Tuesday, April 24, 2018

About That Shop Problem I Mentioned

Yesterday, I mentioned a bizarre problem I had come across while getting my flame licker engine's first part fixtured and cut.  I've resolved the issue and think I'm ready to move on to the original task, but it was honestly a problem that I had never thought could exist.

The problem has to do with my rotary table (RT), which I've written about before.  To use the RT the way I need to, the center of rotation has to be under the center of the mill's spindle.  It turns out the center of rotation doesn't have to be the physical center of the RT.  

I used the same centering method that I used in that "before" link.  I put my MT2 dead center from the big lathe into the matching taper in the rotary table, put a small spotting drill in the spindle and used the Rumblepad controller to jog the table around until the points lined up looking along the mill's X and Y axes.  That puts the physical center of the RT under the spindle.  Satisfied it was in place, I put the work piece (picture of part here) on the table and started working to center that.  I had coated the piece in blue machinist's layout fluid and scratched lines that crossed where the center of the big hole will eventually be.

Now with a center finder (a pointy cone), I was going to put that over the intersection of the two lines then clamp everything down.  I rotated the table so that I was looking down the Y axis, centered the work under the point, rotated the table 90 degrees so that I looked down X, centered, and repeated, going back and forth a few times until it looked like the point was always over the center.  Satisfied, I spun the table farther and when I was on the far side of the original Y axis (table set to 270 instead of 90), the center finder's point was no longer over the center.  It was about .035 away.  Red circles drawn on this photo make the marks I made look obvious.


I spent the rest of the Sunday trying to resolve this issue, eventually going back to the original MT2 center and verifying the problem was there.

On to another way of centering the axis.  I used the open bore in the table, and set each axis so that the coordinate of each edge was the same.  Then I clamped a piece of thin aluminum to the table and using the finest engraving bit I have (which has a .010 tip) rotated the table 360 degrees.  I found that it cut a circle that's about .035 in diameter (center of the path, not the very edges).  If the center of rotation was really the physical center of the hole in the table and directly under the center of the spindle, that should have been one small hole. 

The only conclusion I could make at this point was that since my machine said the center of the bore was 0,0 but the cutter put there didn't just cut a single hole was that the table wasn't rotating around the geometric center.  It was rotating around someplace else - someplace away from the center.

A consultation with some experts online said the trick was to find the coordinates of that spot.  To do that, machine a cylinder by putting a cutter in a fixed spot and rotating the table.  That would machine a cylinder centered around the table's center of rotation, and once I got those coordinates the work would rotate around that spot.

I grabbed a block of scrap aluminum about 2-1/2" square by 1" thick and cut a cylinder.   A convenient diameter to use seemed to be .750. I'd move the mill in X only and use a .250 end mill. This is easy stuff, right? A 3/4" diameter means a 3/8 radius; add a 1/8" cutter radius and the center of the spindle should be 4/8, 0.500", from the center of rotation.  I moved X to 0.500, started the cut, rotated the table 360, dropped the cutter .050 at a time and made cuts to make a ring 0.500 deep.


You might be able to see the slots I started to cut in from the sides in both +/- X so I could fit a micrometer in there, but I used calipers to measure the cylinder (the half inch wide slot is too narrow for my micrometer). 

The diameter is almost 0.700. That means the spindle was really .025 (radius) closer to the center of rotation than it should have been.  Now I switched over to an edge finder, found the coordinates of both sides of the cylinder in both axes and found the midpoints so that I get the same number (+/-) on both sides of the cylinder.

Finally, I put the mill's table to 0,0 lowered a sharp point engraving tool down to touch the top of that cylinder and rotated the table 360 degrees.  There was essentially no movement; it made a hole a couple of thousandths in diameter.  The big wobble I had with my original method was gone.  

Now it looks like the next step is to put the work piece back on it and try to center the cross hairs again.  My test cylinder block has enough clamps to keep my house from moving and I can't clamp more than half the work piece, since I cut half at once.

Up above, I remarked that it turns out the center of rotation does not have to be the physical center of the RT.   It seems that you can approximate how well they'll align by the cost of the RT.  I don't think it's true that all Chinesium tables are crap, but I do see echoes of "you get what you pay for" here.

And I need to figure out a way to not go through all of this the next time I need to use the RT.



10 comments:

  1. You just found the center. Remove that center finding setup, and bore a new center hole in the rotating table. Has to be larger than what it currently is, to leave a clean hole. Fill with a block (glued or pressed in) with centering marks/hole, etc.
    2)Alternatively, calculate if a bigger MT taper would work in place of a thru bore. (Will require a taper cutter)

    BTW, did you examine the table mounts/bearings/gearing system for dirt, slop, broken/worn edges? You REALLY need to verify that nothing is adrift in the table assembly BEFORE making any changes. You think you found the correct center. Is it repeatable? If you dismount it, stand it on edge, bump it a few times, spin the table, remount it, will it still have that "new" center? How did they spin the part when it was made? I would want to see the underside/inside of it.

    A sloppy bearing mounting bore that allows the bearing to shift under stress can drive one to drink. DON'T assume parts are made correctly! Look at EVERYTHING. When in doubt, measure mating parts. EXPECT to find bad parts and/or bad workmanship. Expect MORE than one problem, especially with Chinese stuff.

    ReplyDelete
    Replies
    1. All excellent questions, and other than finding this new center, I haven't done anything to it. Haven't even taken the cylinder off. The table is clamped to the mill's table with three clamps, you can see one on the left and two on the right in that picture.

      What I was thinking of doing was putting my MT2 dead center (or live center, I guess) in the table and getting the coordinates of where the CoR is. Write those in permanent marker on the side (or scratch them into the thick paint) so that when I put it back, I'll find the physical center like I've been doing, but instead of calling that point 0,0, I'll enter the coordinates I get for the physical center now. Sort of an offset from the real CoR.

      The table's body is pretty well sealed, I'll give them that. It has a glossy and tough poly-something-or-other coat on it (that's a guess). It seems resistant to scratching. This is only the second time I've used it as a C-axis (lying on the table with the work parallel to the mill's Z-axis). I'm likely to mount it as an A-axis, parallel to X. Those are pretty common.

      Delete
  2. Frankly, Chinese tools/tooling needs to be considered as a kit, a DIY model that has been pre-assembled to make shipping easier. Always consider that the builder or assembler is trying to pull a scam of some sort, and examine everything with that in mind. Much safer with that mindset.

    The last one I encountered: Brass brushes for using in a drill. Turned out to be brass colored steel wire. Bastards.

    ReplyDelete
    Replies
    1. got it in one. The Chinese manufacture to cost. All else follows. They can and do produce excellent quality- at the same price everyone else does.

      Delete
  3. Will has been around the block a few times...! Good advice there.
    I bought a cheap azz drill press vise from china, had to put it on the mill and machine all the surfaces to get it flat and level. Still a POS, but a flat and level POS.

    Some of the importers like Phase 2, etc, seem to have some quality control in place. Others, not so much- a friend just bought a D-4 collet chuck from some small chinese importer and it had .003 run out. We tried it on two different lathes.

    ReplyDelete
  4. Don't think I've left any comments here before but I've been reading your posts pretty regularly for years now. I was just wondering if there was a reason you didn't just put an indicator on your rotary and sweep a pin in the spindle of your mill? That would get your spindle centered above the center of rotation for the rotary. Would also show any run out you've got in your rotary.

    ReplyDelete
    Replies
    1. Thanks for stopping by regularly!

      What I did was put the indicator in the spindle and indicate the opening in the table. I didn't spin it, but moved the table back and forth in X (I had to turn the spindle around 180 degrees because of the way the indicator worked) and made sure it read the same number. Then I did the same with the Y axis. The spindle was centered by my DTI that reads in .0005" increments. (This is what I summarized as On to another way of centering the axis. I used the open bore in the table, and set each axis so that the coordinate of each edge was the same.).

      That was before making the cylinder.

      I think what was happening was that the RT was rotating around a spot that's not exactly the geometric center. Every now and then I think I can visualize how it could be made wrong and work like that. :-D

      Delete
  5. Suggest you look up some of the machinist Youtube guys, Adam Booth (Abom79), Tom Lipton and Keith Rucker. They have some pretty good videos on metrology and machining. :)

    ReplyDelete
    Replies
    1. And to be well and humbled, Robin Renzetti. He may not be human.

      Delete