Used Lathe Runout?

[Dan Bevilacqua]

I'm always on the lookout for used lathes. Not really interested in old lathes but in a good deal on something more modern. When inspecting a used lathe, is "spindle runout" something that you inspect to make certain it is within a certain tolerance?

When I say "spindle runout", I am talking about measuring with a dial indicator the smooth portion of the spindle just behind the threads and also the inside of the morse taper of the spindle. I turn the spindle by hand and check for any deviation of the dial indicator in the rotation of the spindle.

Although I like to see zero runout, I understand that we are discussing wood turning lathes, and runout is not as critical as on a metal turning lathe. What would be acceptable maximum runout to you? Is anyone aware of an industry standard for spindle runout on a new lathe?

Thanks.

 

[Bill Boehme]

Any runout that can be seen or felt is too much. My first lathe had a bent spindle which caused the chuck to vibrate. When I figured out what was going on, I called Delta and they expedited a replacement spindle plus bearings and retaining clips. That made all the difference in the world.

My opinion is that runout is critical ... maybe not as critical for turning between centers with some caveats, but if you have an eight pound scroll chuck that isn't running true, that is a really big problem because that is vibration that can't be mitigated any way other than replacing the spindle.

This shouldn't be confused with bearing radial free play.

 

[Dan Bevilacqua]

Thank you, Bill. Is there anyway to quantify the amount of spindle runout that would be the cause of the vibration to which you referred?

Also, I am not familiar with the concept of "bearing radial free play". Is that play in the spindle from pressing laterally on the spindle?

 

[Bill Boehme]

Thank you, Bill. Is there anyway to quantify the amount of spindle runout that would be the cause of the vibration to which you referred?

Also, I am not familiar with the concept of "bearing radial free play". Is that play in the spindle from pressing laterally on the spindle?

It would depend the sturdiness of the lathe, the lathe speed, the mass of the chuck, and the degree of precision needed. For example, if I put my Oneway Stronghold chuck on the lathe (jaws closed and no wood mounted) and ran the lathe with a finger lightly resting against the back edge of the chuck, I would consider it unacceptable if I felt any vibration at any speed.

Standard tolerance bearings aren't preloaded and have a little radial play. That is the type of bearing most commonly used in wood lathes. That's not a problem and won't lead to vibration.

Most vibration begins with the wood. Low frequency vibration is caused by the wood being out of balance and it peaks when it happens to be at a structural resonant frequency. High frequency acoustic vibration occurs as the wood gets thinner and begins to flex resulting in the tool skipping across the surface. This is the type of vibration that leads up to getting a catch if you don't recognize what is about to happen.

 

[Gerald Lawrence]

It would depend the sturdiness of the lathe, the lathe speed, the mass of the chuck, and the degree of precision needed. For example, if I put my Oneway Stronghold chuck on the lathe (jaws closed and no wood mounted) and ran the lathe with a finger lightly resting against the back edge of the chuck, I would consider it unacceptable if I felt any vibration at any speed.

Standard tolerance bearings aren't preloaded and have a little radial play. That is the type of bearing most commonly used in wood lathes. That's not a problem and won't lead to vibration.

Most vibration begins with the wood. Low frequency vibration is caused by the wood being out of balance and it peaks when it happens to be at a structural resonant frequency. High frequency acoustic vibration occurs as the wood gets thinner and begins to flex resulting in the tool skipping across the surface. This is the type of vibration that leads up to getting a catch if you don't recognize what is about to happen.

Ok Bill you won me over and I will remove the plastic washer. I had a chuck stuck on my Delta Midi when I started turning (more than once) and upon reflection I think I probably powered up before seating the chuck and then lots of catches then.

 

[Bill Boehme]

Ok Bill you won me over and I will remove the plastic washer. I had a chuck stuck on my Delta Midi when I started turning (more than once) and upon reflection I think I probably powered up before seating the chuck and then lots of catches then.

Gerald, it seems like you may have read Leo's recent post in the thread about using setscrews to hold the chuck since he talked about using plastic washers. My post in this thread on runout was about a bent spindle. However, I don't use a plastic washer and have never had a stuck chuck except on my first lathe that did have a bent spindle.

 

[odie]

It's been many years since I checked the runout on my Woodfast spindle, but it was .002" to .003" when it was new. Now you make me wonder, since I replaced the bearings last year. I'm writing myself a note to check the runout tomorrow, and will report back later.

-----odie-----

 

[Bill Boehme]

I don't think that would have changed unless there was some scuffing going on when you replaced the bearings. With your machinist background, I would discount that possibility.

 

[odie]

Are any of the wood lathe manufacturers using tapered roller bearings for the spindle?

 

[Bill Boehme]

I think that there might be one or two that use angled ball bearings, but I don't recall reading about tapered roller bearings which are normally used for things like car and truck axles where there are extreme loads. I don't think that they would be a good choice for a wood lathe.

 

[odie]

I think that there might be one or two that use angled ball bearings, but I don't recall reading about tapered roller bearings which are normally used for things like car and truck axles where there are extreme loads. I don't think that they would be a good choice for a wood lathe.

There was a centerless grinder where I used to work, that had tapered roller bearings. The runout on that machine was very critical.

 

[Bill Boehme]

That sounds like an application where where tapered roller bearings would be justified since they can be adjusted to eliminate any slack. The downside is the power loss in roller bearings which explains why they run much warmer than the normally used ABEC 3 ball bearings.

 

[odie]

That sounds like an application where where tapered roller bearings would be justified since they can be adjusted to eliminate any slack. The downside is the power loss in roller bearings which explains why they run much warmer than the normally used ABEC 3 ball bearings.

You don't suppose the heat generated by the roller bearings is the result of more contact surface between the rollers and the races than would be with a ball bearing? Of course, the amount of torque applied would be a factor, too. Anyway, a ball would be a single point of contact, while a roller would be a line, which would be a much greater contact surface.

As I recall, there was a set of bearings for this machine that were greater precision than standard bearings. The difference was that each individual roller was mic'd and subject to a higher tolerance requirement. Of course, these bearings came at a much greater cost, but would produce higher consistency in the product.

-----odie-----

 

[Bill Boehme]

There's not much difference at all in contact area since a sphere or cylinder or a truncated cone rolling on another surface has a theoretical contact area of zero. In reality I'm sure that there would be more rolling friction. The big difference in friction is due to the hydraulic pressure as the grease being forced between the roller bearings and the additional back pressure created by the cage that holds the rollers in position.

 

[Hy Tran]

Runout: No such thing as zero runout. With better instruments, you will detect both synchronous and asynchronous error motion. And to really geek out, I much prefer the term spindle error motion to runout; it's more precise (see ASME B89.3.4:2010).

Anyway, I put a clone steb-center into the Morse taper in the spindle of my Nova Comet ii (about 5 months old, lightly used), and set up a dial indicator against the side of the steb-center. Hand rotation shows a total of 0.003" of TIR (all right, all right--that's the wrong term). But this number combines both the spindle error motion, non-roundness on the side of a turned surface (the side of the steb center), and misalignment of the axes of the steb center and the Morse tapers. I was too lazy to use the Donaldson reversal method (look it up, virtual applause if you know what his first name is and where he worked) to actually separate out the non-roundness of the steb-center from the spindle error motion.

 

[Dan Bevilacqua]

Thanks, Hy Tran. I looked up the Donaldson Method and didn't get very far. Too complicated for my brain.

By the way, did you happen to measure what I called "runout" and what you prefer call "spindle error motion" of just the spindle, inside the morse taper and just behind the threads of your Comet.

 

[Hy Tran]

Thanks, Hy Tran. I looked up the Donaldson Method and didn't get very far. Too complicated for my brain.

By the way, did you happen to measure what I called "runout" and what you prefer call "spindle error motion" of just the spindle, inside the morse taper and just behind the threads of your Comet.

Hello, Dan,

Unfortunately, I don't have a "last word" indicator at home, so I really can't measure the spindle error motion that you want to see. I have a magnetic stand (wouldn't work on a Robust lathe...) and a couple of straight dial indicators from a home project several years ago. Based on size of the indicators, the lathe, and the stand, I can't get the indicators inside the taper to measure, and there's not enough space outside the taper to get the indicator to read without bumping into something. I suppose I could chuck up something and turn a nice clean cylinder (using a skew, of course...) and then measure the error motion on the turned surface. That would actually be a pretty close representation to what you want. I'd need a really good hard wood with fine grain so the dial indicator doesn't dimple the wood; all the spindles in my garage are soft maple or cherry or elm (or Home Depot pine).

Hy

 

[Dan Bevilacqua]

Thanks again, Hy. No need to go to any trouble. I was just curious if you had measured it when you measured the side of the steb center.

In any event, I do have a Nova Comet II, and like it alot. I also just looked at your intro post, and noticed you started on a Shopsmith as I did. I still turn on my Shopsmith 10er.

 

[Bill Boehme]

Hello, Dan,

Unfortunately, I don't have a "last word" indicator at home, so I really can't measure the spindle error motion that you want to see. I have a magnetic stand (wouldn't work on a Robust lathe...) and a couple of straight dial indicators from a home project several years ago. Based on size of the indicators, the lathe, and the stand, I can't get the indicators inside the taper to measure, and there's not enough space outside the taper to get the indicator to read without bumping into something. I suppose I could chuck up something and turn a nice clean cylinder (using a skew, of course...) and then measure the error motion on the turned surface. That would actually be a pretty close representation to what you want. I'd need a really good hard wood with fine grain so the dial indicator doesn't dimple the wood; all the spindles in my garage are soft maple or cherry or elm (or Home Depot pine).

Hy

Hy, if you chocked up something and turned it round then it will be aligned with the true axis of rotation and not the axis of the nose of a bent spindle. This is easy to visualize if you have an Escoulen chuck or something equivalent that would simulate a bent spindle. To answer Dan's question, unless the runout is significant enough to cause noticeable vibration when a large chuck or faceplate is mounted on the spindle then it's not a problem.

 

[Hy Tran]

Hy, if you chocked up something and turned it round then it will be aligned with the true axis of rotation and not the axis of the nose of a bent spindle. This is easy to visualize if you have an Escoulen chuck or something equivalent that would simulate a bent spindle. To answer Dan's question, unless the runout is significant enough to cause noticeable vibration when a large chuck or faceplate is mounted on the spindle then it's not a problem.

You'd still pick up the asynchronous error motion. (Or in my case, the spectacular catches!)

That said, we're turning wood--not optics!

 

[odie]

I was able to check the runout of my spindle earlier this afternoon. As you can see, there really isn't a good place to set the dial indicator on the exterior of the spindle.

.....so, I was able to get a good angle on the interior of the Morse Taper. This doesn't necessarily mean I'm getting a good indication of the exterior of the spindle, but it's probably pretty close. Here I was getting about .00075", or about 3/4 of a thousandth of an inch runout.

I then installed a 1 3/4" SC faceplate, and took a measurement of the outside surface. Here, I was getting about .00175", or about 1 3/4 thousandths of an inch runout. This measurement probably includes some amount of error in the integrity of the faceplate.

-----odie-----

 

[Bill Boehme]

Well, I suppose that's good enough for cutting wood.

When you take the reading off a faceplate or spur drive or anything attached to the spindle then we can't really say what the measurements mean other than it's more than satisfactory, especially with the measurements that you got even with the faceplate

The wood is going to distort a lot more than that once it has been finished and and sitting on a shelf for a while.

 

[john lucas]

I just checked my two lathes. A Delta Midi and a Powermatic 3520A. Checking inside the morse taper like Odie showed I had zero runout on either lathe. I checked the face of the spindle which is also a critical alignment and no runout there either. I have checked other lathes and found as much as .003" but didn't have any complaints from the user on those lathes.

 

[Leo Van Der Loo]

Are any of the wood lathe manufacturers using tapered roller bearings for the spindle?

My large lathe has tapered roller bearings running in an oil bath, there is less deflection on that lathe than on my two smaller lathes that use ball bearings, the distance between the bearings does also affect the amound of vibration because of required clearances in the ball bearings, this is usually also not measured when rotating a spindle, as the weight will keep the spindle down, but when turning the spindle wil be lifted up and can give you vibration.

As to the OP’s question, I would think a total runout below 3 thou would be acceptable to me, less is better of course, though there is always some, even the grinding marks will give you some, especially in the fairly coarse machining that we normally find on the low cost far east machinery.

 

[Bill Boehme]

I just checked my lathes with my most accurate wooden yardstick and didn't detect any runout.

 

[Dan Bevilacqua]

I appreciate all the information. It should come in handy if I run across a good deal on a used lathe. I'll bring my dial indicator.

 

[odie]

I just checked my two lathes. A Delta Midi and a Powermatic 3520A. Checking inside the morse taper like Odie showed I had zero runout on either lathe. I checked the face of the spindle which is also a critical alignment and no runout there either. I have checked other lathes and found as much as .003" but didn't have any complaints from the user on those lathes.

Since I do have some amount of runout, even though very small......I wonder if the reason for that is in the manufacturing, a horrendous catch, or worn bearings. What else could be the case? For my Woodfast, the bearings were replaced last year, and my worst catches were on the first two lathes I owned. By the time I owned the Woodfast, I had been turning for about 10 years, and the really big catches I had early on were already past tense. So.....I suspect the spindle was originally manufactured with some amount of runout. It was made when most industrial manufacturers were making a transition to CNC machining.

Although, I was not involved in CNC production myself, I did work in a machine shop during the time it made the transition. My opinion about that, is CNC is capable of precision machining, if the program is made by a competent computer programmer, and sharp tools are maintained. On the other hand machinery made prior to CNC machinery is also capable of the same precision, BUT competent machinists are much harder to come by than competent computer programmers.

Anyway, I suspect the error is in the machining of my Woodfast spindle. Does it matter? As others have indicated, it doesn't matter, when discussing wood turning.....(Well, up to a point!) We all want things to be as mechanically perfect as possible, and in the real world, they seldom are. Things that do add to the accuracy of the finished wood turning, are a minimum of remounts. This thinking is basically why I've chosen to use waste blocks for the finish turning. Once installed, the faceplate is never removed, until everything is completed, and finish applied.......except for removing it and turning the foot. In my particular case, <.002" runout has never been problematic. Now, If I were using a chuck for the finish turning, and multiple remounts are in the picture......that will have a considerable effect on how well the details look to the observer.

Note: I do use a chuck for roughing unseasoned bowls.

-----odie-----

 

[Leo Van Der Loo]

You point out a common reason for inaccurate machining Odie, it is remounting a piece, the best way to get everything right is when the whole machining is done without remounting.

Mind you I have no experience with the newer ways of manufacturing and machining, but that was always the case, and of course cost of tools etc will have the tools used longer and of lower quality, things like that add up as is my experience, and I assume profit is still the main reason things get done or not

 

[odie]

You point out a common reason for inaccurate machining Odie, it is remounting a piece, the best way to get everything right is when the whole machining is done without remounting.

Mind you I have no experience with the newer ways of manufacturing and machining, but that was always the case, and of course cost of tools etc will have the tools used longer and of lower quality, things like that add up as is my experience, and I assume profit is still the main reason things get done or not

Ya.....all true, Leo.

Remounts.....that is a key element to precise machining, and to a degree it does apply to woodturning. It will depend on what level of details are desired. Many turners depend on embellishment done off the lathe, and turn simple, easily sanded shapes for the basis. Nothing wrong with that, and great embellishment is very artistic and admired......but, if the intent is to do detail work on the lathe, the only thing that will support that, is doing everything one can to maintain geometric integrity of the work-piece.......and sanding destroys geometric perfection.....especially in bowl turning, where the grain alternates between long grain and end grain. If one depends on less precision cutting, and more sanding, then the ability to include multiple turned details becomes a smaller option.

---------------------------------------

In the industrial world, profit vs quality is a balancing act that sometimes didn't always work out in the best interests of the consumer. CNC machines would often run 24/7 and without an operator. The operator was basically someone who mic'd parts and changed out cutting tools when specs weren't within tolerance. Leaving a CNC machine running unattended is a practice that was strictly profit motivated.....but, the machine itself didn't know when to stop if parts were not in spec. Although modern carbide tooling lasts many thousands of cycles, some do last longer than others!

-----odie-----

 

[Dan Bevilacqua]

Anyone know whether there is an "industry standard" spindle "runout" tolerance, and if so, what is it?

 

[Bill Boehme]

Anyone know whether there is an "industry standard" spindle "runout" tolerance, and if so, what is it?

That would be a manufacturer's performance spec on metal turning lathes. I never heard of that ever mentioned for any wood turning lathe. The amount of runout typically measured is far less than the elastic and plastic deformations of wood as it is spinning on a lathe.

 

[Mike Johnson]

A runout standard like that would have to be based on the parts that need to be manufactured
on the machine. Plus or Minus accuracy moving to the right of the decimal point adds to the dollars
adding up to the left of the decimal point.

 

[Hy Tran]

Anyone know whether there is an "industry standard" spindle "runout" tolerance, and if so, what is it?

Short answer: No.

Long answer (don't bother reading if you're not interested):

Your various machine tool manufacturers are playing trade-off:

• Satisfying customer needs (whether real or perceived)
• Making sufficient profit (cost to design/develop/manufacture/support vs lifetime profit)
• Strategic positioning (compare, say, Thompson tools vs Benjamin's Best)

ASME B89.1.5 and B89.1.6 (plug gages and ring gages, but I may have these backwards) describe tolerance classes for round objects, typically used in mastering the output of precision machine tools--but since these standards were originally written, there has been a lot of evolution in understanding uncertainty in measurements and in risks. No one makes "triple-X" gages anymore--because they really can't be economically certified when accounting for measurement uncertainty.

The ASME machine tool stds, B5.54 for machining centers and B5.57 for turning centers describe performance evaluation methods. Note that these are called "performance evaluation methods". They don't give tolerance classes. Not speaking either for industry or for ASME, but rather as an opinionated person: Industry really does not want to set standards for spindle runout (btw, please call it error motion)--what they want to do is to establish maximum permissible error in machine tools, of which spindle runout is only one aspect. Considering the system as a whole, you can trade more runout for higher stiffness, damping, and durability, and have a machine that's a speed demon.

B5.54 and B5.57 were written by committees comprising academic researchers, scientists and engineers working in the government sector, folks from NIST (actually, probably NBS at the time, National Bureau of Stds), machine tool manufacturers, and end-users, such as aerospace and automotive manufacturers.