Tormach spindle bearings - assembly fault

Sorry if I misread but some people do literally press separated AC's back together.

The Tormach spindle is a common spindle design used on low cost spindles. The DT bearing doubles your radial and axial capacity on the low end and is "backed up" with a lower capacity upper - all you have is belt lot at top and the opposite axial loading to the lowers.

I would not be surprised if the OEM pressed everything together with special tooling. This is all but impossible to duplicate in the field. He'res how I'd tune up the spindle if it were mine.

1 Carefully scotch brite the shaft and housing bores to get rid of "rash".THey should be clean and smooth when done.

2 tune up spindle until you get a transition fit of upper shaft journal so that you have very smooth and "perfect" slip fit of upper bearing to shaft. Bearing should not fall on, it should just slip on when carefully aligned. 600grit and red scotchbrite should do it..

2 Carefully scotch brite the shaft and housing bores to get rid of "rash". THe surfaces should be clean and smooth when done.

3 SHrink fit upper and lower bearings into place. Press to seat on OUTER RINGS ONLY is OK with good bushing/sleeve once the thing cools and freezes the bearing.

4 At this point, you should let the thing cool and stabilize a bit so the AC's dont fall out. BTW the 7000 series bearings are ALL A/C"s.

ONce the spindle hsg has stabilized, put it back in oven to fit the spindle shaft. Freeze the spindle overnight. Option, pack it in dry ice. When housing is near 220 f, and spindle is COLD froze, IT should drop cleanly into the lower bearing and PLUMET into the upper - which you already modified into a transition fit.

Setting preload: WIth all the bearings pressed at ALL postions, you have assuredly created a nightmare in assembling things without damaging one or more bearings. Besides, you end up with a indeterminate situation as to which bearings are loading and sharing loads.

BTW, DT bearings may or may not carry load equally due to tolerance stack ups. DU's are "matched to 0.0001" or better with respect to axial protrusion so load share is achieved.

What you've achieve by above is the capability to equally load share within the capability of the bearing set. ANd, you can accurately and properly set preload

Heres how. LIghtly oil the bearings. Make yourself a torque wrench out of a flat wrench and a fish scale. Call the oem to see if they have a recommended preload torque of the spindle when tight.

IT would have been a good idea to have recorded this before disassembly but improperly installed bearings precluded this from happening.

Anyway, slowly and carefully tighten the nut while checking rotating torque as you go. WHen you get to the prerecorded value or the OEM spec, stop. Don't go too far as too much STATIC preload can damage balls or cause over heat. IT is rotating torque you check NOT breakaway which varies all over hell with rolling element bearings

Use an industrial syringe to load inner and outer bearing with a light fill of grease- SNdok C is fine and affordable - put a small dab in front and behind each ball on the inner and outer raceway.

Option: you could probably slip fit both upper and lower bearings to spindle shaft. I did this to a 8000 rpm capable dyno shaft and it worked just fine. Made it much easier to rebuild than by having to press the hell out of everything.

If you do the above, make sure you "time" the bearings so you can get the grease into the "inner" piece.

DId the above to a "hobby spindle" of my own, Found some ABEC7 6206's and fitted them to a chinese drill press. Smoothest, quietest drill press spindle I ever heard. ABEC 7's in the motor made it nearly dead quite when running.

Heavy preload on spindle will make for higher stiffness and less spindle/bearing defelction under milling leads - probably should take more time to warm one up, too.

Lower preload makes for higher RPM potential and less heat build up in general. WIth reasonable and proper preload, a spindle operation temp of 180-200 would be considered normal. With any preload, you should warm up spindle at LOW RPM for several min prior to use.

THats how I"d redo a Tormach spindle if it were mine.

NC- Your knowledge of bearings and spindles is really impressive- you must work in a shop that rebuilds them on expensive machines. After reading your posts, it reminded me of a friend from the '70s- he bought a little Fiat sports car with a 4 cylinder motor in the rear. He never called it by its name, but always referred to his " mid engined Italian sports car" He did a similar rebuild- pulled the motor and had it balanced, ported the head, polished the valves etc. etc. and when it was done it was really pretty smooth running, but even after all this work he was still driving around in a Fiat.

Never worked in a spindle shop. Did a tour of duty as machine tool bearing engineer at a bearing company. THey did spindles for Mazak, Toyoda, Mori, Ocuma and many other Pacific rim machine tool companies. I had training at their machine tool spindle shop in Japan.

Otherwise worked mostly in the auto industry. WOrked at 2 OEM's, several support suppliers, 1 aftermarket supplier and a prototype parts/systems supplier.

Have an affinity for "seeing' how machines work. Example: our cam grinding spindle had 50mm bearing that was essentially a drip fed oil sleeve. WHen it would get hot, the spindle would start to oscillate in the clearance space and chatter the daylights out of the parts.

THe last time I got into it, I did some calcs and found the the oil feed path had some dubious cross sectional areas 'tween the oil feed trough and the journal surface. So I added two more feed holes to the bushing and opened up the oil feed hole that fed oil to the holes that fed the oil to the bearing. changed from thin as gasoline "spindle oil" to my ever favorite Dexron III.

Spindle runs a bit warmer due to viscous drag of the thicker Dexron. BUT The spindle now just CUTS when it grinds and the "prior wheel bounce"/oscillation is gone. We now use cheaper Dexron as opposed to Velocite. ANy body need some Velocite - I have some unused Velocite in two viscosities I'd like to get my money out of.

Spindle construction/repair is a combination of bearing knowledge and common sense combined with fastidious attention to detail. I've seen supposed experts forget that bearings don't simply rotate, the inner rings essentially orbit. THey do NOT have a fixed center of rotation. Considering that you have runout in the outer ring and inner ring and a number of microscopically different dia balls, the thing can't/wont. IT takes 15-17 revolutioen for the orbit path to be retraced. Abec 1 and 3's do this, 5's and 7's don't. at least not to the same degree.

Mount two bearings on a concentric shaft and expect the outer rings to stay radially aligned to themselves, it won't happen - probably 5-10 revs after mount, the bearings will loc up as teh orbit paths get to their extremes and take up the clearance. Simple case of radial runout vs clearacne vs non repetitive runout (NRO). Unless you're worked with HDD or machine tool bearings, you've never hear of NRO.

THe Jap tranny bearing expert didn't and he fubar'd a major transmission project because of it when he did mount two AVEC 1 bearings coaxially in tandem. I simply looked at it and said it wouldn't work. Got all kind of flack for it from supers and fellow engineers in dept. WHen the thing locked up, I as the "old timer" explained why - none of the hot shot MBA's did/could and neither did the reverred expert from Japan. Result, mgt was po'd at me for seeing what they didn't/couldn't.

ONe guy did come bye and ask "how the hell did you see that?".

Simple, one of our HDD bearing guys splained NRO to me one day. LIFE IS A LOT OF COMPLEX PROBLEMS SOLVED VIA THE APPLICATION OF SIMPLE PRINCIPLES THAT YOU PICK UP ALONG THE WAY.

Quote:

---

Originally Posted by NC Cams

THe last time I got into it, I did some calcs and found the the oil feed path had some dubious cross sectional areas 'tween the oil feed trough and the journal surface. So I added two more feed holes to the bushing and opened up the oil feed hole that fed oil to the holes that fed the oil to the bearing. changed from thin as gasoline "spindle oil" to my ever favorite Dexron III.

Spindle runs a bit warmer due to viscous drag of the thicker Dexron. BUT The spindle now just CUTS when it grinds and the "prior wheel bounce"/oscillation is gone. We now use cheaper Dexron as opposed to Velocite.

---

That part really caught my attention- when I was stationed overseas we had access to every type of machine oil imaginable for our Shoptask machine. Due to the dust and sand that got into everything we changed things pretty regularly. When I returned stateside and bought a new patriot machine, i asked JT where I could find some good oil for the spindle, without buying a 5 gallon can- he said just get some ATF from your local auto parts store.

Quote:

---

Originally Posted by NC Cams

Sorry if I misread ...
The Tormach spindle is ... "hobby spindle"

---

DNFTT

Thank you for your post 61 NC - it takes time to put all that technical stuff down - appreciate that.

Couple of questions.

The upper and lower bearing sets are about 7" apart. When the spindle heats from ambient temp to operating temp (which is quite a big temp increase on some units - the spindle must grow quite a few thou in length - and so dramatically reduce or even negate the preload? - Is this what you were touching on when you called the head a hobby mill type design?

You say to add the grease after assembly - is this because heating may melt the gearse out or degrade it? Do you wash the bearings with some solvent to remove the packing film - if so and they are heated dry, is there a risk the bearings could oxidise at the elevated temperatures?

[QUOTE=keen;557933]The upper and lower bearing sets are about 7" apart. When the spindle heats from ambient temp to operating temp (which is quite a big temp increase on some units - the spindle must grow quite a few thou in length - and so dramatically reduce or even negate the preload? -[ /QUOTE]

If the bearing housing and shaft were made of the same material (steel) with the same CTE wouldn't the preload not change appreciably with a change in temperature?

[quote=Don Clement;557936]

Quote:

---

Originally Posted by keen

The upper and lower bearing sets are about 7" apart. When the spindle heats from ambient temp to operating temp (which is quite a big temp increase on some units - the spindle must grow quite a few thou in length - and so dramatically reduce or even negate the preload? -[ /QUOTE]

If the bearing housing and shaft were made of the same material (steel) with the same CTE wouldn't the preload not change appreciably with a change in temperature?

---

Hi Don - The bearing housing is a cartridge type in full contact with the head - the head is at ambient temp and will slowly warm up after hours of operation - but it is effectively heat sinked to the vertical slide and anyway has too much mass and surface area to ever reach the much higher spindle temperature.

keen, I think it really depends on how well the cartridge is in contact with the head housing. It is only bolted by its bottom flange, which is adjacent to the bottom bearing pair. The upper part is more/less free to float axially in the head.

If I were worried about tenths-of-thousandths (which realistically is "down in the noise" for most of my machining) I'd be more worried about the fact that the axial position of the spindle is statically indeterminate--it is balanced between the opposing axial stiffnesses of the lower and upper bearing pairs, and any thermal and reaction forces along the spindle axis will be divided between the two sets of bearings in proportion to their internal stiffness.

I would guess (not having worked through any numbers at all) that the lower pair, being of larger diameter/heft than the upper pair, will be dominant in this tug-of-war and will largely determine the cutter position, and the upper pair will bear the majority of any axial flexing.

Best regards,

Randy

Quote:

---

Originally Posted by zephyr9900

keen, I think it really depends on how well the cartridge is in contact with the head housing. It is only bolted by its bottom flange, which is adjacent to the bottom bearing pair. The upper part is more/less free to float axially in the head.

If I were worried about tenths-of-thousandths (which realistically is "down in the noise" for most of my machining) I'd be more worried about the fact that the axial position of the spindle is statically indeterminate--it is balanced between the opposing axial stiffnesses of the lower and upper bearing pairs, and any thermal and reaction forces along the spindle axis will be divided between the two sets of bearings in proportion to their internal stiffness.

I would guess (not having worked through any numbers at all) that the lower pair, being of larger diameter/heft than the upper pair, will be dominant in this tug-of-war and will largely determine the cutter position, and the upper pair will bear the majority of any axial flexing.

Best regards,

Randy

---

Randy you are right in that the cartridge is free to extend longer into the head - it is only a slip fit true.........but what i think is happening is that the heat from the bearings is sucked away into the head etc and so the cartridge is only gaining a small portion of the heat that the poor old spindle is absorbing - it has no comparible heat sink and so probably gains much in relative temp. we could be talking a lot of movement here over 7". If the difference is just 40 deg Cels over 7" that is a quite few thou! If someone has one of those point and read temp measuring guns - please post the difference between the spindle (at R8) and the bottom of the cartridge when it has been running at 4800 rpm for say 1/2 hour.

Quote:

---

Originally Posted by keen

please post the difference between the spindle (at R8) and the bottom of the cartridge when it has been running at 4800 rpm for say 1/2 hour.

---

That is a good point, keen. Given that I machine almost exclusively brass, aluminum and plastic, I almost never run the spindle at less than top speed (4500 for my older machine) for up to 15 hours at a time. I don't have an IR gun or I'd provide numbers myself...

Best regards,

Randy

I've had a Tormach mill for about a year now. I replaced my spindle bearings for 10000 rpm and I also had to cut down the arbor to fit a high speed taper-lok pulley and power drawbar which involved a new keyway and grind the shaft to 1 1/8" diameter to fit the new pulley.

I took the spindle to a national company that makes and repairs spindles, After their ballpark price of $3000 for the work I wanted to do I figured I'd have to do it myself. They do offer a $100 service to break down the spindle and create a repair report including measuring all the tolerances which I decided to do. The service manager was very helpful and agreed to let me watch the technician and ask a few questions.

I have replaced bearings on pumps so I'm not totally ignorant of the operation, but for the super precision bearings installed in a clean room wearing medical scrubs etc. I thought best be left to experts.

The technician was also helpful. The spindle was only 6 months old but when he opened it up the upper bearings (drive end) were contaminated. The lower bearings were all right which was the opposite of what I would have expected. The seals on the spindle aren't really anything more than a washer or spacer. NSKdoes sell a labrinth (spell)seal that fits and if and when I replace the seals again I would use them. That upper seal sits down in a hole under the tormach pulley. Any fine dust or liquid will work its way into the bearing. I guess opening and closing the belt cover door a thousand times a day with dirty hands or maybe wet hands is the problem. I bet most all those bearings are contaminated.

The GOOD NEWS is the spindle is ACCURATE. With highly accurate measuring equipment the technican measures the arbor diameter and housing bore at the bearing locations. My arbor diameter was+ 0.0001 over at both bearing locations. The bore was -0.0001 and -0.0002 under. My R8 taper had a TIR of 0.0006 which was a little concern to the technican who marked the maximum location on the arbor. I took the report to the service manager and mentioned something about Chinese quality and she qujickly pointed out that the spindle was very accurate. She made the comment that she sees big name japanese spindles not as accurate as this. Felt good about that. She quoted me on NSK standard super precision ball bearings which at this size and 2 horsepower she said were good to 13000 RPM with just grease. I ordered from her for $350 for the 2 sets. I also got a 20 minute lecture on conterfiet bearings. I did get another price for the same NSK bearing sets from another distributer for $1500. What? I took the arbor to a grinding shop who cut it down for $100.

In my opinion it is damn near impossible to install the bearings without heat. NSK has very good documentation and instructions on installing their bearings. They give a maximum allowable temperature (160deg f I THINK) I needed every bit of temperature differential including refrigerating the arbor to get the bearing to drop on the arbor. I had to pull out the infared thermometer to get to the temperature limit. I heated the bearings sandwitched between two 1 1/2" thick aluminum plates on top of my gas stove to get a uniform temperature through both bearing rings.

The annular contact bearings are installed in a DT arrangement with both the chevrons facing the same direction. I aligned the chevrons with the taper maximum TIR mark that the tech put on the arbor per NSK instruction. NSK recommends that the taper be re-bored after the bearings have been installed. I didn't do that. The open end of both bearings sets faced towards the cutting tool. In this thread Keen commented 07-04-2008 that he thinks the upper bearings are installed wrong and should be open end up. My original bearings were installed both sets open end facing down and the technican told me that is how I should also install the new NSK bearings which I did

I tried to get more out of the technican on how much to torque the preload nut. All I could get out of him was tighten for good rigidity but not too tight creating extra heat or noise. I have never noticed heat from the lower bearing set and the upper bearing set is in 2"' of cast iron so I wouldn't know. During the breakin period, I did decide that I had the nut too tight. I backed out the nut but the bearings were so tight on the arbor they wouldn't release. With the preload nut loose I ended up carefully and very slowly driving the Z axis (spindle nose) into a wood block to unload the bearing. Once I felt that pressure release, I snugged up the preload nut. The bearings quieted down after 3-4 hours. 3/8" and less diameter endmills in aluminum I almost always run at 10000 rpm which is most of the time for the parts I'm running.

Thanks for that 4Z7 - lots of good detail info. I like the idea of 10,000 rpm. might be my next project.

I can't agree however, that all 4 angular bearings - (2 at the top and 2 at the bottom) should all face the same way - open down. The way the spindle is designed with a bearing preload adjustment nut on the top of spindle/arbor means the bearings have to be open towards both ends so that the nut closes the bearings against the thrust faces. If you have the top set down, you are closing the top set away from the thrust face on on to a tiny land. I emailed Greg at Tormach about this at the time of the rebuild - and he did not disagree. If you come to agree with me, you might want to change your top set (if they are both facing down) before they die.

Hi 4Z7 -that was a very interesting post. Would you mind sharing the part numbers of those high-speed bearings?

Cheers.

duplicated

Keen I have no idea if the upper bearing set should be up or down and I'm pretty sure Tormach doesn't either. I'll just wait it out since I'd likely destroy them trying to pull them off the arbor. When you tighten the lock nut, aren't you pulling the top or closed side of the bottom bearing set up against the lip on the spindle housing and likewise pulling the top bearing set closed side up against the the top seal (or washer/spacer) and lock nut? It seems to me to be the same loading dynamic.
Digits the nsk sets are 7008CTRDULP4Y (tool end) and 7007CTRDULP4Y

Hi 4Z7 - sorry, I cant follow your question.

The Spindle nut draws the two sets of bearings together via moving the inner races together, the outer races are held in position against their bore faces in each end of the spindle housing. If the top bearings are in upside down - the same thing happens, only the loaded surface in the bearing is tiny - less than 5% of the correct surface, and so is unlikely to last.

Draw a simple sketch of the bearings, spindle and housing and it may be more clear to you. In case I was going mad or missing something, I checked my logic with a top Engineer this morning. I have emailed Greg about this, it seems the incorrect assembly may extend beyond a couple of units.

Keen,
What your are describing by drawing the inner rings together is preload within a bearing pair. This spindle is designed with a DT configuration. There is not a gap between the two bearings of a set when placed on top of each of each other in the same direction like there would be if they were placed back to back like in the fixed support of a ball screw. So tightening the locknut on the arbor doesn't preload the bearings or effect any relative movement between the inner and outer rings of a set.

I think the function of the lock nut is basically to provide a resisting force to prevent any rotational slippage and also some axial rigidity. It does this by driving the outer rings against the step in the sleeve and the inner rings against the step in the arbor. I can't remember if the top seal (or washer ) is flat or has a raised ring which acts only on the inner bearing ring on the top bearing pair, but I do believe it to be flat. Without the outer rings of the bearings being forced against an edge and/or against each other in the set they would slip at high speeds. Spindle ruined. Most machine spindles have a top that when bolted to the sleeve act only on the outer bearing ring forcing it to the sleeve step to prevent rotation. The lock nut then only acts on the inner bearing ring forcing it against the arbor step

I do think that maybe the top bearings should be installed with the open end up but for a different reason. The step in the sleeve is supposed to be a specific radius that matches the radius on the closed end side of the outer bearing ring. Same for the radius step in the arbor and inner ring.

Why didn't Tormach give you an answer the first time you asked?

Hi 4Z7 - Yes what I was describing was preload between a bearing pair.

What we have with the Tormach is two DT top and two DT bottom.

DT is duplex tandem, both bearings face the same way - doubling the load bearing capacity, but no change to the thrust direction (DB is back to back, and DF is face to face).

Angular contact ball bearings should only bear load in one direction, so a pair must be installed in opposite directions. If they are DT the two pairs must be installed in opposit directions. The only connection between the spindle or arbour and the machine is via bearings. two double sets (DT sets), that must oppose each other.

The adjustment means is to either draw the inner races of top and bottom sets together, while the other races are held by the lip, shoulder or faces in the housing.

Or to close the outer races etc which is not the Tormach layout so I wont expand this.

The fact that there is two bearings top and bottom does not change this.

When I installed new bearings last year I contacted Greg at Tormach and he sent me a service/fact sheet to help me. I have just dug out my printed copy.

Although the scale is small you can just see that both top bearings are 'open end' to the top - and both bottom bearings are 'open end' to the bottom.

I have not yet heard back from Greg about this yet. My guess is he is either contacting the factory to see if he can get information on how widespread the problem is - or he is on holiday etc. I await with interest!

when properly installed, the DT's rely on each other to accept preload. THUs if you install the lowers at << and the uppers at >>, this takes care of the perload via tightening of the nut. THe bearings preload each other thru the spindle as you tighten the nut.

HOWEVER, this is what is referred to as a "position preload". IT is the cheapest and easiest to affect - hence "hobby grade".

In even a simple machine like a Bridgeport mill, they preload the lower bearings against each other. THe top bearing than is a slip fit in the housing and perhaps even on the spindle. THus, when the spindle heats up, the bottom stays put, the thermal growth is absorbed in top bearing via wave washers or a spring loaded cartridge.

Depending on the bearing retention method of the lower end of the Tormach, one could perhaps mount preloaded lowers in the spindle and perhaps let the top end float. We converted an AC motor from position preload to retained drive end and floated other end. Lot of work, motor is quiet as a mouse and future bearing changes a snap.

THe spinlde guy who won't give out tightening methods was wise. it is very easy to go from too loose to OOPS, too tight once you take out any axial clearance. THis is why the fish scale/torque wrench, rotating torque value is SO critical. WE're not talking half or 3/4turns in from zero clearance, we could be talking MUCH less. THis is also why I WON"T toss out a value here. THis is also why spindles usually have superfine threads for spindle nuts.

Simply adding grease to the bearing as opposed to light oil can make a huge difference in rotating/break away toque. THis is why you grease AFTER you assemble in a position preload spindle..

CLean before assy with BrakeCleen. Do not spin dry the bearings with air. Again, few drops of light oil to assemble. Wash this out with BC before you grease. Residual oil may prevent your actual grease from wetting down the balls/raceways if you leave assy oil in place.

Yes the spindle with .0006 runout is accurate but 0.0000 which a ground spindle taper demonstrates is MUCH better. Wanna bet the high buck spindles they do for the name machine tool accounts have ID tapers ground? One more difference tween hobby and "Machine tool grade" spindle.

Matching the high points of eccentricity for ABED 5., 7 or 9 bearings is almost a waste of time. There is SO little radial runout (check the specs) that you can't hardly fix squatt by messing with select fit assembly - want a great spindle?, grind the ID taper.