spherical on end of shaft 1

[subsearobot]

Hello all,
I am designing a shaft detail that requires a spherical on the end of a shaft. the ball fits into a socket, which allows minimal torsional coupling, while allowing a strong axial connection, and is self centering between elements.

It's a spring biased push shaft that does not couple rotations. no pull, or separation is possible.

The spherical end is perhaps 1/3 of a ball, perched on top of the cylindrical shaft.

I have it in my head that machining a spherical end on a shaft is difficult, and thus expensive. this will be a production part (100k/year). the shaft is ~.080" diam, the radius of the endis .06.

If i were to use a blunted cone instead of a ball, i think it would give me the self centering. The friction would be slightly higher, but i think it woudl be fine. Blunted because there are small impact loads.

In terms of machinability on a screw machine, do either of these solutions stand out as easier and thusly less expensive to manufacture in mass?

cheers,
thank you!

 

[hendersdc]

Can you have a flat on the end of the spherical section? If so I would expect the machinability of either feature to be the same.

 

[subsearobot]

I was trying to avoid the flat, but the flat on flat seemed fine in an informal examination.

 

[tygerdawg]

Trying to design something like these guys make?

http://www.adandd.com

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[btrueblood]

Machining a full ball is tricky (most ball bearings are forged and cold rolled to shape IIRC), but machining a partial sphere on the end of a shaft is not that difficult, especially in this age of CNC lathes. Even old school manual machinists can manage it with a bit of extra kit (ball turning tools). Simpler yet perhaps would be to buy the ball end from a specialist company, with a suitable female thread and spin it onto the corresponding thread on a shaft.

Blunted cone, yeah simple but you lose the off-axis rotation capability.

Not quite clear from your descriptions what you are trying to accomplish, and would expect some sketches from you if you wanted more ideas...

 

[subsearobot]

It is strange, because on several different production examples that i have, the ball was made in two steps. when I say production, only ~100k per year.

I examined ball surfaces with a micrograph. It appears that the ball was machined with a stub on the very apex, which was subsequently removed by a grinding operation. in one case, it left the linear wear pattern, the other a different circular wear pattern from the tool marks.

The joys of working for a medium size corporation. this means I have no idea who will make the parts, so i can't just call the shop up and chat with a machinist.

 

[BrianE22]

There are a bunch of machinists that hang out at:

http://www.practicalmachinist.com

The "General" forum would be your best bet. It's tough to get the nub off the middle since turning the part produces a velocity of zero on the centerline. I'm guessing if you want a nice clean spherical radius with no nub or flat spot on the centerline you would need to do a subsequent grinding operation on a machine that spins the grinding wheel and rotates the part at the same time.

 

[subsearobot]

Nice, thx Brian. I hadn;t thought about the zero velocity issue. of course.

 

[CWB1]

Cutting a radius like described is neither expensive nor difficult so long as the setup is decently rigid. On a screw machine or lathe it would likely be a simple form tool plunged in.

An engineer not being in direct contact with suppliers to quickly resolve any small or large question or concern would be a major problem IMHO, and is unacceptable regardless of the business' size. I would recommend discussing the matter with your supervisor and purchasing dept asap bc I'm not sure how they expect you to do your job to any reasonable standard blindly.