Friction Forged Blades : CATRA tests

[severedthumbs]

sounds like real elite stuff, cant wait to get one. thanks for breaking it down, this stuff is going to kick ass!

 

[db]

Quote:

Originally Posted by Broos


... this thread is a good illustration why most mfg's do not release this kind of info.
Nonsense, the ones who are confident in their results release the information like Glesser. All such analysis is SUPPOSED to be subject to critical analysis.
Science doesn't proceed on faith and blind worship.

-Cliff

Broos I think your correct. How ironic is this statement and thread coming from Cliff who refuses to post his CV or even a list of his self reported dozen or so peer reviewed published papers?
Ref.
http://www.bladeforums.com/forums/showpost.php?p=4532586&postcount=39
I do think it is a little biased and premature to heavily criticize based on the limited info and use of FF, seen so far. Isn’t Cliff doing the same thing he often complains about? Does anyone have a link to info like this that has been released by Glesser/Spyderco?

 

[Cliff Stamp]

It looks interesting, I would be very curious as to the microstructure in detail. In particular the carbide volume.

-Cliff

 

[sputnick]

Regardless of cliffs desire to provide personal information that db has so courteously reminded us all of, his original point is valid.

[...] this is an example of really misleading and biased statistical analysis. [...]

-Cliff

Data is data. Testing procedures are relevant. How someone interprets those things can and is often misleading/biased. In this particular case there are too many variables to simply make the claims they have or to transpose their findings into the likely actual use of the blade. Ironwolf has provided an excellent example. If the claim is that the FF process is superior than a direct comparison should have been made with a same hardness standard process D2 blade. Then take all those blades and test them for bending/breaking strength the way fallkniven has done and some meaningful conclusions as to the properties of the knife as a whole could be made.

 

[Cliff Stamp]

If the claim is that the FF process is superior than a direct comparison should have been made with a same hardness standard process D2 blade.

This would not be possible here as regular D2 is limited in hardness below the FF sample. You could argue for a similar heat treatment in ingot D2 but you could also argue for dissimilar optomization, i.e., ingot D2 would have a more optimal heat treatment which was different than FF D2. If the austenite composions are different this would indeed be the case.

The point is valid however that for an actual scientific comparison the first thing you would do would be to compare against ingot and P/M D2 and even dendretic D2 for completeness. Of course the goal of such work in the above isn't scientific exploration it is simply promotion. Since S30V is a popular steel now you can see why they chose it. Since Crucible is one of the major players it also makes sense promotion wise to take out its highest promoted stainless steel edge retention wise.

If anyone has the raw data I will perform the required analysis on it and show the actual relative behavior of the steel, including confidence intervals on the coefficients and test of fit criteria for the models. This is an open offer to anyone who has such data on their knives and wants an unbiased comparison.

-Cliff

 

[sputnick]

Yup, yup and yup. Although I'm not familiar with the molecular and crystalline makeup of D2 in its various forms, I believe that is what they are claiming they have improved upon.

I don't know how likely it is that we could get our hands on the actual data points they got from testing. That would make coming up with unbiased conclusions too easy for us and since promotion is the goal, why would they want to do that...

The edge retention test seems to show that with identical geometry, finish, and ultimate edge width (verhoeven seemed to call this bur width) that the structure of the FFD2 is a significantly more aggressive slicer (at the start). Sorry but I can't swallow that one. I would be excited to see the corrections required to compare those curves accurately although its quite likely that the testing itself was biased. Of course if they also claim that the maximal sharpness of FFD2 is higher to begin with, I would argue that as the result of differences in hardness which of course brings us back to the intended use of the knife...

 

[db]

Regardless of cliffs desire to provide personal information that db has so courteously reminded us all of,

Sputnick what personal information would that be?

Data is data. Testing procedures are relevant. How someone interprets those things can and is often misleading/biased. In this particular case there are
too many variables to simply make the claims they have or to transpose their findings into the likely actual use of the blade.

With accusatory threads like this and posts like yours how likely do you think they will be forth coming with additional info?

 

[Broos]

Cliff, point me to similar testing done by Spyderco for ZDP-189. Especially similar cutting tests vs. other steels. I have not run across this yet.

This is probably the best test I have ever seen here - even has their testing methodology outlined. It's certainly a valid exercise to criticise and pick apart the data, but in my opinion they should be commended for releasing this, and I knew it was a matter of time before they would be accused of biasing the entire test - hence my comments questioning the benefit of releasing these tests.

I think using the exact same edge geometry and exact same sharpening procedure for each blade is the correct baseline to use to start this test. As the test methodology section explains, the REST tests were done before, during, and after the cutting testing.

The biggest problem is number of data points. More would be much more useful. Really this is not a test of edge retention, but a test of cutting ability using the same sharpening procedures. Doing multiple tests with different baseline sharpening procedures would be interesting.

Some think the test should start by making each knife cut the exact same amount of material before starting the test (this also would result in different REST values for the steels), this methodology would probably be a more pure edge retention test, but would have the knives start the test at varying levels of sharpness, which strikes me as not the best methodology. It would still be a good comparison with the twist of adding sharpening issues to get the steels to the same cutting level.

The amount of material per cut is greater for the FF blade thoughout the test. This test may not be a test of pure edge retention, but it does show that the D2FF blade, in this test, will outcut the other steels throughout the test, out to hundreds of cuts (with the same edge geometry and with the detailed sharpening done). I agree that this is a bit mystifying, but am still looking forward to the release of this steel, and am not ready to assume they are manipulating this data or just testing an untempered knife.

 

[sputnick]

db: I'm not sure what fault you find with my post. It was not accusatory in so much as the article was indeed meant as promotional material whose basis was test data. I'm not accusing them of lying, only of creatively (biased) interpreting results.

Comparing the approach used between the document from FF and say fallkniven test I think it is apparent what I mean.

Perhaps instead of inflammatory remarks, you have an opinion of your own to share in regards to how/why a company should present test data?


Edit: I applaud them for releasing this information in the first place. It is a step in the right direction. But as with all things commercial there is some reading between the lines that must be done.

 

[Cliff Stamp]

I don't know how likely it is that we could get our hands on the actual data points they got from testing. That would make coming up with unbiased conclusions too easy for us and since promotion is the goal, why would they want to do that...

I can digitize the numbers from the graph actually easily to within the noise as it is quite large and obviously a single run sample.

The edge retention test seems to show that with identical geometry, finish, and ultimate edge width (verhoeven seemed to call this bur width) that the structure of the FFD2 is a significantly more aggressive slicer (at the start). Sorry but I can't swallow that one.

Indeed, if anything it should be less.

Cliff, point me to similar testing done by Spyderco for ZDP-189. Especially similar cutting tests vs. other steels. I have not run across this yet.

You have not looked very hard then. I have referenced such work many times on the webpage on steels.

I think using the exact same edge geometry and exact same sharpening procedure for each blade is the correct baseline to use to start this test. As the test methodology section explains, the REST tests were done before, during, and after the cutting testing.

As I noted, you can easily pick the same method which will bias the results in favor of a steel and the sharpness tests were the wrong type for the cutting performed.

The biggest problem is number of data points.

No, I showed in the above how you can perform a proper analysis on such a sample. The biggest problem is that they do not model the results correctly and thus can not make supported conclusions.

The amount of material per cut is greater for the FF blade thoughout the test. This test may not be a test of pure edge retention, but it does show that the D2FF blade, in this test, will outcut the other steels throughout the test, out to hundreds of cuts (with the same edge geometry and with the detailed sharpening done).

This as noted is simply because it started off sharper. If you really don't see a problem with this then I can easily show you how you can make 420HC outcut S90V by similar methods.

-Cliff

 

[Broos]

As stated at least twice, the REST values were the same at the beginning of the test. REST = sharpness. Obviously you are not accepting REST as a defining criteria for sharpness. What is your definition?

And what specifically about their sharpening method favors D2FF, Cliff.

And I'll check your site for links to comparative cutting tests released by Spyderco.

 

[sputnick]

Broos: Of the devices used in the testing, one tests push cutting the other slicing. These are distinct characteristics.

 

[Broos]

They machine sharpened the blades identically and they were all the same level of sharpness, according to the CATRA sharpness tester. There is no better and still practical method for testing sharpness as far as I'm aware, and these testers are standard in the knife biz.
http://www.catra.org/pages/products/kniveslevel1/st.htm

What is the baseline for the start of the CATRA knife sharpness and life testing to begin? You guys are stating that the blades were not the same sharpness at the beginning of the test, yet they provide documentation that the blades were the same sharpness at the beginning of the test, using the industry standard sharpness tester. It seems to me you boys are pretty hard to please. So then what is your definition of sharp, and what methodology will get you where you can be happy?

So you either believe they followed their stated method, or not. I guess it really doesn't make any difference if you are convinced the tests were biased to begin with.

See you fella's later. I'll repeat and say thanks to this company for giving knife nerds some real testing to peruse.

 

[Broos]

You have not looked very hard then. I have referenced such work many times on the webpage on steels.
-Cliff

And after looking for 10 minutes I don't see any links on your site linking to Spyderco's steel/knife testing results. Thanks for the link, not. :thumbdn:

 

[sputnick]

The document mentions a third machine and flex test. Is that available anywhere?


Edit: Any of you guys read cashens "Revenge of the Bladesiths"? Funny stuff...

 

[Cliff Stamp]

As stated at least twice, the REST values were the same at the beginning of the test. REST = sharpness.

As I have noted in detail in the article I wrote on sharpness, slicing sharpness and push cutting sharpness are different. You don't use one to measure edge retention in the other. This would be like weighing a piece of steel to determine its toughness. Ref :

http://www.cutleryscience.com/articles/sharpness_review.html

And what specifically about their sharpening method favors D2FF, Cliff.

Choice of abrasives, time on the abrasive, pressure on the abrasive, many factors can easily favor chromium carbide vs vanadium carbide steels as the latter is much harder. For example, simply use an abrasive which is harder than chromium carbide but softer than vanadium carbide and the high vanadium steels will not have equal sharpness and have much lower edge retention.

You guys are stating that the blades were not the same sharpness at the beginning of the test ...

This should be obvious because if the sharpness and geometry were the same why would the length of material cut be different? There is a huge logic problem here which was not addressed in the pdf because as noted, this is a promotional package not an actual scientific investigation. Do you really think if Crucible released that data that they would argue for the same conclusions?

And after looking for 10 minutes I don't see any links on your site linking to Spyderco's steel/knife testing results.

The webpage on steels links to their CATRA results which specifically compares several steels on about a half a dozen links. It takes like ten seconds to find all the links with a simple in-line page search [ctrl-f] for CATRA.

I also have a page where Buck shows test data they published comparing different steels in full detail. I also showed how to correctly model that data using the equation I developed for blunting which is based on the physics of wear/deformation and chipping. Ref :

http://www.cutleryscience.com/reviews/model.html

As noted in the above, I have expanded the methods I currently used. That page is outdated, I will be updating it shortly with the current statistical methods I use.

I have also discussed in the past why simplistic methods such as used in the above comparison in the pdf are highly flawed and showed in detail why you can not mix push/slicing sharpness together such as in the above as the results will be meaningless.

In fact I did that explicitly for CPM15V many years ago showing how a steel could be low in one aspect but high in another and thus you should not state edge retention without qualifications. Landes confirmed this years later in his book on steels noting the difference between low/high carbide steels. Ref :

http://www.cutleryscience.com/articles/edge_stability_review.html

-Cliff

 

[Wayne G.]

I’ll check in here with my two-bits worth. I haven’t paid any attention to the charts and graphs because I’m a realist. I want to have a knife in my hand in order to determine its value in edge holding, sharpness and tests of strength. I’ve been testing the Friction Forged® blades for close to two months now and they get sharper and stay sharp longer than anything I’ve tested. I’ve cut rope, leather, whittled on wood and skinned a wild boar and cut it up for the freezer with a Friction Forged® knife. I’ve been cutting rope as a test of edge holding for over thirty years. In all those years I never had a knife go over 80 cuts. The first Friction Forged® blade I tested went 300 cuts. I was so impressed with the performance that I asked DiamondBlade to make my traditional hunting knife design in the friction forged steel.
I’d like to add that the four models of Friction Forged® knives in the initial offering are made as precision cutting instruments for hunters. They are thin in the body and thin at the edge, as they should be to perform as expected. I would not expect them to be used for pry bars or cutting through car bodies.
I’ve had questions about whether the edges would be hard to sharpen. The answer is yes without the proper stone. The answer is no with the India/Crystalon made by Norton. I’ll be doing my sharpening class at BLADE Show in Atlanta on Sunday June 10 at 12noon in room 103, which is close to the entrance of the show. I’ll have a deliberately dulled Friction Forged® knife to sharpen as a demonstration.

Wayne Goddard

 

[Cliff Stamp]

In all those years I never had a knife go over 80 cuts. The first Friction Forged® blade I tested went 300 cuts.

That is a fairly strong testament to thier performance. Have you used S90V at 60/61 HRC?

The primary carbides in friction forged D2 are just chroimum so I would not expect any difficulty in grinding. With the high hardness I would expect that it should be easy to get a crisp edge and that combined with a suitable very thin edge should make them easy to sharpen.

I would be interested in how they would compare to CPM S90V and 10V when Wilson does his rope testing.

-Cliff

 

[Wayne G.]

That is a fairly strong testament to thier performance. Have you used S90V at 60/61 HRC?

The primary carbides in friction forged D2 are just chroimum so I would not expect any difficulty in grinding. With the high hardness I would expect that it should be easy to get a crisp edge and that combined with a suitable very thin edge should make them easy to sharpen.

I would be interested in how they would compare to CPM S90V and 10V when Wilson does his rope testing.

-Cliff

CPM 10V and CPM 125V at 63Rc Phil Wilson blades are what did 80 cuts. Phil and I have shared a lot of test results over the years. Our mutual friend Maynard Meadows is the "in the field" tester and he also helps us with our rope cutting. Maynard got me started cutting rope in 1973. We haven't put the friction forged blades side to side with the CPM steels yet but that will come very soon. We do our cutting on the same rope and use a scale for a platform to determine when to stop cutting.

When you say the carbides are "just chromium" could it be that you are overlooking the vanadium in D2?

Wayne G.

 

[sputnick]

Does anyone know if any of the knives are available yet for purchase? I'd pull the trigger to get one in my hands and actually test out, then pass it around.