[wildwestpat]

Hi Mark

The fit of the bearing to the shaft will be covered in the bearing specification. Press fitting the bearing will of course expand the diameter of the inner race and this alters where the balls roll with respect to the outer race. If carried to extreem this impacts on the life of the bearing as does the amount of pre load used. Pre load also keeps the balls in contact with the race way and helps stop then skiding which leads to shortened life as the balls wear with flats on them!

To state the obvious the life is reduced with speed of rotation and the forces applied. There are a host of other factors including lubrication - cage material - type of load etc. In service life calculations are a bit gutsy and some manufacturers have simplified the calculation by providing calculators on line. If you have an evening to spare try looking at the web sites of the big world players in precision bearings and see how all these factors impact on the life that can be expected.

With high speed rotation balancing of the rotating parts and run out are probably going to limit performance life as much as the fit of the bearings. For precision work shrink fitting is common as is the use of anerobic adhesive and certainly no hitting!

Good luck as I want to see some pics of a finished spindle.

Pat

[RotarySMP]

Hi Pat,

Thanks for the answers.

The manufacturer went TU after the German reunification (or at least got bought out). There is no info in these bearing on the web, so I can forget spec sheets.

If you have experience on similar AC bearings, would you mind giving me some ball park numbers on the thickness of the shim stock to space the outer races, and the preload torque on the inner races?

From the SKF site, the minimum load for a normal quality 7204 would be about 460N. For a P2 spindle bearing on a JS3 shaft, SKF recommends a fit in the range of +0.004mm (press) to -0.001mm (sliding), given what I have measured, and the feel of the bearing pushing it by hand about 3mm onto the shaft, I would guess that this shaft/bearing pair are quite well matched.

The fit is not going to be very tight, but will be beyond a hand fit. As best as I can measure them, the ID and Shaft are seem to be within a .0001". If pack the shaft in dry ice, and heat the bearing, I expect I could hand fit it, but that does not help my assembly problem, of not being able to then install the bearing cover screws.

[wildwestpat]

Hi Mark

The fit on the shaft sounds good to me. Rather than press the bearings on to the shaft when the time comes it might be better to freeze the spindle using one of those aerosols used for fault finding on PCBs and also used by plumbers for working on pipes with out the chore of draining down the system by freezing the water with a very very cold spray of gass. Warming to hand heat the bearing but no hotter than hand heat or you could fry the grease, seals and cage.

The shims for pre loading are as far as I am aware specified by the manufacturer and supplied as part of the set. This helps in part to explaing the vast increase in cost associated with matched pairs. I also think the trend is to having the inner race ways sized such that with the outer races back to back the loading is determined by the faces on the inner raceways being in full contact.

Now to your problem the manufacturer has ceased and you have no data. Are you sure the bearings you have are a matched pair or are they two independent bearings that just have the same part number? If the former there is a good chance that they may not require a shim. This can be checked by making a jig consisting of two cutaway thick spacers with a bolt that goes through the spacers to make a sandwich with the bearings. The thick ends of the inner races facing the spacers and the outer rings in contact with each other i.e. the thick side of the outer rings are in contact. Tighten the bolt to barely finget tight. Check the gap between the outer bearing rings. This should be zero if they are a matched pair. Now the question is are the inner bearing rings in contact. This can be found out approximately by measuring the distance between the inner races withe micrometer anvils in the cut outs in those spacers you have made for this jig. Compare the length in the jig with that of the sum of the two race inners measures separately. There should be a gap if they are a matched pair and it should be small. Now increase the bolt up pressure and measure again BUT only bolt up to a relatively low torque and keep checking that the outer race pair which should rotate as one is free to spin. This will establish the shim needed.

This can only be regarded as a rough and ready approximation and depends on the measurements being done accurately. As you have found measuring to tenths is an art not a science with a micrometer!

I am not advocating this as a method for general use and can not stress too much tha bearings should not be removed from their protective packing till the very last moment. This also keeps matched pairs together and avoids any doubt. Bearings that are loose are unlikely to be matched pairs.

There is a category of double row agngular contact bearings that are manufacturered as a single unit. These avoid many of the problems - look at SKF 4201 ATN9 good for 20,000rpm and a nice radial and axial loading as well. (About 20 euros or less for a single which is equivalent to a pair of A/C bearings in SKF.) The series of bearings are also known as 3205 and 5205 depending on source and application details. I am using these for ball screw mounts and can say they are pre loaded and are easy to mount. Just dont make the spindle too big and dont force the outer into a bore that is too small as this also messes up the internal clearances / preloading. I am using Locktite with just finger pressures to assemble. This is also because I want to get the thigs apart and moderate heat should make the Locktite give up it grip. Or so I hope!

Hope this helps

Pat

[RotarySMP]

Thanks for the response Pat.

The bearings I have are not a matched pair. Would you be able to upload a sketch of that jig please?

I guess to address the assembly issue of the spindle design here, I will have to put the spindle shaft between centers, polish down the bearing seat to a nice hand sliding fit, and go a little looser on on the shaft between the bearings. Then I could assemble the bearings into the housings, with the bearing cap bolted in place, and slide in the spindle afterward.

[wildwestpat]

Hi Mark

I am on the road for the rest of the week so will not beable to do any drawing till the week end assuming I get some free time. You might review how you are going to assemble the spindle into its housing as these bearings are expensive an need care in fitting.

Regards

pat

[RotarySMP]

Hi guys,

I have been slowly working on my low speed spindle, but it is not too cold to stand on the balcony for any length of time in the evening.

I found an early use of the method of bearing preload Herbertkabi demonstrated here. Turn out cataract (turned into Hardinge) used it on some of their early Lathes.
http://cnczone.com/forums/showthread...=12184&page=11

Pat, did you get a chance to sketch up that bearing preload measuring jig?

Herbert, do you know of a source for shim stock in Europe? I speak german and work with Austrian engineers, and searched ebay for:
Abstandblech
Distanzblech
Ausgleichsblech
Merklblech
Futterblech

And keep turning up blank. I am going to need to get some thin shim for preloading my bearings.

[Herbertkabi]

Very-very seldom you find the right size shims - I make needed shims myself.
I have tried to explane how I make it sevaral times - seems none understand my DIY English
Lets try to illustrate ...
You need to find different thickness thin sheet materials, like bronze, brass ... kapton ...nickel alloys ... Stack pieces together between two alu disks via centre with adequate bolt (screw) ...
Then using rotary table and milling machine (or just cnc mill) cut out needed size ring - will get a lot of shims with different thickness. First large shims, then smaller ...
Cheers,
Herbert

[JLSD]

Hi Mark,
You could do worse than looking up local model engineering groups, or competitive aeromodellers. I've cut plenty of small cylinder head shims/gaskets over the years, and my material has generally been from model eng. suppliers.
If you're really stuck, PM me, and I'll see what's tucked away.
John.

[RotarySMP]

Thanks for the offer John. I'll need to make pats fixture to measure up what thickness I actually need.

Thanks for the pictures Herbert. Have you also tried punches?

Shim stock must be available here in Europe, I just have to work out what german name is being used. I guess if I can't figure it out, I'll have to drink some canned beer . I am guessing that aluminium drink can is too soft to make ideal shims though.

My Dad is coming out from NZ in a couple of weeks, so I got him to dig around the shed and bring out some shim stock. The thinnest he could find is 0.3mm, which I expect will be too thinck for the P2 bearings I bought.

[Herbertkabi]

Thanks for the pictures Herbert. Have you also tried punches?

.

I hope you meant not punches like used for carton or rubber
Very hard to make adequate "punches" (stamps) to stamp precission shimes from such thin sheets. It is possible, but what is costs ... ... and every size needs own stamp ... inner and outer ... and mother and father ...
cheers,
herbert

[BobWarfield]

What ho? The spindle building game is once again afoot!

Spindles are one of the "pinnacles" of machine work. Very exacting, very interesting, and very valuable when you get a good one.

I've collected a lot of data on this over the years. Eventually I want to build a spindle cartridge for my mill that's capable of 10K rpm or so. It'll be R8 or perhaps NMTB30. My notes can be found here:

http://www.cnccookbook.com/CCMillBeltDrive.html

Meanwhile, a few thoughts:

- Avoid "pressing" bearings unless its to remove known bad bearings for replacement! NCCams always used to say it should never be done. I have seen that corroborated in a number of places. There are spindle rebuild threads on Practical Machinist, this board, and other places on the net that are worth tracking down. It is very easy to brinell the bearings at which point you just bought a new set.

- Always heat shrink the bearings onto the spindle as WildWestPat suggests. You can go hotter than hand warmth though as we'll see shortly. In fact, if you check the spindle rebuild threads, you will find precision machines that don't even require heat shrink. That may be important while you're studying preload shims.

How much preload is the eternal question unless you are fortunate in having a matched set of bearings. The answer is simpler than you would think:

As much as possible without damaging the bearing's operation and life.

But how do we know? There are two approaches I've seen that make sense to me.

First, go get the bearing catalogs and see what they recommend. Even if there is no data available for your particular bearing, use data from similar bearings. OK fine, I know the spec, but how do I "make it happen" with my bearings? On the cnccookbook link are ideas for a couple of test cells to do that. You basically need a controlled way to apply that much preload so you can measure the displacement of the races and then apply a shim equal to that displacement.

That's a lot of trouble, eh? So here is the second way, and one I like a lot better.

The usual way to get into trouble with too much preload is heat. There is a threshold where things are so tight that as you run the spindle it gets hot, things expand, that increases the preload, things get hotter, and you get the picture. Not good!

Well how hot can it be?

The patient stalker of information eventually finds a figure of 160 degrees F. This is why I say you can go hotter than hand warm when heat shrinking! Bearing shops keep a hot plate set to an appropriate temp for doing this.

For experiments, get a handy infrared thermometer. Make yourself a test jig where you can run your spindle on the bench. Crank it up to desired speed and shim the bearings until you can run as close to 160 as possible and the temp is stable. I'd let it run for an hour to be sure it's stable. If it climbs, you've got too much preload. Take it apart (remember the comment about a non-heat shrunk spindle? Here is where you are happy to have one) and reduce the shim size.

BTW, a real bearing shop wouldn't use shims. They'd put the bearings on a special jig with a surface grinder and remove some meat from the race so that it preloads when you torque it down. I've never seen detailed procedures for this, but if you have a surface grinder it could be figured out once you have the desired shim specs.

Phew. Suddenly this seems like a lot of experimentation and work!

Sorry, there is a reason precision spindles are so darned expensive. But, if you like to experiment, it is fun work at least, eh?

BTW, Herbert, love your spacer cutting jig. That's going on my spindle page!

For those that don't want to fool with it, McMaster Carr sells bearing shims too.

Mark, looking forward to seeing your new spindle underway!

Cheers,

BW

[wildwestpat]

Hi Mark

Unfortunately due to the rush in the office I am not going to be near a drawing program on which to make a sketch. I will try and post a hand drawn one some time tomorrow between customer visits if the scanner will play ball! The weekends are a domestic no no in the run upto Christmas so no luck then either.

I am a little concerned about the use of shims as the AC bearings are very ridgid which is after all why they are good for spindles. This means that they will take a very large trust load with little axial movement. Put another way the sensitivity to shiming as a method of pre loading will be very, very sensitive. I have used fine screw threads to run a machined collar and nut down on the race to pre load it. (The pre load being applied by the center raceways with the outers force fit in the housing.) Adjusting the pre load by feel. Tighten until the drag can be felt and lock in place with just perceptible drag chech and adjust until this is achieved. I think shiming will be difficult if not an impossible way of getting the required pre load.

I am also concerned that the bearings should be a tight fit in substantial housings if the best is going to be got out of precision bearings. The force fitting does not of course imply that the balls are under threat as the inner is pressed on separately from the outer. It looks as if the need to have the pre load sensitivity is getting mixed up with the simple expedient of having an adjustable preload via a thread on the inner part of the spindle where a slop fit is required to make the adjustment.

If you are interested look at the fitting instructions for angular contact bearings and also at the fit tolerances on the shaft and housing. This information is for manufacturers of high speed spindles and is not applicable to DIY. The fits specified will require assistance in fitting the bearings either by thermal expansion or use of a press with suitable jigs. Angular contact bearings are easy to fit with a press but can not usually be removed without wrecking the bearings. This is why high speed spindles require specialist services both in terms of technician trainning and in the sophisticated equipment required and explains why there are specialist facilities that offer to service these types of assembly fitting new bearings as a matter of course. To make many of the assemblies practical where there are multiple banks of bearings flush ground matched sets are specified. i.e. clamp them up tight and the pre load will be perfect! All this is not DIY! I have been using the 320X\520X series of double row angular contact bearings with some success on lead screws but have not tried them in a spindle as I am satisfied with good quality tapper roller bearings.

Hope this helps.

Regards

Pat

[Herbertkabi]

hi Pat,
You read a lot - respect - but never tried?
Readings are primary before to do something, of course, but then next step is to try. Fitting of bearings, like many things in relation to Spindle, is question about experience and intellect, ... and also good fortune ... at least my high speed spindles, these ones where I have used super precision angular contact bearings works very well. Some of theses have been used long enough time, so I can say - are long life tools. To be precise about ´intellect´ I meant - just do not foolishnesses
Anyway - viva DIY :cheers:
Regards,
Herbert

[wildwestpat]

Hi Herbert

I think you have put it in a nut shell. In all things there is no substitute for hands on experience certainly bearing fitting being no exception. The current trend is away from the very tight force fits of yesterday to the use of aerobic locking fluids to reinforce with what can best be described as a light force fit.

I have spent some thirty years in satellite and radar research and this has given me a very wide range of experience in both mechanical mechanisms and electronics. During this work I have had the pleasure of working with some of the best fitters and toolmakers in an integrated team. It is no good dreaming up theoretical parts if they can not be made! In this sort of environment where the team costs are very high it is necessary to defend all all design descisions with solid information. This is why I always like to have the manufacturer's data sheets and particularly the application notes as these make a good springboard for a different application.

The world wide web has made access to data sheets so easy but it has also unleashed a lot of opinion based stuff from armchair practitioners who have never made anything or just assemble kits of prepared bits!

You have guessed my guilty secret is insomnia! Hacking metal is not very popular when everyone else is asleep so reading is about the only option.

Your work on high speed spindles fascinates me. In particular how are you balancing the spindle - chuck - drive coupling etc?

Keep up the good work

Regards

Pat

[Herbertkabi]

hi Pat,
I have acquired and took apart seven high speed spindles made by well-known companies - never met problems with removing/inserting of bearings - was easy with help of common means you can find even in most of garages. Yes, good sense is the most of you need for - two first spindles was not possible to assemble again because did not make correct notices and drafts about line up, but next ones was assembled after repairings successfully and sold away (understandably I bought defective spindles).
Presently Im busy with two plane soft bearing dynamic balancing machine ...
also categorized as "no way to DIY" ... but in reality factualness speaks quite the contrary. Found few new friends and with some helps I have progress about. When done then ca $15k saved. OK, I can also buy but why ... at that I like such kind of challenges, I like DIY.
Best regards,
Herbert

PS:
In some points you are correct - military stuff - you never can remove bearings without damages ... as well as another kind of tech. like even common hard drives ...

[wildwestpat]

Hi Mark

I have sketched out the jury rig for measuring the clearance, if any, between the inner raceways of non matched pairs of anular contact bearings. However on reflection I do not think the pre-load can be determined by measuring the gap between the raceways of uninstalled bearings to sufficient accuracy. The problem is that angular contact bearings are designed to be ridgid and resist both axial and radial movement. This means that there will only be a minute movement from just touching to full load. With such a small range of movement between full load and no load, with the balls making contact with the raceways at all times, I have changed my mind. I now think it is necessary to adjusting the bearing pressure as part of the final install process and not as a predetermined pressure applied by shims. In any case the spindle should be run in for a few hours and the pre-load rechecked and adjusted as necessary.

For the above reason I am withdrawing any remarks I made about measuring the gap as I do not think this will work as it is the as installed pressure that has to be adjusted.

Kind regards - sorry about changing my mind.

Regards

Pat

[wildwestpat]

Hi Herbert

Looks as if you have a health desire to do it cheaper and better than any stuff that is available. In my book this is what 'engineering' all about. In many cases the break throughs are achieved by gifted amatures and it looks as if you are one of them.

I do not know where the term DIY has been used to imply limited skill. In my book it means with limited resources in the way of tools and equipment. The active mind is of more importance than all the test equipment and fancy machining gear in the world is in the possesion of people with no imagination or hunger for a better way of doing things.

It is just that when I hear of people running the risk of abusing bearings to the point that they have dented (brinelled) the bearing surfaces I think 'do these people have any appreciation for what they are doing'. It is very obvious that most of the pepole contributing to this thread are not going to do that. But there is a hidden risk with high precision bearings where the determination of pre-load by measurements that are taken prior to the bearing being mounted and the risk that the pre-load calculated will increase too much when the spindle is assembled. This is most likely to show as tram ways and flats on the balls after a short period of running rather than the typical dents caused by subjecting the race ways to a crude brinell impact test. The excessive pre-load would also show as excessive heating as well as nasty noise.

The objectives of military and similar designers is for absolute reliability and removal of all adjustments can be a serious objective on many of these projects. Again it is horses for courses.

Wishing you the best of luck in achieving a means of achieving dynamic balance.

Kind regards

Pat

[RotarySMP]

I really appreciate the discussions here.

No worries Pat, I am not a lawyer, and I am not going to sue you unknown east german P2 bearing on a Chinese ER-16 collet spindle craps out

Since I bought these bearings off Ebay, and the manufacturer is gone, no data is available. To be honest, If I had all the data and knew exactly what I my target preload is, I doubt I would hit that target first time, seeing as I am working with a CNC'ed 7x12 chinese mini mill, out on the balcony in the middle of winter. I am using a lot of time (http://www.5bears.com/cnc16.htm method ) to get tight sliding fits, so that I can assemble by hand without pressing parts together. If I get something to work nicely, I could then look at using a bearing locking fluid (Calling it glue is not engineering cool, right .

Maybe a reliable, high speed spindle is not possible given my tooling, environment and machining experience, but it is fun to try. I bet if I make this spindle and run it till it fails I'll be in a better knowledge position when it comes time to make the next one, and a better all round machinist.

I do have decent measuring tools, (once my father brings out some decent Mitutoyo telescoping guages at christmas, so I should be better able to precisely measure bores cause my current ones are a definate weak point in my tooling).

Yesterday I half assembled my low speed ER-32 spindle and it became clearer what you were talking about in terms of a bearing play measuring jig. I still think using your method is valid as a first step to get a feel for what range of shims will be needed.

Bob made a good post above about using temp rise to confirm preload.

I am guessing that the preload range for these precision AC bearing is going to be less than a thou (0.025mm) from too loose to too tight, as Pat points out. What do you guys use for such accurate shims? McMC seems to only have shims starting as 0.001" (and only for some sizes, for 20mm ID I think the thinnest are 0.003". Beer can is about 0.004" (and I don't like using aluminium in machine parts).

Herbert, did you end up using the thin Aluminium foil from your sandwiches to fine tune the preload? What thicknesses did you end up using on your AC bearing pairs on your last couple of spindles?

Looks like I'll order a bunch of shim packs form McMaster Carr and get a friend to bring them over when they do an aircraft delivery.

[BobWarfield]

I do have decent measuring tools, (once my father brings out some decent Mitutoyo telescoping guages at christmas, so I should be better able to precisely measure bores cause my current ones are a definate weak point in my tooling).

...

I am guessing that the preload range for these precision AC bearing is going to be less than a thou (0.025mm) from too loose to too tight, as Pat points out. What do you guys use for such accurate shims? McMC seems to only have shims starting as 0.001" (and only for some sizes, for 20mm ID I think the thinnest are 0.003". Beer can is about 0.004" (and I don't like using aluminium in machine parts).

Herbert, did you end up using the thin Aluminium foil from your sandwiches to fine tune the preload? What thicknesses did you end up using on your AC bearing pairs on your last couple of spindles?

Looks like I'll order a bunch of shim packs form McMaster Carr and get a friend to bring them over when they do an aircraft delivery.

Some thoughts:

You might want to consider a line boring setup for machining bearing pockets on your lathe if the pockets are at all large relative to the lathe's capacity. Such an arrangement will be more rigid.

RE measuring these bores, I think telescoping gages are going to be tough. It's really hard to achieve the degree of accuracy necessary with them, and they're very subject to technique. A dial bore gage would be preferred but expensive. I haunted eBay for a long time before I got a reasonable deal on my Mitutoyo bore gages.

But there is an easy and accurate solution. Make yourself up some go/no go plug gages. You can measure their OD's with a micrometer. I've used this approach a number of times before I got the bore gages and it works very well.

On the shim stock, you're just going to have to run some experiments to see what is needed. It won't be a lot, but I don't know that it's way less than 0.001 either. If it was me, I'd use shim stock and Herbert's method of machining that into shims.

RE the whole DIY question, machinist's are DIY by definition. A machinist should be able to build whatever they desire with the tools at hand. There are no end of examples of talented machinist's accomplishing wonders ranging from POWs in Japanese camps building machine lathes from nothing (great story from Lautard's books) to the many examples on these boards. It is poor marketing that leaves the DIY moniker in short shrift.

People who build airplanes in garages don't call them DIY or homemade, they call them experimental aircraft. Better marketing!

I would suggest the equivalent here is that your spindle is "shopmade" as you've demonstrated you are a machinist.

Cheers,

BW

[JLSD]

Hi Mark,
I've been head-scratching over this, and attached are a couple pics that may help. One is my trusty setup for shims, just an acetal pad in the chuck and an opposing pad exactly I/D to clamp from tailstock. Tool is just a hand ground inverted HSS knife edge. (Could've made 20 shims while I was typing this!!)
The other would give a basic method of assessing preload from the drag between the unclamped raceways. Not perfect, but at least you'd know whether you had any! Given a pair of bearings with known preload would set some kind of start point.
As to the general principle of spacing raceways, I'm not too confident about the idea of super-thin shims, and one of my baby spindles will be a real cheapo with only two bearings, but inner and outer races clamped, using different length spacers, e.g. 30.00/30.01.
The spindle arrangement will allow setting the loading of the core assy BEFORE committing to the housing, using similar torque arrangement to sketch.
John