What's new
What's new

How would you locate the exact opposite point on a sphere to drill a hole?

Chipp

Plastic
Joined
Jan 24, 2017
Good Evening Ladies and gentlemen. I wanted to pose a quick question on how one would find the exact opposite location of a sphere to drill a hole? I could get pretty close with measuring instruments, string or other rudimentary items but what if I needed to get within a thou or so? This is just a general question and I don't have a particular problem to solve but would like to hear suggestions. Where this question stems from: A buddy of mine wants to drill/tap an 8 ball for his gear shifter. While working on that I was thinking ' What If I needed to be crazy accurate on a different project say for space X for 1 million dollars but needed completed in 2 hours? I don't know how I would get the opposite location with extreme accuracy. I'm very Interested in hearing thoughts or techniques on this. I have done a search but couldn't find a technique for the above problem. Lots of stuff on drilling through the earth but that's for another topic.

A rotary table and some sort of live center with a cup on it that the sphere would fit into and centered below the spindle?
Then rotate it 180 degrees on the rotary table?

Just a fun puzzle.
Regards,
//Chip
 
You would likely have a feature on the sphere that would mate to a pin or something in the jig.

If you are aligning to a patter or print or something, make the hole prior to that step.

I think the rotary setup would make probing hard to find the center. Also, if you are offset along the rotary axis, then spinning it around will make you off by twice as many degrees as you missed to start with.
 
I would probably drill/bore a shallow hole to cradle the sphere in with a press fit pin hole in the center of it. Set the sphere in the hole and let the pin locate center. I think it would be pretty close and who is checking it and how? If the hole in she sphere is deep enough just put a dowel pin in a v-block and sit the sphere on it.
 
I'm assuming there's some feature like a hole that you want to be opposite of. I'd use something like a jeweler's lathe chuck that has a spring loaded center on the headstock end.
 
I'd use a variation on an insulator job I used to make flanges either end of a tubular insulator had to be drilled and tapped identically ''timed'' to the opposite end but with no through holes - so;-

Centre up and clamp a ring (that the sphere will sit in) on a base of some sort under the spindle.

Drill peg hole in the base in the centre of ring same size as hole in sphere.

Clamp sphere and drill 1st hole.

Place peg in base hole and sit sphere over peg

Clamp and drill 2nd hole

Oh yes, ............and if you want it ''within a thou or two'' each hole will have to be single point bored etc etc, as a drill cannot be relied on to be on dead centre.
 
Interesting geometry problem. It's a little more interesting if you're not allowed to drill a hole to reference with a pin. So let's say you have a pencil mark as a reference point, and you need to draw another pencil mark on the opposite side.

My idea:
- Mill two cubes, where the length of each side is exactly the radius of the sphere.
- Put the sphere on a long v-block with the cubes on either side.
- Align the sphere so that the reference mark touches the corner of one cube.
- The corner of the cube on the opposite side will mark the opposing point.
 
Mark the top hole and place it some kind of frame. Let liquid run over it until it drips off the bottom. Some kind of paint or dye would leave a mark. make sure there is no wind.
Bill D
 
Depending on the desired hole size -- if it's large enough -- why couldn't one drill and tap a hole anywhere on the sphere, screw in a stud, then chuck the stud in a lathe and the tailstock center will be directly opposite the stud hole?
 
Don't forget, even scribe lines are only good for .003"

.001 location WITH OUT a significant feature is just about impossible (discounting dumb luck)

But optical methods and precision fixtures could likely make for "acceptable" results.

A simple telescope and precision ROTAB would do it for -/- .002

Then shoot the 8 ball with a .30 cal to make the hole. Hammer it on the shift stick, And for all you Brits reading " Bob is your uncle"!
 
I like a variation on jwearing's idea. Put the ball in a cup with a flat bottom. Shape of the cup is not too critical, three points of contact for the ball would be ideal. Using a surface gage, put your first mark on an equator of the ball, in a plane parallel to your surface plate. Set a scriber on a height gage to the equator plane, and mark the opposite side of the ball. Rotate the ball on an axis passing through the mark and the center of the ball. Use the surface gage and/or height gage again to place the mark precisely back on the new equatorial plane. Use the scriber on the height gage again to mark the opposite side of the ball. Intersection of the two scribed lines will be opposite the first mark.

Now, that's only precise to whatever your scriber marks are good for, which is probably significantly coarser than 0.001". If you want better, I think you are going to have to machine one of the features first and use it with a locational fit, preferably a very light interference fit ("wring" fit), on a fixture you can align with your machine axis.

[Added in edit] I see CalG has already pointed out the limits of scribed lines.
 
It is a good one, the ring above is a good shop way, met lab way was same but giant ring with three matched precision balls in the ring, drop the ball in question in middle, three point support essentially makes a plane thru the sphere, clock over top, the case I did was the spherical concentricity as it were where the ball was then just rotated to see how round it was in great circles as it were, it followed with notes on the most perfect sphere made and the challenge of actually making it, it was fascinating as the damn thing was made by hand, amazing
Roundest objects in the world created | New Scientist
Back to the mundane the ball sits in the ring, gets clamped, flipped over on two 123 or 345 or whatever to clock tdc the other side, ( clamp flat plate with a round hole 2 bolts to the ring below)
The course went on to the 4 ball vee blocks which were fascinating in themselves, altogether an enjoyable week of balls announced on Monday when the instructor said he was going to be talking balls all week
Mark
 
I wonder if any of those double head lathes could handle passing a sphere from one chuck to the other. Just line the sphere to the tail stock, hand off, then drill the other side.
 
This is a nonsense question. You need to state WHAT the hole would be opposite to.

I was once told that any problem can be solved by making a series or ever more accurate (less vague) statements of that problem. All too often we see problems stated here that are in a very, VERY vague state of refinement.
 
Idea:
Make an air bearing from graphite about 2 microns bigger than the ball.
Support ball on air bearing, with a high precision bearing on top to give some radial rigidity, so it spins in a plane.
Just needs some small load, the bearing could probably just be placed in a frame with some mass, but not fixed.
Glue an optical encoder on top, hot melt glue should do.
Use 2 million counts or so.

Should be good to about 1 micron, with a ball about 2" or less in D.

Take maybe 2 days and 300$ for a bearing if You don´t have one handy.
Charge about 20k for it, for a +/- 2 micron location fit.
 
We were shown an inspection ball with a hole edm cut right through, it was less than 1 mm dia, apparently held in a ring and tdc found with a clock or dti.
Mark
 
This is a nonsense question. You need to state WHAT the hole would be opposite to.

I was once told that any problem can be solved by making a series or ever more accurate (less vague) statements of that problem. All too often we see problems stated here that are in a very, VERY vague state of refinement.

He stated that it needed to be opposite of another hole.
 
Make a nice tube with the ID just under ball size. Press ball in hole. Chuck it up and do feature #1. Reverse it and do feature #2. If feature #1 already exists, you have to get the ball mounted so the feature runs true. Not sure the best way but it could be done with a close running hole and cementing the part after alignment.
 








 
Back
Top