Can CATRA Predict Rope Cutting Performance?

[Larrin]

There has been some question about whether CATRA is a useful test for actually predicting the edge retention of a knife. After all, it is a test performed by a robot, not a human, and the knife cuts cardstock with sand in it, which doesn't seem like a very realistic material. So I took three big datasets where people reported their own rope cutting experiments, and then compared those with what we would expect a CATRA test to tell us. https://knifesteelnerds.com/2019/02/11/can-catra-predict-rope-cutting-performance/

 

[marrenmiller]

Very interesting article, Larrin.

On the subject of edge thickness, I think it is worth noting that this factor may not be relevant for tests where cutting effort isn't directly measured. From what I can tell, Ankerson's testing is based around effort taken to cut through a medium (measured by applied force), whereas Pete's testing is just looking at the deterioration of the very apex. I would imagine that slicing paper isn't very dependent on edge thickness, so the edge thickness may not matter for Pete's tests as much. That's only a hunch, though. I'm not familiar with Pavol's tests, but maybe that has something to do with why his edge retention results don't correlate well with edge thickness.

While it wouldn't be entirely scientific, I'd be curious to see the correlation on Pete's data set with some of the obvious outliers removed. I think he adds the results of all of the trials to his Google Sheets page, even if it's obvious that something is wrong with a given trial. IIRC, that low D2 result was certainly not from real D2, and there were a lot of other knives (like S35VN from ZT) that performed like they had botched heat treats. If you asked him, there might be some results that are suspicious enough to warrant being thrown out.

 

[Larrin]

Removing outliers is always tempting, it would likely improve the correlation somewhat. But I end up finding fewer and fewer justifications for removing them so it's usually safer to keep everything in.

 

[pinnah]

In the article you write:
"Based on these kinds of questions I wanted to find out if the CATRA test is actually a useful prediction of slicing edge retention, or if it is just a controlled, scientific test that measures one specific thing but has little applicability when it comes to predicting knife performance, even when limited to slicing edge retention.

The Comparisons

The best way to test the CATRA test would be to test a range of knives with CATRA and then to test the same knives with rope cutting."

This begs the question of what is meant by "slicing edge retention". I think most users of knives will affirm that they can tell when a knife no longer slices the medium they are cutting: the ability to cut through fatty meat cleanly, the ability to make thin slices of a ripe tomato, the ability to make thin, controlled curls with wood.

But there are some confounding factors. Some will argue that "toothier" edges maintain good "slicing" performance in their particular medium (e.g. tomato skins) and others will argue that edge geometry plays an important role in effective working of their medium (i.e. many will argue that convex or chisel grinds are better for consistent curls).

My sense is that all of these rope cutting tests offer some great advantages for those who want numerical answers to performance questions. They can be repeated easily and reduce (but don't eliminate) many variables. In this sense, this result isn't at all surprising to me. I think the slicing use case of the CARTA test and the rope cutting tests are fairly close to each other. I would expect them to be well correlated.

But at the end of the day, we have to ask to what degree to these test related to real world cutting uses cases like meat/game processing or wood working? IMO, not much.

1/4 mile times are useful in evaluating a car, just as CARTA and rope cutting test are useful. But the 1/4 mile times won't predict which car will win on a road course and I'm not all convinced that either CARTA nor the rope cutting tests will predict which hunting knife will skin the most deer, which chef's knife will thinly slice the most tomatoes, nor which bushcraft knife will make the most feather-sticks.

 

[Larrin]

But at the end of the day, we have to ask to what degree to these test related to real world cutting uses cases like meat/game processing or wood working? IMO, not much.

I think that's a big assumption. I know that Phil Wilson has reported to see a good correlation between his rope cutting and deer hide, for example. Of the two, however, woodworking is the most likely to be different since that is usually a push cut. And in hardwoods edge stability could very likely be a significant factor.

 

[Twindog]

Those three "full data" charts at the very end of your article are awesome. People should look at those even if they can't invest the time to digest a long, complex article.

Lots of people think that a typical user can't tell the difference between steels, but that notion is clearly dispelled.

The other thing that those charts show is the complexity of factors that affect edge retention -- hardness, edge width, edge angle. I'd bet that if you could include microstructure of the steel it would get even more complex. Then add in the type of use a knife will see and the toughness of the steel, which affects edge stability, with even more complexity just within the toughness category alone.

Then add in the various heat treats people use. Then look at all the possible combinations of these factors.

It's dizzying.

 

[pinnah]

I think that's a big assumption. I know that Phil Wilson has reported to see a good correlation between his rope cutting and deer hide, for example.

There is more to dressing a deer than cutting through the hide. The knife needs to withstand hitting bone and there is a point at which the blade no longer cuts meat and tissue easily (keen edge), which may or may not the point at which the knife stops cutting through course hair and hide (toothy edge).

Of the two, however, woodworking is the most likely to be different since that is usually a push cut. And in hardwoods edge stability could very likely be a significant factor.

Exactly so. As I said, the question is, how do we define "slicing edge retention" in a way that spans different cutting use cases. I think that's pretty tricky and I'm not convinced that CARTA or rope cutting is sufficient.

Related note on the definitions... I think there is a qualitative difference between metrics about a blade's performance and the blade's physical status and I think things are more clear when we distinguish between them.

I understand "edge stability" to be the blade's ability to withstand rolling or micro-chipping from lateral forces.

I understand "edge retention" to be the blade's ability to maintain a 'V' profile and to resist wearing into a 'U' profile from abrasive cutting.

Neither CARTA nor rope cutting tests are direct "edge retention" metrics. Neither rely on direct measurement of the 'V' to 'U' wear of the apex. Instead, both are performance metrics that measure the blades ability to continue making certain kinds of cuts with various ways of recognizing the failure point. I've no doubt they are closely correlated to actual edge retention.

But I think in most real-world scenarios, people are more interested in "cutting performance" which is the ability to continue making cuts[1]. CARTA and rope cutting tests are good in that they get at performance but they are limited in that they are only illuminating a part of the performance equation.

EDITED TO ADD [1] - Roughly speaking, I think for many cutting tasks (excluding chopping) Cutting Performance = (A* edge retention) + (B* edge stability) + (C* edge polish/toothiness),
where A and B can vary depending on i) the medium being cut and ii) the type of cutting technique by users with decent technique. As noted, for wood working with hardwood, B will be higher compared to other cutting scenarios, like say, tomato slicing. Some mediums want either smooth or toothy edges.

 

[Wowbagger]

Of the two, however, woodworking is the most likely to be different since that is usually a push cut. And in hardwoods edge stability could very likely be a significant factor.

Keeping in mind the silica (mineral) contamination content of wood can vary wildly and even within the same shipment of the same species of wood.

I respect the CATRA for at least attempting to produce a material to be cut with a fairly constant abrasive element. I would think rope can be pretty variable.

I enjoy all this but have spent enough time cutting the materials that present a challenge to super steel alloys, that I cut every week, to know . . . FOR ME . . . what I like a lot, what will get me by and what I can expect to spend a whole lot of extra time touching up and sharpening. Good thing I enjoy sharpening; many of my favorite users fall within the last category

Add to that how the edge is for cutting more delicate and non challenging materials after all this challenging stuff S110V. . . slitting a plastic bag or trimming a fingernail or cutting soft rubber tubing.

coughM4, cough ( )
well it "cuts" the confusion down a bit.
There are some alloys I haven't tried yet so who knows . . . I would like to be surprised.

Coupled with edge geometry, perhaps the most important variable, heck Ankerson even got a Manix in S110V to cut stuff once he thinned out the edge by removing about two thirds of the behind the edge thickness. Anything is possible I suppose.