What's new
What's new

Thiel159 Horizontal spindle overhaul

ian159

Plastic
Joined
Apr 9, 2024
Location
UK
Hi folks,
First time I replaced and adjusted a double row cylindrical roller super-precision bearing. Sharing a few things along the way.

This is what the right amount This is what the correct amount of grease looks like - but do measure it. I weighed itof grease looks like - but I would suggest weighing or measuring with a syringe. Leaving two areas free of grease helps handle it onto the spindle. The reason I had to replace the bearing was damage due to having been overloaded with grease and it having overheated.

P1110052.JPG
It's worth making a couple of tools for the drive-up ring and for the labyrinth/lock ring which also does the job of ejecting the bearing off it's taper.

The service manual is available as a fairly good pdf/scan online, and althoug it only devotes 1 paragraph to the spindle - still it is a great start and it was all I needed in conjunction with the SKF website and various postings here from the good people on PM.







P1110029.JPGDon't overlook the machine table both as an easy way to remove and store the vertical head and as a place to keep cleaned parts safe from dirt and grime (with a nice clean board clamped onto it). Definitely worth thinking through your workspace before getting started. I needed the machine table for storage; a clean table for degreasing and drying parts; a cleanish workbench and vice to hold the spindle, and the end of a surface plate for a bit of lapping on one or two components. It also helps to keep some ultra clean lint free rags for final clean of ones hands before handling a component for re-assembly. I did the final clean of each component with paper and a butane duster (aerosol can) normally reserved for electronics tasks.

P1110040.JPG
Fitting and removing the thrust bearings needed a bit of heat with a hot air gP1110028.JPGun - and a glove to save my hand. Perhaps 50 degrees above ambient was plenty. I covered the shaft with paper insulation to help with removing, but there was probably enough thermal mass there in it that this wasn't important.

The most important thing is, of course the drive-up procedure for the taper bore double row roller bearing. These days, the literature dwells on methods with hydraulic pressure and specialist equipment. One advantage of this machine spindle is that I could set up a 2µ DTI on the spindle nose and adjust the bearing incrementally with it fully assembled in place.

I didn't find anything more useful online regarding the amount of preload, except that
(a) one might expect a radial stiffness of about 190N per µ for a single row cylindrical roller bearing with nominal 55mm bore;
(b) the longest bearing life will be with a condition of minimal but positive preload (say 0µ-1µ).

Apart from this, lots of posts here on PM suggest monitoring temperature rise as the most meaningful parameter - this implies that one might get away with a higher preload on a low speed application like this (2000rpm, as opposed to around 10,000 which the bearing is rated to).

Anyway, this is what I actually did:
P1110055.JPGI setup a piece of mm squared paper with 60mm length on the shaft and used this to eyeball a reading from against a line scored on the drive-up ring. With thP1110054.JPGe 2µ DTI, I took readings of radial play (estimated at 3 points on the periphery of then spindle nose). I tightened the drive up ring a bit and repeated - each time observing the total DTI radial displacement at 3 spindle positions. I applied the radial force by hand on a 300mm bar in a collet chuck secured in the spindle taper.

Note there were 2 attempts at the whole procedure, as I was expecting the last 2µ to disappear (red points asside). I think the 2µ which I could not get rid of was very interesting and useful learning. I believe it is a combination of 2 things:
1) The radial stiffness limit of the bearing once it is preloaded;
2) Elastic deformation of the shaft.
With (2) in mind, i setup a solid 2" bar with roughly the same constraints as the machine shaft (i.e. the length, position of bearings front and back) and clamped it to a heavy drilling machine table at least as rigid as the spindlehead casting. The result was that a roughly a deflection of 1.5µ was still seen with this setup, even in the absence of a bearing, under roughly the same input.

SO - my second attempt (pencil points) aimed for the 2µ line as the actual condition for the onset of preload in the bearing. This is where the job is now. I'm going to run the bearing in next.

Things I learned:
1) Tidy up and organise the workspace before starting.
2) Only handle the new bearing when you are absolutely ready
3) Believe your measurements before your presuppositions.
4) Rigidity/elasticity matters. At least this time, measuring elastic deflections in a lash-up was easier than calculating them.
5) Don't forget to back-off the the lockring before driving up the bearing!


All the best, and I hope this helps someone.

Ian.
 

Attachments

  • P1110054.JPG
    P1110054.JPG
    165.6 KB · Views: 4
Hey there,

well done on the work and the documentation. Lovely these adjustable roller bearings, right? It's great to be able to adjust radial play.

Anyhow, slightly off topic.

Keeping in mind a comment that Peter has made several time regarding the Thiel 159 being the best engineered german toolroom mill, I wonder what it the purpose of the gib on the sliding head:
1715671962696.png
I mean, this is not a moving part when in operation, correct? Why would it need a gib?

BR,
Thanos
 
Increases the contact area! Potential for better vibration dampening. Makes the joint relatively oil and coolant tight. All are good things on machine elements.
Maintains the accuracy of the upper slide so movement or repositioning during a job is OK.....
Cheers Ross
 
Increases the contact area! Potential for better vibration dampening. Makes the joint relatively oil and coolant tight. All are good things on machine elements.
Maintains the accuracy of the upper slide so movement or repositioning during a job is OK.....
Cheers Ross
Hey Ross,

thanks for the reply, not sure I get it though.

In my mind, for a semi-stationary joint like the one of the sliding head on a euromill, all the points that you mentioned can, indeed, be achieved without a gib. You scrape for contact and you're ok, no?
The gib, to my understanding, is normally used when the fit needs to be adjusted while setting up a machine or for compensating for wear, always in a slide where the two elements must...slide one w.r.t. the other. For head on a euromill you just release, re-adjust, re-clamp. It's either fully clamped (so no fit to adjust with a gib) or fully released.

I know it's there for a reason, that's a serious machine, but I can't wrap my head around it.

BR,
Thanos
 
Without the gib you really only have contact on one side of the dovetail opposite the clamps.
Tightening the clamps ( aka Deckel) pushes the slide against the fixed side.
The clamps are in effect single point contacts on one side of the dovetail.
A gib provides full length contact on both sides of the dovetail.
Obviously the factory/ designers believed that full contact was better.
Cheers Ross
 
Hi again,

That question of why a gib on the overarm did cross my mind, too. But, the Ajax00 'clone' also uses a gib in that location for a long reach overarm. It's a much inferior machine to this, both in design and execution (alongside the Dekels and Thiels it looks like a student project!)

My hunch is it was partly a production engineering choice they made. Given they needed manufacturing, fitting and QA setups in place for doing gibs with compound angles on both the X and Y axes, then adding one more gib for the overarm might have been the easiest option as well as the best.

I have been using the machine on some trivial jobs today - the horizontal is fantastically quiet and smooth now it is all reassembled.

I.
 
Without the gib you really only have contact on one side of the dovetail opposite the clamps.
Tightening the clamps ( aka Deckel) pushes the slide against the fixed side.
The clamps are in effect single point contacts on one side of the dovetail.
A gib provides full length contact on both sides of the dovetail.
Obviously the factory/ designers believed that full contact was better.
Cheers Ross
Hi Ross,

thanks for the reply, sorry that I can't, still, get this. The gib is at the opposite-from-the-clamps dovetail side.
So, clamps undone, well, it's all loose anyhow
Clamps tight then you have single pressure points for the clamps on the clamp side and a full dovetail contact for the other side with or without gib! Things would be different if gib were on the clamp side but it's not, to my understanding.
Regarding stiffness, I'd argue that an extra component (the gib) subtracts stiffness rather than adding.
Hi again,

That question of why a gib on the overarm did cross my mind, too. But, the Ajax00 'clone' also uses a gib in that location for a long reach overarm. It's a much inferior machine to this, both in design and execution (alongside the Dekels and Thiels it looks like a student project!)

My hunch is it was partly a production engineering choice they made. Given they needed manufacturing, fitting and QA setups in place for doing gibs with compound angles on both the X and Y axes, then adding one more gib for the overarm might have been the easiest option as well as the best.

I have been using the machine on some trivial jobs today - the horizontal is fantastically quiet and smooth now it is all reassembled.

I.
Hi Ian,

I have another scenario: if they were that serious about their machines, I think they put the gib there in order to be able to center the vertical to the horizontal spindle. On the Deckels, fitting leads to some 0.01 mms of offset between the vertical and the horizontal, not easy to fix if you overshoot :)

Or, on a more practical side, I guess using a gib there makes it much easier to align the overarm to Y.
On a Deckel you have to scrape-fit the whole overarm to be parallel to Y when sliding. On that Thiel, you straight scrape the Y slide dovetail, you straight scrape the overarm dovetail and you fit-scrape the gib for alignment and contact.

BR,
Thanos

PS. wondering, for better understanding, is there provision for gib adjustment on this gib? I mean like on a slide gib. I'd be surprised if there was
 
Last edited:
I was unaware that the gib was opposite the slide clamps. That changes everything.
Think your argument of making it easier to align and correct geometry is most plausible. Good call!
Cheers Ross
 
I was unaware that the gib was opposite the slide clamps. That changes everything.
Think your argument of making it easier to align and correct geometry is most plausible. Good call!
Cheers Ross
Cheers, thanks for confirming!!
 
Just in case the spindle experts here accuse me of not monitoring temperature during breaking-in the bearing.....Screenshot from 2024-05-17 17-36-26.png
And here is where the temperature was measured (Fluke DMM with temperature probe)
Screenshot from 2024-05-17 17-50-54.png
Point 3 was just an afterthought, I noted the top of the column went up by almost as much as point 2 over the course of the test. I did another run the next day at 1000/1500rpm with very similar results.

-------------------#
And, on the gib/overarm discussion, I can confirm that there are no scraping marks on the overarm, top of spindlehead casting or on the dovetails/gib . You can also see the gib is adjusted normally. There is a single extra clamp on the gib's side of the overarm, the manual suggesting to only use it for heavy cutting on the vertical, implying it should be ignored most of the time (and just use the two on the opertor's side). Actually , I have only had any trouble with the main y-axis clamping force - i just needed to make better levers/handles for the cams which clamp onto the y-axis gib....very probably I am just not as strong as the average toolroom guy from the 60s!Screenshot from 2024-05-17 17-52-37.png
One thing that I like about this machine is how neatly the vertical head slides completely out of the way using almost all the space available in that top dovetail. 19 times out of 20 it'll be trammed vertical and you don't want or need to disturb it at all when switching between horizontal and vertical spindles. Chapeau!

Thanks for your interest. I intended to document the work because a few of us on PM are now admitting to having 159s, and the subject of spindle bearings has come up before. But, there are always other things which spark interest. Sooner or later, I'll check what the x-offset is between horizontal and vertical spindle axes.


Cheers,
Ian.
 








 
Back
Top