Differential Tempering

[chknifemaker]

Grizz-don't you heat only the edge though when you "full quench" the blade so the spine is still not turning into martensite????(would drawing the spine still be beneficial with a edge quenched blade??)
here's a video where after full quenching he uses a hot metal block to soften the spine is this a good method?(I believe Wayne Goddard used something similar but with tongs)
at about: 1:40
http://www.youtube.com/watch?v=9phlDH4h-eU

this has been a great thread and have enjoyed learning from it

 

[Bo T]

So, if I understand correctly, toughness is how soft the metal is (as supposed to hard and brittle) and strength is how springy the metal is (so it returns to its shape), and the elastic limit is the same regardless we just perceive it differently because hard steels break and soft steels bend.
Am I way off? I really want to understand

That is a simple explaination and generally true. Before the elastic limit is reached, both will return to shape. The tempered martensite will make a stronger spring. Also, the crystal structure will have an effect. Fine pearlite will respond differently than course pearlite so it gets a little more complex. And there are other crystalline structures like spherzoidite, upper and lower bainite, and carbides that will all have an effect. And we are discussing steel so I don't know if the same applies to other metals like brass, bronze, or aluminum alloys.

 

[lastbrunneng]

Thank you, I'm always trying to wrap my head around these concepts.

 

[Salem Straub]

chknife...
A full quench consists of heating the entire blade, and quenching the entire blade, so that all of it becomes martensite.
Drawing the spine on an edge-quenched blade is done often enough, it is the "belt and suspender method" to make good and sure that the spine is fully soft. This is done sometimes on knives made to pass the abs bend test.
Drawing the spine can be done many ways, I use a pan of water to protect the edge and a welding torch to heat, hot tongs can be used, a hot plate of metal can be used, a torch with no water bath can be used. I like having the water bath because you can get your spring temper good and thorough throughout at least half of the blade width, without worrying about overheating the edge.
Some guys go with a method of triple redundancy, like heat only the edge with a torch, then quench only the edge, then oven temper, then soft draw the spine.

It is an odd concept that elastic limit is not varied by hardness. One day I would like to test that with two identical spring steel jian sword blades- one dead soft, one spring tempered. If the spring temper blade flexes to 90 degrees over its length and returns to true, then so should the spheroidized one...

One other thought is, if you want strength and flexibility in your knives, beyond heat treat, geometry is huge... a thinner blade can be very flexible and yet quite hard. Think of a razor blade...

 

[JacksonKnives]

If I'm getting this right (and I'm no expert by any means...)

assume two knives of the same profile, thickness, etc.

*numbers are entirely made up, and probably not proportional.*

Elastic range (let's say, for the given knife, up to 60 degrees bend with the blade clamped in a vise)

• Stronger (generally, harder steel, eg 61RC) will not bend as far with a given weight -- let's say at 50lbs it goes to 45 degrees.
• Tougher (generally, softer steel, eg 45RC) will bend further with a given weight -- let's say at 50lbs it goes to 59 degrees.

Both knives spring back to true, with no damage.


Destructive range (assume at 60 degrees, the stress in the sides of the steel reaches a threshold)

• Stronger (eg harder, EG 60RC) will snap (but it took 90lbs to get it this far.)
• Tougher (eg drawn to 45RC) steel will start to 'set' and not return to original straightness; the sides of the knife have deformed permanently and while you might be able to bend it back it'll probably break with another bend. Also, it only took 60lbs to bend it this far.

Both knives are broken, though the remains of one are probably sharper.

So you have some different things going on that can change with hardness: you have how much load it will take to get a certain bend, and you have what will happen when it bends too far. (There's also resistance to shock, which is IIRC pretty much directly correlated to toughness/temper.)

Is that pretty much how the science works out? I'm not really clear on the specifics of Young's Modulus (formulas give me a headache and I go back to cutting stuff) but it seems like a reasonable interpretation based on what I've read here.

Many can/have drawn conclusions from this re: differential hardness and ways to get it there, but I've never been all that interested in that kind of thing. I certainly can understand why some makers do an edge quench, but I also see the benefit of through-hardening and differential tempering.

 

[Rick Marchand]

If the spring temper blade flexes to 90 degrees over its length and returns to true, then so should the spheroidized one...

I don't know about this, Salem. The spring tempered blade is tempered martensite and will resist deformation beyond its elastic limit. The spheroidized blade will not.

I like the experiment that Kevin told me about...

Take three identical bars of steel... one fully annealed, one full hard(as quenched) and the other fully quenched and spring tempered. Clamp them about 2/3 off the edge of a bench and start hanging identical weights off them. They will all flex to exactly the same degree until the elastic limit is reached. The full hard blade will soon snap. Once the annealed blade begins to bend it will continue to do so without adding any more weight. The tempered blade will continue to flex with significantly more weight until it finally begins to take set and eventually breaks.

So you have some different things going on that can change with hardness: you have how much load it will take to get a certain bend, and you have what will happen when it bends too far. (There's also resistance to shock, which is IIRC pretty much directly correlated to toughness/temper.)

Is that pretty much how the science works out? I'm not really clear on the specifics of Young's Modulus (formulas give me a headache and I go back to cutting stuff) but it seems like a reasonable interpretation based on what I've read here.

Your first experiment "within the elastic range" both blades will flex exactly the same with exactly the same weight of force.
The second experiment gets more interesting. The harder blade will flex less with more weight until it abruptly breaks. The lower rockwell blade will flex more and begin to take set but eventually surpass the weight carried by the other blade. It is tough and strong.

 

[JacksonKnives]

Your first experiment "within the elastic range" both blades will flex exactly the same with exactly the same weight of force.
The second experiment gets more interesting. The harder blade will flex less with more weight until it abruptly breaks. The lower rockwell blade will flex more and begin to take set but eventually surpass the weight carried by the other blade. It is tough and strong.

Cool, thanks!
so...
Elastic range (let's say, for the given knife, up to 60 degrees bend with the blade clamped in a vise)

• both knives deflect the same amount with 50lbs load, and spring back to true.

Nearing Destructive range (assume at 61 degrees)

• Stronger (eg harder, EG 60RC) will stop deflecting as much per pound (eg from 59 to 61 degrees took 30lbs more, up to 80lbs)
• Tougher (eg drawn to 45RC) steel will start to 'set'; it only took 60lbs to bend it this far.

Stronger blade probably has stress damage in a few places (depending on design), but has come back to true. Tougher blade seems pretty much fine, but has taken a slight set.

Destructive range (assume at 65 degrees)

• Stronger (eg harder, EG 60RC) has broken, under a good deal of weight (eg 100lbs, a further 20lbs from what it took to get 61 degrees)
• Tougher (eg drawn to 45RC) steel is permanently bent; it only took 75lbs to bend it this far, but it will take increasingly more weight to bend it farther, and will still support the weight at 110 lbs.

 

[Willie71]

To simplify, the amount of force it takes to bend steel of a certain geometry does not change depending on heat treatment. The different heat treatment methods change the point that the steel takes a set, or breaks. It will take the same amount of force to bend a knife to, for example, 30 degrees regardless of heat treatment. Whether it takes a set, springs back or breaks is the result of the differing heat treatment methods.

 

[Salem Straub]

Another related idea that interests me:

Take a given set of identical bars, one spring temper and one soft, and both capable of being flexed to 90 degrees at least once without sustaining appreciable mechanical damage.

Clamp bars identically at one end, repeatedly flex to 90 and back until one fails. Which is it?

Seems that a spring temper bar, by virtue of its strength, when bent to 90 degrees describes a much larger arc than the soft bar when bent to 90.

Therefore the stress in the soft bar will be more localized, resulting in enhanced metal fatigue and premature failure at the apex of the bend, with repeated cycles.

Anyone do this test before? Admitting that bending to 90 degrees repeatedly is entirely more academic than practical.

 

[bdmicarta]

From a structural engineering standpoint...

go to this link on wikipedia so we have the same diagram to compare:
http://en.wikipedia.org/wiki/Stress–strain_curve

Mild steel will deflect in its elastic range up to its elastic limit, also called its yield strength, labeled as point 1 in the graph. Soft steel, hard steel, tough steel, doesn't matter at all- up to this point the deflection is the same. Hang different pieces of steel off the edge of the bench, put weight on them, until the yield strength is reached all of them will deflect the same amount. (However significant percentages of different alloy in the mix such as a whole lot of chromium might change the stiffness a little bit).

The harder the steel, the farther it will go before the elastic limit is reached. If the graph contained more than one type of steel, point 1 for one steel would be farther up the line than point 2 for the other steel.

Once the elastic limit is reached then the steel travels along a plateau, increasing stress slightly as it goes, until the end of the plateau where it fractures. Mild steels might have a long plateau, hard steels a short plateau. And for a given amount of hardness, a steel with more toughness will have a longer plateau. (In this way of looking at toughness, it could also be called ductility.) At full hard a steel might not have much plateau at all and might fracture close to where the plateau starts. Traveling along this plateau is where you would say that the knife would deflect without adding much weight.

Superimpose the stress-strain curves for different steels on top of each other and you would see that some plateaus would be higher than others, and some would be longer than others, but in general the higher up the plateau, the shorter.

There is a tradeoff between strength and toughness. As you increase one you decrease the other. Different alloys and heat treatments will change the point at which the tradeoff occurs and will maintain more of one property as the other one is increased.

The hardest steel will have the highest strength, if the blade shape will allow it to develop this strength. At this point there are a lot of variables involved. A fracture could initiate at a stress concentration in the shape of the blade causing the blade to fail at less than its theoretical strength, where a more ductile but less strong steel could continue to be loaded beyond this point. Also a sufficiently ductile steel can start to permanently deform under bending stress and still pick up additional load as greater parts of the cross section begin to yield. For instance a rectangular bar made from ductile steel will have a maximum bending strength of almost 50% greater than the amount of bending at which it first begins to permanently deform, this is the elastic section modulus vs. the plastic section modulus. A stronger but less ductile steel could carry more load before it begins to permanently deform, but could fracture soon after than before it develops its full plastic moment.

 

[bdmicarta]

To simplify, the amount of force it takes to bend steel of a certain geometry does not change depending on heat treatment. The different heat treatment methods change the point that the steel takes a set, or breaks. It will take the same amount of force to bend a knife to, for example, 30 degrees regardless of heat treatment. Whether it takes a set, springs back or breaks is the result of the differing heat treatment methods.

Correct- strength and heat treatment do not change the inherent stiffness of the material. As long as deflecting 30 degrees does not reach the yield strength of any of the materials in the test, it will take the same force to deflect all of them. This is called linear elastic behavior. Once a material reaches its yield strength then it takes less additional force for it to continue to deflect. Any steel will become nonlinear past this point.

 

[JacksonKnives]

just some questions, since this seems to be a good discussion to continue here... (Rick, if you'd rather I just post a new thread or search around let me know)

I'm a little bit confused about the yield strength/elastic limit threshold. I know that this is defined by stress/strain curves, that stress is measured as pressure (on the inside/outside of a bend?), but I'm having trouble applying that to a blade.

The harder the steel, the farther it will go before the elastic limit is reached. If the graph contained more than one type of steel, point 1 for one steel would be farther up the line than point 2 for the other steel.

Is the difference you're talking about here strictly a function of difference in steel structure (due to alloying elements) and thus only applicable when comparing blades of different alloys, or is the elastic limit for a given steel different with a different temper?

 

[mete]

You've got it . Elastic limit is particular to the metal not the HT or anything else .We ,even we metallurgists , sometimes misuse terms . It's a stiffness measure fe, al ,ti etc each have their own E.

 

[Rick Marchand]

The different heat treatment methods change the point that the steel takes a set, or breaks. It will take the same amount of force to bend a knife to, for example, 30 degrees regardless of heat treatment. Whether it takes a set, springs back or breaks is the result of the differing heat treatment methods.

Not exactly.

From what I understand that is only true within the elastic limit. It the elastic limit is not exceeded by the 30deg mark any steel of the same geometry will behave the same. Once it enters the plastic range(lets say 31deg) heat treat will dictate the outcome. High hard will continue to resist deformation until it breaks. Tempered steel will resist up to a point then yield, take set and eventually break. Annealed steel will deform rather than break but not be able to offer much resistance.

Again, as I understand it, when steel bends it forms dislocations within the structure. Think of steel as stacked marbles and deformation as them shifting around. Red hot steel allows the stack to move rather freely(forging). Cold steel moves less freely and this uneven shifting creates pile-ups and spaces(dislocations). In annealed steel the shifting occurs more readily, while in hardened steel the stacking is much more rigid. If the atoms can't dislocated they separate and the result is fractures. Repeated bending creates more and more dislocations until there is a "traffic jam". This is called work hardening and can be demonstrated by bending a wire clothes-hanger back and forth until it breaks. Here is a neat little web page that shows a bit of what I am trying to say. If I am saying, right. I'm still trying to wrap my head around it, too.

From a forger's perspective, dislocations can be fixed with thermal cycling... fractures cannot. That is why we forge hot and run normalizing cycles before heat treating.

 

[Rick Marchand]

bdmicarta and Mete... beat me to it. I hope I didn't offend you guys too much with my caveman logic. let me know if I am off the mark.:thumbup:

 

[Willie71]

Not exactly.

From what I understand that is only true within the elastic limit. It the elastic limit is not exceeded by the 30deg mark any steel of the same geometry will behave the same. Once it enters the plastic range(lets say 31deg) heat treat will dictate the outcome. High hard will continue to resist deformation until it breaks. Tempered steel will resist up to a point then yield, take set and eventually break. Annealed steel will deform rather than break but not be able to offer much resistance.

Again, as I understand it, when steel bends it forms dislocations within the structure. Think of steel as stacked marbles and deformation as them shifting around. Red hot steel allows the stack to move rather freely(forging). Cold steel moves less freely and this uneven shifting creates pile-ups and spaces(dislocations). In annealed steel the shifting occurs more readily, while in hardened steel the stacking is much more rigid. If the atoms can't dislocated they separate and the result is fractures. Repeated bending creates more and more dislocations until there is a "traffic jam". This is called work hardening and can be demonstrated by bending a wire clothes-hanger back and forth until it breaks. Here is a neat little web page that shows a bit of what I am trying to say. If I am saying, right. I'm still trying to wrap my head around it, too.

From a forger's perspective, dislocations can be fixed with thermal cycling... fractures cannot. That is why we forge hot and run normalizing cycles before heat treating.

You are correct. I was trying to oversimplify things.

 

[Michael Rader]

Hi Grizz. My humble suggestion for you is to use a breadpan like mentioned above, but fill with peanut oil instead of water. Use a submerged spring-clamp or two to hold the blade erect. (Sand and oil might be the ticket here too. Never tried it.) The steam from the water will cool the spine a bit right where you want it to stay hot. I currently like to use two Mapp-gas torches (one held on each side of the blade) and I can watch the colors. That "silver" color is about where I stop. Can get a laser-thermometer to know exactly if you want.

The technique I used in my performance test knives was to weld thick steel bars to some tongs and heat to very dull red in the forge and clamp on the spine for a few minutes (edge in the oil of course.) I would repeat that 4-5 times. I wasn't using 52100, but you can give it a try.

Could always use Fowler/Burke's method of using a torch on the edge only and quenching only the edge. Then the discussion about tempered martinsite vs. pearlite comes back into play...

 

[Bo T]

So at this point it seems;

Identical blades of martensite but tempered to different hardness will all reach the elasticity limit at the same degree of bend if they are the same alloy?

It will take more force to reach the elasticity limit for harder blade?

Once the plastic region is reached both blades will deform (not return to true), the harder blade will deform less?

Once the plastic region is reached the harder blade will break sooner relative to the degree of bend, i.e. the softer blade will bend further before breaking?

What in the heck is true stress and true strain?

 

[bdmicarta]

So at this point it seems;

Identical blades of martensite but tempered to different hardness will all reach the elasticity limit at the same degree of bend if they are the same alloy?

It will take more force to reach the elasticity limit for harder blade?

Once the plastic region is reached both blades will deform (not return to true), the harder blade will deform less?

Once the plastic region is reached the harder blade will break sooner relative to the degree of bend, i.e. the softer blade will bend further before breaking?

I'm going to say no, yes, yes, yes. But I'm not sure what qualifying it with "martensite" does to the results.

If you take a give steel such as 1095 and you anneal it, you reduce it to its lowest yield strength. If you heat treat it to full hard then you increase its yield strength to its maximum. If you draw down the hardness a little bit then the yield strength lies somewhere in between. Each of these 3 states will reach the elastic limit at different amounts of bend.

 

[Bo T]

So, in the stress/strain diagrams is the strain correlated with the amount of bend?