pic: Stress plots related to broken output shaft
 

Re: pic: Stress plots related to broken output shaft
 

Interesting plot and good to know the ramifications of a mis-fitted snap ring. I'm curious to know if the stress loading is the same when a properly-sized snap ring is in place?

Re: pic: Stress plots related to broken output shaft
 

By properly sized, I assume you mean cut all the way down into the hex? On hex shaft, you can get away with doing the next size larger snap ring, it just won't handle as much load. Somewhat off topic.

Say the groove did go all the way around, it would have similar effects. Stress will concentrate at the sharp cornered junction; the decrease in diameter is much less "damaging" than the sharp corner of the change. Varying diameters and radii (of that diameter change) will all create different reactions. It's really a simple guy to calculate if you know all the physical properties.

I found this book on google that has one of the tables you would use (this for torsion, for other types of loading a different chart is used).

Re: pic: Stress plots related to broken output shaft
 

Edit: misunderstood your question. Adam got it.

I would imagine that cutting the groove even deeper into the shaft would only exacerbate the problem. The stresses would have an even greater discontinuity to work around.

I should be clear that this shaft is experiencing torsional loading.

Re: pic: Stress plots related to broken output shaft
 

One option that comes to mind might be to use a tool that has radii in the bottom corners (or completely rounded on the bottom and cut your groove a few thousandths deeper) rather that sharp points. Sharp edges are stress concentrators, so the blends might remove some of the stress points.

Now, one might say a completely rounded groove bottom might also take away too much of the vertical side walls of the groove and defeat the primary purpose--holding a snap ring and supporting an axial load...but some snap rings have rounded cross-sections, so...

Re: pic: Stress plots related to broken output shaft
 

I really think the solution is to not use snap rings in the torsionally loaded section. We've used snap rings extensively the last few seasons, and have never needed to put one centered like that.

You're talking about a groove .030-.040 wide and only .020ish deep for most snap ring sizes, not much opportunity to radius the edges there.

Re: pic: Stress plots related to broken output shaft
 

Personally, I'm a fan of just using Delrin spacers in the high torsion areas of shafts to transmit any axial loads to whatever is being used for shaft retention (stepped shafts, tapped ends with washer, snap rings, etc) at the non-torsionally-loaded ends of the shaft.