Coupling taper angle û relating to locational accuracy 2

[reubencon]

I am trying to find quantify the locational accuracy of two parts coming together, i.e. a wheel onto a shaft. A coupling with no taper (slip coupling) has concentric precision but no rotational precision. While two parts butted together, have no concentric locating but have rotational precision. Some where in between these two extremes should be an optimal taper angle. Machine collets have a taper angle of 15 deg with a large L/D contact surface. The taper angle of 15 deg seems to be common choice, I am wondering what mechanical theories are behind this choice or did this angle come about from trial and error?

I appreciate any info or guidance on where to find a source.

R

 

[tygerdawg]

I've seen a variety of methods to do this, really don't know what's the best:
(1) my disk brake rotors on my VW are located onto the hub concentrically by a boss on the hub slipping into a hole on the rotor. The rotational location is taken care of by a M8 FHCS that has a countersink in the rotor. The torque is handled by four lugs through the rotor to the hub. Finally, parallelism is handled by flat face to flat face.
(2) many robot manufacturers have a similar arrangement for their tool flanges for hanging end effectors. Boss-thru-hole for concentricity, dowel pin into a hole for rotation, & flat-to-flat.
(3) I designed some equipment once that used some of those 3-lobed couplings you've probably seen some where (imagine a tapered arrangement, but with three or more lobes for torque carrying). Those things were manufactured with a certain "pitch" where the two parts mated along the length of the male & female coupling parts. Pretty cool method of locating round stuff accurately, but expensive to manufacture.

TygerDawg

 

[gaufridus]

The choice of taper angle on a machine tool collet etc is governed by whether it needs to be self locking or not. If the taper self locks it can also transmit the drive torque. A non locking collet requires some kind of drive dog to transmit torque.

The collet you mention is probably an 'International' sized collet of which there are a few depending in the size of the tool being used.

The International taper series does not self lock and requires a draw bolt to hold the male half in position to get the concentric accuracy. The non locking feature is useful in the autochange mechanism of CNC machine tools.

A Morse series taper(commonly used on taper shank drills/reamers in the UK), on the other hand has a much smaller taper angle and is self locking i.e. the female part grips the male part!!. A drift is required to separate them.

There are various other tapers used in the machine tool industry:
e.g. Jacobs (used to hold 3 jaw chucks to drill spindles) - self locking. R8 taper is used on Bridgeport turret mills - semi self locking - needs a draw bolt but is easy to unlock.

Machinerys Handbook lists most types and their taper details. All these tapers have a reference dimension so that axial location of the two mating can be specified (subject to wear of course.


I guess they all developed in an ad hoc sort of way. I suppose the same applies to taper dowels etc.

 

[GregLocock]

Bear in mind that your axial location is poorly controlled on a taper.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[JJPellin]

The fit mentioned in the third point by gaufridus is called a polygon fit. And he is correct; it is one of the most certain taper drive arrangements but is by far the most expensive. I work in an oil refinery. We use taper fits for critical couplings. They rate steep tapers in inches per foot. The most common is probably 3/4" per foot or about 3.6 degrees per side. This taper is easy to install with relatively little pull-up which gives good consistent axial positioning. But it has limitations. At a taper that steep, they have to retain the hub against popping off. They also usually have to have a drive key as a back-up in case it slips. We use a lot of 1/2" per foot or 2.4 degrees per side, also. It is still probably retained with a nut and has a key for a backup. If we want a positive drive without a key, we usually use a hydraulically mounted hub with a 1 degree per side or 1/2 degree per side angle. These hubs have to be expanded with hydraulic pressure and driven up the taper with a ram. This allows for very, very high interference with no key. But, especially for the 1/2 degree hubs, the pull-up is long and the axial positioning can be tricky. We mounted a 6 inch diameter hub on a motor shaft with 1/2 degree per side last week. The pull-up was about 0.800" giving us about 0.014" interference on the diameter. With over 0.002" interference per inch of diameter, this mounting can take extremely high torque, start/stop cycles and requires no retaining nut or key. A polygon fit would be better. It could drive the same torque will less interference and would be much easier to mount and dismount.

 

[ajack1]

Why would a taper give poor axial control?

Am I missing something as machining centres run using a taper and they run true within microns?

 

[GregLocock]

axial (lengthwise) location, not axial runout. sorry.


Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[reubencon]

Thanks for the good info about machine tool collets.
I mentioned the machine collets only as an example of a 15 deg taper. I am specifically interested in the theories used choose taper angles such as those listed in the Machinery Handbook.

R