Edge retention testing compared - analysis

Cobalt

Cobalt

Platinum Member
Joined
Dec 23, 1998
Messages
17,908
So I figured on starting a new thread based on testing done by some forum member. I figure I could analyze this data, since if the data is for all to see.

Take a look at the numbers below and tell me what you see. It is the difference that matters here since the starting sharpness will vary from knife to knife. I guess if the tester had wanted to make it a really good test he would have made both knives equally sharp, but instead he left INFI requiring twice the force to cut in the beginning.


Cut INFI 420 ATS FF LAU

005 080 060 070 040 040
050 080 080 090 060 060
100 110 080 090 070 070
200 130 100 100 090 080
300 110 140 110 090 090
400 120 140 130 110 100
500 110 170 130 110 100
600 120 200 140 110 110
700 130 260 140 120 ---
800 130 --- 150 120 ---
Click to expand...


From Cut 5, which is probably the true original sharpness of each steel, INFI has a sharpness of 80 and FFD2 has a sharpness of 40. There is a considerable difference in the starting sharpness of both steels. Yet, when you follow the continued cutting you will note that FFD2 requires 80 more grams to cut than it did at the start. Whereas, INFI required 50 more grams of force to cut than it did in the beginning. INFI actually closed the gap in cutting force from 40 grams difference to FF to 10 grams difference to FF.

I figured I would use his data to show how he is actually incorrect in his deductions.

Funny thing is that I proved him wrong when he originally came out and stated how 420HC had beaten INFI in his thread. I told him how if he actually analyzed his results he would see that INFI's degradation was much less than 420 which meant that INFI would be the winner in the long run and I was proven right.
 
Cobalt said:
So I figured on starting a new thread based on testing done by some forum member. I figure I could analyze this data, since if the data is for all to see.

Take a look at the numbers below and tell me what you see. It is the difference that matters here since the starting sharpness will vary from knife to knife. I guess if the tester had wanted to make it a really good test he would have made both knives equally sharp, but instead he left INFI requiring twice the force to cut in the beginning.





From Cut 5, which is probably the true original sharpness of each steel, INFI has a sharpness of 80 and FFD2 has a sharpness of 40. There is a considerable difference in the starting sharpness of both steels. Yet, when you follow the continued cutting you will note that FFD2 requires 80 more grams to cut than it did at the start. Whereas, INFI required 50 more grams of force to cut than it did in the beginning. INFI actually closed the gap in cutting force from 40 grams difference to FF to 10 grams difference to FF.

I figured I would use his data to show how he is actually incorrect in his deductions.

Funny thing is that I proved him wrong when he originally came out and stated how 420HC had beaten INFI in his thread. I told him how if he actually analyzed his results he would see that INFI's degradation was much less than 420 which meant that INFI would be the winner in the long run and I was proven right.
Click to expand...

Excellent points. It could also be added that there was a good degree of room for error with his testing method, though he obviously tried to be as scientific about it as possible. From watching his video, it appears that he obtained his data by visually monitoring the highest point the needle reached on a postal scale. A slight variation in any number of variables could cause at least a 10 gram deviation, in either direction. I could point out several possible variables offhand, but don't think it really necessary.

Setting up and conducting such tests in an a truly accurate manner is extremely difficult. Especially without laboratory equipment in a mechanically controlled environment. That said, the forum member's subsequent misinterpretation of his own data, and the conclusions he drew therefrom is another issue.
 
Someone with more time on thier hands should graph the test results. I believe the slopes of the lines would show the rate of edge degradation. It'd probably be a little more clear that infi stands out if it was in graphical form.

Just a suggestion:)
 
I remember that tread. His testing method was quite interesting IMO, but because he tested with the orinigal ege, it was more a comparison between those factory edges than purely on steel quality.
Problem is that it can also be affected by the fact that the busse probably will have had a convex edge, while the 420HC had a V-grind, etc
So this difference can also state the superiority of a convex edge over the long therm, while a V-grind has better initial sharpness.

It would be great if someone with lots of time would do that test again, with blades of exactly the same initial geometry(if that would be possible)
Then we could start making real scientific conclusions.

I didn't think the starter of that thread stated that he was certain about his conclusions(just first deductions), but I can't find the tread back since I can't search anymore.
Do you have a link please?
 
He that can't be named has noted that blunting is not a linear process and different causes of blunting act in different ways such that blunting curves are simple representations of complex phenomena and are dificult to analyze. For example, all of the steels tested show an exponential decrease in sharpness, followed by a linear decrease in sharpness. Unfortunately, the test was not taken out far enough to see a statistical difference between all but 420. This one, by far, blunted the most over the course of the test. The others do not show a difference that I would call above the background error of the experiments, which is easily +/- 10g.
 
hlee said:
He that can't be named has noted that blunting is not a linear process and different causes of blunting act in different ways such that blunting curves are simple representations of complex phenomena and are dificult to analyze. For example, all of the steels tested show an exponential decrease in sharpness, followed by a linear decrease in sharpness. Unfortunately, the test was not taken out far enough to see a statistical difference between all but 420. This one, by far, blunted the most over the course of the test. The others do not show a difference that I would call above the background error of the experiments, which is easily +/- 10g.
Click to expand...


:eek::eek::eek::eek::eek: My head hurts!! That was way over my head.
 
At cut 500, the data shows INFI and FFD2 at the same force. INFI required only an additional 50 grams of force where FFD2 needed an additional 80 grams by the end of the test. Personally, I like D2, but this data shows INFI as the more consistent steel.
 
The short answer is that the manila rope cutting test, at least done in this fashion and taken to this number of cuts, is inconclusive in distinguishing the steels tested (except for 420) and needs to be taken out further to draw any conclusions. It is also possible that this test is not able to distinguish these steels regardless of cut count.
 
I agree, please don't think I'm validating these results in any way. I'm just interpreting the data as if it were valid.
 
Actually, I was responding to...

"My head hurts!! That was way over my head" from Chuddy Bear.
 
hlee said:
The short answer is that the manila rope cutting test, at least done in this fashion and taken to this number of cuts, is inconclusive in distinguishing the steels tested (except for 420) and needs to be taken out further to draw any conclusions. It is also possible that this test is not able to distinguish these steels regardless of cut count.
Click to expand...


that may be true, however I suspect that if taken out to 2000 or so, you will see the same thing that happened to 420 happen to D2. It will degrade beyond INFI and INFI will continue cutting at a moderate level. In fact by graphing the 800 cuts you can probably see how INFI is actually levelling off and you can predict where it will go to a point and I am sure that is well beyond D2.
 
hlee said:
The short answer is that the manila rope cutting test, at least done in this fashion and taken to this number of cuts, is inconclusive in distinguishing the steels tested (except for 420) and needs to be taken out further to draw any conclusions. It is also possible that this test is not able to distinguish these steels regardless of cut count.
Click to expand...

Much better. :D I dont even need my dictionary to understand all of the words. Thanks.
 
A comparison of knives with different edge geometries, edge thicknesses, weights, etc. is completely worthless IMHO and is little more than a waste of time. I do not doubt the good intentions and honest beliefs of "he who shall not be named", but good intentions do not make his tests results or methods valid.

Who here could even reliably reproduce his results? Hell, take a second set of the very same knives he tested and even he would likely get different results.
 
Quoted from Cobalt above...

"In fact by graphing the 800 cuts you can probably see how INFI is actually levelling off and you can predict where it will go to a point and I am sure that is well beyond D2."

Unfortunately, all of the curves (except for 420) pile on top of each other and, due to intrinsic error in the experiments, extraplate to roughly the same end. I doubt that an extrapolation of this data would track well with actual data taken out to 2000 or more cuts.

On_the_edge is on the right track. With out starting at the same sharpness, with the same edge geometry, and similar blade styles- there are too many variables to parse out the steel properties.

He that shall not be named is Cliff S... You almost got me.
 
Pretty sure it's not Cliffs work.

http://www.bladeforums.com/forums/showthread.php?t=509097

I would like to point out that Cobalt's original analysis used pretty fuzzy math in my opinion. (In this case, with regards to 420HC)

If you look at the average for the first six cuts, you will notice that INFI was averaging 20 more than 420(80 for INFI and 60 for 420HC), which means that INFI started off more dull than 420. THen if you look at the final number of 140 for 420 HC and 120 for INFI you will notice that INFI deteriorated by 50% while 420HC deteriorated by 133%. That means INFI performed 83% better than 420HC which is a landslide IMO.​

While I agree with your conclusion that 420HC deteriorated at a faster rate that INFI, that doesn't automatically prove that INFI is better. Since 420HC started sharper, it "could" have still ended sharper than INFI, even though the rate of dulling was greater. Now, in the end, this wasn't the case and INFI proved it's value, but using "rates" instead of "absolute" numbers is a good way to get yourself in trouble.

As far as INFI starting duller than the 420HC. I've seen some of the sharpness video's this guy puts out, and he knows how to put an edge on a knife. I'm not exactly sure of the initial conditions of the edges on the knives, but if he put the edges on himself, it would seem to show that you can put sharper initial edge on his 420HC than with his INFI. But the tests also go to show that INFI will hold a better edge for a longer amount of time.

It should also be pointed out that dulling is a non-linear process. So claiming that INFI is superior after just the first 100 cuts would seem to me to be premature at best.
 
ThunderUnderKilt said:
It should also be pointed out that dulling is a non-linear process. So claiming that INFI is superior after just the first 100 cuts would seem to me to be premature at best.
Click to expand...

premature declaration....




i dont understand what any of you guys are talking about. :D
 
ThunderUnderKilt said:
Pretty sure it's not Cliffs work.

http://www.bladeforums.com/forums/showthread.php?t=509097

I would like to point out that Cobalt's original analysis used pretty fuzzy math in my opinion. (In this case, with regards to 420HC)

If you look at the average for the first six cuts, you will notice that INFI was averaging 20 more than 420(80 for INFI and 60 for 420HC), which means that INFI started off more dull than 420. THen if you look at the final number of 140 for 420 HC and 120 for INFI you will notice that INFI deteriorated by 50% while 420HC deteriorated by 133%. That means INFI performed 83% better than 420HC which is a landslide IMO.​

While I agree with your conclusion that 420HC deteriorated at a faster rate that INFI, that doesn't automatically prove that INFI is better. Since 420HC started sharper, it "could" have still ended sharper than INFI, even though the rate of dulling was greater. Now, in the end, this wasn't the case and INFI proved it's value, but using "rates" instead of "absolute" numbers is a good way to get yourself in trouble.

As far as INFI starting duller than the 420HC. I've seen some of the sharpness video's this guy puts out, and he knows how to put an edge on a knife. I'm not exactly sure of the initial conditions of the edges on the knives, but if he put the edges on himself, it would seem to show that you can put sharper initial edge on his 420HC than with his INFI. But the tests also go to show that INFI will hold a better edge for a longer amount of time.

It should also be pointed out that dulling is a non-linear process. So claiming that INFI is superior after just the first 100 cuts would seem to me to be premature at best.
Click to expand...

I would agree that if the original tester had put an upper limit on sharpness, as in what is actually considered dull in his test, then the numbers would have changed, but we did not have that. We do have those numbers now, however, as one of the steels in his test did drop off at just over 260 grams(say 280 grams is considered dull), so making percentage estimates would be easy enough. if dulling is a nonlinear process, but do you know why? and why it differs?
 
You must log in or register to reply here.
Back
Top