Centec 2A Rebuild

[mm289]

So after taking the Richard King scraping course a few weeks back I intend to continue the refurbishment of the Centec partly to get it running nice, but also partly as practice before attempting to rebuild the Beaver VBRP I have in pieces for the last 2 years

Just as  bit of history, I bought the Centec locally a few years back, it was in decent condition but as I have started to learn how to use it I have found it has a lot of slop in the table especially and if you try to take this out with the gibs it binds at each end of travel.

So, as it is a sensible size it also seemed a fair contender for a "project" 

Below are a few pics of what it looked like when I picked it up and a couple with it in its new home.

(vtsteam edit: reduced all photos from 3000 pixels wide to 800)

[mm289]

Just as a bit of history the first job I did was to replace the tapered bearings on the shaft in the vertical head. There was a fair bit of run out and, although I am still not convinced it is good as it should be, it is better than it was.

There is very little info on the web about the inner workings of these machines that I can find apart from a Yahoo Group, but navigating the Yahoo UI is soooooo painful I tend not to frequent that gathering.

So just for posterity sake, a brief overview of how to overhaul the head is as follows (from what I can remember )

The vertical head mounts where the horizontal mill overarm normally goes and is driven by belt of the horizontal arbour pulley. It rotates to allow milling at angles and has a MT2 taper as per pic 1 & 2 below.

The shaft sits in two taper roller bearings at the top and bottom and is driven from a bezel gear off the pulley shaft. The bottom of the head has a plate held in by 3 screws, remove this and you expose the lower bearing. Pic 3

The shaft also has a preload arrangement for the bevel gear pic 4. Having removed the head from its mounting you can loosen this off and then I used a hydraulic press to push the shaft downwards and out of the top bearing. Removing the bearings and cups was then just a case of using the press and various bearing pullers I had in the auto shop.

With everything stripped I could then look at the bearings involved.

(vtsteam edit: reduced oversize photos)

[awemawson]

Looking good Paul 

Do you think you can possibly edit those pictures down to 800 x 600 max size to reduced storage capacity. They will still have more than enough detail

[mm289]

Will try to do that Andrew, don't suppose there is a quick way other than editing every individual picture?

Back to Centec, the pic below gives the specs of the bearings that came out - useful to know as I couldn't find this info anywhere on the web.

Ordered up replacements, but I know understand there may be different "grades" of bearing with high precision bearings needed for these sort of applications ? 

Anyone able to tell from these bearings what standard they are? Couldn't see any other options when I bought the replacements.

Cheers,

Paul.

[mm289]

So prior to starting to scrape stuff I looked up Connelly and a few posts on here. First job was to check the true ness of a datum point, namely the horizontal arbour, so you can then use this to set the knee flat to the shaft.

I have included the relevant pages as attachments below but basically you are checking for run out on the spindle end face, what I would call "end float" on the spindle (in/out movement) and then run out at various projections from the spindle.

The bad news was that although my spindle run out with "no load" was only about 1 thou, when any side load was applied it was up to 7 thou and the end float when applying load (pushing/pulling the pulley) was 10 thou.

This was with the bearing pre load way higher than I would like so I think the bearings are knackered 

So before I can start scraping in the ways I need to get some new bearings ordered. Pulled the old ones out which are Timken and interestingly marked as "Precision 5" - so I guess I have answered my own question about precision bearings being used in these machines.

I now need to find a supplier of precision bearings 

The project continues (slowly)

[Pete.]

Can't see much of the inside face of that taper cup but it doesn't look anything like bad enough to give the lash figures you're quoting from wear alone (in fact I can't imagine any complete bearing giving 10 thou play).

How tight was the spindle to pull out of the inner bearings? If it was very tight it might simply mean that friction was making it impossible to adjust.


By the way image resizer I use:

http://www.bricelam.net/ImageResizer/

You can resize the whole bunch in one hit, either creating new files or re-sizing the originals.

[PekkaNF]

At work we use mostly pretty big carb bearings, some taper roller bearings, and seldom angular contact bearings....some stuff I have picked up. If you can measure fault, bearing is really rough. Three most common problems are: Assembly problems, lubrication faults, counterfeit bearings. New brand name standard bearings have better running accuracy than older bearings. Some new deep groove ball bearings are pretty damn close P5. Problem is the bearing clearance and mounting accuracy. It is hard to keep tolerances on check. When we are talking spindle bearings, bearing/part/fit tolerances are conspiring against you. You can find bunch of perfect bearings and some of them will slide on shaft and some needs press. This sort of changes internal clearance. Also it would effect on preload, or how you can adjust it. Sometimes you have one shot at 80 decgrees C while it is still moving and then run intermitent until it setteles and stabilizes.

This thread has lot of ramblings and some things I learned. Not sure if it applicable to you, you have diferent bearings.
http://madmodder.net/index.php/topic,12269.msg146280.html#msg146280

It's still working, not sure if it success, let's see few tens more hours of running. Now it is quiet, but still can't measure real problems with original bearings.

Pekka

[timby]

The bad news was that although my spindle run out with "no load" was only about 1 thou, when any side load was applied it was up to 7 thou and the end float when applying load (pushing/pulling the pulley) was 10 thou.
This was with the bearing pre load way higher than I would like so I think the bearings are knackered.

If the bearing surfaces are reasonable I suspect that you need more pre-load .

Try it and do it with the bearings dry it makes a difference.

Your 1 thou. run-out  is just over the Timken tolerance for standard bearings, see the PDF link below page 19.

https://www.timken.com/wp-content/uploads/2016/10/Timken-Tapered-Roller-Bearing-Catalog.pdf

You could probably reduce the 1 thou run-out  by changing the relative position of the bearings on the shaft.

When I used to use high precision Timken bearings long ago they had a dot on them that indicated the maximum run-out and we assembled them with the dots opposing.

[Pete.]

The worst radial runout I see for precision inch taper bearings is .0003" page 25.

[mm289]

Interesting input thanks guys. I will reassemble and run a set of tests with both bearings dry, then loosely packed with grease and take some new measurements. couple of thoughts/questions:

1. On the preload the reason I felt it was pretty high was because of the drag it was putting on the spindle rotation, it wasn't running freely and felt like it had a fair bit of resistance. Now I am more used to bearings in automotive applications where overtightening is a recipe for disaster so maybe I am being too cautious?

2.  The run out with no load I can live with at 1 thou. What concerned me is the deflection under load. To test this I was just gently pushing/pulling against the side of the chuck with a dial indicator on it and this is where I was getting the 7 thou. Is this a reasonable test?

3. Likewise the end play is the way I test end play on cranks in engines, set up a Dial Indicator then push/pull the shaft and read max deflection - in this case 10 thou. Again is this a reasonable test?

The cups look "ok" but not spectacular but the inner races feel very sloppy to me. The other thing I noticed which I have tried to show in the assembled pics below is that the inner race is not concentric with the outer. If you look at pic 1 you can see that the gap between the inner sleeve and  and the cage is less at 6 o'clock than 12 o'clock. In pic 2 I have rotated the cage and now the gap is less at 12 o'clock than 6. I may be missing something but doesn't that mean the bearings will run eccentric or do they only centre under load?

SLow progress so I have started swopping out the single phase motor for a 3 phase and VFD 
More on that later.

Cheers,

Paul.

[mm289]

Wow, having read Pekka's thread and links I now know more about bearings than I ever wanted too 

Just out of interest, if you want to check the tolerances on Timken bearings you can use this page on the website http://www.timken.com/engineering-tools/bearing-tolerances/

You can out in your bearing reference and then select the bearing class and it will give you the tolerances 

Paul.

[PekkaNF]

Felt pretty much the same....bearings are one thing - fit is another and measuring the shaft and bearing cavity reliably is pretty damn near imposiblle in typical home work shop. Best tu use a gauge - oversize or undersize bearing ring next best thing....most of the bearing shims and such are way to coarse to spindle - maybe fine for a gearbox.

Pekka

[Pete.]

Interesting input thanks guys. I will reassemble and run a set of tests with both bearings dry, then loosely packed with grease and take some new measurements. couple of thoughts/questions:

1. On the preload the reason I felt it was pretty high was because of the drag it was putting on the spindle rotation, it wasn't running freely and felt like it had a fair bit of resistance. Now I am more used to bearings in automotive applications where overtightening is a recipe for disaster so maybe I am being too cautious?

2.  The run out with no load I can live with at 1 thou. What concerned me is the deflection under load. To test this I was just gently pushing/pulling against the side of the chuck with a dial indicator on it and this is where I was getting the 7 thou. Is this a reasonable test?

3. Likewise the end play is the way I test end play on cranks in engines, set up a Dial Indicator then push/pull the shaft and read max deflection - in this case 10 thou. Again is this a reasonable test?

The cups look "ok" but not spectacular but the inner races feel very sloppy to me. The other thing I noticed which I have tried to show in the assembled pics below is that the inner race is not concentric with the outer. If you look at pic 1 you can see that the gap between the inner sleeve and  and the cage is less at 6 o'clock than 12 o'clock. In pic 2 I have rotated the cage and now the gap is less at 12 o'clock than 6. I may be missing something but doesn't that mean the bearings will run eccentric or do they only centre under load?

SLow progress so I have started swopping out the single phase motor for a 3 phase and VFD 
More on that later.

Cheers,

Paul.

1. Not sure
2. It's a reasonable test but the results are far from reasonable. I would expect it to be less by a factor of between 5 and 10.
3. The gap to the cage is unimportant. All it does is space out the rollers.

[hermetic]

Just a little aside, the surface you are checking spindle runout on is not necessarily "exactly" concentric to the spindle, You really need to check the runout on the taper (Centec are int30?) Check the taper for chips, bumps and dings, carefully ream if neccasary, then fit an adapter or even better a test bar, lock it in with the drawbar, and check the runout on that.. This observation stems from realigning (or trying to) an Indian motorcycle crankshaft with the DTI fitted with an elephants foot and running on the outside rim of the flywheels. After about an hour of getting nowhere we swapped the DTI onto the shafts of the flywheel, and got instant results! Provided the oil level has been kept topped up, it is unlikely there is any problem with the main bearings unless this machine has had an exceptionally hard life.

[timby]

1.    I would try to pre-load a bit more just to see if you can eliminate the deflection.

2.   I would expect to see almost no radial  deflection.

3.   I would expect no axial deflection, your 10 thou axial movement would cause chatter when cutting with a helical cutter.

The roller cage is only a spacer, check to ensure  it is not rubbing after assembly, you can true it up a bit in situ with a screwdriver or something else as a lever.   

One inspector that I met used to check the pre-load by trying to move the rollers  with a small screwdriver or a scriber,  he expected only a slight sideways movement.

Check spindle runout as advised in post 13  above inside the tapered bore.

[PekkaNF]

The preload should not be too great. Do you have acess at the both ends of the bearings and make sure that  bearing rings are seated right? A small chip somewhere on the spindle shaft or bearing seat would set the bearing askew, cause too big preload and cause runout + play.

When you turn the spindle (only spindle, no belts or gears in mesh) manually, does it rotates smoothly whole round? Or does it has one (or two) spots of harder resistance or can you feel "cogging".

How this bearing arrangement does the preload? Two taper roller bearing and nut on the shaft pushing inner ring of the bearing?

*Fatty fingers&dystolexia

[Fortis64]

 I also had a bit of faffing about to get the bearings set up . The rear bearing was my problem,the bore for the bearing shell was in my opinion too tight of a fit to allow proper adjustment of the the bearings .

This is what I fitted to my Centec 2
Timken
Outer 1930 2-0
Inner 1985 1-6

Outer 1729 1-6
Inner 1775 4-6

[mm289]

Success 

SO the short story is I now have runout on the horizontal spindle under control 

For the sake of anyone trying to do the same thing, or if you enjoy a Friday night read, the long story is as follows.........grab a cuppa 

Had a long chat with Pete this morning and in amongst other ramblings the discussion got round to how it would be almost impossible for a set of taper roller bearings to be so far out unless they were literally falling apart - which mine weren't. I was all for just slapping a new set in and getting on with doing some scraping, but Pete got me thinking so I went back to basics.

First off - eliminate the spindle as a source of error i.e. is it straight? So over to the lathe which I had helpfully just taken the 4 jaw off and replaced with the 3 jaw yesterday  so swapped the 4 jaw back again (which is mahoosive, note to self must get smaller 4 jaw chuck  ). So en route to setting up the spindle it gave me a chance to strip and clean the 4 jaw which i hadn't done since aquiring the lathe 18 months ago!

45 mins later... got the chuck set up and popped the spindle into the chuck then centred it. At first look I had a bit of a panic as it looked like the spindle had .45mm runout from one end to the other - bugger I thought - bent spindle (pic 1)  At this point I will claim some sympathy for being a bell end as I haven't got any formal machinist or engineering background  - obviously it would help when taking these sort of measurements if you had the other end of the spindle supported as well 

So ran a live centre into the nose of the spindle and ran the tests again..... much better - runout down to  +.04mm. Checked at the chuck end and the runout was -.03 though which seemed a bit odd. (Pic 2,3 & 4).

Initially I had the spindle gripped in the chuck near its end, so I moved the spindle in to where the rear bearings would sit and re did the test (again). Bingo - runout reduced to between .01mm and .015mm (4 to 8 tenths) which is good enough for me. (pic 5)

My conclusion being, the spindle between bearings is running true but is slightly bent between the rear bearing and the end of the shaft (where the pulley sits). Possibly only about .04mm or 2 thou so I am going to ignore it for now and press on 

More to follow...... just grabbing some grub!

[mm289]

OK, back from dinner (homemade meat pie made from Christmas leftovers  )

SO having convinced myself the spindle was OK back to re-assembly. Old bearings were dropped in some Jizer for a while to get them as clean as possible, meanwhile man from Royal Mail arrived with my RS Components delivery of an enclosure, 3 way switch and potentiometer to tidy up the 3 phase motor and VFD upgrade I am doing (more of that later  )

So the way the spindle bearings are mounted is they are situated in the casting and held on each side by a retaining plate/cover. This is bolted together with 4 bolts - but it became apparent that when these were machined they must have been done in situ or as matched pairs. Reason for saying this is on mine at least, they would only bolt up in one orientation, the bolts would bind as they went into the covers otherwise, so both sets were test assembled to find the "best" fit then centre punched to mark orientation (pic 2)

I then drove the bearing cups into the case so they butted up against where the retaining plate would sit and reassembled the shaft with dry bearings. I just did the front bearing first and was puzzled that as I tightened up the retaining plate on the front of the knee the bearing became tight to the point of stopping the shaft moving. AT first I thought it was pulling in crooked and binding on the nose of the shaft so despite having aligned the plate/cover I tried all 4 possible orientations but still had the same problem.

Eventually I engaged brain and broke out the depth gauge. Now this is going to be hard to explain without a diagram - and your not getting one of those!! - but, it seems that if the bearing cup is driven in flush with the retaining plate (which sits inside the case) then when you tighten up the cover it effectively preloads the bearing to the point of clamping it. This is because the cover has a flange which will press on the bearing cage. To fix this I drove the cup in further so as to give me 1mm clearance between the cover and the bearing.

SO all tightened up nicely and shaft rotating freely. Repeated process on the rear and exactly the same problem! Again pushed the cup in a bot further and the bearing ran freely.

Moment of truth... wound in a small amount of preload (1/8th turn) and put the Dial Gauge on the nose and..... 0.015mm runout 

Rebuilt the spindle again having lightly packed with the NGLI 3 bearing grease (Pic 3) equivalent of the original Centec recommended lubricant and then ran a series of tests at different pre loads as per Pic 4.

Result, finished spindle showing only 0.015mm runout at nose and when under load only deflecting 0.006mm (~3 tenths). Also end float is now zero. I then measured runout in the taper rather than on the nose (as suggested earlier) and it reduced to 0.01mm.

RESULT 

SO bearings were OK (thanks Pete  ) - I can only think that somehow the cups had got moved around that that whilst I was winding in pre load on the shaft it wasn't actually loading the bearings hence why I was getting so much movement in my original test. Anyway, now I have done this I can move onto setting up for measuring and scraping ( a day(or 4 ) later than planned 

Cheers,

Paul.

[mm289]

Have now removed the knee from the base and setup on workbench to make scraping easier.

In doing so I have discovered a few things:
1. The horizontal arbour when mounted in the overarm is uber accurate, only .01mm of runout at the far end of the arbour which is 6" from the nose of the shaft 

2. My clarkson collet chuck is pants  the chuck body gives about .02mm runout but when I mount a straight bar in it it is giving me .07mm run out, got the same result mounting it in the 4 jaw in the lathe so figure the chuck is rubbish.

3. My "spare" collet holder doesn't take any of my Clarkson collets (not sure what collets it needs  )

4. My Jacobs chuck on an MT2 taper is my most accurate work holder with a runout of only .03mm -

5. Trying to figure out how to check the alignment of the knee from the horizontal arbour when none of your workpiece holders are accurate is a PITA

Oh well, tomorrow is another day 

Cheers,

Paul.

[Pete.]

If I'm reading this right the cover pulls the bearing in flush with the outside edge but it has another mm to go before it stops against the seat? That would explain the results you were getting. You would have forever been adjusting the spindle until both cups were fully seated.

Glad you got that sorted.

[timby]

Success!, you were lucky with the   horizontal arbor I had visions of it running out badly.

[Mover]

Pleased you got your bearings sorted Paul, also a very good description of your trials and tribulations.

Just one thing, when you mentioned you tried to mount a pice of round bar in your clarkson collet chuck.  If it’s an autolock type the tool needs a threaded shank to work. The nut does not tighten the collet but as the cutter screws into the collet during cutting, against the pin in the chuck body the collet is pushed downward against the nut. Self tightening the collet hence the name autolock.

[mm289]

Just one thing, when you mentioned you tried to mount a pice of round bar in your clarkson collet chuck.  If it’s an autolock type the tool needs a threaded shank to work. The nut does not tighten the collet but as the cutter screws into the collet during cutting, against the pin in the chuck body the collet is pushed downward against the nut. Self tightening the collet hence the name autolock.

Thanks for that Ross - that explains it, could never figure out how these were meant to work 

Still gonna have to buy myself a new "ordinary" MT2 collet chuck and some ER32's or something, I have loads of plain cutters as well as the threaded ones for the Clarkson.

Cheers,

Paul.

[timby]

Just one thing, when you mentioned you tried to mount a pice of round bar in your clarkson collet chuck.  If it’s an autolock type the tool needs a threaded shank to work. The nut does not tighten the collet but as the cutter screws into the collet during cutting, against the pin in the chuck body the collet is pushed downward against the nut. Self tightening the collet hence the name autolock.

Thanks for that Ross - that explains it, could never figure out how these were meant to work 

Still gonna have to buy myself a new "ordinary" MT2 collet chuck and some ER32's or something, I have loads of plain cutters as well as the threaded ones for the Clarkson.

Cheers,

Paul.

It is possible to grip plain shank cutters in Clarkson Chucks by fitting a threaded  and centred plug in the end of the Collet, cut a Screwdriver slot  in the other end and adjust to suit .