[martinw]

Dear folks,

I have bought some double row angular contact bearings from an e-bay supplier.

They are 5201ZZ types and are unbranded (yep, you can guess where they came from). I entend to use them at each end of a 12mm dia. leadscrew. Max speed say 700rpm, minimum thrust, and a very light pre-load. Given the inaccuracies of the MDF 3 axis router to which they will eventually be incorporated, I am blissfully unconcerned about their suitability, so please do not rip me to pieces!

My only question is this..

On the inner race on one side of the bearing, there is a mark that looks like a "minus" sign. My guess is that indicates a preferred orientation.

Is this correct? and if so, which orientation?

I am donning my suit of armour as I type!

Best Wishes

Martin

[martinw]

On the inner race on one side of the bearing, there is a mark that looks like a "minus" sign. My guess is that indicates a preferred orientation.


Martin

I mean ring not race

Martin

[Willbird]

Well I'm sure the experts will sign in soon, but I'll share what I understand.

You would mount two bearings preloaded and opposing one another at one end of the leadscrew(assuming they are single row angular contact), and at the other end just a radial bearing.

If you hold the bearings race to race you should be able to see the preload, the typical way they are used the inner races would be slightly apart when the outers are in contact as I remember it, the nut on your leadscrew will pull the inners together, and you can shim the outers further apart to create more preload, in a perfect world you would send them to a bearing guro to disassemble them and grind the races for greater preload and re-assemble them, you and I will probable just shim them :-)

http://www.skf.com/portal/skf/home/p...&newlink=1_3_3

there is some info there.

edit > I just re-read your post, and saw that I missed the double row part

http://wanrui.en.alibaba.com/product...s/showimg.html

is this how your bearings are made ??

Bill

[martinw]

Well I'm sure the experts will sign in soon, but I'll share what I understand.



edit > I just re-read your post, and saw that I missed the double row part

http://wanrui.en.alibaba.com/product...s/showimg.html

is this how your bearings are made ??

Bill

Dear Bill,

First of all, I'm not worried about the experts "signing in" , I'm worried about them devouring me (hence my suit of armour). My major concern is NCCams emerging from his abode breathing fire! (Only joking, honest)

Yes, they are double row angular contact bearings. I understand that they can take axial loads in either direction, but I read somewhere that the ball loading is done in one direction when they are made, hence a preferred orientation. Now, in my future application, my guess is that these approx. 1/2" bearings are "way over the top", hence my relaxed attitude to loadings and speed etc. However, if I do not need to chuck away accuracy on a minor point of ignorance, why do so? Hence the original question.


Thank-you for taking the trouble to reply. I appreciate it.

Best wishes

Martin

[Willbird]

I'm on the same point of the bearing learning curve as you are :-). I had not looked much at the double row ones like you are referring to, the ones in the link look like you are locked into the preload they are built with ??

Just looking at the picture in that link would not leads you to believe orientation would matter.

Bill

[martinw]

Bear Bill,

I think we may have to wait for the "Big Beasts" to wake from their slumbers.

Best wishes

Martin

[NC Cams]

5201's as double row angular contacts BUT they are not and can not be preloaded. They are simply a prepackaged bidirectional thrust carrying bearing that are lower in cost and lower in stiffness than a true A/C bearing.

They are not nor should they be confused or misapplied with with the stiffness and thrust carrying potential of a 7201CTYDUL or DUM or DUH style TRUE preloaded A/C bearing which will have essentially the same OD and ID but not the same width - check it out in the specs.

5201's are not typically sold as ABEC7's as I recall - too many tolerance stacks for them to meet the ABEC7 runout specs, especially the axial runout specs of a class 7.

EDIT Until or unless you can find a way to "move the raceways", you aren't going to be able to axially preload the 5201's. These tend to have RIC (radial internal clearance) built into them - they are not designed or intented for PRECISE axial positioning due to an inability to preload the raceways.

You want preloaded raceways, get some 7201's and buy them preloaded or shim them but forget about axially preloading a 5201.

END EDIT

[martinw]

[QUOTE=NC Cams;231314]5201's as double row angular contacts BUT they are not and can not be preloaded. They are simply a prepackaged bidirectional thrust carrying bearing that are lower in cost and lower in stiffness than a true A/C bearing.

QUOTE]

Dear NC Cams,



I appreciate that there is no pre-load within an individual 5201.

If you have two 5201 bearings separated by say 4 feet on an entirely rigid structure, and the two bearings are connected by a rigid shaft, and the shaft is tensioned against the inner ring of both bearings by, say, a nut at one end,
I would imagine that each bearing is pre-loaded. In other words, the shaft is pretty well fixed axially, all other matters being ignored.

Do the markings on the inner ring of the bearings have any importance.


Best wishes

Martin

[NinerSevenTango]

If you put them at opposite ends of the shaft, your preload will depend on the temperature. Whatever else you do, you should probably think about putting both of them at one end of the shaft, and just support the other end with a radial bearing.

Sorry I can't help on the markings, it's been too long since I used the 5200 series.

--97T--

[NC Cams]

The mounting scheme outlined in post 8 is effectively a "position" preload. Essentially, you're relying on the outer set of bearings in each end bearing to accept axial thrust.

This may be sufficient for the application - I dunno. HOWEVER, as soon as you start to accelerate HARD in one direction or the other, it may be possible to unload the bearings and thus experience compliance. Time will tell if your mounts and methods are adequate.

Or, if you try cutting circles, you'll start to see flats or dwell marks at the 12, 3, 6 & 9 o'clocke positions due to hysterisis in the system if your mounting (regardless of the apparent robustness) is of sufficient rigidity. Keep in mind that hysterisis can creep in at any moving joint. Thus, any hysterisis could be coming from anywhere BUT the 5201's.

Folks have tried to tension a screw to optimize positional accuracy - some swear by the procedure, some swear at it. No matter, it has to be done properly or else you'll not be satisfied with the performance. Too much stretch and you'll have axial lead problems creep in especially in long cuts. Too little "stretch" and the screw will have too much resilient spring.

My main reason for explaining the lack of "preload" of double rows is to quash the misconception that they are RIGID, non-compliant a/c bearing assemblies - something they simply are not nor should they be expected to be.

Based on your original question (and known practices of misapplying them), one would not expect that you were "stretching the screw" as you indicate in a subsequent post.

Under the circumstances, you could get by just as well with a single 7201BYG properly positioned on either end. Same OD and ID and a true A/C. Besides, since you're stretching the screw, the "inner" bearings in each 5201 is essentially unloaded and not doing much for you anyway.

If the 5201's were cheap, so be it. However, for machine tool applications, knowing what I know about double rows, I'd chose something a bit more application suited - a 5201 isn't noted as being primarily being used to support ball screws to any appreciable degree.

Essentailly, they are an inexpensive way to get 2 @ 6201 or 7201 bearings into a common, smaller package size. You'll never find them used in a properly designed ball screw application due to their lack of axial and/or radial rigidity.

EDIT: Markings: these are not universally standard within the industry. True A/C's often have their eccentricity high points marked but these are either chevrons < or > or they could be little circles "o" etched onto the face of the inner ring.

Only the "no name" bearing company who made the bearings or the rework facility that marked the inner rings would know that the true/real purpose of your marking are.

END EDIT

[martinw]

If you put them at opposite ends of the shaft, your preload will depend on the temperature. Whatever else you do, you should probably think about putting both of them at one end of the shaft, and just support the other end with a radial bearing.

Sorry I can't help on the markings, it's been too long since I used the 5200 series.

--97T--

Dear NinerSevenTango,

Thanks for that. The original idea was to try and keep the screw in tension at all times in order to reduce whip, but I do appreciate that the coefficient of thermal expansion of the screw might call for a really monstrous pre-tension. I have not done the calculations yet but I think your suggestion is an excellent one.

Best wishes

Martin

[martinw]

Dear NC Cams,

Thank-you very much for your trouble. You cover a lot of ground, a lot of which still has me scratching my head in ignorance, despite your excellent explanation.

I appreciate that 5201 bearings may be quite unsuitable for precision applications. Although I do not wish to recklessly chuck away precision, I am quite prepared to take a few risks by using 5201s. They were ludicrously cheap, and even if I use better and more appropriate bearings as you suggest, my whole machine is likely to be pretty rough and ready ( to say the least!) compared to a race car crankshaft, or a commercial machine tool.


If you don't make mistakes, you don't make anything.

Best wishes and a very Happy Christmas,

Martin

[Zumba]

I assume you'll only be using one of these bearings on each axis, the motor side. The opposite side of the screw will be "floating" in a standard radial BB, no preload. In that case, you're going to have to choose... is your axis moving to the left more important than moving to the right? Or vise versa?

Yeah, it's silly. I'd say flip a coin. In your application, you're so far away from maximum load capacity of the bearing that there's really no reason to worry about it.

It's probably similar to a double-acting hydraulic cylinder... most cylinders push more force than they can pull, but some are the opposite. As long as your load is far below the maximum capacity in either direction, who cares.

[Zumba]

P.S. NCCams is probably the most knowledgeable bearing expert I've ever known of. If I were repairing a precision machine tool, I'd probably try to dig up every single post he's ever written.

However, we're talking about woodworking here. Most people report "excellent results" with machines that have Porter Cable 7518s as their spindles. Those routers have about .005"+ of runout in the spindle from what I've heard. I haven't actually measured it myself, though I think I may just have to do that later today.

The point is, .001" of radial runout (which is an extremely pessimistic estimate) in the 5201 bearing that Martin bought is negligible. How much of that runout would actually translate to axial inaccuracy? Far less than the ballscrew backlash, WAY less than the spindle runout.

BTW, did you buy your 5201s from VXB for $9.95? Black seals? Yea, I have the same ones. They sure are sweet for the price.

[martinw]

I assume you'll only be using one of these bearings on each axis, the motor side. The opposite side of the screw will be "floating" in a standard radial BB, no preload. .

Dear Zumba,

It's all up for grabs at this stage. I'm now thinking of using two bearings at each end.(1) Two angular contact bearings "pre-loaded" at the motor end to take out some axial play in these bearings and to stop axial loads on the motor, and (2) two standard BBs at the other end with the non-motor end of the screw floating.

One reason for using two bearings at each end is to increase the critical speed at which resonance occurs, and hence reduce whip. I have seen a factor of 2.23 mentioned if you use two bearings at each end. Quite how far apart each pair has to be is unknown to me at the moment. I'm looking into it.

Best wishes

Martin

[martinw]

P.S. NCCams is probably the most knowledgeable bearing expert I've ever known of.
BTW, did you buy your 5201s from VXB for $9.95? Black seals? Yea, I have the same ones. They sure are sweet for the price.

Dear Zumba,

I totally agree with you about NC Cams. I've learned an awful lot from his advice over the last year. Unfortunately, I have a long way to go.

Thanks for your comments about tolerances. Believe me, the radial run-out figures you mention will be swamped by the other short cuts I'm likely to take. At the risk of provoking general uproar all round, I will not say what these are likely to be!

Yes the 5201s I got were from VXB, but they have metal shields not rubber seals. I thought that the seals might introduce more friction, but I could be wrong. Same price I think but that was back in April.

Best wishes

Martin

[NC Cams]

Accuracy is relative. Relatively easy to get yet hard to maintain. Sooner or later, EVERYONE tries to do something more precise than they ever imagine/consider and THAT's when the compromises that are made come to bite you.

The lowly wood router starts to machine intricate patterns and then hysterisis creeps in. Duh what? What do I do now??? Go thru the whole she-bang again or preload it "right" the first time? I'm not in a position to judge so I tell the "machine tool way" and you can let slide what you wish.

HOWEVER, as soon as you do NOT pre-tension the (ball) screw, the hysterisis (internal slop/clearance endemic to a single deep groove radial double row bearing at one end, a regular ball at other) will start to become MORE self evident. Your dwell marks MUCH more self evident.

If you know the reason now, you can choose to ignore it or address the issue. Otherwise, you can say "I was told so" later on when the problems do start to develop and/or become more intense.

Its not like no one has asked how to solve this problem anytime before.

[martinw]

Dear NC Cams,

Thank-you for trying to save me from myself.

I am probably being stupid but I do not understand why, if using the configuration in post #15 above, the screw is not pretty much constrained as regards axial movement.

Leaving aside the screw-motor coupling arrangement, I am now thinking along these lines...

At the motor end, I have two 5201 bearings separated by say 3 inches. The outer ring of each bearing sits in a housing which prevents the outer ring moving in either direction axially. The two housings are connected together rigidly. The screw goes through each inner ring and is fixed with two nuts as follows, starting at the motor end..

Nut one
Spacer one
Bearing one
Bearing two
Spacer two
Nut two

Please note, the spacers each bear only on the inner rings.

Suppose Nut one is somehow locked to the screw. If nut Two is tightened, surely the axial play between the inner and outer rings of both bearings is eliminated, or at least severely reduced?

At the non-motor end there are two deep groove radial bearings, say 6201s also separated by 3". The screw is allowed to "float" axially through these. The purpose of having two bearings at this end is to reduce the effective length of unsupported screw.

Now, I fully appreciate that it is not desirable to have a threaded screw running against the inner ring of any of the bearings, and these are issues that have to be resolved somehow, either elegantly or otherwise.

I would be really grateful if you could give a brief comment on this arrangement.

Best wishes

Martin

[NC Cams]

I personally avoid commenting on a design based upon word pictures. They don't always convey the intent and/or my semi-dyslexic mind doesn't pick up the vagaries that need to be addressed and I then give bogus info.

In the mean time, until or unless you preload a deep groove ball bearing, the RIC (radial internal clearance) and/or any axial clearance that may result from the raceway curvature DIFFERING from the ball radius, you'll have compliance/hysterisis/deflection in the bearing. How much? It all depends.

RIC is easy to understand - axial play isn't.

However, if you picture a bowling ball rolling/waddling down a gutter you can see what's going on however when it comses to axial deflection/clearance potential. The ball rolls down the gutter all right but it can't/won't roll perfectly straight - not unlike what you're trying to prevent with a ball screw support bearing.

When you preload the bearing. you FORCE the balls to take a fixed, repeated path against the side of the raceway. You also prevent the bearing from deflecting appreciably under any further load (this is the A/C, axial thrust bearings - I don't care about any supplemental balls or bushings to prevent whip - at this or any other point).

By preloading the bearing, you remove any residual slop/initial deflection that RIC and/or "ball waddle" can create - This is often represented as "backlash" but in reality it is the initial bearing compliance which is technically NOT backlash.

Although dealing with backlash you often need to address axial compliance in bearings but not always - you could have ZERO deflection in the ball screw bearings and a sloppy ass ball nut or lead screw and still have backlash!!! Same effect, WRONG solution/remedy point.

Anyway, as you load a A/C bearing, it deflects as a funtion of RIC, ball contact angle, and raceway curvature. When you achieve enough load/preload, it stops deflecting axially and this is the point you're looking to achieve. Namely you are trying to take up the initial deflection potential via preload.

This is done as opposed allowing any internal deflection from occurring, which eventually stops growing once you load the bearing axially to a sufficient degree (read the prior statement slowly and carefully to make sure you understand EXACTLY what I'm saying).

Shims and spacers are just that, shims and spacers. Until or unless you truly and properly preload the bearings axially, they move and the deflection shows up readily at direction changes during cuts. The shim and spacers are merely packaging ploys to achieve a decent fit (not unlike a padded bra - it may look nice but it don't always create the real thing).

This issue of bearing compliance can be a lot - it can be a littl - it depends on each bearing and how it is mounted. Moreover, it can NOT be compensated for by any fancy back lash comp chicanery. Once you've ever cut with a machine that is deadly accurate, you're simply spoiled.

Need ideas? Check out some machine tool ball screw mounts - no sense reinventing the wheel. A section drawing of one can be found in most Bridgeport mill manuals. Simple, effective, KISS mount and easy to duplicate with a 7201CTYDUM or DUH bearing pair.

By the way, I have some 5210's in a follower wheel that runs on our cam grinder. If you load them wrong, you can get a non repetitive runout of upwards of 0.002" - makes it hard to hold cam base circle runout of 0.001" TIR with that much slop in a "precision" bearing.

WHen we can find the time, we'll engineer up a ABEC7 DUH duplexed something or other but the thing is simply too much of a PITA to machine/remachine. So we take our time and deal with the runout issues these "precision" bearings create. Too bad that the makers of an $85,000 machine couldn't have spent a couple hundred extra bucks and designed in the "right" bearings for the application rather than merely "convenient" ones.

Yes, protection from one's self is often a good idea. I learned that the hard and expensive way I'm sorry to say.

As is always the case, the member is free to follow or ignore any or all of the afore mentioned suggestions.

[Zumba]

Hey NC,

Lets say that instead of spending $10.00 on the 5201, I decided to up my budget to maybe $50. What kind of A/C bearings would be available to me and where do I get them? ID of around 15/32" btw. Thanks.