Overhaul of Haas spindle cartridge

I was chasing down a slight bearing howl in the head of my VF3. The noise occurs after the machine warms up and is running somewhere above 1000 rpm. The sound does not really become more intense with speed increase, it either howls or it doesn't.

I removed the belt from the transmission, and it seemed as though the noise disappeared, with only the motor and transmission running.

I hummed and hawed over whether I should just "do it" and buy a factory remanufactured unit ($2900 Cdn) or try it myself. Part of the reason I wanted to do it myself, is because I'm not dead sure that the problem is in the spindle. Plus, I was dying to see what was inside

So, guess what I decided?

I got the old cartridge apart without wrecking anything. It is a different assembly method than I have seen before. Basically, everything is permanently assembled with heat shrink collars. While not very user friendly to take apart and put together, I guess that the collars would have the benefit of holding the races very parallel to one another. I had experience with another machine that someone had rebuilt, and they had overtightened the bearing retaining nut and actually produced some runout of the spindle!

Everything is a wee bit on the loose side, yet it doesn't appear to be worn loose. The lower pair of angular contact bearings have a half thou clearance in the housing, as do the upper set. I would have thought that the main angular contacts should be an "on size" sort of interference fit.

One of the old bearing inner races is also a slip fit on the spindle. The other one seems closer to size on. I don't know firsthand if there is enough pull in a 1300 lb drawbar to actually expand the mouth of the spindle by a half thousandth. I did read a warning somewhere in the Haas service cd that running the spindle without a toolholder in place was detrimental to a new set of bearings.

The original A/C bearings and spacers are all serialcoded to be a matched set of Barden bearings. The inner spacer is a full .001" shorter length than the outside. Before disassembly, the bearings did roll over with a light preload. However, inspection of the inner spacer and bearing races shows me that there was some relative spinning going on between the spacer and the bearing, so the preload collar was not clamping the inner races tight against the spacer. Could this be the cause of the howl?

I ordered new bearings flush ground with a light preload. I plan to lap the outer spacer to the same length as the inner spacer. I don't know if light preload is enough, but I am wary of getting the thing too tight, because I like to run it fast more than slow.

My plan is to carefully loctite the inner races of the a/c bearings to the spindle. Same in the housing. By careful, I mean so as to not have an excess of that crap running around causing trouble. A wee bit of loctite near the back shoulder and a wee bit on the outer race of the front bearing as I slide it together. The upper bearings will not be loctited whatsoever, as I realize they must float.

When all assembled, then I plan to regrind the taper a wee bit. It was bruised a bit anyway, but I didn't do it! This will be done on the lathe, holding the front nose of the housing in the steady rest, while driving the spindle with the chuck.

Comments by those who have tried this themselves are welcome. Also, from the pros, of course.

Man I would never have the balls to take a spindle apart, I'd have a serious case of pucker that I would never get the thing together again.

Grinding my spindle taper in mylathe....fuugetaboutit.

Snap some pics I'd love to see the guts of your spindle.

So far I think you will be fine, if you can, use a rotary surface grinder liks a Heald to grind the spacers, both at the same time if possible.

Another thing you did not mention but should be aware of is the little burnish dots on the brg races, they should be aligned on the spindle and the outer races should be aligned in the housing.

Keep it clean and you should be ok...

What class brgs? 5 or 7 or 9??
Barden is good and will help you on the phone.

How a about brg lubrication? is it oil or is it grease?

Everything is a wee bit on the loose side, yet it doesn't appear to be worn loose. The lower pair of angular contact bearings have a half thou clearance in the housing, as do the upper set. I would have thought that the main angular contacts should be an "on size" sort of interference fit.

Click to expand...

That troubles me, I see it also troubles you. I agree, the fit should be a slide to tight fit, not a force fit.

How are the brgs preloaded?
The usual way is a nut on the spindle drawing everything together seating the brgs on the spacers.

As for grinding the spindle ID, ummm, do you have the proper equipment? a Heald Universal Internal maybe? you mentioned doing it on a lathe?.. yikes, I dont think so...

What is the Outside Dia of the Brgs?
Are there 2 brgs at the spindle nose?
How far apart along the spindle or how long are the spacers?
What is the min and max RPM?

Let's see some pics

I'll take some pics, of the parts and of the grinding. I should have explained, I will assemble the spindle into the housing and run it on its own bearings. The housing will be held in the steady rest, but not allowed to turn in the steady, if that makes sense to you. For the operation, the lathe chuck will simply be turning the spindle. It should not be difficult to set the pulley end running true. The trick will be getting the wheel on center.

I can always fly Walt in if I can't do it, I guess

The spindle is oil mist lube.

The new bearings are RHP brand, class P4 which is about Abec7, 95mm x 60 mm. Yes, I did learn about aligning the marks on these bearings. I checked to see if they had the marks lined up on the outer races when I disassembled the originals. They were. The inside races were not, and I suspect the races had crept anyways.

No nuts are to be found anywhere on this spindle. Everything is retained with heat shrunk collars. They did have the foresight to design the shrink collar with a groove in it so you have something to grab the collar by when pressing apart. This way, if necessary to take the spindle apart again, there is no need to press the bearings and the collar off at the same time. The main bearings are spaced about 1.5" apart, one on each side of the spacer. This is a 7500 rpm spindle.

The top bearings are 6010 series ball. In my effort to try to do a good job of this, I ordered some with a "G" designation for "low running noise" and found when they came that they were also super precision bearings and cost more than the angular contacts! Oh well. This pair of 6010s are spaced with a wave washer, actually rather tightly compressed, so I suppose they work against one another with maybe 10 or 20 lbs pressure. They are retained on the spindle with a heat shrunk collar. The drive pulley is shrunk on immediately on top of this.

I rebuilt a 30 taper 10k rpm spindle a couple of years ago, it also used interference fit collars for bearing retention. I was going to hire someone to do the rebuild, but none of the four experts who looked at it knew how to dismantle it. For all the dire warnings I got from people (need a 'clean room', special grease, exact amount, etc, etc.), it was a surprisingly easy task.

The Torrington/Fafnir superprecision bearing catalog has an excellent tutorial on spindle bearing replacement.

Hu,
How did your drawbar springs look? Do you plan on replacing them while your at it?

The drawbar springs look okay, but I will replace them anyways, I guess.

Hu
About the heat shrunk collars...

How do you get them off?
How do you put them on?

How can you be sure the brg preload is proper once the collars are installed?

I used a bearing seperator to grab the collars. For those who may not know, a bearing seperator is a wedge type tool associated with gear pullers and the like. It has two hemispherical wedges that can be clamped guillotine style on a groove or in a narrow gap. Then, the puller (or press) is used to pull the bearing seperator, which drags the part along.

I also used a little dash of high heat. There is no substitute for oxy-acetylene and a half inch rosebud. Put a bit of pressure on the parts, apply the heat, and then pull or jack as fast as you can. This keeps the collars from seizing and scoring.

How to put them back on? Well, the interference is in the neighbourhood of .002" and its spanning a 60mm diameter, so it doesn't take a lot to get it hot enough to slip on. I suspect that after the final assembly, I will have to 'jog' the collar down tight with the press to accomplish the preloading. This is because one cannot rely on the position of the collar after it cools down, and shortens.

So, I better have the spacers length correct, or I'm going to have to pull the collar again I will test the preload before assembly by stacking the bearings, and spacers in the press, and pressing on the inner race/spacer/inner race stack. Then, roll it over and go by feel. I don't anticipate too much difficulty that way, the manufacturer states that the bearings are flush ground for light preload, so if I make the spacers identical in height, the preload should be correct.

Understand and agree...
It's about all you can do with that design.

You might try measuring the preload. All you need is a fishing scale, some string and a little tape.
Get it loaded the way it wil be when installed.
Tape a pc of string about 2 ft long to the Outer dia of the brg stack.
Wind it up like a fishing reel.
Measure the pull with the scale as you pull and unwind the string.

It would be ideal if you could measure that before installing the brgs, spacers, collars etc... and after all is installed on the spindle,

If you have any trouble, that information would be usefull to the Brg Manufacturer.


Be real carefull with the heat, but you know that.

Yes, indeed, I could measure the preload that way. I wonder where the preload values can be found? I don't see any real definition of what light or medium preload is, in my ********* book. In the long run, heat rise is going to be the ultimate judge, I guess. It would help to quantify the adjustment to be made to the preload, if the temperature rise is excessive.

Well, it's so long ago that I actually did that sorta stuff, all my records going back 45 yrs is long gone. But you said the brg lube is oil mist?... that should carry the heat away. Did they get hot before you started this project? how about when it was new? Do you run it at max RPM under heavy load for long time periods? Get a little round case thermometer that you can glue on to the spindle carrier right near the brgs and keep your eye on it.

Well, the cartridge is now reassembled.

I did a "dry run" of the preload. I carefully cleaned up and around my arbor press and then faced off some spacers to go above and below the new shopspacer/bearing/precision spacer/bearing/shop spacer stack. On top of this, I placed a spherical washer assembly so when pressing down firmly with the ram, the stack could settle without cocking.

Initially, I had lapped the outer race spacer to match the inner one. But, in the dry run, it would move very easily, indicating that the inner one was too long. So, I carefully lapped the inner spacer down and tried it repeatedly, until I could feel some drag on the outer spacer with full pressure on the ram. Turns out, .001 difference was about right, just as it was with the original set.

Then I warmed the bearings slightly and went for the permanent assembly. As expected the heat shrunk collar did not preload the bearings just by dropping it in place. So, I did have to give the assembly a final bedding via the press. The collar creaked once, and the bearings settled nicely into position. The preload is tight enough that it takes a light hammer tap to move the outer spacer around. I think it is peachy, at this point. Not too much and not too little, IMO. We'll see when I do the run in.

Next up, regrind the taper.

Now is time to use the fishscale to measure the torque... then again after the time spent in the lathe...geeze your brave to grind that ID in a lathe...I would stone the spindle ID to remove high spots and let it go at that.

But if you must...
Do you have a NEW NEVER USED TOOL HOLDER...???
to use as a master and some redlead to use on it to show the fit?

I feel like I'm watching a soap opera in cyberspace.

"Tune in tomorrow to find out, will the spindle run without a glitch? Or will it make noise and make Hu's life miserable? Find out in the next episode of, As the Spindle Turns". Right here on the Practical Machinist.

Hu,
I think you got mail. But I ain't sure
If you get an E Mail, OK if not post here and
I'll try again. Really had some good stuff for
your enjoyment. And your setup as well.
Regards, Walt

Thanks Walt, I have been getting your messages and helpful advice. I have been replying, as well, maybe the messages are getting lost? Anyways, thanks very much for your helpful pointers.

Thanks to Machtool as well for an informative private message.

Gary, yes I have some brand new toolholders. The problem, as I see it, is this spindle has integral drive keys, which makes testing the taper fit a bit difficult. I can jam it in, but not twist it very far at all. Best I can do is insert a plain bar in the other end of the toolholder, and test for wobble, kind of using the wiggler effect. I have used this sort of test in the past to detect looseness in the taper when it is well below "feel level" and bluing, well I don't have much luck with that because its so thick. I tend to rely on felt marker ink It is thin enough to never pile up.

I don't think there is much risk in grinding the spindle in the lathe. It will be turning in its own housing, so no problem with eccentricity. It will be totally a matter of accurate setting and good stone dressing.

Sacrafice a toolholder by carefully and slowly cuting the flange off so that the keys will not interfer.

Bluing is terible because it is thick as you mention, you need redlead to determine seating contact. Dont have redlead?, cant get redlead? I would not grind it, period, end of story. IT IS THAT IMPORTANT.

Your risk is more than not being ecentric, if not done properly the ground socket could be pointing not parallel to the spindles axis of revolution and the surface could be not a true cone. As I mentioned before, if you do not have a good Heald internal Grinding machine, skip it and use a stone to clean up the inside of the spindle.

Gary,
If the spindle is spinning, it is not
possible to grind it eccentric. You can
maybe grind it in a bow tie condition but
you cannot grind it pointing anywhere but
where the centerline of the rotation is pointing
If the tool path on a lathe is traveling
from high to low, left to right, it will
cut shallow on each end and deeper in the
middle where the tool aligns with the center
line of the part. Same thing with ID grinding
of a spinning cylinder and or cone.
Regards, Walt