Analysis of Recent M390 CATRA Testing

[Larrin]

The minor controversy with hardness and edge retention of different M390 knives led to two different small CATRA studies on M390 knives, and one in S30V. I went through both of them including what was measured and a couple small surprises. https://knifesteelnerds.com/2019/10/07/catra-tests-of-m390-knives/

 

[Mikel_24]

I read through it and well... focused on the summary, which basically states that the results were well within the expectations, although didn't match the initial estimations due to the different stroke length used in the test (25mm VS the standar 40mm...). Good to know!

I truly appreciate those with the interest, skills and time to carry out these tests and summarize them for us.

Only think I have to say is that I concurr when they state that diamond sharpening media provides a coarser edge... I see that every day I need to slice and dice some ripe tomatoes in the kitchen... If I hone the edge with the ceramic rod... no dice. The knife might very well shave, but doesn't do much on the tomato skin. However, a couple swipes on the diamond rod and... it slices like a laser.

Mikel

 

[Larrin]

I read through it and well... focused on the summary, which basically states that the results were well within the expectations, although didn't match the initial estimations due to the different stroke length used in the test (25mm VS the standar 40mm...). Good to know!

I truly appreciate those with the interest, skills and time to carry out these tests and summarize them for us.

Only think I have to say is that I concurr when they state that diamond sharpening media provides a coarser edge... I see that every day I need to slice and dice some ripe tomatoes in the kitchen... If I hone the edge with the ceramic rod... no dice. The knife might very well shave, but doesn't do much on the tomato skin. However, a couple swipes on the diamond rod and... it slices like a laser.

Mikel

Thanks!

 

[chiral.grolim]

The minor controversy with hardness and edge retention of different M390 knives led to two different small CATRA studies on M390 knives, and one in S30V. I went through both of them including what was measured and a couple small surprises. https://knifesteelnerds.com/2019/10/07/catra-tests-of-m390-knives/

Your comment on 'ICP' vs sharpness in interesting, a warning that initial apex geometry is annihilated almost immediately by the silicon-impregnated card. This is one of those points which has always fascinated me - the concept that if we assert "initial sharpness" (defined as the inverse of the amount of effort/force required to penetrate into the medium to some depth X within the edge-bevel width of the blade) to be 100%, we might lose 90% of that in the midst of completing that first cut, and everything that we measure thereafter is the gradual loss of the remaining 10%, our "working edge".

Weird that they bothered with the 25 vs 40 conversion factor... are you typing that the "measured" 'TCC' as reported is really a converted value using this imputed constant?

It would be helpful to know the BET on the samples IF the cutting depth proceeds beyond the edge-shoulder for any of the blades tested. If, however, each card is cut and falls away prior to the bevel-shoulder depth, it is irrelevant, no?

I didn't read much about the "small controversy" you mention, but the tests do seem to show (albeit with VERY limited sample size), that Lionsteel, by leaving their M390 soft in this particular knife, managed to provide customers with lower performance than likely would have been achieved. Did they claim it was to increase "toughness"?

Thank you for the write-up.

 

[Larrin]

Your comment on 'ICP' vs sharpness in interesting, a warning that initial apex geometry is annihilated almost immediately by the silicon-impregnated card. This is one of those points which has always fascinated me - the concept that if we assert "initial sharpness" (defined as the inverse of the amount of effort/force required to penetrate into the medium to some depth X within the edge-bevel width of the blade) to be 100%, we might lose 90% of that in the midst of completing that first cut, and everything that we measure thereafter is the gradual loss of the remaining 10%, our "working edge".

Well it depends on what 90% you are referring to. 90% of total sharpness no. 90% of "hair whittling ability" yes.

Weird that they bothered with the 25 vs 40 conversion factor... are you typing that the "measured" 'TCC' as reported is really a converted value using this imputed constant?

Yes the number that CATRA reported is a pre-converted value. The raw numbers were not given.

It would be helpful to know the BET on the samples IF the cutting depth proceeds beyond the edge-shoulder for any of the blades tested. If, however, each card is cut and falls away prior to the bevel-shoulder depth, it is irrelevant, no?

Behind the edge thickness was found to be relevant in this study: https://knifesteelnerds.com/2018/06/18/maximizing-edge-retention/

I didn't read much about the "small controversy" you mention, but the tests do seem to show (albeit with VERY limited sample size), that Lionsteel, by leaving their M390 soft in this particular knife, managed to provide customers with lower performance than likely would have been achieved. Did they claim it was to increase "toughness"?

I didn't pay enough attention to the controversy to answer your questions either. Higher hardness does indeed lead to better edge retention. It depends on what the manufacturer and customer are looking for. I wouldn't necessarily call 59 Rc M390 low performance or even suboptimal. But I have my own preferences.

Thank you for the write-up.

You're welcome.

 

[chiral.grolim]

Well it depends on what 90% you are referring to. 90% of total sharpness no. 90% of "hair whittling ability" yes.

Yes the number that CATRA reported is a pre-converted value. The raw numbers were not given.

Behind the edge thickness was found to be relevant in this study: https://knifesteelnerds.com/2018/06/18/maximizing-edge-retention/

I didn't pay enough attention to the controversy to answer your questions either. Higher hardness does indeed lead to better edge retention. It depends on what the manufacturer and customer are looking for. I wouldn't necessarily call 59 Rc M390 low performance or even suboptimal. But I have my own preferences.

You're welcome.

Found it again - their card stock thickness is supposed to be 0.31± 0.02 mm.
I remember that other article and seem to recall commenting on it that because the 0.010" BET blades cut deeper than the height of the bevel in the 50' inclusive edges whose bevel shoulders entered the cut at ~0.27 mm depth for the thinner blade, not accounting for minuscule (but important, relative corresponding increase in apex width) loss of edge-height due to degradation during the cutting. But I cannot quite figure out why a difference in thickness for the 34' inclusive edges would appear since each card cut should fall away just before the bevel shoulder is reached on the 0.010" edges (cutting depth 0.31mm using angle of 34' inclusive requires 0.2mm wedging = 0.008"). I suppose that it is because reality isn't quite so clean

Your thoughts? I ask here (instead of in the previous thread which could probably be dug up again) because the spread between knives here appears to fall within the range of variation between disparate BETs, but as long as these knives are thicker than 0.008" at the shoulder (and I cannot imagine that NOT being the case), each card should be cut well within the edge bevel of each knife, none of which are sharpened wider than 34' inclusive. My brain says, "each card is cut within the edge bevel, so thickness beyond that bevel cannot be relevant to the test." Am I wrong to think that? While the previous test shows a difference, does the difference reach statistical significance by a basic t-test? If it does not, isn't the apparent difference just noise?

Regarding "low performance", I am curious what aspect you think that customer and manufacturer would desire in producing an M390 blade with noticeably lower edge-retention despite a slight advantage in edge-angle (the dominant factor in CATRA testing)?

ETA: Regarding "90% of total sharpness", I am not sure what that would even refer to - the cummulative amount of a given material that can be cut before the edge is considered "dull"? Of course not. But perhaps you can remind us - the CATRA device cuts until a certain "length" (# of cards) are cut or it requires >50N to cut through a card, correct? Well, if we could measure the force required to cut the first card, and we have measurements for subsequent passes, we can see how much "sharpness" (apex acuity) we lose in those first few strokes. In the first 3 'strokes', each knife cuts >100, but by the next 3 stokes how many are cut? After 60 strokes, only some 600-800 are cut (assuming those numbers are raw rather than converted). The chart from CATRA leaves the first stroke out of the graph of declining performance. What percentage of sharpness (i.e. ease of cutting, mechanical advantage) is lost in that first stroke?

 

[Larrin]

Found it again - their card stock thickness is supposed to be 0.31± 0.02 mm.
I remember that other article and seem to recall commenting on it that because the 0.010" BET blades cut deeper than the height of the bevel in the 50' inclusive edges whose bevel shoulders entered the cut at ~0.27 mm depth for the thinner blade, not accounting for minuscule (but important, relative corresponding increase in apex width) loss of edge-height due to degradation during the cutting. But I cannot quite figure out why a difference in thickness for the 34' inclusive edges would appear since each card cut should fall away just before the bevel shoulder is reached on the 0.010" edges (cutting depth 0.31mm using angle of 34' inclusive requires 0.2mm wedging = 0.008"). I suppose that it is because reality isn't quite so clean

Your thoughts? I ask here (instead of in the previous thread which could probably be dug up again) because the spread between knives here appears to fall within the range of variation between disparate BETs, but as long as these knives are thicker than 0.008" at the shoulder (and I cannot imagine that NOT being the case), each card should be cut well within the edge bevel of each knife, none of which are sharpened wider than 34' inclusive. My brain says, "each card is cut within the edge bevel, so thickness beyond that bevel cannot be relevant to the test." Am I wrong to think that? While the previous test shows a difference, does the difference reach statistical significance by a basic t-test? If it does not, isn't the apparent difference just noise?

I’ve only seen a handful of CATRA tests in person so it’s hard for me to comment. If I had a machine and many test specimens maybe we could figure it out. The result in the 154cm study was an average of many tests; it is statistically significant.

Regarding "low performance", I am curious what aspect you think that customer and manufacturer would desire in producing an M390 blade with noticeably lower edge-retention despite a slight advantage in edge-angle (the dominant factor in CATRA testing)?

I’m referring to all performance factors not just edge retention. That includes things like grindability for the manufacturer and toughness. Edge retention isn’t the only desirable attribute.

ETA: Regarding "90% of total sharpness", I am not sure what that would even refer to - the cummulative amount of a given material that can be cut before the edge is considered "dull"? Of course not.

Exactly. I don’t know what you mean by 90% of sharpness lost.

But perhaps you can remind us - the CATRA device cuts until a certain "length" (# of cards) are cut or it requires >50N to cut through a card, correct?

No. The 50N is fixed. It cuts with 60 strokes and then the test is complete. The knife could still be very sharp or it could be cutting nothing at the end.

Well, if we could measure the force required to cut the first card, and we have measurements for subsequent passes, we can see how much "sharpness" (apex acuity) we lose in those first few strokes. In the first 3 'strokes', each knife cuts >100, but by the next 3 stokes how many are cut? After 60 strokes, only some 600-800 are cut (assuming those numbers are raw rather than converted). The chart from CATRA leaves the first stroke out of the graph of declining performance. What percentage of sharpness (i.e. ease of cutting, mechanical advantage) is lost in that first stroke?

I don’t know how much is lost in the first half-stroke. With 33-34 degree edges the measure of the first stroke (forward and backward) can vary from about 25mm to 45mm. I have several representative CATRA curves here where you can try to infer what you are looking for: https://knifesteelnerds.com/2018/11/26/steel-edge-retention2/ But with current information inferring is as good as we can do.

 

[chiral.grolim]

I’ve only seen a handful of CATRA tests in person so it’s hard for me to comment. If I had a machine and many test specimens maybe we could figure it out. The result in the 154cm study was an average of many tests; it is statistically significant.

I forgot that I asked you to send me the data, I created this chart from it:

I ran 1-way ANOVA for all the groups as well as T-tests comparing just 34' at 0.010" to 0.020" and comparing 50' at 0.010" to 0.020". The P-value was not significant for the 34' blades (which, it so happens, are the ones where each card-cut is completed within the edge-bevel).

I’m referring to all performance factors not just edge retention. That includes things like grindability for the manufacturer and toughness. Edge retention isn’t the only desirable attribute.

In an M390 blade? I get grindability from the manufacturer stand-point - makes it easier/cheaper, but noticeably vs a couple points higher? It won't affect sharpening for the end-user. Higher toughness? Again, hard to believe. While edge-retention isn't the "only" desirable attribute in knives, I would frankly require evidence from someone that the lower-hardness result gives any advantage to anyone at all other than competitors.

Exactly. I don’t know what you mean by 90% of sharpness lost.

I told you what I mean multiple times. My apologies, I will try again below.

No. The 50N is fixed. It cuts with 60 strokes and then the test is complete. The knife could still be very sharp or it could be cutting nothing at the end.

I don’t know how much is lost in the first half-stroke. With 33-34 degree edges the measure of the first stroke (forward and backward) can vary from about 25mm to 45mm. I have several representative CATRA curves here where you can try to infer what you are looking for: https://knifesteelnerds.com/2018/11/26/steel-edge-retention2/ But with current information inferring is as good as we can do.

I added emphasis here, and thank you for that. I added the emphasis because of the comparison to other sharpness testing I have seen measuring the force required to complete a cut rather than measuring the amount of cutting achieved with a set maximum force, my brain-fart. Watching the video linked in your link is helpful too, seeing the card-stack ride up the blade-edge during the stroke with only 50N applied downward force If cutting nothing at the end, 50N is insufficient force to complete a cut... really a 'mash' at that point!

"Sharpness" = mechanical advantage, i.e. the ease (or lack of force/effort) with which a cut is completed. CATRA, rather than measuring the force required to complete a cut measures the amount of media cut on a given stroke at a given force. While it is difficult to determine "sharpness" from data that does not measure force, the ratio of media cut in the first stroke to the next or the third shows the loss of "sharpness" i.e. edge-retention represented in the bottom-right quadrant of my image above. While the edge is theoretically at "relative 100%" sharpness BEFORE the first cut, we can't measure force OR amount of media cut at a given force before making a first cut However, if edges are sharpened identically, we might infer identical "100%" sharpness levels in advance of the cut. Anyway, performance is down to <80% and in some of the cases <40% of the first-cut level by the end of the 3rd cut.

In your image below, from the link provided, by the 4th or 5th cut the exponential decline in performance begins to level out. By the 10th cut in the steel with lowest wear-resistance, the edge is performing at near 10% of where it was in the first cut. That is an example of what I am referring to in regard to the "working edge" being at only ~10% the initial sharpness/performance - doing 9x less work or requiring 9x the effort to do the same work. It is wonderful to see materials fighting this, however, providing us a sharper edge through more work

 

[MolokaiRider]

Thanks again for sharing this work.

 

[DeadboxHero]

Higher hardness eats more abrasives and can take extra steps to the HT process to achieve. It also requires more R&D to make it stable at higher hardness since in reality it's the microstructure that is important not just a hardness number but how we get that number with the structure we make.

A lower hardness will extend the plastic curve/ductile region but at consequence to the yield Strength, resilience and wear resistance. Don't kid yourself, it is more cost effective to mass produce a run of 500,1000, 10,000 knives with lower hardness and less steps, play it safe, also play it safe for bad knife users.

People forget that it's not just their special knife being made but it's a business and these are being made at volume and must generate profit to pay for cost and overhead.

It's akin to whining that a fast food burger isn't a a Michelin 3 star meal.

With the CATRA the load and cutting volume is fixed so we can just see how much volume is cut thanks the to the robotic control. They have standards for the ICP So a knife has to pass in order to continue with the testing so it actually makes it through the 60 cycles, a cycle by the way is one back and forth cut.

It's not the sharpeness you're thinking of, because of the back and forth saw motion and that we are measuring cards cut with fixed load and fixed amount of cuts. So it will have a toothy edge bias. A hair whittling edge will not do as good as a toothy saw type edge on CATRA testing. technically I feel sharpeness is the finest apex shape possible. Doing the least amount of work for a given task can vary with edge finish and how the user is using and for what.

I forgot that I asked you to send me the data, I created this chart from it:

I ran 1-way ANOVA for all the groups as well as T-tests comparing just 34' at 0.010" to 0.020" and comparing 50' at 0.010" to 0.020". The P-value was not significant for the 34' blades (which, it so happens, are the ones where each card-cut is completed within the edge-bevel).


In an M390 blade? I get grindability from the manufacturer stand-point - makes it easier/cheaper, but noticeably vs a couple points higher? It won't affect sharpening for the end-user. Higher toughness? Again, hard to believe. While edge-retention isn't the "only" desirable attribute in knives, I would frankly require evidence from someone that the lower-hardness result gives any advantage to anyone at all other than competitors.

I told you what I mean multiple times. My apologies, I will try again below.


I added emphasis here, and thank you for that. I added the emphasis because of the comparison to other sharpness testing I have seen measuring the force required to complete a cut rather than measuring the amount of cutting achieved with a set maximum force, my brain-fart. Watching the video linked in your link is helpful too, seeing the card-stack ride up the blade-edge during the stroke with only 50N applied downward force If cutting nothing at the end, 50N is insufficient force to complete a cut... really a 'mash' at that point!

"Sharpness" = mechanical advantage, i.e. the ease (or lack of force/effort) with which a cut is completed. CATRA, rather than measuring the force required to complete a cut measures the amount of media cut on a given stroke at a given force. While it is difficult to determine "sharpness" from data that does not measure force, the ratio of media cut in the first stroke to the next or the third shows the loss of "sharpness" i.e. edge-retention represented in the bottom-right quadrant of my image above. While the edge is theoretically at "relative 100%" sharpness BEFORE the first cut, we can't measure force OR amount of media cut at a given force before making a first cut However, if edges are sharpened identically, we might infer identical "100%" sharpness levels in advance of the cut. Anyway, performance is down to <80% and in some of the cases <40% of the first-cut level by the end of the 3rd cut.

In your image below, from the link provided, by the 4th or 5th cut the exponential decline in performance begins to level out. By the 10th cut in the steel with lowest wear-resistance, the edge is performing at near 10% of where it was in the first cut. That is an example of what I am referring to in regard to the "working edge" being at only ~10% the initial sharpness/performance - doing 9x less work or requiring 9x the effort to do the same work. It is wonderful to see materials fighting this, however, providing us a sharper edge through more work

 

[Revolverrodger]

Thanks for this excellent article
I have good experiences with nitro-v
I was wondering if you had toughness data for 80crv2. It’s not too expensive and has worked quite well for the bigger knives I have made

 

[Larrin]

Thanks for this excellent article
I have good experiences with nitro-v
I was wondering if you had toughness data for 80crv2. It’s not too expensive and has worked quite well for the bigger knives I have made

I think we have some 80crv2 samples coming. There is a bunch of stuff that has very slowly been trickling in.