Exploring Run Out On The Rf-31

mikey

Platinum
Staff member
Rest In Peace
Joined
Dec 20, 2012
Messages
9,422
My new-to-me Taiwanese RF-31 had excessive runout. I knew that because I could see the end of a drill bit wobbling as it ran. I thought, “Hmmm, that doesn’t look right …”, so I put an indicator into the spindle taper in an attempt to put a number to that wobble and it turned out to be around 0.002-0.003” and possibly more. I say “around” because the indicator needle was bouncing so bad that I couldn’t tell if it was just bad or really, really bad. I asked what other guys were seeing but got little response so I figured I would give it a go and report it here.

This led to a complete tear down of the mill, partly to clean and recondition it and partly to sort out where that runout was coming from. This is how I discovered that the design and components found in this machine are exactly what you will find in most Asian drill presses, including a round column. It’s a drill press on steroids but it is still a drill press and we all “know” what “they” say about milling with a drill press.

To be fair, there are a number of things that make it suitable for milling – it has a solid 3” quill that runs through a 5” X 5” part of the cast iron head stock; it has tapered roller bearings supporting a solid steel (but un-hardened) spindle and a solid cast iron drive sleeve supported by spaced ball bearings that are held in place with a ring instead of a simple Circlip like a drill press has. It would be world’s better if the spindle/quill ran in a solid, machined cartridge but it doesn’t. On the other hand, the design makes bearing changes easy and I suspected the runout was coming mainly from the bearings that have lived here for over 20 years.

If you look at the design of the head, there are only a few things that can contribute to runout:
1. The spindle itself can be bent, the R8 taper can be really worn or there may be a defect inside the spindle taper that the indicator needle is bouncing off of (not likely).

2. The spindle bearings might be worn, loose or the preload may be inappropriate. The tapered roller bearings in this machine felt tight but they were full of dirt and dust particles and were never, ever maintained so this was a good candidate for much of the runout.

3. The drive sleeve, called the spindle taper by Rong Fu, can be worn inside where the spindle splines ride but this is not a precision fit. Mine felt smooth and spindle ran with very little play so I didn’t think this was it. However, the bearings that support and locate the sleeve do influence runout and if they are worn or are moving in the mount then that has to be addressed. I noticed that when I yanked on the main drive pulley I got an additional 0.006” of movement so I suspected those bearings were pretty much shot.

4. The main drive pulley is a press fit onto a taper on the drive sleeve so any vibration from the belts will transfer directly to the drive sleeve and thus to the spindle. While this doesn’t affect runout directly it does get transmitted to the spindle and can affect finishes, especially if the drive sleeve bearings are bad. On this particular machine, when run at 1800 rpm there is some harmonic that causes the rubber belts to vibrate badly; it almost sounds like a loud groaning (the scary kind, not the sexy kind. The hair on the back of my neck actually stood up when I first heard it!). In contrast, the jackshaft pulley bearings are tight, smooth and will be left alone.

5. The quill itself is another issue that can cause runout; to fix it you need a new quill. There were no scuff marks on the quill to indicate a problem with the quill or the passageway so replacing the quill was the last resort if I couldn’t fix the runout any other way.

I decided to look at and address each of these concerns and check static runout after each change to see what, if anything, makes a difference. My guess was that I should be able to improve the runout but getting anything under 0.001” … wasn’t sure.

The first thing I did was to change the tapered bearings to sealed angular contact bearings using the original spindle. This changed the runout very little and there was still a lot of bouncing of the indicator so I ordered a new spindle from Enco and installed it after switching the new spindle bearings over.

With the quill placed in blocks and clamped firmly on my work bench I measured 0.0001” TIR inside the new spindle taper. This confirmed that the new spindle is straight and is running reasonably true in the quill and bearings. Since my quill runs nicely inside the head I was relieved that I got away with not having to replace this quill.

I then installed the quill into the machine and again found the indicator needle bounced a lot but with less range. This confirmed that the drive sleeve was definitely having a negative influence on runout.

The next step was to pull the drive sleeve and change the bearings. The sleeve itself is in pristine shape – no wear was seen and all the machining was still sharp. I replaced the OEM 6009-ZZ bearings with Nachi Quest 6009-2RS bearings. These are deep groove bearings (can handle high radial and axial loads), sealed on both sides and permanently lubricated. Once the drive sleeve bearings were replaced the bouncing was reduced dramatically and runout decreased to a little less than 0.0001” TIR, the same as I found at the bench. I checked runout with the quill elevated to several positions and got the same result … finally.

With the runout addressed I decided to replace the rubber belts with Fenner Power Twist Plus belts. Fenner belts are relatively expensive when compared to rubber belts but they will not wear nearly as fast as the rubber belts, they stretch less once broken in and they run smoother and quieter. The RF-31 uses a B-type belt, which is 5/8” across. You need a little over 6 feet of belting to make the two required belts.

After changing the belts the machine does run noticeably quieter and smoother. That nasty harmonic and groaning at 1800 rpm is now gone so I’ll consider this handled. If the machine won’t finish nicely now then I don’t know what else can be done.

So, it turned out that my spindle was worn. I suspect that the spindle bearings were shot and that accelerated the wear inside the taper but replacing both sets of bearings was necessary to get the spindle to an acceptable level of accuracy. I'm not sure which was more important but if I had to guess I would say the spindle bearings make the most difference. I took some pictures and will post some “how to change your bearings” threads in the near future in case someone who hasn’t done it before wants to give it a shot.

At least now I know that the RF-31 spindle runout can be reduced to about 0.0001” or less. As to how accurate the machine will be, that remains to be seen because measurebation is one thing; cutting accurately is another.



Mike
 
My new-to-me Taiwanese RF-31 had excessive runout. I knew that because I could see the end of a drill bit wobbling as it ran. I thought, “Hmmm, that doesn’t look right …”, so I put an indicator into the spindle taper in an attempt to put a number to that wobble and it turned out to be around 0.002-0.003” and possibly more. I say “around” because the indicator needle was bouncing so bad that I couldn’t tell if it was just bad or really, really bad. I asked what other guys were seeing but got little response so I figured I would give it a go and report it here.

This led to a complete tear down of the mill, partly to clean and recondition it and partly to sort out where that runout was coming from. This is how I discovered that the design and components found in this machine are exactly what you will find in most Asian drill presses, including a round column. It’s a drill press on steroids but it is still a drill press and we all “know” what “they” say about milling with a drill press.

To be fair, there are a number of things that make it suitable for milling – it has a solid 3” quill that runs through a 5” X 5” part of the cast iron head stock; it has tapered roller bearings supporting a solid steel (but un-hardened) spindle and a solid cast iron drive sleeve supported by spaced ball bearings that are held in place with a ring instead of a simple Circlip like a drill press has. It would be world’s better if the spindle/quill ran in a solid, machined cartridge but it doesn’t. On the other hand, the design makes bearing changes easy and I suspected the runout was coming mainly from the bearings that have lived here for over 20 years.

If you look at the design of the head, there are only a few things that can contribute to runout:
1. The spindle itself can be bent, the R8 taper can be really worn or there may be a defect inside the spindle taper that the indicator needle is bouncing off of (not likely).

2. The spindle bearings might be worn, loose or the preload may be inappropriate. The tapered roller bearings in this machine felt tight but they were full of dirt and dust particles and were never, ever maintained so this was a good candidate for much of the runout.

3. The drive sleeve, called the spindle taper by Rong Fu, can be worn inside where the spindle splines ride but this is not a precision fit. Mine felt smooth and spindle ran with very little play so I didn’t think this was it. However, the bearings that support and locate the sleeve do influence runout and if they are worn or are moving in the mount then that has to be addressed. I noticed that when I yanked on the main drive pulley I got an additional 0.006” of movement so I suspected those bearings were pretty much shot.

4. The main drive pulley is a press fit onto a taper on the drive sleeve so any vibration from the belts will transfer directly to the drive sleeve and thus to the spindle. While this doesn’t affect runout directly it does get transmitted to the spindle and can affect finishes, especially if the drive sleeve bearings are bad. On this particular machine, when run at 1800 rpm there is some harmonic that causes the rubber belts to vibrate badly; it almost sounds like a loud groaning (the scary kind, not the sexy kind. The hair on the back of my neck actually stood up when I first heard it!). In contrast, the jackshaft pulley bearings are tight, smooth and will be left alone.

5. The quill itself is another issue that can cause runout; to fix it you need a new quill. There were no scuff marks on the quill to indicate a problem with the quill or the passageway so replacing the quill was the last resort if I couldn’t fix the runout any other way.

I decided to look at and address each of these concerns and check static runout after each change to see what, if anything, makes a difference. My guess was that I should be able to improve the runout but getting anything under 0.001” … wasn’t sure.

The first thing I did was to change the tapered bearings to sealed angular contact bearings using the original spindle. This changed the runout very little and there was still a lot of bouncing of the indicator so I ordered a new spindle from Enco and installed it after switching the new spindle bearings over.

With the quill placed in blocks and clamped firmly on my work bench I measured 0.0001” TIR inside the new spindle taper. This confirmed that the new spindle is straight and is running reasonably true in the quill and bearings. Since my quill runs nicely inside the head I was relieved that I got away with not having to replace this quill.

I then installed the quill into the machine and again found the indicator needle bounced a lot but with less range. This confirmed that the drive sleeve was definitely having a negative influence on runout.

The next step was to pull the drive sleeve and change the bearings. The sleeve itself is in pristine shape – no wear was seen and all the machining was still sharp. I replaced the OEM 6009-ZZ bearings with Nachi Quest 6009-2RS bearings. These are deep groove bearings (can handle high radial and axial loads), sealed on both sides and permanently lubricated. Once the drive sleeve bearings were replaced the bouncing was reduced dramatically and runout decreased to a little less than 0.0001” TIR, the same as I found at the bench. I checked runout with the quill elevated to several positions and got the same result … finally.

With the runout addressed I decided to replace the rubber belts with Fenner Power Twist Plus belts. Fenner belts are relatively expensive when compared to rubber belts but they will not wear nearly as fast as the rubber belts, they stretch less once broken in and they run smoother and quieter. The RF-31 uses a B-type belt, which is 5/8” across. You need a little over 6 feet of belting to make the two required belts.

After changing the belts the machine does run noticeably quieter and smoother. That nasty harmonic and groaning at 1800 rpm is now gone so I’ll consider this handled. If the machine won’t finish nicely now then I don’t know what else can be done.

So, it turned out that my spindle was worn. I suspect that the spindle bearings were shot and that accelerated the wear inside the taper but replacing both sets of bearings was necessary to get the spindle to an acceptable level of accuracy. I'm not sure which was more important but if I had to guess I would say the spindle bearings make the most difference. I took some pictures and will post some “how to change your bearings” threads in the near future in case someone who hasn’t done it before wants to give it a shot.

At least now I know that the RF-31 spindle runout can be reduced to about 0.0001” or less. As to how accurate the machine will be, that remains to be seen because measurebation is one thing; cutting accurately is another.



Mike
Good post, Mike! A typical tolerance for maximum spindle runout is .0002". I have only messed with knee mills, which are made a little differently. Another thing that can cause runout at the tool, even when the spindle checks out OK, is a problem with the anti rotation pin. If is is too long, bent, or buggered, or the keyway in the collet is not wide and/or deep enough to accept it, that can throw off the runout at the tool.
Here:
is how to repair a spindle after you do all the tests you did to make sure the machine itself is in solid, tight condition. Stan is a well qualified millwright, not a hack, and he is doing this job in a careful, well thought out manner. Very little is needed in the way of tools and tooling to accomplish what Stan did here. Just make sure you understand that this is the LAST thing you do to correct mill spindle runout, start with what Mike did in this thread first...
 
Mike, tapered roller bearings will take much higher loading and more abuse than angular contact ball bearings will...
 
Mike, tapered roller bearings will take much higher loading and more abuse than angular contact ball bearings will...

My choice has more to do with not having to maintain/re-grease the bearings. The RF-31 has openings all over the place and dust, grit and chips will get into the upper bearing. In addition, load bearing capacity in angular contact bearings depends greatly on the contact angle of the race and the bearings I used have a 40 degree angle that allows much higher load bearing. I'm not sure that the loads a milling machine spindle sees is anywhere near the max load limit of these bearings; I suspect that it is only a fraction of what these bearings can handle.
 
Good post, Mike! A typical tolerance for maximum spindle runout is .0002". I have only messed with knee mills, which are made a little differently. Another thing that can cause runout at the tool, even when the spindle checks out OK, is a problem with the anti rotation pin. If is is too long, bent, or buggered, or the keyway in the collet is not wide and/or deep enough to accept it, that can throw off the runout at the tool.
Here:
is how to repair a spindle after you do all the tests you did to make sure the machine itself is in solid, tight condition. Stan is a well qualified millwright, not a hack, and he is doing this job in a careful, well thought out manner. Very little is needed in the way of tools and tooling to accomplish what Stan did here. Just make sure you understand that this is the LAST thing you do to correct mill spindle runout, start with what Mike did in this thread first...

Very cool, Bob! Impossible to do on an RF-31 due to the construction of the machine but I bet one of the other guys will greatly benefit from this. Thanks!
 
Very cool, Bob! Impossible to do on an RF-31 due to the construction of the machine but I bet one of the other guys will greatly benefit from this. Thanks!
Oh, I see... The head does not swivel on those machines.
 
No, the head is fixed. To tram it you have to tram the column and if you move the head elevation, you lose that tram. Just like a drill press, along with all the main faults of that design. Over time I plan to find its limits and strong points and perhaps report them here if any fixes work. I got this thing for free and will use it to honor the guy who gave it to me.
 
Oh, Bob, I meant to say that you're right - the anti-rotation pin can cause runout if it sticks too far into the collet passage. I removed both of the OEM ones and will rely on a spindle wrench and a brass hammer to get the collet loose. I didn't want to hassle with those pins for the exact reason you stated. Besides, to adjust or mess with them you have to pull the spindle out of the quill and I didn't want to back in there a third time. Thanks for pointing out a very important omission.
 
Back
Top