ya on that one. I was running the okamoto surface grinder tonight at work and I was grinding some hardened tool steel (62-65 Rockwell I think) with a CBN grinding wheel. We only dress the aluminum oxide wheels and not the CBN or diamond wheels after we start the machine. (Probably because CBN and diamond wheels are expensive, I don't know they told me to only dress the aluminum oxide wheel.) So with the CBN wheel not dressed after it was mounted it had some repeatable run out, and wow did affect the surface finish. Running at .007" crossfeed, the surface finish was really wavy and bumpy (.005" depth of cut). I slowed the crossfeed to .004", and it helped the surface finish but there was still and wavy/bumpy finish there.

Later when I had to go to an aluminum oxide wheel I put the tenth dial indicator on the taper of the spindle and it had .0001" runout, I thought i saw the indicator go to .0002" once but that might of been the surface finish on the taper affecting the indicator. Maybe the reason for the runout was the taper wasn't machined perfectly true with the bearing journals on the spindle, or maybe the reason for the runout is the spindle is bent ever so slightly?? (someone did blow apart a grinding wheel on the machine at one time while dressing the wheel too hard). Pretty sure it was .0001" runout because I thought the surface finish of the taper was affecting the indicator, so I moved the indicator and it read .0001" once again. I couldn't feel no side to side movement in the spindle, but I didn't check endplay movement. (The machine is about 7 years old).

The other machine I run is a Wasino, and that has spindle bearings that are starting to be very noisy. (just started this past friday when i was running it) I can feel slight side to side movement but i haven't indicated it. For some reason though we don't see bumps in the finish on the optical comparator though? (Machine has an opitical comparator) Maybe because like you said Adam, the feedrate is slow enough and the wheel speel is probably 20k+ rpm, and even though it is non repeatable runout, it is smoothing out the high and low spots??

chuck run out

your comments are quite interested guys, but nobody mention ways or tricks to fix these problems, i have a mori seiki cnc lathe, it don't matter what i do, its always running out .002 or .003. I'd replaced couple parts, and machined soft jaws in different ways, but its always off, any suggestion??