Maximizing Edge Retention – What CATRA Reveals about the Optimum Edge

[Mo2]

 

[Pomsbz]

I've been seeing a bunch of tests recently showing the effect of a 'thin' edge angle on edge retention. Has anyone done a study to find the optimum compromise with 'real world' cutting such as ham-fisted blokes like myself might be doing with EDC type usage? I assume it's thinner than factory but wider than entirely optimum for a controlled machine test.

 

[Mo2]

I've been seeing a bunch of tests recently showing the effect of a 'thin' edge angle on edge retention. Has anyone done a study to find the optimum compromise with 'real world' cutting such as ham-fisted blokes like myself might be doing with EDC type usage? I assume it's thinner than factory but wider than entirely optimum for a controlled machine test.

Depends on the steel type. Generally between 15 and 17 deg is good for edc. Some steel would work fine at say 12 deg but may chip or roll easier.

 

[Larrin]

Controlled studies looking at the effect of edge angle on how easily they are deformed or chipped are in relatively short supply. But even with a very good study you still have to determine what edge angle is appropriate for your knife, steel, heat treatment, and use.

 

[Larrin]

To summarize: Using a Spyderco knife in XHP at a 30° angle, he tested push cutting cardboard with a coarse finish and also a polished edge finish. He saw a 15% improvement with the polished edge. I believe that it was one test each so I'm not sure how much variation there would be if he were to do three tests per condition, for example. He also had text stating that he stopped cutting when it would no longer shave, I believe. Regardless, there is evidence out there to suggest that a polished edge is better for push cutting edge retention and this is in line with that.

 

[Twindog]

Controlled studies looking at the effect of edge angle on how easily they are deformed or chipped are in relatively short supply. But even with a very good study you still have to determine what edge angle is appropriate for your knife, steel, heat treatment, and use.

This is really the key point. What you want in a knife is the most acute geometry possible, provided that that geometry can be supported by the steel and the heat treat for the kind of use you need.

I think you start with good blade geometry and then adjust the edge angle/microbevel to be as effective as possible without suffering damage.

The advantage of the newest high-tech steels -- with a good heat treat -- is that they allow you to use a more acute geometry without suffering damage.

 

[chiral.grolim]

Really really liked this article

A couple things:
1) Can we all agree that "obtuse" means >90' and simply refer to angles <90' as "more" or "less" acute? I have this tick where I bang my head against a wall every time someone (myself included) uses "obtuse" to refer to an angle <90'.

2) If you calculate the mechanical advantage obtained by using different edge angles as a measure of cutting efficiency, one can expect the the 20' inclusive edge to penetrate ~2.8X more easily than the 50' inclusive edge when both are pristine. In this test after the first cut, the 20' edge is cutting ~3.5X more than the 50' edge (cut length), indicating that the fine 2-4 um apex has degraded sufficiently to alter that advantage further in favor of the thinner blade. To then see the graph of the performance equalizing after 60 cuts, really neat

3) My favorite part of the study was the end measurement of the apex diameters, so great that that was included One can imagine that the 50' edge reached 16-17um relatively quickly (the first 10-15 cuts?) because that thickness is achieved within a fairly short distance back from the original apex, ~17 um. In order for the 20' edge to achieve apex thickness 23um, it was worn back some 65 um! From modelling the regression of that first chart, you could probably determine just how much of the apex has worn away and the thickness of the apex after a given # of cuts. After 60 cuts and some 900 cards, the 20' edge lost some 65 um of edge while the 50' edge, had lost only ~17 um. It would have been neat to see how much of the 50' edge could be ground away if forced to cut at a higher, well, force then the set newtons, or how many cards it would complete before being ground down those last few microns to match the apex-thickness of the 20' edge.

Again, great study and thank you for presenting it

Here is a series of graphs I made a long time ago presenting edge angle (in dps) according to "strength" (stiffness) and "efficiency" (mechanical advantage), as well as what different edge profiles might look like:

 

[chiral.grolim]

Follow-up questions:

Larrin Am I reading that correct, that there were only 2 blades tested at 20', and both were sharpened 120-grit? I must admit, I am disappointed that the data on 20' is based off of n=2. Perhaps you do not know this, but I am guessing that those 20' blades were run near the very end just as a quick test to optimize performance on the CATRA machine, as they utilize the both the lowest edge-angle and the coarsest grit?

It would be nice to have the graphed-out performance of the 27' edges to compare against the 34' & 50' edges, as that seems to have been the intent of the study... or to have that final table in excel or some other format to quickly/easily transfer into graphing software so as to run the analysis with the limited information there given (obviously losing all of the data points of the curve of cutting performance).

Finally, did the authors come up with a mathematical description of the relationship between edge-angle and CATRA performance similar to what we have for mechanical advantage and edge-strength?

 

[Larrin]

Follow-up questions:

Larrin Am I reading that correct, that there were only 2 blades tested at 20', and both were sharpened 120-grit? I must admit, I am disappointed that the data on 20' is based off of n=2. Perhaps you do not know this, but I am guessing that those 20' blades were run near the very end just as a quick test to optimize performance on the CATRA machine, as they utilize the both the lowest edge-angle and the coarsest grit?

There were only two tested at 20° and those were at 120 grit. They were indeed run at the end.

It would be nice to have the graphed-out performance of the 27' edges to compare against the 34' & 50' edges, as that seems to have been the intent of the study... or to have that final table in excel or some other format to quickly/easily transfer into graphing software so as to run the analysis with the limited information there given (obviously losing all of the data points of the curve of cutting performance).

If you email me I can send you the final table in Excel.

Finally, did the authors come up with a mathematical description of the relationship between edge-angle and CATRA performance similar to what we have for mechanical advantage and edge-strength?

I created a regression equation as part of creating the table with the effect of different parameters on R2. I don't think I saved it though.

 

[chiral.grolim]

There were only two tested at 20° and those were at 120 grit. They were indeed run at the end.

If you email me I can send you the final table in Excel.

I created a regression equation as part of creating the table with the effect of different parameters on R2. I don't think I saved it though.

Email sent via your webpage!

 

[chiral.grolim]

Thank you for that table
Here is a set of a few charts from the data. I grouped the runs according to edge-angle and thickness.
I do not know how thick each card is and such information is important to understanding how edge-thickness figures into a test - if the edge is sufficiently tall that it cleaves the card prior to reaching full edge-thickness, then it does not require energy to push apart the card to the full BET. Is it possible that CATRA uses 0.010" card-stock? It looks even thinner than that in the video...
From the images in the article and the schematic I presented earlier, one can easily understand why it would require more effort per card for a wider edge angle. At the set force (50N = ~11.2 lbs) and feed-rate (50 mm/s), the wider edge cannot cleave as many cards as the thinner edges. Indeed, at a given edge-depth of the 20' blade, the 50' blade is ~2.65X thicker. If you think of "total cards cut" as a measure of productivity, it is clear that at a this low level of force and set rate, a worker could cut through 5-10X more material over that 15 minutes using a thinner blade, or he would require a LOT more effort (force) to accomplish the task with the thicker blade!

As Larrin stated, cryo/RC/PM had almost no noticeable impact as you can see from the groupings at each angle/thickness. One might even speculate that, since each edge is hand-ground, the level of variation may account for the lack of difference between 32', 33', and 34'. And again, edge-thickness only matters if the card is sufficiently thick to require the full edge to penetrate prior to cleavage
Larrin also noted how performance dropped precipitously after the first cut and then stabilized (image included in a post above). I was only able to chart the 1st and 3rd cut-length values from the excel chart, but i think it is important to note what is seen in this:
While 20' gives the highest number of total cards cut, as well as the highest cut-length for both cut 1 and cut 3, the drop in performance is substantial from cut1 to cut3, cutting at only ~60% of what it had accomplished in the first cut, whereas the 27-34' edges averaged ~70%. It is important to put that into perspective - the 20' blades appear to have lost more edge than the other blades BUT they also cut more cards.
If you divide the loss in performance from cut1 to cut3 by the total mm cards cut after 3, you come to the final 2 charts below, the second simply excludes the 50+' edges. In those first 3 cuts, the 20' and 27' edges are losing ~0.25% per mm of card cut (the latter shows the high scatter of increasing n) while the 32-34' edges are losing closer to 0.35% per mm, and the fat 50+' edges are near to full 2% loss per mm cut on their way to stabilizing.

After looking at this data, I cannot help thinking that this study has less to do with "edge retention" than it has to do with basic cutting efficiency as an effect of mechanical advantage. To me, "edge retention" describes the rate at which the edge itself abrades/cracks away is use. One would expect the rate of wear on the edges to be the same per mm cut if not for the confounding factor of mechanical advantage allowing the thinners edges to more easily remove cards from the action. Am i mistaken in thinking that this is what is happening? As the machine slides the cards over the edge of the blade and downward, the fatter edge, requiring more effort to penetrate and cleave each card, spends more time being dragged through each card (or rather each card is dragged across the edge) and so cannot cut as many cards as the thinner edge from the very beginning. It would not surprise me if, should this test be repeated with a pristine tungsten carbide edge at 50' and the same steel edge at 20', the thinner edge would "outperform" the thicker edge AND, when the edges are examined afterward, the 20' edge would have a significantly wider apex-diameter than the fat carbide edge. The fatter edge simply could not cut through as many cards per pass, as many cards per minute, as the thinner edge due to simple mechanical advantage.

What do you think? This test actually tested "edge-retention" in the different edge-angles, or was testing something else entirely, or that it doesn't matter since the end result is more cutting accomplished with less effort?

 

[FortyTwoBlades]

Though it’s not the way the industry has been moving, there have been people out there preaching the merits of acute edge angles and easy-to-sharpen steels for a while. My personal experience has also been leading me to the same conclusion, although without scientific rigour.

Thanks for posting this, Larrin.

*Raises hand*

Thanks for posting this, Larrin.

As far as remarks regarding durability for "regular folks" go, that's the challenge. A tool that's 100% maximally efficient will fail the instant it experiences strain outside of its precisely-defined design parameters. In most cases, efficiency and resiliency are opposing factors. So you start from a place of an ideal form and walk yourself back to the point where enough "insurance factor" for the intended context of use and anticipated user is achieved.

 

[hardheart]

The 20 degree edges were more of a spur of the moment inclusion, not planned.
For years, the idea of less acute edges providing longer cut life (more media cut for the same effort) was accepted by many. Under the circumstances of the test, this was not demonstrated, and testing that was the reason for some of the variables. Really, that concept suggested that increased edge strength led to more cuts, even though thinner edges would be more efficient on the beginning. This test seems to show something maybe more intuitive, the edge that begins more efficient remains more efficient.

And I'm honestly not sure how to separate the common concept and testing of edge retention from efficiency. It's mostly tested or gauged with the idea of "how much can I cut until it's too hard/I'm too tired."

As for adding some variations in the angle the blade is presented to the media, I'm not sure what it would do other than reduce the total card cut. None of the cuts are percussive, the force isn't very large, the media is thin, flexible, and soft, and even the thin edges are too thick to be deformed by a stack of paper. The idea that the results might not track the same needs a theory on why less optimal angles against paper would cause that. I personally can't think of one, though that doesn't mean there isn't a good one.

 

[Twindog]

The loss of cutting ability can be caused by damage to the apex -- either rolling, denting or chipping or corrosion -- and dulling through abrasion.

The CATRA cards are unlikely to cause damage, so dulling through abrasion would be the dominant cause of decreased cutting ability. It's difficult for me to see that the edge angle would make much difference.

But in the real world, edge damage can be the dominant cause of lost cutting ability. Edge angle can have a major effect on vulnerability to edge damage. Edge angle and shoulder width can also have a huge effect on cutting material that is difficult to split, such as thick, tough cardboard.

Knife performance seems so simple, but in reality there are countless variables and combinations of variables that come into play: steel alloy, cutting task, edge angle, edge width, heat treat, corrosion. Each of those variables come in many different flavors. Think about all the steel alloys available. Then how many ways each can be heat treated. Then all the ways the edge can be used. All the different edge angles, not just angle acuity but also type (V edge, convex edge, hybrid, etc.). Then think of all the combinations of these variables.

It seems simple to test edges by controlling all the variables, but there are just too many variables to control.

This combination of variables will make a huge difference in the outcome and make generalities difficult.