edge stability vs edge retention : 52100 vs 1.2% forged carbon (Zubeng)

[Cliff Stamp]

I decided to compare my MEUK in 52100, forged and heat treated by Caffrey to an edge hardness of 58 HRC, to the small Zubeng utility hunter in cross forged 1.2% carbon steel at 67 HRC. It should be obvious that the higher carbon steel and higher hardness Zubeng would have better extended slicing aggression, but the finer carbide size of the 52100 might give it an advantage in terms of edge stabiilty and thus it might do better for high sharpness push cutting.

For the edge stability check, both blades were sharpened at ten degrees per side, finishing at the UF Spyderco benchstone. Six rounds for each blade, with a complete sharpening for each round starting from x-coarse DMT. They were used to push cut through strips of 1/8" cardboard perpendicular with one cm of travel laterally through 10 cm. The sharpness was tested by push cutting thread, four masurements were taken through the one cm of edge exposed to the wear. Thus each final data point in total came from averaging (median) over 24 points. The results :

The results came out exactly as Landes indicates, the higher carbide steel has a lower blunting rate long term (b is lower) but initially blunts much faster (a is higher) due to edge stability issues. You can also see it visually even under 10X magnification as the Zubeng shows edge damage sooner, I am talking about small effects here, on the scale of 10 microns. I was surprised though at the extent of the difference, I would not have expected it to be so large. The graph on the left hand side displays some statistical information about the graphs, the residual ratios and the intersection which basically says how much more one knife can cut than the other.

I was also surprised at how fast they both dropped in sharpness early so I later did one test run where I finished with 5 passes per side on 0.5 micron chromium/aluminum oxide and it made an improvement on both, and especially so on the MEUK but I didn't have enough cardboard for a full run, I might do this later. It might also be that the abrasives used are not ideal for the Zubeng because it was visible at times under magnification that the coarse diamond hone tore pieces out of the edge so it isn't unreasonable to assume that something similar may be happening on a smaller scale at the finer grits. This should be repeated with similar grit waterstones.

Ok, now for the part where the Zubeng shows its dominance, an extended slicing comparison. The blades were again compared with the same edge angles which were very close. After one trial for example the edges were measured to be 0.010x0.030=9.5 degrees for the Meuk and 0.012x0.032=10.6 for the Zubeng. The grits were x-coarse DMT, fine DMT, medium Spyderco benchstone. Both could push cut newsprint over two inches from the hand (it was greater for the first finer run) and shaved readily. The sharpness was measured by slicing light cord under 45 grams of tension. 1/8" cardboard was sliced through 3 cm of edge. The results :

The MEUK was consistently sharper and also had a small advantage in edge retention, just at the level of significance. The cutting advantage was 1.12 (4) which meant it could cut 12 (4) % more material before reaching a similar state of blunting. This was very surprising. Note that the cutting advantage here is calculated very differently that I have done in the past. Before I used the two fitted curves but that method has several problems. Recently I switched to just using the data directly.

In order to bound the confidence intervals for the results I used a monte carlo simulation and generated 100 data sets with normally distributed data and reran the calculations for each one of them and averaged all of these results to produce estimates for the uncertainty in the cutting advantage points. This was actually done as DOS batch file which called a bunch of gawk scripts. Awk isn't a language you would use for math but I was just curious as to how bad the code would look. It is much easier in MATLAB.

The code and raw data is available if anyone wants to see it. I'll include the raw data in the webpage writeup eventually. I have to rewrite the model webpage to reflect the change in cutting advantage and I'll also likely put the relevant key code up there as well. Ref :

http://www.physics.mun.ca/~sstamp/knives/model.html

-Cliff

 

[gaj999]

Interesting indeed that they were so close on slicing. I'm curious about the rate of sharpness loss versus edge angle and carbide size for a given steel. Might there be some critical angle for a given carbide size where performance really begins to suffer due to carbide tear-out? Hmm, given that the actual width of the edge is so low in comparison to the carbide size, it may not matter at all, except for steels with small carbide size already. Boy, I hate it when I shoot holes in my comments before I'm even done typing.

Thanks, Cliff, interesting as always.

Gordon

 

[db]

Interesting given you have been saying edge stability mainly effects very polished and thin edges, and not really slicing like Phil W."s test.

 

[hardheart]

It should be obvious that the higher carbon steel and higher hardness Zubeng would have better extended slicing aggression, but the finer carbide size of the 52100 might give it an advantage in terms of edge stabiilty and thus it might do better for high sharpness push cutting.

The results came out exactly as Landes indicates, the higher carbide steel has a lower blunting rate long term (b is lower) but initially blunts much faster (a is higher) due to edge stability issues. You can also see it visually even under 10X magnification as the Zubeng shows edge damage sooner, I am talking about small effects here, on the scale of 10 microns. I was surprised though at the extent of the difference, I would not have expected it to be so large.

Ok, now for the part where the Zubeng shows its dominance, an extended slicing comparison.

The MEUK was consistently sharper and also had a small advantage in edge retention, just at the level of significance. The cutting advantage was 1.12 (4) which meant it could cut 12 (4) % more material before reaching a similar state of blunting. This was very surprising.

Doesn't seems so surprising. If the high carbide steel blunts faster, then the slicing edge finish would be lost sooner. The larger carbides and quicker edge damage would contribute to longer term slicing aggression once the edge began to break down after blunting. The fine grained and stable 52100 benefitted more from the edge finish initially, but then the steel properties worked against it in the long term, the edge smoothing over but not breaking down.

Unless I'm reading this wrong.

 

[db]

Hardheart that sounds reasonable, but I think it ignores the hardness difference. In fact I'd thought the hardness difference between the 2 would have made a pretty big difference between the 2. Then again maybe it did I can only guess with what I know of this test.

 

[hardheart]

The way I'm imagining it is that lower wear resistance/hardness of 52100 in comparison would cause the edge to smooth over time while slicing, reducing the ability. The Zubeng could have higher wear resistance, but because it is not as stable, the coarser (med Spyderco instead of UF Spydie used in push cutting) finish was chipped away because of low stability and larger carbides - allowing the edge to break down while subject to the forces created by slicing. The medium finish was degraded to a smoother finish (the initial advantage of the 52100), but then continued to degrade, into a more aggressive profile again (the long term advantage of the Zubeng). The 52100 blunted, and continued to wear without fracturing, which would have recreated a coarser finish to improve slicing.

The 'teeth' of the Zubeng would not wear away as quickly, but could break away, leaving other irregular 'teeth' to continue the slice. The 52100 'teeth' would wear away more quickly without replacement.

Again, I could be looking at this wrong.

 

[Cliff Stamp]

Interesting given you have been saying edge stability mainly effects very polished and thin edges, and not really slicing ...

Edge stability as defined by Landes, is a measure of ability to hold a high sharpness at acute angles. Johnston has the same defination for edge integrity. The above shows the 52100 blade has a very large advantage in the short term when push cutting at a high polish due to its higher edge stability. The graph also shows how the 1.2% carbon steel is starting to gain ground as the cutting progresses which would be expected because as the edge thickens in blunting it obviously becomes less sensitive to edge stability and more dependent on wear resistance. The graph implies in the long term the higher wear steel would overtake the 52100 blade but extrapolation is always risky. The second graph shows the same effect because more coarse finishes give inherently thicker edges and thus the 52100's edge stability is compensated by the enhanced wear resistance of the 1.2% carbon steel.

Doesn't seems so surprising. If the high carbide steel blunts faster, then the slicing edge finish would be lost sooner. The larger carbides and quicker edge damage would contribute to longer term slicing aggression once the edge began to break down after blunting. The fine grained and stable 52100 benefitted more from the edge finish initially, but then the steel properties worked against it in the long term, the edge smoothing over but not breaking down.

Yes that is exactly right, however I expected that the intersect of the curves would have been sooner. Now it may be the case that in long term the higher carbide steel would pull ahead significantly but if you note from the curve the slope is almost zero towards the end so both of them are starting to plateau which means they have both stabilized.

Basically every point you want to add towards the end basically doubles the cardboard you need so the stock demands rise quickly fast. I'll check this on carpet as this is more abrasive so doesn't require as much material cut. I intended to use a D2 blade right alongside of these two but simply didn't have the material. that should have showed clearly the effect of a low edge stability and high wear resistance as D2 pretty much benchmarks both of those in the extreme.

Interesting indeed that they were so close on slicing.

Yes, I expected it to be a blowout as the 52100 is both softer with a much lower carbide content.

I'm curious about the rate of sharpness loss versus edge angle and carbide size for a given steel.

Landes is exploring this in detail in his new book where he is being more quantitative in regards to these properties for the steels.

Might there be some critical angle for a given carbide size where performance really begins to suffer due to carbide tear-out?

Exactly right, I would expect if the edges were honed at 20 degrees per side the performance of the coarser steel would be greatly enhanced and the performance of the 52100 would not be and thus the relative performance would change significantly. You would also expect that if the angles were reduced it would favor 52100 (in the short term) and if the edges were made more coarse it would favor the 1.2% carbon steel in the long term. Thus for example x-coarse at 20 degrees would radically be expected to alter the performance.

-Cliff

 

[db]

Hard heart I cann't really guessas to how these knives dulled based on the results. You could be right on but I think without knowing more we are just peeing in the wind. It is too bad Cliff didn't look at the edges and watch how they were dulling, since that seems to be the purpose of this little test.

 

[Cliff Stamp]

It is too bad Cliff didn't look at the edges and watch how they were dulling...

Learn to read :

"The results came out exactly as Landes indicates, the higher carbide steel has a lower blunting rate long term (b is lower) but initially blunts much faster (a is higher) due to edge stability issues. You can also see it visually even under 10X magnification as the Zubeng shows edge damage sooner, I am talking about small effects here, on the scale of 10 microns."

As for the purpose, it was mainly to expand the use of the model I developed for cutting lifetimes and to develop specifically how the coefficients behave in regards to comparing edge retention vs edge stability as noted in the above. Secondly it was to work with the advancement in determining the cutting advantage. Thirdly it was to benchmark the performance of the Zubeng utility against a known steel.

-Cliff

 

[db]

Learn to read :

"The results came out exactly as Landes indicates, the higher carbide steel has a lower blunting rate long term (b is lower) but initially blunts much faster (a is higher) due to edge stability issues. You can also see it visually even under 10X magnification as the Zubeng shows edge damage sooner, I am talking about small effects here, on the scale of 10 microns."

And I think it is a real shame that you didn't simply observe for yourself how and why these 2 knives were dulling as they were doing the work. It could have been very informative, instead of trying to guess why based on the results. Instead you rely on another"s work, Landes, to explain your results.

As for the purpose, it was mainly to expand the use of the model I developed for cutting lifetimes and to develop specifically how the coefficients behave

Ok then if your main goal is to make graphs, models, use mathematics, and charts, then fine. More power to ya. I thought learning was the main purpose.

Thirdly it was to benchmark the performance of the Zubeng utility against a known steel.

I would think watching how it dulls as it is dulling would help very much in determining how it performs to the known steel. Maybe next time you could look and observe the edge as its dulling and maybe we all can learn something instead of guessing from it.

 

[Cliff Stamp]

And I think it is a real shame that you didn't simply observe for yourself how and why these 2 knives were dulling as they were doing the work.

I did observe them to verify the edge was degrading as noted. It would be interesting to see the edges under really high magnification (1000X) however the equipment to do so isn't inexpensive. You also need to expose the edges to reagents to actually bring out the grain/carbides which makes it impossible to do that as part of an edge retention trial.

Instead you rely on another"s work, Landes, to explain your results.

This has to be the most absurd statement I have heard. You don't simply use only the work you have done or verify every existing principle. Landes didn't invent metallurgy from scratch, he drew on an huge body of work done by others as would anyone trying to study anything. The exact same conclusions on steels were reached by Johnston who didn't even attempt to measure anything precisely. He just used knives and had other use them and based on what was known in the ASM texts came to the same conclusions. It is just much faster when you have better equipment because you can do less trials and each trial is much faster. Plus if you have enough money to buy the equipment you just get a grad student to do the experiment for you anyway. What is really amusing is just how similar their viewpoints are, I asked Landes for example for a steel optomized for high edge retention at high sharpness and he recommended 1.2562. Ask Johnston and he what does he say very commonly, O7.

I thought learning was the main purpose.

Developing a model to replicate behavior is fundamental to experimental work. There are individuals who actually do nothing but that and just use data gathered by experimentalists. Once this is done you can use said model to predict behavior and developing the model also allows development of the underlying theory. The graph provides a direct visual representation because most people could not picture the two curves from the two equations. Similar you also plot the residuals because the average value won't tell you if there is a systematic variance, and there is one in the above which is significant which I'll deal with later. Measuring the sharpness directly also obviously tells you the responce of the edge to the cutting which is obviously in itself informative and bottom line what is of most practical importance to most people.

-Cliff

 

[db]

I see, that clears up alot. Thankyou very much.

 

[yoda4561]

Cliff, do you plan on doing these edge stability tests with a more steels in the future, I'm just a little curious as to how infi will stack up with these new testing procedures.

 

[db]

I also see you again edit a post of yours after I reply to it. Very cute.
Btw around here people explain their own work and then if needed will use another's work to verify it if necessary.
You did not observe how the edges dulled, you just checked that they did get dul. Big difference. I think if you do more testing like this it would be very informative to observe how they dull, it could explain alot I think but I'm no expert like you are.

 

[Cliff Stamp]

You did not observe how the edges dulled, you just checked that they did get dul.

As I noted several times in the above, I did observe the edges periodically to verify how and to what extent they were dulling visually. I have done that for years and the details are noted in various reviews. I in fact did this back in the late ninties in discussions with Swaim about steeling. Yes, it would be of benefit to examine the edges on a microscopic scale which I would do if I had the equipment, however as noted you can't actually do this directly and observe the critical details of carbide/grain because you need to expose the steel to reagents to see them and this effects the edge.

As well the research as to how they dull by carbide fracture has already been done by Landes so there is no need to repeat that because it is published independent work. This is no different than when I use the principles of gravitation or how tension acts in a string to determine the force on the blade in the sharpness measurements. Or many of the statistical principles when I am doing the analysis. What I am specifically interested in is the physical characteristics, specifically the push cutting and slicing sharpness and how they can be modeled because as I noted previously, once this is done you can actually predict a geometry/steel for a given desired performance. There is of course no need to do what he has already done and has published.

The edits, as I have noted before are simply grammar related, if you are so concerned that some conspiracy is taking place then just quote the posts. Bladeforums will do it automatically for you. If observing the edgs is of such a high interest to there is also nothing stopping you from buying the equipment and adding to the work done by others.

Cliff, do you plan on doing these edge stability tests with a more steels in the future,

Yes, it is just really time consuming. There is a massive amount of variablity from one run to the next which means you have to do a lot of trials to be able to reach any definate conclusion. This really puzzeled me for a long time because I could not see it being due to the variation in angles, stock materials, or just quality of sharpening because they were too small.

Finally I discussed it with Landes and he noted that is to be expected because it is just the influence of the carbide breakdown and once I saw some micrographs of steels it became obvious because it is so random. The ironic thing was that the same thing is evident in the BUCK Catra data which I had seen a long time ago but since they plot the data differently I never connected the two.

I may automate this at some point, not the cutting, but measuring the sharpness because that is what takes the time. I have avoided doing it for a number of reasons. Mainly because I was curious as to what you could actually do by hand and I wanted the work to remain very simple to basically allow anyone who read it to duplicate it at basically no cost.

-Cliff

 

[db]

As I noted several times in the above, I did observe the edges periodically to verify how and to what extent they were dulling visually.

To bad you didn't share that info with us.

I have done that for years and the details are noted in various reviews.

Too bad you didn't see the need to note it in this one. All I said is it would be good info to have instead of guessing like Hardheart had to do in this case.

I in fact did this back in the late ninties in discussions with Swaim about steeling. Yes, it would be of benefit to examine the edges on a microscopic scale which I would do if I had the equipment,

I say even a 10x look would give alot of info that would help.

The edits, as I have noted before are simply grammar related,

Now Cliff you know that is a lie and a bad one at that. This isn't the first time you've changed your posts after I've replied to them.
http://www.bladeforums.com/forums/showpost.php?p=3810799&postcount=62
You've done more than just grammer and you know it. Just like before in this other post you have changed the content, and you know full well you did then and now. Pretty poor in my opinion.

if you are so concerned that some conspiracy is taking place then just quote the posts.

I don't think there is any conspiracy, it's just you that does it. I'm not going to repost every one of your long winded posts. This for sure is the second time you've changed your post after I have replied to them. Kind of peddy on your part. Wouldn't you think?

 

[Cliff Stamp]

To bad you didn't share that info with us.

I did, you just didn't read it. I even quoted it, you then quoted it, and you still didn't read it.

I'm not going to repost ...

Of course not, because if you did you could not make implications and spread misinformation. There is also no need to actually repost it if you are so concerned about bandwidth, you could simply save it, or even retrieve it from a cache. The very idea that I would edit a post in responce to a later post to avoid a direct responce is absurd, you really think I have a problem with making a point directly to someone.

-Cliff

 

[Larrin]

Now if only there was a source for O7. I've only found one source that looks a little promising, but have to find out if they'll even sell it. Also, I think it might be the same place that Chuck Bybee bought what he thought was O7, and then when they sent him the steel, the composition was suddenly quite different. If you look at the O7+ at alpha knife supply, it has way too much tungsten and chromium to be O7.

Edit: Of course, if I was rich, I could just order my own custom melt of O7, but since I'm not, it's not much of an option.

 

[Cliff Stamp]

Johnston has been looking for F2 and O7 bar stock for a long time. But it is the class rather than the exact steel which is of interest, a cold work steel which has a high hardness and a fine distribution of tungsten carbides to enhance wear resistance. Hitachi's Super Blue for example is that type of steel.

http://www.paragoncode.com/temp/YSS_HCC_spec.pdf

The white steels have higher edge stability and the blue steels have higher long term edge retention. It would be interesting to see Spyderco use some of these cutlery steels in a few of their knives.

-Cliff

 

[db]

Of course not, because if you did you could not make implications and spread misinformation.

No misinformation or implication on my part. The facts are just as I posted. Your lies are clearly shown.

very idea that I would edit a post in responce to a later post to avoid a direct responce is absurd,

Not absurd at all since that is just what you have done more than once now.

you really think I have a problem

No comment.