[Chrisjh]

Hi,

Don't know if this is the right place to post this. Please let me know if it needs to be posted elsewhere.

I have designed a simple R8 Spindle Assembly for milling machines and thought I would let others know my thoughts.

The design is intended to be able to be made by hobbyists with time on their hands but not much cash and is not intended to be world's best.

Here is an overview:

http://www.cjh.com.au/spindle%20design%20overview.pdf

Regards
Chrisjh

[arizonavideo]

I believe that the spindle bearing you picked is not pre loaded so will have backlash which would not be usable for a spindle.

Two standard AC bearing would be fine.

[PBr]

Also i don't see a locker screw which would prevent the main shaft from falling down.
You should use a pair of AC bearings as mentioned above and tie them together with a lock screw. The O or X configuration of them is a subject to further discussions.

[mactec54]

Chrisjh

The spindle looks good, The bearing choice is fine & I'm sure it works well, 2RS instead of the ZZ would of been better

Guys the front bearing that Chrisjh is using are preloaded, I use the same type of bearings as this, for my cheap spindles & ballscrews thrust bearings

The VBX is not the best brand to use, But they work, & are the cheapest, I mostly use SKF & NSK which have better tolerance

You do need though to lock the front bearing,Outside & inside, Just let the top bearing move for expansion

[RICHARD ZASTROW]

GMN has sealed angular contact bearings in 15 or 25 degree contact angle. Seals work better than shields in retaining grease, especially in a vertical orientation.

We use them regularly in our spindles. They are installed back-to-back with lock-nuts securing them against a shoulder on the spindle shaft.

For increased stiffness, a spacer is placed between the pair of bearings that is equal to the width of the bearings or wider. The spacers are very precisely ground to be equal length within .0001". Actually, the inner and outer races of a same sized bearing can make good spacers as long as they are equal lengths.

This spacing can also spread and dissipate heat generated by the bearings.

Dick Z

[arizonavideo]

Chrisjh


Guys the front bearing that Chrisjh is using are preloaded, I use the same type of bearings as this, for my cheap spindles & ballscrews thrust bearings

Where can I find out some info on the bearing stating that they are preloaded?

I have some of the smaller ones of 12mm ID and they had .0015 backlash.

[RICHARD ZASTROW]

Bearing angular contact mfg's catalogs (some are on line) usually give you a choice of light, medium or heavy preload.

The normal ball bearings (deep groove etc.) aren't normally made to be preloaded, especially axial preload. The "tightest" level is less than normal clearance, not without clearance or with preload.

Dick Z

[handlewanker]

Hi Chris, one thing with the R8 spindle end....the R8 collet end is about 25mm diam and if'n you compare this to the spindle diam you'll find that with a spindle wall thickness of about 5 mm you'll need at least 35mm int diam A/C bearings in back to back configuration.

I went down this path when I redesigned the spindle for a Beradi jig borer,( looks like a small vertical mill), and due to having an existing quill with a 2 Morse spindle and and no draw bar and wanting to use R8 the design got "iffy" when the wall of the spindle got to 3mm because the available bearing int and ext dimensions only allowed a spindle of a certain size.....I couldn't bore the quill internal diam out to allow a bigger spindle diam.

I was stuck with a quill with an end bearing housing that had been bored out to the max to fit A/C bearings, but the int diam of the quill became the problem regarding the spindle diam.

With a spindle you really want the R8 collet to be up as close and into the bearing area as possible, and R8 is 25mm diam at the end and being fairly long.

I could have dropped the R8 collet taper seating down a bit but it still had the collet end inside the bearing internal diam and this made the collet taper seating a fair distance from the bearings, not too much of a problem with decent sized A/C bearings and an extended spindle end, but not desirable design wise, so I decided to go for Int 30 taper in the spindle....problem solved, the short dumpy taper of the Int 30 has an end that gives plenty of spindle wall and is right inside the bearing area.

Except I now have to get (already got) an Int 30 collet chuck with ER32 collets and anything else like Morse adaptors, boring heads etc with Int 30 tapers.

Fitting an R8 system to the spindle means you have to get the spindle hardened and ground.......the spindle end does take a fair bit of wear and tear, being an active member of the collet assembly, one of the reasons I advocated going to an inactive collet chuck system.....the chuck being hardened and ground takes all the wear and tear.

BTW, the upper bearing set-up needs to be two sealed radial bearings spaced apart about 10mm and allowed to float axially, IE, not restrained.

The bottom A/C bearing set will need seperate lubing and sealing from ingress of dirt and loss of lube, and I prefer a labyrinth seal with .02mm clearance instead of the lipped seal contact type.
Ian.

[RICHARD ZASTROW]

Ian reminded me of the front seal.

We use Nilos metal seals. We fill the void between the bearing seal face and the Nilos seal with grease. It helps the sealing function until the Nilos metal seals "seat".

The main function of the Nilos seals in our applications is to prevent debris from migrating and contacting the bearing seals.

Dick Z

[Chrisjh]

Hi All,

Thanks for the feedback re my design attempt.

The feedback has prompted me to rethink several things. Mainly making sure that the lower bearing is indeed suitable ( I have found from researching that there is confusing (and in some cases, conflicting) info out there on the net. The objective of this design was to keep costs down so that hobbyists could afford such a spindle without having to resort to high spec, high cost bearings.

The main reason I chose the 5207ZZ was for the integral shields to negate the need for additional seals and I got the impression from confusing data sheets that they were lightly preloaded.

I am currently looking at a 7207BDB bearing set but am awaiting quotes from local suppliers ( I can't find one price on the net). I have asked them to comfirm if the 7207BDB pairs are available with pregreased and with integral seals, but I don't like my chances.

I need to investigate and research the Nilos Seals suggested.

Your feedback has also got me thinking about adding a simple "labrinth" to help prevent ingress of foreign material.

Regards

Chrisjh

[philbur]

Do you have a reference for this.

Also I think the top bearing needs to be preloaded otherwise the spindle will be capable of pivoting about the bottom bearings.

Phil

Guys the front bearing that Chrisjh is using are preloaded

[handlewanker]

Hi Philbur, not quite sure what you mean by top bearing "pivoting" round the bottom bearing, the top bearing is a radial load carying ball race (it could also be a roller bearing, but needs sealing) and as such only has to resist side loads only like the pull of a belt, and as a ball radial bearing you would be hard pressed to find any side to side movement in the top bearing, but to allow for thermal expansion the top bearing is allowed to move up or down on the spindle or in the housing, approx a few thou is sufficient.....the bearing being a sliding fit......this also allows the bearing to be lubricated and sealed for life, as it's very difficult to make a seperate lube system and seals in this area.

The distance between to and bottom bearings is such that if'n you gripped the spindle at the top and attempted to move it sideways the radial fit of the bearing would resist it as it was designed to do.

All the thrust and positioning of the spindle is taken by the bottom angular contact bearings placed with cups facing outwards and back faces in contact.....being held together in the housing by the front bearing retaining cap, which also carries the seals, and the spindle has a screwed collar to hold the inner races against a shoulder on the spindle.

In this manner they will be idealy place to run with out further preloading.

The labyrinth seal is not something to be lightly considered as it has to be very closely fitted with clearances of max .02mm on labyrinth faces.

The labyrinth ring attached to the shaft acts as a "flinger" ring initially and prevents the grease from running out of the bearing when it gets warm, at the same time the static part of the ring attached to the bearing retaining cap prevents the ingress of coolant and swarf by the fine clearances.....it also ventilates the bearing cavity so preventing air pressure expansion when the bearing becomes warm from forcing the grease out.

A normal lipped rubber seal will do an adequate job too as long as it has a smooth diam to run on....alternatively a felt seal can be fitted in a groove in the bearing retainer cap that presses against the spindle end diam.

If'n you are going to lube with an oil supply then it become pretty complicated to keep the oil in.....the secret is K.I.S.S.

BTW, I hope I haven't got the impression that being "designed for hobbyists with time on their hands" interprets to making it with bits and pieces lying around without a lathe or mill....ain't never gonna happen...trust me.....you need to have very good lathe experience and probably some milling too.
Ian.

[RICHARD ZASTROW]

FYI, FAG bearing 3307-B-2Z is a split inner ring, sealed both sides, double row angular contact ball bearing. More specification is required to determine preload.

5207 double row, angular contact bearings in C2 clearance (less than normal) can have as much as .00062" axial play.

Dick Z

add: More FYI The Nilos seals I mentioned above are metal seals that are in physical contact with the inner and outer races of rolling type bearings. They are clamped by one bearing ring and "seat" against the opposite ring. After a short "run in" period they seal the bearing. We've installed them for years and work fine for greased bearings. They're great for a primary seal to keep "stuff" from getting to the seals on the bearings. I suggest "googling" Nilos Seals for your personal edification and education. LOL

[philbur]

It may be a radial load carrying ball race but if it has negative preload (meaning some radial float) then the spindle will be able to act as a lever, where the pivot is the bottom bearing(s). This in turn will mean the cutter (at the other end of the lever) will have some radial freedom.

The top bearing also needs some preload.

Phil)

Hi Philbur, not quite sure what you mean by top bearing "pivoting" round the bottom bearing, the top bearing is a radial load carying ball race (it could also be a roller bearing, but needs sealing) and as such only has to resist side loads only like the pull of a belt, and as a ball radial bearing you would be hard pressed to find any side to side movement in the top bearing, but to allow for thermal expansion the top bearing is allowed to move up or down on the spindle or in the housing, approx a few thou is sufficient.....the bearing being a sliding fit......this also allows the bearing to be lubricated and sealed for life, as it's very difficult to make a seperate lube system and seals in this area.
Ian.

[philbur]

Seems they have only negative preload, that is they have internal clearance:

Accuracy Double row angular contact ball bearings 30..-B-2Z

Unless I misunderstand something, which is always possible.

Phil

FYI, FAG bearing 3307-B-2Z is a split inner ring, sealed both sides, double row angular contact ball bearing. More specification is required to determine preload.

[RICHARD ZASTROW]

5207 double row angular contact does have internal clearance, they have no preload. 35 mm ID C2 (less than normal) has up to 16 micrometer (.00062") clearance.

The 3307 split inner ring angular contact bearings have light, medium or heavy preload, 15 to 40 degree contact angle. The preload removes the clearance.

The same method is employed in the 2-way angular contact thrust bearings SKF uses in some of their spindle bearing sets. Ex., a pair of cylindrical roller bearings separated by a 2-way A/C thrust bearing. This combination takes greater radial and axial loads than angular contact bearings.

Dick Z

[philbur]

This page says nothing about preload, all bearings have an internal clearance.

Accuracy Double row angular contact ball bearings 33..-B

I don't understand why there would be a specified internal clearance which is then negated by a specified preload.

Do you have a reference that describes the preload options.

Phil

The 3307 split inner ring angular contact bearings have light, medium or heavy preload, 15 to 40 degree
Dick Z

[philbur]

There is no reference to preload in this document for double row A/C bearings(starts page 22), only internal clearance.

http://krpower.com/images/Angular.pdf

Phil

[handlewanker]

Philbur, re post#14.....I think you are nit picking when you "see" a spindle pivoting around at the bottom end due to "no radial preload".

You just possibly could (highly unlikely, but possible) get some cutter side deviation if'n you have the spindle designed with the cutter too far from the support of the bottom A/C bearing set-up due to the spindle being extended and a collet chuck introduced as a tool holding component, and also, but not as well, if'n you are using a long series end mill of smaller diam......but this is not a bearing configuration failure, but a spindle design flaw, common where the R8 collet has to fit up deep inside a spindle of less than desirable diameter.....as in the case of my mill spindle modification that I previously described.

Just how would you attempt to introduce preload to a single row sealed radial bearing when the bearing is designed to run in the manufactured condition without any artificial extraneous loading contrary to the bearing design, E.G. spring loading the radial bearing component in the axial plane to attempt to give it a sense of pseudo "angular contact" properties while at the same time attempting to keep it free floating axially?

The design I advocated was in the SKF bearing catalogue, a common spindle configuration widely accepted for stability and rigidity, and was described as for a high speed milling spindle.

BTW, the aspect ratio of cutter end point to the centre of the A/C bearing set, in relation to the distance of the top bearing to the centre of the A/C bearing layout is probably in the order of 6:1, so if'n you had a wopping .02mm side deviation in the upper radial bearing (extremely and highly unlikely) then you would have 1/6 of this apparent at the cutter tip, and as you would normaly get cutter deviation under load due to the flexing of the cutter, it would be impossiblt to measure....but extremely easy to hypothesise on and lose sleep over.

I would go so far as to state that if'n a preload is applied to a radial ball bearing to make it behave in the manner of an A/C bearing, then the radial component factor will fail due to the curvature of the raceway being unduly worn from side loading contact from the balls, and this will leave it prone to movement if and when the radial preloading application is overcome by heavy cutter side loading.

By all accounts you are very obviously a draughtsman or design person who theorises on aspects of possible spindle failure and overlook the prime purpose of a design engineer in that the design must be kept simple, when all other aspects are equal.

BTW, you could also fit a second radiaL bearing half way down the spindle if you really wanted to "stiffen" the spindle from deflecting under the "massive" loading of this DIY spindle design, but it would take a genius to be able to bore the bearing seating in the quill or spindle housing, and also measure it to give a floating fit to the bearing, but then pigs have been known to fly.
Ian.

[philbur]

Philbur, re post#14.....I think you are nit picking when you "see" a spindle pivoting around at the bottom end due to "no radial preload".

The difference between a spindle that is susceptible to chatter/noisy bearings and one that isn’t is all in the nit picking.

You just possibly could (highly unlikely, but possible) get some cutter side deviation if'n you have the spindle designed with the cutter too far from the support of the bottom A/C bearing set-up due to the spindle being extended and a collet chuck introduced as a tool holding component, and also, but not as well, if'n you are using a long series end mill of smaller diam......but this is not a bearing configuration failure, but a spindle design flaw, common where the R8 collet has to fit up deep inside a spindle of less than desirable diameter.....as in the case of my mill spindle modification that I previously described.

You make lots of unsubstantiated statements throughout you post. When somebody doesn’t know you from Adam your reassurances have no meaning.

Secondly we are discussing the OPs spindle design. Which if you look has a very short distance between top and bottom bearings.

Worst case Radial clearance for a 6202 is a C4 with 0.033mm or nearly 1.5 thou. Radial clearance.

Ref: Accuracy Deep groove ball bearings 62..-2Z

The aspect ratio of the Ops spindle is approximately 1:1 which gives the possibility of uncontrolled radial movement of the cutter of 1.5 thou. If the OP where to specify a class C2 bearing this would be reduced to something like 3 tenths of a thou. Of course the belt tension will maintain a none symetrical preload but will leave the balls on the other side floating.

Just how would you attempt to introduce preload to a single row sealed radial bearing when the bearing is designed to run in the manufactured condition without any artificial extraneous loading contrary to the bearing design, E.G. spring loading the radial bearing component in the axial plane to attempt to give it a sense of pseudo "angular contact" properties while at the same time attempting to keep it free floating axially?

Of course it is very difficult to preload a single DG bearing if it needs to float. That’s why I would not use such a design for a machine tool spindle, at least not without understanding the compromise I was making.

The design I advocated was in the SKF bearing catalogue, a common spindle configuration widely accepted for stability and rigidity, and was described as for a high speed milling spindle.

I have searched for a machine tool spindle design in the SKF catalogue that uses a single floating deep groove bearing at the top end and I cannot find one. Do you have a reference.

Here is the reference I did find, note that page 11 includes 4 different machine tool spindle designs, none of which includes a single deep groove at the top end.

Ref: http://www.skf.com/files/514629.pdf

BTW, the aspect ratio of cutter end point to the centre of the A/C bearing set, in relation to the distance of the top bearing to the centre of the A/C bearing layout is probably in the order of 6:1, so if'n you had a wopping .02mm side deviation in the upper radial bearing (extremely and highly unlikely) then you would have 1/6 of this apparent at the cutter tip, and as you would normaly get cutter deviation under load due to the flexing of the cutter, it would be impossiblt to measure....but extremely easy to hypothesise on and lose sleep over.

The OPs aspect ratio is closer to 1:1.

I would go so far as to state that if'n a preload is applied to a radial ball bearing to make it behave in the manner of an A/C bearing, then the radial component factor will fail due to the curvature of the raceway being unduly worn from side loading contact from the balls, and this will leave it prone to movement if and when the radial preloading application is overcome by heavy cutter side loading.

Isn’t the belt tension equivalent to radial preload. Deep groove bearings can still function adequately with axial loadings between 0.25 and 0.5 of Co

Ref: Single row deep groove ball bearings - SKF.com / Products / Interactive Engineering Catalogue/Rolling bearings

By all accounts you are very obviously a draughtsman or design person who theorises on aspects of possible spindle failure and overlook the prime purpose of a design engineer in that the design must be kept simple, when all other aspects are equal.

A second rate draughtsman might accept your unsubstantiated statements at face value but a good engineer would require/provide credible references Also it is not the primary purpose of a design engineer to keep things simple as you claim. The primary purpose is normally to produce a design that will achieve fitness for purpose for the lowest possible lifecycle cost.

BTW, you could also fit a second radiaL bearing half way down the spindle if you really wanted to "stiffen" the spindle from deflecting under the "massive" loading of this DIY spindle design, but it would take a genius to be able to bore the bearing seating in the quill or spindle housing, and also measure it to give a floating fit to the bearing, but then pigs have been known to fly.
Ian.

I’m not sure what this silly statement has got to do with anything. You make a silly proposal then you point out that it is silly.

Phil