What's new
What's new

How to turn a complete sphere

Lotaxi

Aluminum
Joined
Jun 9, 2020
Hello,

I'm trying to figure out how to turn a complete sphere on a lathe. The front side is easy enough, but I need to keep a neck on the back side. After I part it off, how exactly should I hold it to machine the neck off? I'm most worried about re-clamping ruining the alignment.
 
I have never done it.
Could you cut off the ball and super glue it backwards into a slight “socket” cut into bar stock to allow access to the “nub”.

I feel getting it centered would be next to impossible
 
I have done this with soft jaws. It worked okay once the blend was comped out a bit.

What if keeping shape was pretty crucial? I would like to get the surface matched with as little blending as is possible.

I have never done it.
Could you cut off the ball and super glue it backwards into a slight “socket” cut into bar stock to allow access to the “nub”.

I feel getting it centered would be next to impossible

I've seen an idea of making a fixture that would be clamped into the jaws separately. Idea would be that the mating surface gets turned true after the fixture is clamped, and then the sphere is mounted into the fixture. Seems like a lot of work, though.
 
If you made a set of soft jaws with a below center spherical pocket you could realign the backside nubbin with one of those concave live centers. Set your z zero by putting the sphere off center and touch off from the cut to size sphere section.
 
What is the tolerance, and what do you have available to do it?

±.030 is pretty damn easy, and not a big deal.. Turn maybe 200 and
something degrees, flip into some soft jaw, or even hard jaw..
Turn again. done.

Half a thou spherical.. yeah.. that's a bitch. It amazing how far a
.0002" variance in a tool radius can make in that situation.

Why do you need a full sphere? I've made tons of stuff, I have yet to
need a full sphere for anything. If I need a ball bearing I buy it.
 
Sub spindle lathe with soft jaws bored to match in sub spindle. Bit of fiddling with offsets and you should be able to get it pretty darn good.
 
Hmmm, if it’s a sphere of constant radius does it really matter if the nib is in the center upon flipping? Blue it up and creep in on it in Z?
 
Pretty easy. Just turn more than 60% on the first side. Have radiused jaws grabbing on 55% and finish the back on op2.
 
Pretty easy. Just turn more than 60% on the first side. Have radiused jaws grabbing on 55% and finish the back on op2.

Not easy. Have you done this? Very difficult to get it to blend perfectly. Even a tenth of runout becomes obvious, so the feasibility of this method is very application dependent. Then there are issues with the diameter of the ball (and position in the jaws) moving a bit as things heat up.

I made a production run of 2in spheres that had to be held within .001in diameter and needed a cosmetic finish. For my part I wound up turning about 75% of a sphere in OP1 using a button insert. Soft jaws with and undercut for the part to sink into for OP2, then off to the scotchbrite wheel to blend the slight mismatch. They moved all over, and it was a PITA. Like I said, for that application even a tenth here-or-there makes it tricky.

AFAIK, if you really need a sphere, the only "normal" way to achieve it is double disk grinding.
 
Not easy. Have you done this? Very difficult to get it to blend perfectly. Even a tenth of runout becomes obvious, so the feasibility of this method is very application dependent. Then there are issues with the diameter of the ball (and position in the jaws) moving a bit as things heat up.

I made a production run of 2in spheres that had to be held within .001in diameter and needed a cosmetic finish. For my part I wound up turning about 75% of a sphere in OP1 using a button insert. Soft jaws with and undercut for the part to sink into for OP2, then off to the scotchbrite wheel to blend the slight mismatch. They moved all over, and it was a PITA. Like I said, for that application even a tenth here-or-there makes it tricky.

AFAIK, if you really need a sphere, the only "normal" way to achieve it is double disk grinding.

Yes, I have done it. Several times. Not very difficult getting it to blend. Do you know how to bore jaws? If not, I can give instructions on that. 75% on OP 1 sounds good. That is close to what I would do. I would use a full radius grove tool rather than a button. I have never found button tools to be good.
A blend on one part is no different than any other. If you are having alignment issues try to set up your op2 jaws so the center of the sphere is in the same place as op 1. Are you using a ring to bore jaws? I just ask this because here we have a 65 year old machinist who claims to have 40 years experience that just cut his jaws without a ring or pin, then just moved them in a notch. Then, he could not understand his concentricity problems. Sorry, just have to ask.
 
Virus,
I don't get it either, boring soft jaws with no ring or pin. It seems so intuitive, but I see it: older even younger guys, despite being good machinists in other respects.
 
If only those Union and Confederate troops had that technology. All they had was a campfire and some tongs ...

Not correct. Union troops were issued cast Minie balls (actually hollow base bullets with lube grooves) and Southern troops, who generally had more of a marksmanship background, used everything from pre-cast Minies to round balls cast in tong molds by a campfire. Many Southern troops carried the English made Enfield rifle, which like the Union Springfield was a rifled musket instead of a smooth bore musket or conventional rifle.
 
I found that it is much different chasing a blend on a sphere, when the growth of the part will affect its z position as well. Normally on a shaft growth doesn't cause it to push or pull out of the jaws.

The issue isn't making jaws (yes I know how). The issue is that on a sphere any slight deviation will cause the tangency point with the jaws to change. I was easily seeing .0002in of variability in height for OP2. Easy to blend one or two parts - but not so easy to setup and run a batch.
 
I found that it is much different chasing a blend on a sphere, when the growth of the part will affect its z position as well. Normally on a shaft growth doesn't cause it to push or pull out of the jaws.

The issue is that on a sphere any slight deviation will cause the tangency point with the jaws to change.

If I understand you correctly, the back of the round part(op2) is sitting against the the back of the jaws? If that is the case, that may be your issue. Your jaws should have no support for the back of the sphere. Think of a tube with a shallow id radius. That is how your jaws should be. You are just gripping on a small section of the od...nothing else. That should keep your sphere centered in the same spot, even if your OD changes size slightly. It is like you are just grabbing the part between three fingers on the OD instead of cupping it in your hand.
 
What is the tolerance, and what do you have available to do it?

±.030 is pretty damn easy, and not a big deal.. Turn maybe 200 and
something degrees, flip into some soft jaw, or even hard jaw..
Turn again. done.

Half a thou spherical.. yeah.. that's a bitch. It amazing how far a
.0002" variance in a tool radius can make in that situation.

Why do you need a full sphere? I've made tons of stuff, I have yet to
need a full sphere for anything. If I need a ball bearing I buy it.


I wasn't clear in the OP, sorry. 'm not making a sphere, I'm trying to finish a spherical surface.

I'm trying to make solids of constant width. When used as a bearing surface, they act as spheres, but they aren't.

My issue is that turning them around to machine the flatter surface seems... difficult. I don't know how to do it correctly. I'm fairly inexperienced, so I'm trying to learn as I go and figured I'd ask more experienced people.

Why do you want the headache, just leave a little flat spot and sand it out...phil

I've seen that done, yeah, but I'm more interested in learning how to be precise when batch machining. Hand finishing is fine, but if I make a bunch of something I'd rather it come off the machine just about finished and dimensionally identical to the others (within the machine's ability, anyway.).
 
Ah, now I get it. Interesting shape. Hadn't heard of that before.

Possibility... Indicate the bar in, machine the "flatter" side first, plus a bit past the "ridge" . Part off with enough stock remaining to finish the part and then some. Cut off, rechuck end-for-end, held by the bar stock, reindicate, face and center drill.

Turn soft jaws to hold the shape over the "ridge" (could be as simple as a vee in the jaws), grab the part and help center it with a center in the tailstock. Turn the remaining shape.
 








 
Back
Top