View Full Version : Understanding the results of testing as it applies to HT

Josh Dabney
01-31-2011, 09:30 AM
I thought I'd start a new thread on this even though this is the same blade from my hamon attempt thread.

Having recently got Parks 50 I'm working our my HT as it applies to Aldo's 1095 and want to achieve full confidence that I've got my HT nailed both as a challenge to my own abilities and a service to my customers.

This blade was created as a test blade but having been quenched 3 times (twice for hamon and finally fully hardened) it's not what will be my final recipe for this steel but I wanted to put it through some paces and see just what it would do then break to check grain structure.

I tempered at 400 2 hours x2

I finish ground the blade to .017 give or take .0005 then convexed the edge and sharpened on stones.

Throughout this testing I've realized that I'm not 100% confident in reading what the blade is telling me as it applies to HT.

I am very interested at this point of better testing methods with more meaningfull results so I'd LOVE to hear suggestions.

1. I tried "brass rod testing the edge on the corner of a brass bar but couldn't get the edge to deflect and the blade bit into the brass a bit. So now I'm thinking that the edge is still fairly hard at a 400 degree temper and perhaps thinner edge geometry would be better.

2. Whittled an chopped a little on a preasure treated 6x6. No damage or dulling

3. Chopped on an old dry antler as hard as you can chop with a 4" blade and no handle. Again no chipping, flat spots or dulling of the edge.

4. Batoned the blade through the antler. Now I got a big old wave in the edge. The edge again didn't chip, flatten or dull but did sustain permanent deformation. The waved section would still cleanly slice paper afterwards.

So now I'm thinking that I must have plenty fine grain for no chipping but contrary to the brass test perhaps the geometry is too thin. I wouldn't expect edge stability to be a problem at this thickness so maybe this "test" just goes too far. Again though I'm thinking that that a hotter temper may add some toughness.

Following that I figured I'd cut some cardboard and see what kind of edge retension I've got so I grabbed a box and cut it up into the smallest pieces I could make relatively quickly using only the section of blade from the wave forward. Cardboard is a fairly abrasive medium to cut so it usually accels at dulling an edge (usually)

To still be shaving sharp after that many cardboard cuts seems excellent performance to me which once again leads me to believe that a perhaps it could be a little softer for a better balance between edge retension and toughness

Satisfied with everything other than the wave on the edge I clamped the blade in the vise to break it. I've got the rough jaws on my vise and figured I'd just clamp it tight to break the blade which it did easily right at the vise jaws. Too easily, but now I'm not sure if it broke easily because of hardness or because I clamped it tight with sharp vice jaws instead of using smooth jaws. Thought I'd add that PVC is NOT a good leverage handle as the tang blew right through the back side when the blade snapped. Guess I'll get a piece of pipe for next time.

Now looking at the grain it appears to be pretty fine but I've seen better so I'm thinking i'll normalize 2 more times at 1425 to attempt to maximize grain refinement. The break does look fairly rough (especially toward the spine) like it tore apart as opposed to snapping clean. I'm honestly not sure what to think about that. A product of grain being bigger than what it could be ? Perhaps a sign of a problem in the grain boundries ? Or just a product of tempered steel that wants to stay together until it's literally torn apart ?

A couple of my buddies do have Rockwell testers so eventually I'll get this blade tested which will assist in determining the tempering temp needed. But in attempting to figure this out without a tester the method of tempering low and raising the temp till you get no chipping would lead me to believe I could temper at a LOWER temp and still be ok which is contrary to my suspision that the blade is still relativly hard.

As long as the edge will holds up I'd have no problem spring tempering the spine to add overall toughness to the blade.

James Terrio
01-31-2011, 11:31 AM
If you had to baton it through an antler to get the edge to deform, and still didn't chip it, I'd say you've got plenty of both toughness and hardness. Good golly what more do you want? :)

I've heard of people using a torque wrench to measure how much it takes to snap a blade in a vise, but I'm not exactly sure how one would go about that. I definitely think you should put wood inside your vice jaws, the rough/sharp edges may be causing issues that are hard to control.

Doug Lester
01-31-2011, 01:23 PM
I would agree that the toughness is fine, it's the strength that MIGHT need to be increased. Toughness is the ability to resist breaking; strength is the ability to resist deformation. If you feel that the grain was a little large, then I would agree to correct that with multiple normalizations. I immagine that the difference that you were seeing between the spine and the edge of the blade was due to the shallow hardening of the steel. Verhoeven illustrated that when a blade made from shallow harding steel that the martinsite will only form to a thickness of twice the depth of hardening. The sections thicker than that will form pearlite. With proper grain size control with the 1095 you will end up with automatic differential hardening of the blade. Refining the grain will actually decrease the depth of hardening in the shallow hardening steel while increasing toughness. In the end, you can only work within the limitations of the steel that you use.

Doug Lester

Josh Dabney
01-31-2011, 02:01 PM

Just trying to make the best knife I can ;) Not to mention the whole HT - Test - Destroy thing is kind of addicting but once I'm satisfied I've got all I can get out of the steel then it'll be time to stop breaking them and start finishing them. I do realize that batoning through an antler is abuse but normal activities weren't causing any damage ;)


I had not thought of that but now that you mention it it makes perfect sense. In studying the grain again it really looks to be very fine grained in the lower half but also up the sides to a depth of maybe .040 in from the bevel and across the corners of the spine. The area that was giving me concern is limited to a triangle section in the center of the top half of the blade.

The blade certainly performed like it was fine grained but the "torn area" was throwing me a curve.

I just scrolled up and realized I didn't post the pic of the grain-
It's not a very good pic but here it is

I'm a little paranoid about having large grain so I've been forging down to 1800 then straigthening at 1600 in the forge and following that up by normalizing in the evenheat at 1600, 1550, 1500, 1450, and 1425 which I would think should do the job just fine.

This blade is a little crap shoot after that but I would think that multiple hardening cycles would refine the grain even further.

I have read and heard that it's possible to get the grain so small that the blade will no longer harden. I certainly didn't get to that point yet so I'm gonna stick to my plan of normalizing at 1425 3 times and quenching once.

James Terrio
01-31-2011, 03:16 PM
I do realize that batoning through an antler is abuse but normal activities weren't causing any damage ;)

I understand, and that's a good thing. Testing a knife in a way you would hope never to have to use it can give you extra confidence that it's more than tough and strong enough for anything resembling normal use. Heck look at my avatar... a 4" utility knife wouldn't normally be my first choice for cutting through a 2x4 but it worked :D

Kevin R. Cashen
02-01-2011, 09:31 AM
Josh I think you have a great title for this thread due to the fact that perhaps 90% of the misconceptions in the bladesmithing is the result of incorrect interpretation of results. Bladesmiths are some of the most honest and decent people I have met so much of the misunderstanding probably comes from the fact that they indeed have seen things with their own eyes and are just mistaken as to what those things were saying. And it is this issue that makes "testing" so darned tricky.

Testing in a lab is not held in such high esteem because it is conducted by eggheads with a string of letters next to their name which makes them smarter that us, well perhaps that is why many mistakenly hold it in high esteem, but the real value is in the controlled conditions that environment can provide. When one realizes this then accurate testing becomes possible for anybody so long as they have tight control of the variables and know what it is they are measuring. So long as there are unkowns we will never be certain if were are measuring what we want or some other unaccounted for variable.

The brass rod test is a very good example. It may seem that I am rather hard on that particular test, but I only scrutinize it so much because it is so widely accepted as definitive when it is one of the best examples of misinterpretation of what is being reveled. Any form of flexing will circumvent the accurate measuring of true strength (hardness) in a metal until you completely exceed the limits of flexibility, and then it is rather a moot point. How much, or how easily, a blade flexes in a vice or on a brass rod is solely the function of it thickness and shape, no matter how you heat treat it. And yet flex based tests are the most widely accepted in our field. I too have seen magazine articles showing torque wrenches on blades in vices and the only thing it told me was how little the readership and the “experts” in our field really know about the physical properties they are hoping to measure.

The most accurate test commonly available to knifemakers is the Rockwell hardness test. It is more commonly available than other lab type tests because it eliminates the need for “interpretation” by well trained eyes, other hardness tests (Brinell, Vickers etc…) are trickier. Simply put you sample on the tester and apply the load, the equipment does the rest in providing you with a very precise and very specific reading about one single property. The draw back is, of course, that it is just one tiny piece in the big puzzle. A Rockwell test will tell you exactly how hard the blade is but it won’t tell you why. It can’t tell you about grain size, edge geometry, carbide dispersion etc… But the more things a “test” tells you in one shot the more cross contamination of information, misreading and misinterpretation there will be.

I believe the best overall tests for average knifemakers is to simply use the knife hard for the purpose it was intended, not parlor tricks that grab attention, but hard use that duplicates exactly what the knife was meant for. But even then one must be careful not to over-think it, many would say that cutting rope or cardboard 50,000 times replicates cutting but not real life cutting, and one can make these cuts in a way that will dull the knife in half the time just by the pressure and direction of cut. Where the human hand is involved a test is never truly objective.

Another area that our eyes fail us in the interpretation of fractured grain surfaces. I have seen the differences in smoothness from edge to spine pointed to as proof that edge packing works, when all the unfortunate interpreter was really seeing was the different modes of fracture from martensite to pearlite. The only thing they were observing was the depth of hardening.

I realize that you understand this and you seem to be drawing the correct conclusions from your broken sample. It is after all 1095 you can expect appreciable pearlite formation anywhere from 1/3 to 2/3 up from the cutting edge. The edge will give an accurate idea of the grain size since the martensite will fracture via cleavage at the grain boundaries, while the spine will give a much rougher appearance due to the stretching shear type tearing of the pearlite. Aslo be aware that cycled normalizing will decrease hardenability (a necesarry evil) and move the pearlite even closer to the edge.

This brings us to another quirk in steel that leads to confusion, steel deforms very differently under gradual loading than with sudden loading, thus pressing on a brass rod will give you an entirely different outcome than whacking a brass rod. I like to impact the edge, and from different directions, to test the geometry versus the toughness/strength ratio.

Doug is on the track that I would take, and have. Find a geometry that works for you and then test to see if I have reached the brittle limits. If my edge is still plastically deforming under sudden load I can bump up the hardness or I can increase the thickness to add either resistance or strength. If you really want cutting efficiency I will go with higher hardness, fine grain size and thinner edge. I personally don’t give a rip about how it bends in a vice, I want to know how it can suddenly encounter a hard object.

Josh Dabney
02-01-2011, 11:25 AM
Great to have an honest, civil, dialog on this subject and I thank you gentlemen for that !

I had considered that perhaps I was over normalizing and needing an even faster quench to harden my blade.

I had NOT made the correlation that in effect this fact makes the steel even shallower hardening.

I have been wondering how grain size correlates to creating hamon. Now I'm thinking the finer the grain the easier it is for pearlite to form which translates to less clay needed and higher hardening temps needed to get enough martinsite for hamon to form with any level of detail.

Assumimg I'm interpreting this correctly I can see how differing normalizing procedures and less than optimum heat control for normalizing would have a profound effect on the outcome of a clayed HT. Could be a large factor in explaining the "mystery" of this type of HT and why what may work perfectly for one smith may not work in anothers shop.

Thinking further about the Rockwell test- Really the only place I could test this broken blade is on the ricasso where both sides are flat and parallel but I'm not sure if that would even tell me anything about the condition of the edge where the steel is fully hardened. I would imagine that with the flats of the ricasso being basically surface hardened with a core of pearlite that this could skew the results in comparison to the condition at the edge of the blade. Not to mention the areas of differing mixes of martinsite and pearlite may very well react differently to the temper providing mixed results at different areas on the blade.

Boy this does get complicated in a hurry huh.

Adressing testing methods-

I'm forging out these smallish blades just for testing purposes and figure that the results of my HT will correlate also to larger blades. I'm still fairly slow at forging and figured that small blades would be economical use of steel for this purpose. I'm starting to think that this is a false economy because it's just difficult to push this little blade hard enough doing things normally reserved for larger knives like chopping with this little guy.

I think real world testing has valid results even considering the "human factor" involved. Without making scientific claims based upon the results the comparison value remains for one maker comparing his own blades.

That being said though I'm still at a bit of a loss trying to come up with a set of tests that will push the blade hard enough to reaveal flaws in HT & geometry without being downright abuse.

I'd love to develop a testing routine that is both hard and objective.

Gotta run for now but want to thank you guys again for your assistance !