开云体育

I pulled the headstock and pressed out the spindle and bearings today.

My original plan was to make an elaborate expanding mandrel. This was mandated by the need to press the bearings out of the housing, but having to clear the 25 mm bore to do it.

But happily I woke up this morning with a simpler idea. I turned some bar to about 27 mm to clear the 28 mm spacer and 25 mm bearing bores. I then cut off a 2-3 mm slice and cut that in half. That let me drop the 2 halves through the bearing bore and then set them on top of the inner races. I've included phtots of the half disks and with them in place prior to pressing the inner races out.

I then used a socket and extension to press on the half disks and push out the outer cones and the plugs. To remove the outer races I used another socket and extension.

The last photo shows the housing with the bearings removed and makes clear why both had to be pressed outwards from the opposite side.

The headstock housing appears to be undersize, but I'm not confident measuring it with telescoping gauges, so I've ordered a 0.0001" dial bore gauge. After a lot of searching I found that the spindle and housing tolerances for ABEC 7 bearings are the same as the bearings. In the case of the 7205C-P4-DGA bearings that's +0/+0.0002" and +0/-0.0003" respectively.

Aside from being *very* difficult to measure that closely there is the question of how I'll remove the metal to bring them to tolerance. However, I'm very happy I don't need to plate the headplate bores.

I measured the spindle journals and they are 0.0001" and 0.0002" oversize based on 5 measurements of each journal.

After investigating tool post grinders, I've decided that is too risky for such small cuts. Present plan is to flatten the face of a small corundum sharpening stone and glue it to an angle plate, mount it on the compound and then press it against the spindle while it is spinning between centers on my 10 x 20 Clausing 4902. That will be *very* slow which in this case is a big advantage. I'm allowing myself a tenth of overrun.

The housing bore is more difficult. Present concept is to make an expandable plug with 220 grit wet or dry paper and just turn it by hand or perhaps electric drill to open up the bores.

In the event I fail on the ABEC 7 bearing installation, I may simply make tapered bronze bearings and lap them to the outer races of the OEM 30205 taper roller bearings. That's the traditional bearing for an instrument maker's lathe as it is the only bearing that has *no* rotating errors. You rough out the taper and then lap the bearing to the race.

My 4902 has plain bearings, but they do have the downside that you *must* be attentive about oiling them often or fitting large oil cups.

So the adventure continues. There are high spots at the exterme ends of the front way. As the ways are ground, I'm going to glue some 220 wet or dry paper to a plate which has been ground on a surface grinder and sand them down a tenth or two.

Have Fun!
Reg

After cleaning up I miked the spindle bearing surfaces. I took 5 readings rotating the spindle by 20-30 degrees between measurements. The bearing nominal dimensions are 25 x 52 x 15 mm. In inches that's 0.98425" for the ID. I've ordered a 0.0001" dial bore gauge as I can't measure accurately enough with telescoping gauges. That's the resolution, not the accuracy. Mitutoyo states the accuracy of their 0.0001" gauge as 0.0008". Mine is a Chinese gauge at 1/4th the price. So I'll be checking the dimension with an outside mike. The resolution is needed to be sure I have it aligned correctly.

For the rear journal I got 0.9843" 3x & 0.9842" 2x So I'm going to call that 0.98425"

For the front (spindle face) I got 0.9844" 3x and 0.9845" 2x. So I'm going to call that 0.98445".

ABEC 7 fits are "line to line", the average fit should be within the +0.0/-0.0002" range of the ID and the +0.0/-0.00025" range of the OD. Thus for the rear journal I'm at ~+0.0 and for the front journal I'm at ~+0.0002".

This means that I want to remove about 0.00005" from the rear journal and 0.00030" from the front. In the case of the rear journal I'll stop as soon as I see any effect. The front journal will get miked as soon as I notice an effect and I'll continue based on the readings.

The journals are ~5/8" wide. To keep taper within 0.0001" across the width I have to maintain parallelism within 33" of arc. That's ~1/120 of a degree!

There's no way I can set or maintain that with my 60 year old lathe. It was never intended to do such things. However, human touch is a marvelous thing. I can easily *feel* that a small whetstone is parallel to the existing surfaces simply by how much force the rotating spindle puts on the stone. Parallel will transmit the most thrust.

The planned set up is to mount the spindle on the MT 2 test bar, put a piece of copper pipe over the bar to protect it and then indicate it to a tenth in a 4 jaw independent chuck. That will have a lot of overhang which I don't like, but I know from traversing the carriage ways that I have about 0.003" of wear on the carriage bed.

I've never measured the tailstock ways. So I'll have to indicate them from the carriage to see if they are worn. They probably aren't in which case I'll be able to support the outboard end of the spindle. The ultimate decision will be governed by the TIR at the outboard journal. Even with the bar dead true at the chuck it is highly unlikely that the bar and spindle will be aligned with the axis of rotation. Even a collet would be challenged and a chuck even more so.

I will dress the faces of the whetstones on a sheet of silicon carbide wet or dry paper stuck to my surface plate with water. I flattened them recently, but for this I'll do a quick touch up. I've routinely used this arrangement for lapping for many years. But having seen Robin Rennitz's "magic sanding plate" I'm going to make one of those for routine work with 220 grit and reserve the surface plate for cases where I want to run through successive grits to reach a high polish. For flattening whetstones I use 80 or 100 grit. A surface grinder would be better, but I don't have one and holding a stone on a magnetic chuck would be challenging and dangerous if it came loose.

For large composite corundum stones I've simply wet down a flat spot of concrete and done figure 8s rotating the stone. That is an effective way to remove the hollow that use puts in the stone. If you dress your whetstones as soon as you notice a hollow developing it's a lot less work to get it flat again. And your knives will appreciate it.

Have Fun!
Reg

After lots of investigation and thought I think I'm ready to "grind" the spindle.

I've determined that I need to remove 0.00005-0.0001" from the rear journal and 0.00025 - 0.0003" from the front journal to bring them within the +0.0/-0.0002" ABEC 7 spindle tolerance.

Because of the extremely light cuts I've concluded that a tool post grinder is neither needed nor desirable. Instead I intend to hand work a small whetstone against the journals while the lathe is running much as one would a file. The rear journal will just get the medium Arkansas stone. The front will start with the corundum stone and finish with the Arkansas stone. Both stones have been lapped flat using silicon carbide paper on a surface plate. I plan to run at minimum speed and measure often. When the front journal gets to a tenth oversize I'll switch to the Arkansas stone.

The spindle is mounted on an MT 2 test arbor which has been wrapped with copper wire and placed in a 4 jaw independent chuck. It will be indicated to a tenth or less TIR. The chuck jaws look to be in excellent shape with no discernible wear.

I considered a number of other approaches and this seemed the most controllable. Maximum drag corresponds to being parallel to the existing surface.

Well, that didn't work worth spit. The Arkansas stone removed an unmeasurable amount of metal. Perhaps, maybe a half tenth.

Despite attempting to use the tailstock to align the test bar in the chuck, the copper crush resulted in excessive run out at the spindle journals. So I didn't try the corundum stone.

While evaluating how much metal I removed I discovered I have a tenth out of round at 90 degree orientations. I got the same values of 0.9844" and 0.9845" at the two orientations alternating orientations by turning the chuck. Much more careful measurement than my initial evaluation.

So lapping seems to be the way to go.

How will you make your lap? What material, shape, etc?
I attempted to reduce the diameter of pilots used in valve guides of automotive engines by using abrasive cloth years ago and all I did was put superficial scratches on the surface. This was before I had any machine tools available and I was probably a teenager.

I've roughed out the lap per plan. It is now ready to drill and tap for adjustment screws before slitting. Almost ready to go.

I just *love* cutting grey iron.

I've completed the lap except for tapping and buying screws.

I decided to try a set of first tier OEM type bearings, in this case the Federal-Mogul National line which O'Reilly carries as they were most easily available. Made in Thailand to my surprise.

Once I've lapped the spindle, I plan to fit the 30205 taper bearings, lapped to fit the spindle if needed, and test their performance. Then I'll install the ABEC 7 7205 angular contact bearings and test those.

It would cost about $100 to build up a headstock using plain taper bearings. As it's the only option with no periodic rotational errors, having such a headstock is quite desirable even if infrequently used. If I build a 2nd headstock I'll redesign the outer housing with the switches to be more convenient to work on, i.e. hinge out of the way to change headstock.

FWIW I strongly recommend watching:



I had not previously considered the impact of the spacers, etc. I've also discovered that mine are not right which will require remediation.

I'm considering making a class 100 glove box for doing spindle assemblies, probably a folding arrangement using muffin fans, coroplast (plastic "cardboard") and a multistage wet bath air filter feeding laminar flow across the work area. For me it has additional utility for working on hard drives if the need arises. I've got a lot of scrap computer fans ;-)

I installed a pair of National brand bearings, lightly greased with a brush. After running a a few minutes with about 0.1 end play I adjusted it down to about5 thou or so and ran the bearings in th distrubute the grease, 10 minutes in 4 speed steps each direction with reverse following forward at each speed.

I then aligned the headstock. during this work the cover was sitting on the side so I had access to the headstock adjustment screws. It's a *lot* of work to thread the wires through under the spindle. If you ake one apart, glue the wire labels in place with hot melt glue as you disconnect each wire.

I've run the machine for over an hour at full speed each direction. It gets warm, but not hot. Endplay using a lever and significant force is ~0.0002"

I mounted an Amazon Chinese/Indian EW32 collet chuck on the spindle and aligned it.using an indicator on the inside of the chuck.the result is a 3-4 tenth maximum TIR in cycles of 0, 1, 2, 3 tenths swing.as the center of rotation shifts in the ABEC 1 30205 bearings.

I concluded after much study that the ABEC 7 bearings need a new headstock design such as used by Levin in which the spindle is a cartridge assembly that mounts in the headstock. It's easier to make a new headstock than it is to modify the OEM design. Fortunately a suitable piece of ductile iron is under $40 with machined faces. That also allows me to *use* the lathe while I'm working on the headstock.

I strongly recommend watching the videos these guys have on YouTube:



Have Fun!
Reg