Steel Edge Retention testing

[papathud]

Post them the way you see them when you test them.

All of this takes effort.

Thank you for the Effort.

I am just here to read what you post.

 

[HoB]

LOL. It appears my quick analysis at a glance from a hotel room was more accurate than your attempt at it in tons of long paragraphs.

I am happy to see that you haven't lost your good humor. I prefer to be wrong than see a thread like this glide off into sillyness. As long as I am learning something, I don't mind being wrong.

Btw. I haven't made an attempt at an analysis at all. All I have done (or tried), is to point out that an attempt at an analysis is pointless if you don't have a reasonable point of reference to which you can relate the data to.

As you say, this is really the key:

at what gram weight would you consider a knife barely able to cut. I think this is crucial in your test. There is no way to quantify and compare the results unless you have a gram point at which a blade can barely cut thread, which would be the same point for all knives.

Seems like we are all on the same page now.

 

[HoB]

After hundreds of rope cuts, his knives are about the same level as most quality factory fresh blades.

Excellent point!

I had a shaving edge with a median of 190g, from 120 grit.

I have been wondering how the grit at which the blade is finished at affects the thread cutting. I am pretty sure that it doesn't matter much if you finish at 4000 or 8000#. But are there differences between really coarse edges and really fine edges. Does for example barely shaving edge at 8000# come in at 120g and one at 120# at 190g for example, or vice versa? I have no answer, I have just started to work with coarse edges, but I could see that on the really coarse finish the individual strands of the thread will find spots between the "teeth" and either will find flat (blunt spots) and increase the force or will get wedged in between the teeth which might decrease the force.

This also reminds me when I asked what cutting was anyway, and no one had a great answer. The thread parts at 1.5kg, and it parts at 30g. When is it no longer being 'cut'.

Well, I think in the extrems at least this question is pretty easy to answer. When you test the strength of the thread with a plastic ruler which has pretty rounded edges, you can really see the strands fibers of the thread, so the thread is rather torn than cut. The force component that makes the thread fail is essentially all along the thread. The thread would likely fail in the same way if you would hang a weight from it. When cutting the thread, the strands are terminated pretty sharply. The thread seem to fail largly due to a tranverse force component. But I agree with you that the question where cutting starts and tearing ends is probably largely philosophical.

 

[OldPhysics]

You can't interpret results that way. You can say that this test is not completely accurate because they didn't start out at the same level of sharpness, but only through statistical nonsense can you say that INFI performed "83% better than 420HC," unldess you're just making sarcasic comments, in which case I don't think you're adding anything to the discussion.

Interesting point of view.

I like Nozh's tests ... they are very practical/pragmatic. I take them as a general guide to behaviors...perhaps an 'order of magnitude' comparison. They are not, in any sense, rigidly scientifically controlled.

If you genuinely want scientifically valid tests, particularly tests that can be used to make predictions in an absolute sense, you will have to work much harder even than this.

"Sharpness" is a poorly defined quantity. Nozh has taken a reasonable approach to better define "sharpness," but I do not believe he is ready to publish in a scientific journal (of course, Nozh makes no such claims himself). If it matters, I actually am a scientist -- a materials scientist -- and I would find it necessary to do considerably more work before creating a valid physical parameter for careful testing that would serve as an adequate analogue for the term of art we describe as "sharpness." And this is merely a necessary first step before approaching anything so rigorous and challenging as making absolute measurements. Truly valid and valuable measurements of absolute quantities, even apparently very simple ones, are much more difficult than you would believe -- and usually very expensive. For such measurements to have statistical validity, you must know a great deal about the nature of the variables involved and you must take statistically valid samples. This sort of challenge is what has kept me paid -- and well paid -- for many decades. I don't think I'll be out of work any time soon.

It is much, much easier to make useful measurements that can be interpreted in a 'relative' sense (bigger v. smaller, percentages, rates of increase, that sort of thing...). That appears to me to be what both Nozh and Cobalt are attempting to do -- to make relative comparisons of Nozh's measurements. Frankly, in the absence of equivalent initial conditions I would also be tempted to make the comments Cobalt has made. I can't be sure those comments are correct, since I lack detail on precise experimental conditions, but I would lean in that direction.

If Nozh's stand-in for "sharpness" was a more carefully controlled physical variable, I would be strongly tempted to prepare samples with equivalent initial "sharpness" and repeat the tests. Not once, not twice, but a statistically significant number of times on a statistically significant array of samples. Frankly, that's time consuming, expensive, and rather boring...and no one seems anxious to pay for it!!

BTW, when I am speaking of 'absolute' measurements, I am referring (in a loose sense) to single measurements with units and associated numerical descriptions of error. An example would be: "The temperature is 90.6 degrees Centrigrade +/- 0.4 degrees Centrigrade." Some of my peers would even insist that I should make a statement about how the errors propagate when I make the bold claim of an absolute measurement (something such as: "normal distribution of errors assumed, error quoted at one standard deviation.")

Sorry for all the technical jargon ... generally, I simply prefer to admire Nozh's clever measurements and to thank him for his effort. :thumbup:

 

[sodak]

So what's the point, besides assuming that we don't understand the "technical jargon"? You might be surprised at the education level of many of these posters.

Nozh is trying to perform some meaningful tests, and he's getting some interesting results. I seriously doubt that he has the time or patience to test as many samples as you are proposing. So should we all just throw up our hands and say it's too hard? It'll never work?

He's attempting to add to the overall knowledge of this forum in a very transparent way, which is good. He has repeatedly offered for others to independently verify what he's been doing, I haven't seen anyone take him up on it yet.

When all is said and done, I'll bet he has something that we can draw some conclusions from and increase our understanding. I'm waiting for the final results, but I'm sure he needs to let his hands and arms rest for a while.

 

[nozh2002]

This is physic approach or even more engineering - you learn math as a tool and in this terms limited. In short I do not think that here we are talking in the same language. Of course standard material lab method not applicable here for absolutely clear reason, but this simple means that different methods need to be used.

I am not physic and can not talk about mechanical properties on thread and rope and edge. But I see numbers and this abstract numbers follow physical property of the edge in certain pattern. So sharpness and median of 21 tests on rope tied together pretty well and experiment shows this also - I cut quite a bit of threads to came to this conclusion and now absolutely confident.

I am not scientist-profi and not publishing articles etc, but this results are pretty reliable and scientific, if you like.

And honestly this is best known so far (with all respect to official scientists) and most precise and reliable way to measure sharpness. And I am not defining it but measure it.

If modern science know something better - let me know.

Now can we say 40 is twice sharper then 80 - not of course - sharpness is not defined!

Thanks, Vassili.

 

[hardheart]

Does for example barely shaving edge at 8000# come in at 120g and one at 120# at 190g for example, or vice versa?

I didn't try to get it any sharper, that was what I got right after I took the bur off, maybe it could improve further. With 8000 grit (or any, really), I could hone at 90 degrees, reducing sharpness and eventually (not anytime soon) end up in the primary grind. From there I could increase sharpness all the way up to an edge maybe 0.3-0.4 microns across. I don't know how narrow I could get with 120 grit, but it will start breaking out the edge as sharpness increases/edge width decreases much sooner and more drastically than the 8000. This of course limits ultimate sharpness to a much greater degree. IOW, I could get a 190g edge on 8000 grit (without the rough profile), but I don't know yet if I can get a 120g edge from 120 grit. I am thinking of doing some more testing, I want to try a few different knives and measure each at at least half a dozen or so grits, being able to shave arm hair at the least before cutting the thread.

My old method of sharpening was one step, 280 grit on the Lansky. slightly toothy,

 

[gunmike1]

I applaud the amount of work and information we are seeing from this testing. I did only 3 runs with 64 cuts in each run with 2 knives and wore my butt out trying to fit that into my hectic schedule. I can't imagine how long all of this cutting has taken to do. This is a lot of work with a lot of data to sift through. The starting sharpness I would think would ideally be equal, but what if 1 steel just plain gets sharper than another? How noticeable are the actual differences in cutting? Maybe a sharper knife with a thicker edge actually takes more force to cut the 1/2" manilla, as that is what I found when I was testing on a bathroom scale the actual force to cut the rope in my testing (and why I love Tom Krein regrinds). That is when I decided to do an independent test to try to measure the slicing sharpness that took the blade profile out of the mix and only measured the edge sharpness, as Nozh is doing, though his test is much more precise than mine. The more testing we see like this the more we can share methods and try to learn more about better testing procedures and the performance of knives. Keep up the good work, Nozh.

Mike

 

[nozh2002]

It all started while ago:
http://www.bladeforums.com/forums/showthread.php?t=346429)

But this final testing I began 17 Oct 2007. First steel was most boring, because you have nothing to compare. Anyway I develop this as a routine exercise wich I do little by little every day. However, when I start second steel 420HC it turns very excited, because now I can see how different steels compete each other, now I have to limit myself, to not to hurt myself too much...

Again I am very proud of rediscovering 420HC as well as latest proof of INFI as a hardworking steel which can not give up sharpness - really crazy thing, as well as ATS-34 stability... Who knows what I may see next!

I feel like I am on frontier discovering new horizons and this feels soooo goooood! Pretty excited stuff and so I am not tiered at all (so far) and not thinking about how much work I done...

Thanks, Vassili.

 

[theonew]

I look forward to seeing your results if and when you test D2, that steel in my personal experience behaves curiously. Not matter how fine a grit I use to sharpen it, it still always feels very toothy. It then holds its toothy edge for quite awhile, but when it starts to dull it seems to get really dull very very quickly. But then this is based solely upon my experience using it, not formally testing it.

 

[nozh2002]

I look forward to seeing your results if and when you test D2, that steel in my personal experience behaves curiously. Not matter how fine a grit I use to sharpen it, it still always feels very toothy. It then holds its toothy edge for quite awhile, but when it starts to dull it seems to get really dull very very quickly. But then this is based solely upon my experience using it, not formally testing it.

But if you look at the pictures on the previouce posts you may see why. It has huge carbides which may form this teeth you feeling. And they may stand while surrounding iron wear out - and this will keep it toothy, but eventually they all fall out and expose not sharpened rounded carbides...

Think about old castle stone wall sharpened and then wearing out...

Of coures this is just hypothesis but I will eventually test D2 - I have Dozier as well as Microtech Currahee.

Thanks, Vassili.

 

[theonew]

But if you look at the pictures on the previouce posts you may see why. It has huge carbides which may form this teeth you feeling. And they may stand while surrounding iron wear out - and this will keep it toothy, but eventually they all fall out and expose not sharpened rounded carbides...

Think about old castle stone wall sharpened and then wearing out...

Of coures this is just hypothesis but I will eventually test D2 - I have Dozier as well as Microtech Currahee.

Thanks, Vassili.

That does make perfect sense.

Being someone who sharpens his knives regularly, I really love D2, stays wicked sharp for a lot of work over a weekend camping trip, and gets touched up at home long before the edge degrades.

 

[theonew]

Hey Vassili I know you're already spending a lot of time, money and effort running these tests, but I think it would add valuable information to your results if you had each knife used in your tests Rockwell tested. Just a thought.

 

[HoB]

I didn't try to get it any sharper, that was what I got right after I took the bur off

Sorry, that I've been so unclear, but that's not what I mean. What I tried to say is: If you compare to edges that are able to perform the same task about equally well, such as both being roughly shaving, but one being sharpened with a coarse finish, the other with a polished finish. Will they both come in at the same weight in the thread-cutting test, or will they be different?

That is regardless of the fact that you could get both sharper.

theonew: Larrin posted a while back a series of micrographs, which, I believe, were provided by Sandvik. The three images I posted above are from that series. This is another image from that series. This is a worn 440C edge. 440C has a similar microstructure as ATS-34 does. The really large carbides are slightly smaller but it has more small carbides. I think you can see from the worn edge what will happen with the large carbides (the image is looking down on the edge):

 

[hardheart]

hmm, a question for me is what would qualify the finish? I thought about stropping between grits, but then that 120 grit edge wouldn't have been a 120 grit edge. Even one pass per side would have refined the edge somewhat, but then it wouldn't measure the same as one pass per side after 10000 grit. So, is a stropped edge a stropped edge? I think the history of the edge, the condition coming off the previous grit is important.

If an edge can shave at 200g, I think you could take a 250g edge, make a few passes on a polishing grit, and have it test at 200g and shave. But, how do you define the edge finish? There's conventions for surface finishes, but I've only seen a few people who check their edges under magnification strong enough to quantify it. And does that translate directly to the edge width? How many passes on an 8K stone do you need before you have an 8K edge?

 

[db]

I tried repeteing some of Nozh's work. Nozh wow. I had a very hard time just trying to get a reading on the scale at the point the thread is cut. For anyone who doesn't think Nozh is doing alot of work just try measureing force to cut some thread. As for slicing rope if it is done right at the end of the rope I didn't notice or could tell any difference between a thicker blade and a thinner blade with the same edge. As said before the threads in the rope just fall away as they are cut before the blade thickness difference is seen by the rope.
Thanks Nozh for all your work even if I don't understand all the tech stuff.

 

[Broos]

Thanks for the results, Noz.

My understanding is that Noz is testing thread with a push, but previous thread tests I've seen are measuring it with a slicing motion. Is everyone in agreement that a test that uses a slicing motion on thread is a test for sharpness? I was under the impression that those thread tests were measuring "slicing ability"? Many variables are added with the addition of a slice to the thread test, and I think this push cut version would be more repeatable from tester to tester.

This same test method - measuring (true or push cut) sharpness with the wear or dulling caused by slicing - was vociferously objected to with some previous test results posted here, if you recall. No objections here, but thought it was worth mentioning, when there is opinion out there that a knife that slices like a mother can get poor results with a push cut test. For this type of knife/steel may not perform well in this test, or just never reach the sharpness levels of other knife/steels will reach using this test (yet still slice like a mother). something to consider, anyways.

All the statistical questions are to me best answered by looking at a good graph of the results. In my opinion trends and correlation are much easier to see that way then the numerous statistical analyzations that are possible.

 

[Larrin]

Thanks for the results, Noz.

My understanding is that Noz is testing thread with a push, but previous thread tests I've seen are measuring it with a slicing motion. Is everyone in agreement that a test that uses a slicing motion on thread is a test for sharpness? I was under the impression that those thread tests were measuring "slicing ability"? Many variables are added with the addition of a slice to the thread test, and I think this push cut version would be more repeatable from tester to tester.

This same test method - measuring (true or push cut) sharpness with the wear or dulling caused by slicing - was vociferously objected to with some previous test results posted here, if you recall. No objections here, but thought it was worth mentioning, when there is opinion out there that a knife that slices like a mother can get poor results with a push cut test. For this type of knife/steel may not perform well in this test, or just never reach the sharpness levels of other knife/steels will reach using this test (yet still slice like a mother). something to consider, anyways.

All the statistical questions are to me best answered by looking at a good graph of the results. In my opinion trends and correlation are much easier to see that way then the numerous statistical analyzations that are possible.

It looks like he's cutting the rope with a push cut to me.

 

[hardheart]

the tests have varied. I think the most common is slicing the rope and then push cutting another material to check loss of sharpness. I recall objections to the D2 slicing tests because it was being done by hand, the length of slice and amount of force applied was considered by some to be too variant for such a light media to cut.

 

[nozh2002]

It is all on videos - it is "normal" cut which is combination of push and cut. If you see how mark was wiped out - diagonally. So I think this is way everybody use knife with power cut pushing and slicing at same time.

Thanks, Vassili.