[Oldmanandhistoy]

I would like your thoughts and opinions about grinding skate bearings with a radius to suit a round rail.

Would the benefit of full contact of the bearing with the rail make this a job worth doing?

I believe also that I could increase the pressure on the rail which would help increase over all rigidity.

There maybe other benefits that I don’t know of that some one may be able to enlighten me to also, or possibly reasons why this would be a bad idea?

If this subject has been covered else were could some one please point me to it? I have been unable to find anything.

Thanks in advance, John

[sendkeys]

In my opinion


Im sure it would help but i think it would be very hard to get any type of precision or a waste of time/money. I would just add more bearings to give it more support.

But time and money aren't everything, if you go for it maybe you should think about using a sleeve over the bearing. If you go this way you're not going to have to deal with any type of over heating or cutting away to much metal. You could also use a softer metal that would save the rails.

[Oldmanandhistoy]

Thank for your reply

I have already considered the idea of adding a sleeve to the bearing but decided this would be too much extra effort.

I have grinded a bearing just to see how thing went and had no problem with overheating.

As for precision I thought that the rail would quickly wear in so to speak and match the bearing?

What I really need to know is would there be any benefits to using this idea?

[NC Cams]

Hard to tell. Why not try it and see???

Approach seemed to work with the skate board bearings (as a former bearing engineer, I"d have advised agains it BUT you can't argue with success - at least at this point).

[joecnc2006]

I doubt the wall of the bearing would be thick enough to get the radius you are looking for.

[sendkeys]

Well if this is a router presision isn't going to matter i was speaking more for tir type error(out of round). If you have this type of error it will be hard to preload them without using some type of spring.

as for benfits. The main one would be a longer lasting surface. you wouldn't have to adjust them as often and over time would have more presision. Should also be a little quieter.

you would be able to add more preload if you used a sleeve but without i would guess you have to use about the same seeing you're taking away metal from the outside of the bearing.


Might just buy v bearings that should give 2times the surface contact. as long as the rails aren't to big or the bearings to small.


p.s this just my opinion only way to really tell is to test

[ViperTX]

Generally the bearing outer surface is hardened steel......

[Oldmanandhistoy]

Thanks all for your input.

I have calculated that grinding the bearings would remove 0.36mm if I don’t cover the full width. Do any of you guys know how much metal I have to work with?

I was thinking of running them on drill rod (silver steel over here in UK) but thus far I have been unable to find a supplier. Any one know of a supplier in the UK for 25mm diameter rod?

As I said earlier I had no problem grinding a bearing.

So would the improvement I would gain from grinding the bearing justify the time it would take to grind them? Would the time it takes to grind them be less than time I would spend adjusting them if not ground?

I think that with out a certain answer to this question and as I have made the bearing holder for my lathe already I should grind a couple more and do some simple tests. I was just hoping that some one might have tried this idea and could save me some time either way.

With the bearing I did grind, I ran it along a pipe in my hand and it seemed to almost grip the pipe. Having said this do you think that there may be a problem if I don’t get them set up dead straight with shear action?

Really wish I was a qualified engineer lol.

[WhiteTiger]

If the axis of the bearing varies from true 90 degrees to the axis of the rail, there won't be any sheer, it'll just mean that the ends of the radius make contact with the rail and the center doesn't (assuming you have true matching radii). The depth of the curve in the bearing remains the same, but when running out of alignment the outer edges are effectively closer together.

It can be an advantage as you would double your point contact on each bearing, but getting the skewed recess of every bearing exactly centered over your cylindrical rail would be a booger.

If you have full contact, "dead on" alignment of matched radii, then the lowest point in your bearing groove is smaller than the rest of the groove, and you end up with travel differences across the whole band of contact.

Basically it becomes a problem of a big wheel and a little wheel trying to travel the same distance in one rotation, and what was a rolling contact becomes a wiping/sliding contact across most of the involved surface.

This is why manufacturers of v rollers and ways tend to warn that they are best used in light load situations and will wear faster than plain surface equivalents.



Tiger

[Oldmanandhistoy]

Thanks WhiteTiger

A lot of sense in what you say and I understand your points.

Having read what you have said, it gets me thinking about the bearings I’m grinding.
Other than the point about the axis being 90 degrees the other points would come into affect in the way these bearings are constructed. A round ball running in a curved groove would have the same travel distances effect. Is something done to counter act this or is this an excepted effect.
Also the roller thrust bearings I have come across would also have a wiping/sliding action. Obviously the end of the roller closest to the centre would want to rotate slower than the opposite end. Seems to me they need to be conically in shape to avoid this effect.
But in both these cases it is obviously an excepted effect?

Having said all this would you say my idea is a bad one? Or would the sliding/ wiping effect be acceptable as is in the roller thrust bearings I mentioned.

Like I said in an earlier post I have now done three bearing. Measuring the roundness it is with in 0.01mm (can’t measure more accurately than this with the instruments I have).
Heating was not a problem.
I will set them up tomorrow evening and see how they run. Nothing to loose now.

John

[NC Cams]

As a former bearing engineer, without a drawing, it is hard for me to see exactly what's going on and/or desired with regard to this bearing application.

Thus, if one starts to prescribe a rocket science solution, the inevitable "this is only a what-its that doesn't need to be that accurate" will probably be a reply sooner or later.

Basics to consider are this: For smoothness and running accuracy, one can and prefers to apply a 1%-2% of the radial capacity in the form of an axial preload. "Snug" is an irrelevant and dimensionless term.

You do want the load axis to run perpendicular to the bearing rotation axis. Some side thrust can be tolerated but only the internal geometry of the particular bearing used will limit the amount - NOBODY in the bearing industry gives out their internal geometery info.. This is why they say to "contact the factory" for cases where there is/will be severe axial thrust.

As stated previous, I never authorized any of my clients to modify bearings - they are too easy to damage internally. Folks who ignored the advice/recommendation usually did and usually did (modify and damage, respectively).

There are specialists out there who the bearing companies work with all the time. But, I assure you, someone who calls in and asks for info on how to grind skate board bearings at home on a lathe won't get a lot if any support....

The unequal speed issue is true - although a tire/wheel combo is running at a given rpm, the ft/sec (meters/sec) of the instantaneous surface speed is one thing at the axle and a whole lot more at the OD.

"Skidding" (unequal relative surface speeds) is a real problem when you deal with bearings that are being asked to make sliding contact with an adjacent moving component - this is especially thru with non-equal radii of contact situations. The V groove thing is correct and proof that you can't always "wing it".

Now, having said that, the "but i've already done this or that" with respect to the device in question has clearly occurred. This thread is surly NOT transpiring in real time as I doubt the member is sitting waiting for a reply before he acts....

In light of the success outlined so far, the design may simply be uniquely intolerant of engineered/misengineered excesses.

Or the user is lucky

Or he has created a very unique and tolerant design that is exploiting the daylights out of the bearings he's chosen.

I"m in no position to judge...

When it comes to the use of ball bearings, there are MANY situations where they are used incorrectly - sometimes well and sometimes with disasterous results. From the looks of things so far, the speeds and loads of this application are probably such that real bearing engineering calcs don't matter all that much if at all. The bearing may simply be overdesigned for the load conditions it is seeing.

After all, the bearings were modified (against what I"d recommend) and they worked. ANother example of the perverse nature of inanimate objects.

I'd be more inclined to ask the member to keep us posted with the results of his experimentation. He's clearly not needed an engineer's input and has done just fine without it. HOwever, when or if he does have problems, any improper useage of the bearings that is endemic to and a critical part of the design could make it difficult to utilize traditional fixes upon.

Then again, he's using skate board bearings which are techically the low cost 'loss leaders' of the industry. It will probably be easier and cheaper to feed it bearings than to re-engineer it, regardless.

I say to our member, "press on, enjoy, and keep us posted". Many times in my career, when a machine doesn't perform according to my theory, I"ve had to believe the results and invent a new theory.... This could be another example of that....

[WhiteTiger]

But in both these cases it is obviously an excepted effect?

Having said all this would you say my idea is a bad one? Or would the sliding/ wiping effect be acceptable as is in the roller thrust bearings I mentioned.

Yup, bearings seem to be no different than any other mechanical product. Production economies play as big a part in design as function, most times. Roller thrust bearings in most common sizes have a negligible surface inch of travel discrepancy, and where it becomes significant in relatively larger sizes, you find operating speed and load spec is cut proportionately way down.

I have to agree with NC Cams; you already have a process that is working to your satisfaction, and parts in hand. Might as well finish up and put it through it's paces. Worst case I can see is you end up replacing with unaltered bearings... best case, it turns out to be a "gee whiz this is cool".

"All the reasoned opinions in the world aren't worth one case of been-there-done-that"



Tiger

[Oldmanandhistoy]

I have added a quick drawing to show the bearing arrangement. The blue object is drill rod and the red objects are bearings.

Any suggestions to how I might test this set up?

Thanks guys I really appreciate your input.

[WhiteTiger]

The only low tech test that comes to mind would be to finish the assembly to the point it's ready to function and then use a weight and pulley arrangement to see what it takes to make the thing move a given distance. Replace and repeat with unaltered bearings and see if there's any difference in ease of movement.

If you can lay hands on a gunsmith's recording trigger pull gage it would make things simpler. These are examples of the gages I'm referring to.

Pretty hard to get any meaningful test data when you don't have any control standard to compare to, so you'd need to do the same test(s), whatever they are, with both altered and unaltered bearings, I'd think.

Maybe someone more awake at this hour than I am can think of something else.

EDIT> backing up to the ballscrew question earlier that I managed to skip replying to, higher grade ballscrews don't use a simple circular chord segment profile, they use offset radii (or some variant), so the ball trough has a profile like a V with the sides bowed outward. This gives two point contacts for the balls at theoretically the same distance from the axis of rotation.



Tiger

[NC Cams]

Options: instead of 3 bearings mounted at 90 deg intervals as illustrated, you can mount 2, 1 each mounted 60 deg off of vertical. This assumes the load is at 6 o'clock position.

I"d not recommend mod'ing the bearings - see prior post - BUT you've already done that.

I'd recommend fitting a collar to OD with desired radius - BUT you've already ruled that out.

Point is, you seem to be doing quite well with "looks ok, give it a try" applications engineering, I'd see no reason why you should change your approach now - nor expect that you would if the suggestion/recommendation would involve scrapping out what you have made that is working.

Regarding testing, no traditional method for bench testing bearings will suit BECAUSE you are running the bearings at low speeds and have a real strong chance of developing oscillatory motion. Bearing testing is done at a rotating speed under a prescribed load. There are provisions for speed/load variations in bearing testing but it does NOT typically provide for oscillation potential which you will have to some extent or another.

WHY no oscillation??? ANY rolling element bearings tends to have problems with "false brinnelling" problems that occure with oscillatory motion. This problem occurs because the rolling elements roll back and forth in a small arc of travel and then "punch/wear" a hole into the grease/lube in them.

When the rolling elements make metal to metal contact, minute wear particles are generated - the wear phenomenon looks like brinnel marks in the raceway - hence term "false brinnell".

Solution: run it until it smokes and see how long it takes. At the price of skate board bearings, it would not be a very expensive learning process.

You could program in a step where you force the machine to move the "trolley" from end to end on a regular basis. This could help equalize wear and reform the lube film. Since you're talking low speed use, you could heavily lube the brearing with a light viscosity EP grease (not oil). This may help prevent false brinnelling.

Bottom line is this: the solutions to the problems that you may/will have add complexity and cost to the project if should you try/decide to build/engineer in a solution (which may not be possible due to the operating environment).

I don't know what the selling price of skate board bearings is/are anymore but you will hardly find any alternate bearings that are as affordable... It will simply be easier to feed it bearings....

[NC Cams]

Reply to Tiger's edit: the "offset radii", 2 point contact thingie he's referring to is called "gothic arch" configuration.

Sort of looks like this: \O/

when it is done but the \ & / are curved slightly. And the RADIUS of curvature of the raceways is much GREATER than that of the ball so as to force 2 point contact.

This drastically increases rigidity and enables you to preload balls to reduce slop.

[Mcgyver]

I have already considered the idea of adding a sleeve to the bearing but decided this would be too much extra effort.

you are increasing the load bearing area so it seems like a good idea.

I'm a little baffled by the above statement, how exactly are you grinding this radius that you feel it would be more trouble to turn and press on a sleeve?

I'll speculate (apologies if I'm wrong) that based on the fact you're not sure where to get tool steel that you are not a 25 year machining vetran that being the case, I'd also speculate that you don't have a radius wheel dresser, cylinderical grinder or a tool post grinder. This is what you'll need to grind a radius on the outer race of a hardened bearing.

I agree with the previous post that it will make a mess of things without said fancy tackle. Also, the race is engineered so by removing material you're messing with its integrity.

It would be FAR less effort imo to makle a sleeve. good excuse to make a radius turning attachment for the lathe. Make sure your turning is concentric, use a suitable bearing press fit (or even loctite) some and make it out of drill rod or chrome moly, heat treat and your done. If THATS to much bother or you don't (yet) have a lathe, I'd go with the suggestion of adding more bearings to handle the load

btw, re drill rod, I'm not in the UK, but most industrial suppliers or the (lucky you) great selection of model engineer suppliers will have it. drill rod is tool steel, in a annealed state. ideally you want hardened rod, ie linear shafting. you DO NOT want this component to 'wear in', you want it hard to resist wear.

[Oldmanandhistoy]

I have been unable to do anything further as yet due to other commitments.

Mcgyver

The bearings I have already prepared for testing I did free hand with a Dremel.
Not really I have a lathe and a tool post grinder.

I must say your post makes me feel that being a keen amateur and not a fully qualified engineer; I should not be attempting such a project and should leave such things to the experts. So should end this thread here or I might give other keen amateurs the wrong idea and they may attempt such projects.

If I was to use linear shafting I would just purchase linear bearings.

The reason I am looking at this idea is that I am building a cnc router for a friend on a low budget. Also having spent quite a large amount of time on this forum I see that a lot of people use skate bearings and pipe in their builds. Obviously if this idea could improve the operation of said method of linear motion it may be of interest to others.
I’m not suggesting other should grind their own bearing but may know a 25 year machining veteran like your self who could do the job for them.

As for the integrity of the bearings they will not be running at 10,000 rpm and likely to explode in my face.

Lucky me has been unable to fined Silver steel (as it is called over here so I am led to believe) much over 300mm.

John

[Mcgyver]

John, from a fellow keen amateur (with a lot less than 25 years experience) I can tell you the intent wasn't to offend or discourage, it was to help someone with whom i misjudged the experience level - someone who's keen and gets the info & direction they need will stay keen. sorry for mistaking your experience level.

The other reason I’d sleeve rather than grind is the old joke that the best way to use a tool post grinder is down the hall in a separate room from the lathe. I'm not saying I never do it, but it is a pita. how are you radius it? do you have a radius dresser or are you swinging the TPG?

why would you go with linear bearings if you used precision ground shafting? the shafting is maybe 1.5-2x the tool steel price, but you still avoid the big cost with is the linear bearing. Meanwhile you get a hardened way with better tolerances.

on the other hand, if you're not hardening it, pipe would make a lot sense. I wonder if you can get it with suitable finish and tolerance, or for that matter pay some jobber to run it through a centreless grinder - you'd get the same finish as drill rod (oops silver steel). hmmm larger dia 4130 mechanical tubing hardened and ground...

I can't speak for the UK, but know the metals distribution business in NA. Any industrial supply place will bring in the drill rod you need but failing that a metal's distributor (called service centres here) would bring it in. These guys for example
http://www.metalsupermarkets.com/default.htm

should be able to get you what you need. I'm sure there others but they come to mind as a friend founded the business so I knew he'd expanded in the uk (no longer owns it though). thats just a for instance, metal prices are a seemingly random walk so phone around

[NC Cams]

Pleasantries asside, we seem to have devolved as follows:

A member inquired about doing avante garde things to bearings and asked for opinions about the feasibity.

The "things" were experimental :idea: so as to facilitate creation of a low cost linear guide for a router (low risk loading situation - this is a critical caveat in this instance).

Suggestions were made to do things in alternate ways contrary to what was planned/proposed/deemed acceptable by original member.

In spite of suggestions to the contrar, member pretty much went and did them the way he intended anyway.

More discussions pro and con with recognition given to avante garde techniques that "shouldn't work" but did :wee: (things like that happen - sometimes they don't, however ).

More questions posed about testing and answers given why this would/would not be feasible.

The net result seems to be that the avante garde idea(s) are working. IMHO, probably because of the built-in robustness of the bearings in concert with the relative light duty that the modified bearings are seeing/will see .

Feelings got hurt because:

a. the member feels that comments were made charging that his integrity as a machinist and creative thinker have been insulted :argue:

and

b. the contributors who made genuine and well intended contributions felt that their suggestions were categorically ignored - the member seemed to do what he pleased to do in the first place regardless.

c. the after the fact reportage that the concept work gives a "see, I did it anyway and it works - now what do you say about that???" (chair) sort of impression to the contributors.

I think that about covers it the way I see it.

POINT: I have a STRONG AVERSION to ANYONE but an EXPERT modifying bearings - most DIY'ers don't have the tools or skills to modify them PROPERLY. However, in this case the mods are effectively benign due to loads and speeds. Great = we all learned something and the user got by.

HOWEVER, the conveyance of the idea (hacking on bearing rings), when improperly applied or done (and it can and does happen VERY easily) can result in shattered raceways and people getting hurt in more critical situations.

Why? Thru hardened 52100 simply doesn't like to be treated like the member has treated it (long story)..... Having done bearing applications engr work for a number of years, I"ve heard/seen the horror stories of "neat ideas" gone horribly bad. :frown:

I have nothing against creative people doing innovative things. Been there done, that. However the unavoidable conflict :boxing: of A and B followed by a perceived C above can't help but result in feelings getting hurt ....

Before this degenerates further, at this point, I intend to step asside :tired: and may continue to watch the progess of the project from the sidelines....

I wish the member well and would urge him to consider making some radius'd sleeves to fit the bearing OD's in lieu of hacking on them. IF for no other reason, it will be easier to replace bearings later by R&R'ing the sleeve :wave: