Self-locking taper angle 1

[dgeesaman]

I've been looking into a self-locking taper connection. My ideal outcome is 'slightly' locking so therefore I've dug up the formulas.

The math that predicts that a taper locks when it has more frictional force than reaction force, i.e. (half-angle of the taper) = arctan(coefficient of friction).

The other agreed nugget is that well-machined or ground steel tapers begin to lock when the half-angle is at or shallower than 7 degrees. (

Combining them, the 7° rule of thumb corresponds to a frictional coefficient of .123.

Book friction values for dry steel vs. steel range from .4 to .8. Less than 0.13 would require grease or high polishing. It's well known that steel tapers should not be greased unless you want them permanently fused or you're asking for the hub to burst.

Can anyone shed light on this discrepancy?

 

[Doug Hunter]

It seems to me that you're attempting to play in a dangerous area. I've seen taper-fit joints used for locking parts securely, but never "slightly locking". I think that it will be extremely difficult to control all the parameters adequately to get an Engineered amount of "locking". I'm not familiar with the 7 degree rule, but I think that the tangent of 7 degrees is significantly lower than the steel friction coefficients in order to provide a safety factor to ensure a locking joint.

 

[dgeesaman]

It's not dangerous per se. This is a joint that will require frequent assembly and disassembly and the working loads are in the direction that drive it closed. My intent with this taper angle is to have a joint that disassembles easily with minimal wear, with enough locking to keep things in place and well-aligned in between operations, but preferably not so much as to require full press force to disassemble. Worst case is inconvenient.

To your point I see commercial tapers are typically very steep (quick change machine tool tapers) or very shallow (fully locking machine tool tapers). This thread discusses the 7° half-angle threshold. https://www.eng-tips.com/threads/locking-taper-angle.108966/

 

[3DDave]

Machine tapers are also subject to radial/bending loads that would tend to loosen at wider angles. Even a cylindrical interference fit can have the pin migrate out under those circumstances - which is why I just had to shove a door hinge pin back down where it ejected itself against the action of gravity.

Self locking is an interference fit with much less sliding.

 

[Doug Hunter]

Dgeesaman, what kind of forces does the joint need to carry?

 

[dgeesaman]

A couple tons of axial force. No bending or torque.

I have a reference design that uses about 3 deg/side that releases with consistent, moderate force. I purchased a ball joint taper reamer that is slightly steeper and will make a prototype using that taper.

When the math doesn't line up, you just have to prototype and test. I may try to measure the axial force and measure the various geometric factors to back-calculate an effective coefficient of friction. But at that point it's mostly irrelevant.

 

[Doug Hunter]

Seems like a good plan.

 

[LittleInch]

Ah friction, your unreliable friend.

This source uses Cof numbers similar to yours for screw threads.

Coefficients Of Friction - Roy Mech

roymech.org

 

[Artisi]

From memory a similar question / thread was discussed some years back, could be worth looking in the old posts.

 

[Jboggs]

A very hard-learned lesson on taper fits: The aspect ratio (length/width) MUST be correct. Some high speed rotating equipment in our plant was designed 30 years ago with rotary unions on each end of a shaft. These unions are changed out on a regular basis.

The engineer thought he would follow good practice and designed the rotary union connection joint similar to a Morse taper, a very common thing in machine tools to achieve high levels of precision. He got the right angles, and the right level of precision in the drawings. But made the joints TOO short because of limited space. The result, that our company struggled with for 20 years, was "wobbling" rotary unions. Even if all the parts are perfect, a certain length of contact is required to prevent misalignment, thus "wobbling", thus vibration, thus bearing wear, etc.

There was no room to redesign this joint to use the appropriate length of taper contact, so we redesigned it to use flange mount rotary unions. Problem solved.

He thought he was doing the right thing.

 

[lucky-guesser]

From memory a similar question / thread was discussed some years back, could be worth looking in the old posts.

Design of tapered shaft friction coupling - Mechanical engineering general discussion

I'm designing a wheel that can be moved along a machined tube. I have chosen an 8deg taper between my split collar and clamp plates but that was just a guess. Are there any guidelines on split friction clamps that might reduce the number of design iterations I have to build and test...

www.eng-tips.com