On Wed, 18 Aug 2010 00:02:45 -0400, Joseph Gwinn
wrote:

In article ,
Ned Simmons wrote:

On Mon, 16 Aug 2010 23:28:56 -0400, Joseph Gwinn
wrote:

In article ,
Ned Simmons wrote:

On Mon, 16 Aug 2010 09:55:27 -0400, Joseph Gwinn
wrote:



For the record, the two outer races ("cups") are each Timken 19268; and
the two assemblies of inner races ("cones"), rollers, and cages are each
Timken 19150.
The stated ID of a 19150 cone is 1.5000".

The quill shaft is about 0.0011" larger than the stated cone ID, so it was a
light press fit. As others have noted , this cripples the running preload
adjustment by turning a ring nut on the spindle, and press fits are too
hard to get exactly correct.

By the way, the interference may be half that. I found the the zero of the
micrometer (B&S 84) I am using is off by about 0.0005". This will be fixed over
the weekend.


Is there a standard solution? I'm tempted to polish the spindle shaft down
until the (new?) bearings are a (hand) push fit. Has anyone done this?

That's near the high end of the recommended shaft diameter, which is
.0007 to .0012 (+18 to +30 microns) over nominal. The ID of a class 0
precision cone is nominal +13/+0 microns, with a resulting 5 to 30
micron interference. Personally, I wouldn't mess with the bearing
seats. Even if you can maintain cylindricity, you're risking fretting
of the seat if the fit is not tight enough.

D.C.Morrison says that the bearings should be a "tap fit" on the spindle, not a
press fit, and endorsed my plan to polish the quill down a bit to achieve a
tap fit.

I dunno who DC Morrison is so can't decide whether their opinion
should carry more weight than that of Timken's engineers in this case.

DC Morrison is the company that supports Millrites, having bought the wreakage
of Burke. http://www.dcmorrison.com/ You must call for Millrite manuals and
parts.

The issue is to ensure that the preload can be set accurately enough. This is
impossible if the bearing closest to the preload ring nut cannot move because
it's too tight a press fit.

That's one of the complications of tapered rollers. Angular contact
bearings can use equal length ID and OD spacers to control preload,
and the same spacers will work with a new bearing pair. Not so with
roller bearings.


Does Timken mention alignment of high spots and low spots to reduce runout?

Not that I see in the literature I have, but "Timken Bearings for
Machine Tools" is mentioned in the section on precision bearings, and
may have more info if you can find a copy online.


"Tap fit" is a rather subjective way to characterize a bearing fit,
depending on who's doing the tapping and the size of the tapper, but I
assume it means a transitional fit with minimal interference. The ID
of your cones can vary by 5 tenths, so a tap fit on one bearing may
result in excessive clearance on another.

This may be what is happening. And things may have corroded and so grown a bit
over the years. I think I'll do most of the polishing on the bearing ID, rather
than the shaft OD, although I'll clean corrosion off the shaft.

I certainly wouldn't hesitate to cautiously remove any corrosion or
high spots you find.



Be aware that the specs on standard tapered rollers are very loose
compared to inexpensive ball bearings, and precision class bearings
are quite expensive. Expect to pay around $270 for each cup/cone ass'y
in class 0 vs. $45 for a standard bearing.

Which class is a "standard bearing"? I'd guess 4, for wheel bearings.

Yes.


What are "airplane bearings", like from skygeek.com? They are priced at the $45
level for the cup and another $45 for the cone et al.


I don't know what "airplane bearings" are, but Motion Industries also
references "aircraft bearings" in their database, with no indication
of what distinguishes them from standard bearings. But they are listed
as "standard precision."

What is a typical price for Class 3 set of one 19268 and one 19150?

Looks like about $125 for the cone and $68 for the cup. Those are my
approx prices from Motion Industries.



What class is needed for a MVI anyway?

Class TIR
4 .0020
2 .0015
3 .0003
0 .00015
00 .000075

The runout of Class 0 roller bearings is approximately the same as the
ABEC 7 angular contact bearings typically found in machine tool
spindles.

DC Morrison says that Timken Class 3 are what was specified for the Millrite
spindle, and the existing bearings are in fact marked 3.


BTW, the first thousand or so J-head Bridgeports used tapered rollers
on the spindle; the subsequent couple hundred thousand have angular
contact bearings.

Burke used taper roller bearings for the standard R8 spindle, and ball bearings
for the high-speed spindle. Taper roller bearings are more robust than ball
bearings under shock loads (like interrupted cuts).

True, but BP spindle bearings will take a lot of abuse if they're
lubricated properly and kept clean. Tapered rollers will carry a
substantially heavier load than angular contact bearings in a given
volume. But there are tradeoffs, such as added complexity, expense,
and less tolerance of contamination.

All true, and Burke apparently valued robust more than Bridgeport. Perhaps the
intended audiences were different. Not that I've heard of Bridgeport spindles
being so fragile. But I don't know why Burke didn't provide more protection
against the entry of dirt and coolants.

BP spindles are not especially well protected, either. The nose pair
is protected only by a labyrinth formed by close clearances between
the various parts of the spindle and quill. Indiscriminate use of air
to clear chips can force crud up into the bearings.

--
Ned Simmons