How Stress Risers Lead to Broken Blades

In my destruction tests, most of my blades broke around the first hole on my handle. Regardless of steel type. So what I’ve been doing lately is chamfering the holes to hopefully release some stress. I haven’t tested to destruction in a while, so I’m not sure this has helped much. But I’m guessing it will! They all broke, either in a flex test, or a hammer side impact test. All right on the hole. These are all drilled before heat treat of course, usually 1/4 inch.
 
Thanks for the article, Larrin!

Regarding chamferring holes:
I ran a couple quick simulations using 1/8"x3/4" flat with a 3/16" hole, under bending. I then looked at chamfers and hole shape changes. The ellipse is obviously impractical unless waterjet/CNC*, but an interesting point of reference. All chamfers are at 45°; increasing the angle didn't change much.

Round hole, unchamfered: 100%
Round hole, chamfered 0.01": 99.8%
Round hole, chamfered 0.02": 90.4%
Round hole, chamfered 0.03": 93.4%

Elliptical hole, 1/4" long, unchamfered: 91.3%
Elliptical hole, chamfered 0.02": 84.3%

Slotted hole, 1/4" long, unchamfered: 92.7%
Slotted hole, chamfered 0.02": 80.2%

Increasing the length of the slot did not futher reduce stress.

Two overlapping holes had similar stress reduction to a smooth slot, at 90% unchamfered and 80% chamfered.

*A waterjet hole might have worse stress concentrations due to surface finish.
 
Last edited:
Could you tell me what the percentages are again?
 
Could you tell me what the percentages are again?
Yeah, I wasn't particularly clear. That's maximum stress as a percent of the stress for a simple round hole. That is to say, stress levels around a plain round hole peak at 100%, while chamfers might reduce that stress to 90% as much.
 
Yeah, I wasn't particularly clear. That's maximum stress as a percent of the stress for a simple round hole. That is to say, stress levels around a plain round hole peak at 100%, while chamfers might reduce that stress to 90% as much.
Thanks for the numbers!
 
Stress risers, or "that uncommon case when removing material in a smart way can actually make a part stronger".

I've always thought of stress concentrations as being due to any sudden changes in geometry, as internal loads on a material don't just instantly distribute themselves evenly throughout a body when the cross section of that body changes, but rather distribute (somewhat) evenly over finite distances. The internal loads need gradual changes to redistribute evenly; for this reason, larger radii fillets do a better job gradually changing geometry, and therefore are preferable to sharp 90 degree corners.
 
Last edited:
Thanks for the article, Larrin!

Regarding chamferring holes:
I ran a couple quick simulations using 1/8"x3/4" flat with a 3/16" hole, under bending. I then looked at chamfers and hole shape changes. The ellipse is obviously impractical unless waterjet/CNC*, but an interesting point of reference. All chamfers are at 45°; increasing the angle didn't change much.

Round hole, unchamfered: 100%
Round hole, chamfered 0.01": 99.8%
Round hole, chamfered 0.02": 90.4%
Round hole, chamfered 0.03": 93.4%

Elliptical hole, 1/4" long, unchamfered: 91.3%
Elliptical hole, chamfered 0.02": 84.3%

Slotted hole, 1/4" long, unchamfered: 92.7%
Slotted hole, chamfered 0.02": 80.2%

Increasing the length of the slot did not futher reduce stress.

Two overlapping holes had similar stress reduction to a smooth slot, at 90% unchamfered and 80% chamfered.

*A waterjet hole might have worse stress concentrations due to surface finish.


I have been making slotted and chamfered holes in my tangs to give me more flexibility in pin placement. I accidentally created a tougher tang. Cool.
 
Something I am curious about is if the rate of stress change for a given change in geometry is different for materials of different stiffness. For example, do you require more gradual geometry changes for a more stiff material to arrive at the same stress concentration (Kt) factor? For an axial load testing specimen, would you require a longer tapered part for that stiff material as compared to a less stiff one? Or do I have that backwards? Generally Kt is always considered to be the same for a range of engineering materials, but that might just be a broad generalization.
 
Last edited:
Something I am curious about is if the rate of stress change for a given change in geometry is different for materials of different stiffness. For example, do you require more gradual geometry changes for a more stiff material to arrive at the same stress concentration (Kt) factor? For an axial load testing specimen, would you require a longer tapered part for that stiff material as compared to a less stiff one? Or do I have that backwards? Generally Kt is always considered to be the same for a range of engineering materials, but that might just be a broad generalization.
I don't know the answer to your question. However, brittle materials are more susceptible to fracture in the presence of a crack (fracture toughness). A crack is essentially just a very severe stress concentration.
 
I don't know the answer to your question. However, brittle materials are more susceptible to fracture in the presence of a crack (fracture toughness). A crack is essentially just a very severe stress concentration.
I found a paper discussing how steel hardness can effect stress concentrations, but it wasn't very clear how that was the case. Just seems odd that it's so much more substantially a geometry based problem than a materials based one.
 
Great info. I'm curious how much effect these factors have at the "micro" level. I've read that pitting/rusting acts as a stress riser in structural applications. Relevant/significant to knives? Would the final grinding grit cause a measurable difference in fracture initiation? Ex: Would a 60 grit finish fracture easier than 2000 grit?
 
Great info. I'm curious how much effect these factors have at the "micro" level. I've read that pitting/rusting acts as a stress riser in structural applications. Relevant/significant to knives? Would the final grinding grit cause a measurable difference in fracture initiation? Ex: Would a 60 grit finish fracture easier than 2000 grit?
Ohh they would absolutely have effects, and probably noticeable in the right circumstances.

For fatigue life calcs in engineering, if at all possible you include a factor for the surface finish to help make your estimates more accurate. It makes a real difference.

Likewise, rust damage on a blade tang could act as a stress raiser. Chopping with a blade that's got a rusted up tang could easily be a recipe for a snapped knife.
 
I found a paper discussing how steel hardness can effect stress concentrations, but it wasn't very clear how that was the case. Just seems odd that it's so much more substantially a geometry based problem than a materials based one.
I saw some references to higher Poisson's ratio equating to higher Kt. Seems reasonable to me. I'm not sure that it has a significant influence on this scale (in that steel Poisson's ratios don't vary much). One instructional presentation stated that more crack-resistantc materials are "less affected" and that the full Kt doesn't always need to be applied.

Reviewing some stress concentration literature - ideally, of course, notches (including round holes) are eliminated or feathered. Another option, however, is to add some smaller cuts leading up to the man notch. In some cases, this can be a series of smaller holes drilled to create an effective taper.

This means that lightening/epoxy holes, when properly placed, could reduce stress at the pin holes (and vice versa). My preliminary checks show that a couple % reduction can be gained with some strategic 3/32" holes.

I'd like to look at tang cutouts & holes in a bit more depth. Does anyone have any particular configurations they'd like checked? My thought is to do a baseline (just a rivet hole), and then a variety of comparitive examples with images to show max stress level & location (assuming the same load on all cases).
 
Maybe my favorite article so far (with very useful connections to other great articles as well!)! I’ve read it, at least, three times.

The more I read your articles, Larrin Larrin , the more I understand that knifemaking is no joke! It’s not just giving a guy a piece of steel and a drawing and ask for a knife exactly as in the draw. There’s too many variables to consider, and it takes just an error in just one step to end up with a piece of garbage.
 
Last edited:
Maybe my favorite article so far (with very useful connections to other great articles as well!)! I’ve read it, at least, three times.

The more I read your articles, Larrin Larrin , the more I understand that knifemaking is no joke! It’s not just giving a guy a piece of steel and a drawing and ask for a knife exactly as in the draw. There’s too many variables to consider, and it takes just an error in just one step to end up with a piece of garbage.
The engineering of any tool goes as deep as you want it to. Always more things to learn. :)
 
Back
Top