Selection of bearings for two critical low slop/low runnout applications.

[LaserSteve]

I have a desire to make two lifetime want to have / need to have machines. 

 1. A lapidary saw / spindle for sawing and grinding.

2.  A tiny table top glassblowing lathe.

SO, MEs,  I hear I need "Angular Contact" bearings for the required precision shafts. Please school this old man on what I really need, and like all things here cost is an issue.  Loads and forces involved will be small, in the 10s of Pounds.  So what grade do I need?  I know you will need more details.

End goal is small scale glass laser tubes, so angle has to be tight to 0.1 degree on finished product after any hand reworking.

Believe it or not I have the rest of the materials, the isotopic gasses, the glass to metal sealing technology, and the optics.

Steve






Steve

[Gyro]

In terms of rigidity and smoothness / lack of chatter, taper roller bearings with a small amount of preload will be superior to angular contact ball races. From a quick check, you can easily get them down to 15mm bore but you might want larger if you are planning to pass glass tube through the spindle.

[Benta]

What options do you have for making these machines? Lathe, mill etc.?
And I've absolutely no idea what a "glassblowing lathe" is. Speed, size etc.

[coppercone2]

i think if you seat them in a hardened ground sleeve too its good for them

not sure why, but I vaguely recall something that seating bearings inside of a hardened thing that is inserted into the material might be better then inserting bearings into a material. Maybe because you can seat a precise ID ground pipe in a material with higher force then you seat the bearing. It might be also more precise then the hole you can make in some weird part, maybe the bearing sleeve compensates somewhat for imperfect holes

[jpanhalt]

For a lapidary saw, I suspect you want decent bearings.  Accuracy like a Myford, Harding or Monarch (some old brands) is not needed.  Of note, even some of those lathes used adjustable sleeve bearings instead of roller or ball bearing.  I would consider bearings that would be resistant to the water cooling that will probably be used, such as all stainless.  The bearings on my septic aerator  motor used to go out about every 2 years.  I replaced then with all stainless, and its been going for at least 8 years.  When I looked at Internet sources for them, the price was shocking.  However, Cleveland, OH happens to have a huge distribution center for ball/roller bearing, and a local business was able to get them for me quite reasonably.

For a glassblowing lathe (attached), I suspect a wood lathe level of accuracy would be more than enough.  Every glass blowing lathe I have seen used a three-rod chuck, often with a heat-resistant soft sleeve, like asbestos (1960's -1970's).  I have no advice on getting the angle right, as I never did glassblowing to that level of accuracy.

[LaserSteve]

I'm outside of Cleveland, Ohio in Akron.  Personal Tooling:  Two 7x10 Lathes, one with a three jaw,one with a four jaw,  One surplus  Old Enco Mill with Boring head and Rotary Table.

Skill Level: Former R&D Technician,  Lasers, Analog Electronics and Lab Instrumentation, Major University. (18 years then Downsized).  I used to make nearly everything out of Aluminum and Brass, rarely steel. I have basic glassblowing skills for lab glass and vacuum grade glass.

I'm used to making quick and dirty one-off devices for labs.  My machining skills reflect that, I can do my own setups, (NO CNC).

The Akron-Cleveland  area is amazing for obtaining machine parts.  I can throw a frisbee from my desk here at work and hit two places that sell drills and cutters.

On a glassblowing lathe, both chucks move in exact synchrony... Most home made glass lathes use two steppers to drive the chucks. I need a common shaft design, Litton "F" model comes to mind, but I do not need the huge scroll chucks.

Somehow in 18 years, my bearing tolerance education got neglected, that might have to do with working for Polymer Scientists, Electrical Engineers, Chemical Engineers Biologists, and Physicists.  ME was handled by another team.

I'm at work, I'll post more this evening.,

Here's a video of an "F" being abused by a newbie, so much wrong in that technique.

https://youtu.be/S4rNxa8KqPM

Steve

[langwadt]

unless you buy the absolute worst junk from ebay et.al. I doubt regular ball bearings would the limitation

[CatalinaWOW]

My background is much like yours, except spent much of my career dealing with precision gimbals.

I agree with Langwadt, bearings are not going to be the limitation as long as you are willing to spend a few dozen dollars.  My ancient Atlas lathe has two bog standard tapered roller bearings mounted about ten inches apart.  Runout is lousy by lathe standards, a couple thousandths, but that translates to angular errors on the order of a about 0.01 degree.  Bigger question is direction and magnitude of applied loads.  I don't think your application calls for large axial loads, but if using angular contact ball bearings you will want to look into contact angle and axial loads.   In general the roller bearings will support larger axial and radial loads, but it is possible that even your saw application is well within load capacity of both bearing types.   

The real issue will be setting up preloads and alignment.

I would suggest two bearings in each end stock.  You first need to align those two bearings.  Both axis and direction.  Best way is to turn spindle to fit bearing inner bores with shoulders on each end.  That will align inner races.  Then outer races need to be preloaded and mounted.  One clever trick is to make the bore you are mounting in slightly oversize and apply the preload with springs, weights or whatever.  Rotate the spindle a few times to make sure everything has settled in and then fix the outer races with Loktite or other appropriate adhesive. 

Next step is aligning the two spindles.  Lots of ways to do it.  Most pretty obvious.

If you have only a few tube diameters to fuze I would look into collet chucks for a combination of accuracy and moderate cost.  You can price to the appropriate precision or build your own.

As a last thought, if your hobby tools are used like mine, very, very lightly by industrial standards some of the design concerns about bearing type may be moot.   How much do you really care if bearing A wears out in 3,000 hours and bearing B will last 50,000?

[ace1903]

From my hobby experience in machining, I think that two servo drives is way to go for glass lathe.
I have put servo drives on my milling machine, crappy ones from overseas and results are satisfying.
I think that some better ones like from clearpath ( the ones that popular YTubers are using like Markoreps and Clough42) can outperform mechanical linkage and be way cheaper.
Absolute encoders in 17bit resolution can make miracles. If I need machine like that I would convert one of the existing lathes to be optionally driven by servo drive and will replace tailstock by custom
cast from epoxy granite. Electronic to drive both servos in sync should be straightforward.
I replaced angular contact bearing on leadscrew on my lathe with SKF 15Euros part and errors are not detectable with hobby measurement tools(I use Insize micrometers and electronic indicators).
My experience with SKF bearings is that they always outperform spec and 0.1deg will be easily achieved with careful preload with torque wrench. 

[jpanhalt]

@LaserSteve
The distributor I mentioned is Bearing & Drive Systems, 1488 Foltz Industrial Parkway, Strongsville, OH.  The dealer I use is on Brookpark (Hwy 17) in Parma.  I doubt that's worth the drive, but a retailer in Akron may have an account with that distributor.  Is Timken walking distance?

Of course, I forgot to mention that a glassblowing lathe must have the headstock and tailstock turn in synchrony under power.  Trying to do that with just a ball bearing tailstock would make a mess.  I agree fully that bearing precision is not the limiting factor.   I would just use pulley drives with a single shaft under the bed.  For others who may not be familiar with them, they also are low speed.

BTW, I am about 20 miles due West of Medina off Route 18.

John

[LaserSteve]

Timken is 4.5 miles from back yard as a crow flies..  Akron Bearing is on the way home.  Krome hard Twist Drill is right next to work for the precision tooling.

There is just something scary about there being no spline.  Here is a stepper based unit.

https://youtu.be/kkNrYnDewH0

On a Laser tube, we often have two window stems cut at Brewster's Angle (55 degrees 32 minutes) for near zero loss for one polarization of light.  I make those stems before joining as they are polished to the point the glass stem will join the window under an electric field and some heat, a process known as "Optical Contacting".
Aligning the stems is done by looking at the reflected spots on the ceiling from another laser, and then passing a polarized laser down the bore through both windows and measuring the intensity as the stems are bent/spun/ tweeked.  That is is why I'm weary of dual stepper drive, a twist is just evil.

Most tubes are 1B tubes, meaning one end has a sealed mirror and the other a Brewster window.  Its the 2B tubes I'll need precision on.

OK, so tapered roller looks good. 

Does ABEC grade come into the picture?

Thanks,

Steve

[benseno]

There are different grades of angular ball bearing starting from 7xxx standard series till high end matched pair angular ball bearing for spindles. Angular roller bearing are used for high load and probably will be an overkill. Besides that there are plain bearing for low rpm, air bearing, even magnetic bearing. afaiu, you are targeting low budget design. Is there kinematic scheme of your machine which can be shared or example video (like you posted for lathe)?

[jpanhalt]

Considering what you are doing, I would go with 3-jaw scroll chucks that are cheap.  There are loads of them for wood working and you obviously don't need the rigidity of a metalworking chuck.

The very common 2080-type aluminum extrusions may be completely adequate for your frame, if you go with dual stepper/servo drives..  Based on the picture you showed, simple flame annealing should work fine rather than getting a  kiln for that. 

[ace1903]

Just want to add that there is significant difference in positional accuracy in stepper, hybrid stepper and servo drive.
Good quality servo drive with encoder with suitable resolution, powerful DSP with implemented algorithms for tuning and vibration suppression,
output stage with precise current measurement costs money and can deliver positional accuracy.
200steps per rev crap steppers can not deliver 0.1deg at any practical speed and shouldn't be discussed for this application.
I follow one guy that makes nixie tubes on YT and he showed some glass blowing lathe(Dalibor Farný) and some problems he had with it.
Think that it was based on splined shaft to sync both sides. What is the expected tolerance that those lathes can deliver?
Do they have some mechanism to compensate for backlash in the splined shaft mechanism?