Design of tapered shaft friction coupling 1

[kyleShropshire]

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?
Specifically regarding the taper angle and any heat treatment the split ring might need to allow sliding on the tube without lubricant.

Clamp must resist 12kNM wheel torque without slipping.

I have x16 M16 SHCS at 75%Proof to accommodate the joint C factor of 1.
I started with the tapered wedge calcs and radial bearing moment of friction from Gieck Technical Formulae 9th to determine the amount of friction-moment required to resist the wheel torque.

 

[kyleShropshire]

The estimated 237kN impact hits at 34deg angle from vertical.
I was worried about the impact exceeding the friction and the bolts supporting it.

 

[TugboatEng]

We impact loaded one of our tapered couplings recently. It bent the shaft, the coupling never slipped. Tapered fits are incredibly strong. Our propellers are fitted using an adaptation of the SKF OK coupling system.

https://www.skf.com/us/products/power-transmission/couplings/ok-couplings

I strongly suggest incorporating hydraulic mounting and dismounting for a coupling in your size range. I also strongly recommend hydraulic mounting for galling prone materials such as titanium.

I do recommend using a single taper. These tapered fits can be axially loaded in both directions without separating.

 

[SwinnyGG]

I was worried about the impact exceeding the friction and the bolts supporting it.

Did you account for the fact that the vertical component of your impact force increases the normal force between the coupling faces, drastically increasing the amount of (lateral) force it will take for them to move?

We don't have a lot of detail on your application but I'm going to assume that if the coupling slides under any condition it's failed.

237kN isn't that huge of a force for a part this large, or a tapered coupling with this much area. Have you calculated face-to-face pressures of your coupling yet? I'm betting they're huge.

 

[kyleShropshire]

No, I had not thought about the impact loading increasing the force it took for the coupling to spin on the shaft.

 

[geesaman.d]

If it's a self-locking taper angle, then yes the friction and preload in the taper will transmit all loads including shock. I work with a different kind of propeller and we see only self-locking tapers. QD and Taperlock are self-locking angles. Draw it down and technically you don't even need the thrust plate anymore. They always require an extraction method - jacking screws, 3-jaw puller, hydraulic release, etc.

If it's a non-locking taper angle, then the bolts will carry and absorb those loads. Loosen the bolts and it comes apart.

I haven't checked the math but coefficient of friction may affect the threshold between locking and non-locking taper angle.

I think one of the biggest risks you'll need to address is the actual friction of your titanium surfaces in the operating and installation conditions. You didn't mention temperature or other environmental factors but for many applications of titanium, they aren't kind and titanium does annoying galling things. Hopefully you can use a Ti grade and heat treatment that will allow it.

 

[lucky-guesser]

Regarding the angle, rule of thumb I have seen is that you angle should never be less than your mu. EX: mu of .20 should have at least 20deg taper.

 

[BrianE22]

Close, for self-locking, tan(theta) < mu.

 

[SwinnyGG]

If it's a non-locking taper angle, then the bolts will carry and absorb those loads. Loosen the bolts and it comes apart.

Not really - a self-releasing coupling is no different than a bolted flanged. Loads in the plane of the flange are carried by the friction between the faces. The bolts don't see them unless there's a huge amount of displacement between parts, which would mean the interface has already failed anyway.

 

[lucky-guesser]

Close, for self-locking, tan(theta) < mu.

Dang, you're right, haven't had to use if for a while.

 

[geesaman.d]

Not really - a self-releasing coupling is no different than a bolted flanged. Loads in the plane of the flange are carried by the friction between the faces. The bolts don't see them unless there's a huge amount of displacement between parts, which would mean the interface has already failed anyway.

Thanks, that's what I meant but you said it better.