Proper bearings for a mill spindle rebuild?

[psionik]

I'm rebuilding a hand-me-down mill and I've been looking at different bearing types, trying to decide what to use to improve the accuracy and repeatability of the machine.

I am not bothered with adapting the current hardware for any type or combination of bearings, but what I don't know is just what would be ideal for this machine, even fabricating an entirely new shaft and line-boring the head as I have a lot of metal lying around and access to large machines via a relative.

For what it's worth, the current spindle shaft seems smooth and concentric on all the important surfaces (with various diameters along the length in the up to near the 2" range) and seems to be in good condition, but the bearings themselves seem to be totally shot as they have visible rust and make a horrible racket. It uses an R8 holder but I'm thinking of making a new shaft for a different taper, or perhaps ER collets. With the current motor and belts the mill operates up to about 2500 rpm but if I follow my impulse to do a cnc conversion later I might want to push for more power with some servo motors I have on hand and higher speeds.

My concern is the fact that a mill will experience radial and thrust loads and the bearing choice should reflect that. While tapered roller type does well in this regard, it has rather low rpm limits and necessitates designing for preload.

Angular contact bearings seem to be interesting although I suspect I might want to combine it with another bearing type to ensure it can handle all types of loads.

And from what I gather, ordinary ball bearings won't support thrust loads that might arise from heavy milling although from what I can tell this is exactly what this mill already uses.

There have been a few descriptions of commercial machines boasting about double or triple bearing spindles, which I guess would help average out imperfections and multiply load capabilities and I would be happy to consider this plan if worthwhile.

So I'm wondering, what do the experts here have to say on the subject?

 

[Bakafish]

I don't claim to be an expert, but know enough about bearings that I'd advise you to replace the bearings that are there with the same exact type that the machine was designed for. The first step is identifying what's actually in there. Even if it is just using deep groove ball bearings, if they are good quality and properly installed, they can handle a lot of load and are available in precision versions that are likely higher than an old machine actually justifies.

 

[psionik]

Yes, I'm sure you're right but since I was thinking about changing the machine later I wanted to consider building in some extra oomph since I'm already neck deep in the rebuild and expect this project to take quite some time. I suppose part of it is my interest in what is considered when designing these things in the first place.

I was looking at angular contact bearings and the prices go from affordable to insane, so I was wondering why so much difference.

 

[IceCzar]

design from scratch is a bit different than starting in the middle
I'd imagine you have a hardened spindle?
Replicating that isn't particularly easy compared to regrinding or adapting.
2 back to back angular contact bearings with a spacer for a given preload plus and additional bearing near the spline is pretty typical for a given era.

but you havent supplied much information about the mill

 

[Bakafish]

I was looking at angular contact bearings and the prices go from affordable to insane, so I was wondering why so much difference.

Materials and precision. Testing and validation after all the complex manufacturing costs a lot of money because of the time and equipment required to ensure it is accurate. The best manufacturers have very extensive testing regimes with the bearings properly loaded and lubricated to ensure they perform to the ratings they claim. Off brands may claim the same precision, but unless you have the ability to validate it, you only have their word, you pay a lot for a brands integrity. Remarking or forging bearings as a name brand or higher precision is a big thing as well, so a bargain should be treated with suspicion.

All that expensive precision is wasted on a normal machine, even a good machine, as it isn't designed to take advantage of it.

 

[psionik]

I have access to some quite large equipment at a relative's shop, including lathes, mills, surface grinders, large line boring machine, and even a wire edm and I've used them all with confidence.

This mill is one of those sacrificial machines you learn on if you know what I mean. I wanted to get a notch on my belt rebuilding it, making it better than it was new and maybe even converting it to cnc just for the value of the experience. For me this is kind like restoring a classic car without the car. Perhaps I'm polishing a turd here but I'm retired and have nothing better to do. I've set my budget at $1k for rebuild parts but I see I can easily spend that much alone on just the bearings. What I'm curious of is the sweet spot of cost vs quality for a machine that was perhaps 10-15k new probably built in the 80's or 90's at the latest. I don't know what brand it is since it is beat up and has so much use nothing is readable on it except on the motor plate which is 220V 3 phase at 12 amps. I have a rotary phase converter already fyi.

I have a lot of steel laying around, some pretty heavy plate (1.5" I think) and I've put a little thought in what I might do with the head since it's probably cast iron and I can't weld that but I could cut it and bolt on sets of machinable bearing plates which would make my options wide open... just rolling around a lot of ideas in my head before I settle on a plan...

I do have some experience with case hardening and grinding and some various alloy bars lying around just begging to find a new home so I'm not frightened by the scope of possibility here even if I have to create a spindle from scratch. My 360 skills are starting to develop and I'm getting a sense of what goes into proper design after a lot of reading and video watching and I'm eager to start, just trying to get some second opinions here on bearing choices.

 

[Luke Rickert]

Spindle design is no joke, but for an inexpensive manual (low speed) machine it isn't that big of a deal. As mentioned two angular contract bearings at the nose and one in the opposite direction is a standard solution. If you draw up the current configuration that would be helpful in understanding the best approach (which may very well to leave everything along and just replace the existing bearings)


Luke

 

[EmGo]

I'm rebuilding a hand-me-down mill and I've been looking at different bearing types, trying to decide what to use to improve the accuracy and repeatability of the machine.

Give it up. You're kidding yourself. Unless it's a Devlieg and you know what you're doing, you're not going to improve doodly by playing armchair machine tool designer.

Just get some new ones of whatever it had and put them in there. Install properly.

That's the best you're going to do.

 

[IceCzar]

There is a lot to be said for being an arm chair tool designer. It led me to working for a real engineering firm. Even failure teaches.

 

[Bakafish]

There is a lot to be said for being an arm chair tool designer. It led me to working for a real engineering firm. Even failure teaches.

I agree, but precision bearings are a 'low hanging fruit' trap for beginners. They seem like a modular thing that can be swapped out to improve performance, but are in reality so dependant on the other parts of the system that they likely will provide little value without radical changes to other parts of the machine. Like putting giant fuel injectors on an engine that is still equipped with an inadequate stock fuel system, intake, exhaust, etc. It's just going to make it perform terribly as the rest of the systems can't leverage the increased flow. The improved tolerance of high precision bearings gets diluted, often completely swamped, as they are constraining parts with far greater eccentricity errors than the better bearings have eliminated. And even if a change in 'architecture' might theoretically offer more force handling, it is very likely the rest of the system cannot, as it wasn't designed to (give the old school engineers some credit.) So in a crash situation rather than replacing some reasonably priced bearings you create a much bigger and more difficult to address failure.

But the best point made above is we don't even know that the bearings that are in there even need to be upgraded. It is very likely that they are properly sized and spec'ed and some fresh name brand replacements won't be deemed completely acceptable to all... This is premature optimization, another famous trap to the new player.

 

[gbent]

Replace with original bearings then start making stuff that can't be bought, for needs no one else has. You haven't told us what mill you are fixing, but it does no good to build a spindle as stout as a 5" boring mill and then hang the spindle on a rubber band machine frame.

 

[mottrhed]

No offense meant here, but you are in over your head if you need to ask some of these questions. A few others have given good advice, which I can put a finer point on. Figure out what bearings came out of it, and then buy the highest quality replacements you can find, dont expect them to be cheap. This still is not where your struggle ends, but its a start. Then you will likely have to adjust for capture, set preloads, potentially bring in runout etc etc.

Making a new spindle shaft is a tall order even for a competent shop, and if you want it to survive any length of time while producing tight tolerance parts, the shaft must be damn good. Your final runout, (+ vibration, heat, repeatability) and therefore part quality is dictated by the stackup of tolerances that will be massively influenced by the shaft. If you can design the shaft, machine it, grind all critical surfaces, and balance the final product, then you would be a person who doesnt need to work on this machine..

If you want actual help from people with experience, enlighten us with the following information.
Machine make/model
Bearing mfg and pn
Bearing arraignment
A print of the spindle would be even better.

 

[michiganbuck]

I agree that designing and making a spindle is often above most people's talents.
A 2500 RPM spindle with replaced original bearing is pretty fast, and converting to CNC might be better served with just buying a CNC machine.

 

[Conrad Hoffman]

There are mechanical engineering texts that cover bearings and spindle design fairly well. IMO, the best ones will be some decades old. If you're up to doing this, you're up to finding them!

 

[jim rozen]

You will probably find that extra 'oomph' in the spindle really means "larger bearings" which leads to the connundrum 'where does the extra meat of the larger bearings GO?' to which the answer might be 'just open up the receptacle the bearing seats in.' Doing this reduces the rigidity so the oomph factor might go down, and not up.

I suggest you bush up on FEA software as a start. Serious, not a jab.

For the record, the original bridgeport spindle bearings were "SAE205" automotive bearings. Yep, plain open radial bearings, where the outer races were speciall ground so they'd run in approximation of angular contact - where the OD and ID spacers were dead length identical.

 

[Rob F.]

Maybe the new guy/OP read the rules of what machines can be discussed, so he knows NOT to mention the machine brand? Just a possibility...

 

[CarbideBob]

Angular contact go from cheap to expensive due to runout accuracy and how to preload.
"Precision" ones are ground on the sides to give a specific preload when mounted and have much less runout or "cycling".
You just tighten them up and done. Lower levels and you grind a spacer or shim to get the loading you want or tighten some nut just right.
Then classes in precision. Here is runout and ability to take speed.
Add in contact angles so a precision ballscrew bearing is different than a spindle.
Normal base setup is a rigid mount pair up front and even just a simple single out back.
I would never consider single deep groove both ends for a spindle on anything other than a bench grinder.
Original inside and outside spacers on a B-port where not the same size. Ground to fit at the factory and we are talking tenths or less here.

Many will say do not try but I learned by trying and I did many bad things along the way that worked... sort of.
A spindle that has to mill 24/7 to microns is different than one that runs on the weekends and needs a thou.

FEA, for me I thought a good idea at one time but no help at all. File this under duh and worthless in my opinion.
More useful in modeling chips and such but even here it fails massively.
Maybe a drawing of what you want to try?
I am all for go at it. Works is great. Not so well and a life lesson.
Bob

 

[JST]

.................. but the bearings themselves seem to be totally shot as they have visible rust and make a horrible racket. ...................

That would tend to suggest replacing is a decent idea...........

Find the numbers, and then look for them from the major manufacturers. They usually have cross-references.

Use the best ones you consider to make sense given the value of the mill itself.

 

[psionik]

Hey folks, I appreciate the warnings and they have been heard.

Though I may not be an expert I am the kind of guy that achieves by increasing the mass of my effort until it outweighs the weight of my incompetence. It's been my recipe for success for decades and every time I've heard someone tell me I can't do it I become that much more determined. It's never a big deal for me to fail; I've always had the attitude that failures are steps on the road to success. I take what everyone else considers a doubt to actually be a dare. So now I've been double dog dared to try.

Every achievement I've made in life, though it may seem at the time to be worthless, has been added to my vocabulary of understanding which I use to leverage success later. I am not calculating whether my test machine "is worth it" but rather what I will gain in understanding along the way.

I heard the same stories when I first got into the idea of grinding my own telescope mirrors. Telescope mirrors are not micron accuracy, but angstrom. I was told I was in over my head, etc etc and yet as far as I can tell at star parties, my mirrors beat anything commercially available and even run head to head with some boutique names.

Same story when I started school, published my first book, rebuilt my first transmission, got my first patent. Naysayers are a dime a dozen.

I'm a polymath and an intellectual omnivore. I eat books for breakfast, so I suppose my next task is to devour some of your recommendations and get to work.

I have nothing but time and money now, and my shop is my playground. You don't retire at 46 in this economy without getting some things right along the way. If I have to blow a big pile of money on granite surface plates, ovens, more machine tools and tooling just to say I can do it, I will. Roar.

 

[mottrhed]

If you want to chase this down, go for it!

The thing is, you are in very charted territory, many of us do this exact thing 5+ days a week. We know exactly how to address this for the best outcome given a particular set of performance requirements, no research or guesswork needed, we will nail it the first time, with equipment we already own.

On the other hand you may have to spend $50k-150k man/hrs researching, trying, learning, tooling up to do this "right" on the 10th try. And the results will likely be a compromise, as youve got too much time in, so at some point its "good enough".

Im not trying to dissuade you, if you have the time and the money to tackle it, get after it! It just doesnt make much sense if you value your time, this could be rebuilt better than new inside of x hours by someone with experience, while you do something enjoyable.