Heat treating and Normalizing confusion

G

Greenhead

Active Member
Hi all,
I am wanting to try my hand at heat treating some 1095 for my folders. However, I am confused. I am trying to get a hamon and I understand how tedius that can be but.

My confusion is that I see all this stuff about normalizing first. Is this just if your forging the stock? I was just going to rough grind it and then heat treat it. So, what's the deal with normalizing?

Thanks in advance!

Jamie
 
Normalizing is a must after forging but may or may not be necessary for stock removal. Personally, I always Normalize the steel before hardening. I feel more confident that I'm getting the most from the steel by doing so.
Normalizing sets the steel up to be hardened properly. I find 1095 to be very finicky so I don't use it unless a customer specifies it. When I do use it, this is how I treat it.
1. Heat to 1575, soak for 5-10 minutes, & cool to black.
2. Heat to 1475, soak for 5-10 minutes, & cool to black.
3. Heat to 1375, soak for 5-10 minutes, & cool to room temp.. At this point the steel is ready to quench to full hardness.
4. Heat to 1475, soak for 10 minutes, & quench.
5. Temper.

There are definitely other ways to do it but this is what works for me and my equipment.
 
Depending on the source of the 1095 you may or may not wish to normalize. Some batches have some pretty heavy segregation that can be improved even in stock removal knives. Also even with stock removal there could be uneven strain effects that could result in warping, but for this normalizing is overkill when a simple stress relieve after machining/grinding will suffice- heat to 1200F, soak and air cool. You may also see some effects from normalizing of you are shooting for a hamon. Normalizing can increase the responsiveness to austenitizing and thus shortening your soak times, and if cycles are introduce to refine grain size (technically not normalizing) you will lower hardenability and make your clay more influential.
 
^^^ Looking forward to exactly how Mr. Cashen explains this. It should provide confirmation for me on how I'm thinking about some things. Not too long ago, he helped with a different thread and I "think" I have at least one or two more ducks in a row because of it. I was confusing/interchanging thermal cycles with normalizing. I think the grain refinement falls under thermal cycling and isn't necessarily normalizing (a bit higher temps starting out and descending in temp, thus "refining" the bits inside).

Most of the time this all can make my brain hurt...but soooo glad for guys like Mr. Cashen. I try to separate what exactly the procedure is trying to accomplish in order to keep some of it all straight in my head. And I'm sometimes successful at that ;).

Jeremy
 
Normalizing allows the steel to change phase and back to remove stress from the steel and smiths put a lot of stress in there blades by beating on them with a hammer. Picture the matrix of iron crystals being twisted and and pulled out of alignment by the hammer blows. This changing of the iron phase (change in crystal shape) allows the new crystals to reform an the boundary of the old crystals and form a new alignment without the stress that was produced by the hammer blows. This happens both as the steel is heated above the austinization point and when it is allowed to slowly cool back to normal temperature.

Doug
 
Knifemaker.ca said:
I know I'm going to regret asking :34:but could you elaborate on that a bit please? :les:
Click to expand...

Ohh, you may regret giving me that much latitude:3:. Normalizing by definition needs to be above Ac3 or Accm so we are talking much higher temps to homogenize than what you would want for grain refinement. You see, in order to homogenize carbon content all carbides are put into solution, and when all carbides are into solution there is nothing holding the grain boundaries in place any longer; so actual normalizing will almost necessarily involve grain growth. But the upside is that the growth will result in a more uniform grain size, and uniformity is equally as important as the size itself. So the grain may be huge but if they are huge but all the same size, and you then refine the size with the same uniformity, you will be much better off than if you just went for finer size without uniformity. Always remember that large grains grow at the expense of smaller grains (they devour them) so if you have uniform grain size the grains will reform at a uniform rate, but if you get out of balance on size the larger grains will grow larger by eating their smaller neighbors and you will have a vicious circle.

Add to this that carbide size is even more critical to a fine and stable edge than grain size and proper normalizing becomes much more important. So you go ahead that heat that steel to 1600F or better to get actual normalizing temperatures and refine those carbides, and also homogenize the grain size (albeit larger, but even). The following cycles that bladesmiths do are far to low in temperature for it to actually be called "normalizing" but instead are simple thermal cycles for the purpose of refining the grain size, e.g.- anything below 1500F, or cooling any way other than air (like quenching) cannot technically be called normalizing with any steel except for 1080 or 1084, and even then it must be air cooled or it doesn't qualify.

So much of this low temp nonsense that you may read in magazines is just that, and is oblivious to the all powerful carbide which much be tended to properly to make a really fine edge. Grain size is the most easily fixed issue with the steel, but you really need to have you act together, with some knowledge to back it up, to deal with carbides. Excessive low temp cycling takes a tunnel vision focus on grain size while ignoring carbide size which will suffer from such treatments.
 
So, when I think I'm normalizing I'm using something like 1650 - 1550 - 1450 - all with air cool, I'm actually covering normalizing with the 1650, and doing grain refinement with the 1550 and 1450 right? My problem was not so much with process, but with nomenclature?

Rob!
 
Knifemaker.ca said:
So, when I think I'm normalizing I'm using something like 1650 - 1550 - 1450 - all with air cool, I'm actually covering normalizing with the 1650, and doing grain refinement with the 1550 and 1450 right? My problem was not so much with process, but with nomenclature?

Rob!
Click to expand...

The 1650F or the 1550F could be normalizing, depending upon the steel. Normalizing is heating above Ac3 or Accm in order to dissolve everything and achieve total austenite solution, something that you would never do with any other heat treatment of a hypereutectoid. Also for it to be normalizing cooling must be in still air. Heat treatments below Ac3 or Accm may look exactly like normalizing but technically can only be accurately described as "thermal cycles". Normalizing has no obligation to reduce grain size in order to be called normalizing, but it does rearrange carbide. Grain size does matter, quite a bit in fact, but I down play it a bit for a few reasons. The first is the even greater importance of carbide condition. The second is the fact that if the grain size is not uniform it almost cancels out whatever size it is, And third is how it is so misrepresented among knifemakers that it has became what I call the bogeyman of knifemaking as if it is this horrible beat that only the most skilled of smiths can tame, when in fact controlling grain size is such a no brainer I don't see it as a skill as much as something we can just take for granted that anybody can mange it. It is quite simple- don't overheat your steel. And if you do, just reheat to the correct temperature and allow the steel to fix itself.
 
I've been waiting for this to pop up somewhere. How, if at all, does the need for uniform grain size relate to the notion that an array of grain sizes is needed for the highest performance levels, particularly in low alloy or plain carbon steels? I have seen the array of sizes pushed as every bit as important as fine grain size. Just for clarity, I have stated in other places that the grain size will be an array to some degree no matter what, thus the need for the rather complicated ASTM average grain size procedure. However, I don't feel there is a need for a wide array of sizes, and if we could dump a bag of 1" marbles into the normalizing machine and get all 3/8" marbles out, it wouldn't hurt anything, and may be beneficial.
 
me2 said:
I've been waiting for this to pop up somewhere. How, if at all, does the need for uniform grain size relate to the notion that an array of grain sizes is needed for the highest performance levels, particularly in low alloy or plain carbon steels? I have seen the array of sizes pushed as every bit as important as fine grain size. Just for clarity, I have stated in other places that the grain size will be an array to some degree no matter what, thus the need for the rather complicated ASTM average grain size procedure. However, I don't feel there is a need for a wide array of sizes, and if we could dump a bag of 1" marbles into the normalizing machine and get all 3/8" marbles out, it wouldn't hurt anything, and may be beneficial.
Click to expand...

This is a question that was often brought up in the old edge packing debates- if that really fine grain made from smashing or squeezing iron atoms will give you a stronger and tougher edge, why not do it to the whole blade? What you are describing is folks who really have a very limited understanding of austenite grain size and functions revealing their information deficits. Uniform grain size is very important for several reasons, the first is encountered in the grain refinement operations themselves. Large grains grow that the expense of smaller grains so that if you have a large grain in the mix that is surrounded by smaller grains you have created a monster and surrounded it with all the food it needs to grow on the following heats. This is covered quite well by Grossman and Bain in several of their texts where it is referred to as “exaggerated coarsening”, “uneven coarsening” or in some older texts “germination”. Addressing the edge packing nonsense specifically certain rates of deformation must be maintained in order to not create this problem, which sort of set the stage for guys who lightly tap their edges below Ac1 to once again create problems that they can then address with all kinds of quirky things to impress less knowledgeable audiences.

Are grain sizes ever perfectly uniform? No, but you can get them in the same ballpark with proper treatments. The real reason for the ASTM methods of determining grain size, rather than just snapping steel and just looking at it like bladesmiths do, is that even if the grain size was perfectly uniform the grains are arranged in a 3 dimensional array. Toss all of your ½” marbles into a bucket and fill it with epoxy. Now break the column of marbles in half and you are going to get a view of the curved outer surface of the marbles on the fractured end, very hard to measure the diameter of (especially since grains are not perfect spheres like marbles), but you will also get little faces of marbles peeking out from between the others that are partially covering them so that the light is going to be scattered in every direction, making the who thing hard to see much as less measure. So they came up with the idea of cross sectioning the piece and measuring the area of the exposed, level faces at 100X magnification and formulas to come up with the average grain size throughout. I have heard some bladesmiths say they know their ASTM grain size, not only do I doubt that they do, I doubt they even know what is that they are claiming to be able to do since they have no knowledge the actual ASTM process. Water cooled diamond saws, proper mounting of specimens, reagents capable of revealing grain boundaries in martensitic samples (a real killer), a metallograph set up for the measuring, and a knowledge of the various methods of determining the grain areas involved, it all puts it a bit beyond the average bladesmith shop. Some could say that they have a proper lab do it, but I have never seen any specific labs cited, for some reason that is always a secret, but more problematic is that the descriptions always seem to resemble the old Shepherd fractured grain size determinations which would indicate that in reality we are just back to breaking steel and looking at the fractured end. Hint- there are no special limits around 10 in ASTM, but the Shepherd standard set only goes to 10, very telling.:3:

Edited to add- Hmmm, I just had to go back in and add spaces between every word on this post after it was entered, but it kept reoccurring until I had to delete the post and start over. never had that happen before at KD??
Edited again- hmmm still isn't fixed..
Edited yet again to add- Finally got it to take the spaces by switching from I-Explorer to Firefox, perhaps this is something the Admins need to know about??
 
Last edited:
Kevin, First of all, let me say THANK YOU!!! for all the great help you've been thru your posts. Just amazing to have a resource such as yourself for help. Now, you mention looking at grain structure at 100X. Would it tell me anything if I took two samples, one a known sample of fine grain, then along side put a sample of steel I'm working with to compare grain structure?

Would a known fine grain be an old Diamond or Nicholson file broken to give a view of broken edge to compare to broken edge of working steel sample?

Thanks again for sharing so much of your knowledge,

Ken H>
 
KenH said:
Kevin, First of all, let me say THANK YOU!!! for all the great help you've been thru your posts. Just amazing to have a resource such as yourself for help. Now, you mention looking at grain structure at 100X. Would it tell me anything if I took two samples, one a known sample of fine grain, then along side put a sample of steel I'm working with to compare grain structure?

Would a known fine grain be an old Diamond or Nicholson file broken to give a view of broken edge to compare to broken edge of working steel sample?

Thanks again for sharing so much of your knowledge,

Ken H>
Click to expand...


As always, you are most welcome. Yes Ken, don't get me wrong, just viewing the fractured end grain can be very valuable it just can't give you an official ASTM grain size. A quality file is a pretty darned good control sample. The thing about naked eye or low mag fractured grain examination is that once you get past around 5 or 6 you really aren't examining grains as much as you are observing what their overall effect has on the smoothness of the fractured plain, so then you are looking for overall texture. But one needs to remember that since you are looking at texture, the phase (or more precisely-hardness) will play as large a factor as the grain size. Softer steel will fracture involving slip mechanisms and give you a rougher looking surface than a fully hardened piece of steel. This is also another error in interpretation that contributed to the edge packing confusion. Simple steels tend to form unwanted pearlite in the thicker areas away from the edge, then guys would "edge pack" and then break the steel and see the very smooth fracture along the edge as compared to the spine and scream eureka. There is so much mythology in knifemaking that has been spawned and reinforced by simple misinterpretation of the results, and this is what makes so much of it so hard to correct- who are you going to believe, me, or your lying eyes?:3:
 
Kevin, there is not doubt who I'll believe - YOU!!! Not my eyes, because I KNOW I don't know enough to actually "know" what my eyes are seeing.

The thing about naked eye or low mag fractured grain examination is that once you get past around 5 or 6 you really aren't examining grains as much as you are observing what their overall effect has on the smoothness of the fractured plain
Click to expand...

Are you saying the small "grains" (pebble looking surface) isn't the true grain of the steel, but is actually the "texture" which is a factor of both grain and phase?

I'm understanding you to say the observed texture is important, so while breaking the sample will distort the actual grain, since both the file and sample are broken, the resulting visual texture will be a good guide to compare the sample to know file texture. So, perhaps it's important to use the term "texture" rather than "grain" when commenting steel blade?

When you say "5 or 6", are you referring to mag of 5X to 6X power?

What would be a good mag level to visually make a judgement on the grain size with broken file and sample side by side under microscope?

But one needs to remember that since you are looking at texture, the phase (or more precisely-hardness) will play as large a factor as the grain size.
Click to expand...

If I measure the "pebble" size at 45X (or a reasonable power), is that any relation to the actual "grain" size?

Thank you again for your comments,

Ken H>
 
ASTM grain size numbers do not refer to the magnification, all the numbers use the same 100X magnification, but instead is a number representing the average grain size based upon how many grains are in a given area. So it is like gauge measurements in sheet stock or shotguns the smaller the number, the larger the grain. So ASTM grain size 1 is HUGE, while 9 is much finer and nicer.

No, those individual pebble looking things are indeed that outer surface of a crystalline grain, but if you can make it out as “pebble like” with the naked eye or low magnification it is a grain size that is too large for what you want in a knife. Think of your “pebbles” as a paved road with gravel in it and you are 100 feet above it looking down on it. If the gravel is fine enough all you can really make out is the overall effect of how the pebbles scatter the light that is reaching your eyes, but if you can actually see the pebbles from 100 feet away then somebody put rocks in the pavement instead of pebbles, and that road will not last through a single Michigan winter (and won’t be much fun to drive on in short order). 45X magnification would mean that you are now 55 feet above the road surface and perhaps you can make out the individual pebbles, if they are not too fine. But what can really mess you up is if the pavement has a smooth packed surface or if it is hot and gooey and the surface is all roughed up from the tar sticking to tires. This is the difference between fully hard steel snapping clean and pearlite snapping in some places and stretching or shearing in others. If you can’t cleanly identify the individual grains the texture is determined as much by the hardness as it is the grains themselves; softer steel does not break as cleanly.

But now we have looked at the pavement from different distances, and without a standard reference each vantage point will have us seeing a different pebble size despite the fact that their actual size has not changed regardless of our perceptions of size or texture. So in order to actually determine their size we are going to have to get really close to measure them. Since we can’t lay a ruler on the road in this case, then we must all agree on a distance that all of our observations will be from. ASTM chose 100X for this, so in our analogy of our normal distance being 100 feet away, we will do all of our actual observation at 1 foot away from the road surface. At a 1 foot distance we can then see the size of almost any pebble, and what we do is mark out an area (say 6 inches by 6 inches) and count the number of pebbles in that area to get an idea of the average pebble size in the pavement. Where our analogy differs is that if the fractured road surface were like the fractured face of steel there would be voids, pits and jagged peaks covering a height or depth of three feet and we could only see the very top or bottom of them, the rest is blurred or in darkness, and if the stuff is soft this problem would be even worse. So in order to actually count the pebbles/grains we pick a section and grind it completely flat and polish it so that we can see the outline of all the pebbles.


If you are using the broken file trick any magnification that will assist you in seeing the surface will work since you have a standard (the file) by which to evaluate the grains, but if all you have is you broken blade to try to assign a standard grain size then you need to use 100X.


P.S. In truth you shouldn't believe me or your own eyes, but instead tirelessly pursue the facts until you have the correct answers. (hint- the really good facts tend to crush your beliefs!)
 
Last edited:
ok, i have read and pretty much understand heats/times/cooling and what it will yield. If the steel(O-1, thickness<3/32") i am using is from a known maker and is fully annealed and completely spheroidized from the factory, am I going to gain anything doing pre-hardening heat treat?
scott
 
scott.livesey said:
ok, i have read and pretty much understand heats/times/cooling and what it will yield. If the steel(O-1, thickness<3/32") i am using is from a known maker and is fully annealed and completely spheroidized from the factory, am I going to gain anything doing pre-hardening heat treat?
scott
Click to expand...

I think as long as the steel is not heavily spherodized, as in Aldo's 52100, you won't gain anything. More than likely it has already seen the cycling at the mill. If it is heavily spherodized...then the pre hardening normalizations are practically a necessity. For me it all boils down to...do I know what the internal steel structure is when received. If I don't know.....it gets the whole routine. If I do know...and trust the source....I go right to hardening. The OCD in me almost demands that every blade gets the cycling.....letting go of that on some steels is hard for me to do!
 
Only if forged. Why would being heavily spheroidized require normalizing before hardening? PG 01 comes ready to harden and temper. After grinding, it should be stress relieved, but mainly to help prevent warp. Spheroidizing has nothing to do with grain size, so why would you cycle it? All the PG 01 I have ever checked had a grain size as good as one could want, HTed as is. Spheroidized carbon is slower to go into solution than is lamellar carbon, but 01 needs a good soak before quenching anyway.
 
You must log in or register to reply here.
Back
Top