Brian,

I just read that thread and I'm as confused by it as you are. Over in the ABS University forums, Kevin has very graciously posted a series of topics on the different steels in a sub-thread called http://www.americanbladesmith.com/ipboard/index.php?/forum/60-heat-treatment-and-metallurgy/

Here you will find a list of primary blade steels and their HT specs, including proper annealing techniques. For my Damascus, I follow the annealing specs for the HC steel in these threads.

So if I am using 1095/15N20 for my Damascus, I anneal the billet as listed for 1095.

It is a way to break up carbides that may have formed even up structures and get a material that is soft and workable with out need to control cooling rates or take the time of a traditional anneal. A spheroidal anneal VS traditional also tends to machine much easier, is less prone to work hardening. My understanding is that the spheroidal structure forms most easily from martensite thus the advice to harden the part first.

MP

The two words I see tossed around the most and either misunderstood or misused are anneal and normalize.

I haven't done an "annealing cycle" on a forged blade in a decade. After a three or four step post-forging, reducing heat "normalizing" procedure it's totally unnecessary.

And all of the reasons explaining those statements would fill a volume.

Just one more reason to sit in on a heat treating seminar.

|quoted:

The two words I see tossed around the most and either misunderstood or misused are anneal and normalize.

I haven't done an "annealing cycle" on a forged blade in a decade. After a three or four step post-forging, reducing heat "normalizing" procedure it's totally unnecessary.

And all of the reasons explaining those statements would fill a volume.

Just one more reason to sit in on a heat treating seminar.

I went to the Intro school in Tex Ar back in 2017. I anneal in vermaculite after thermo cycling the blade. But I just ran across this and thought I would ask. I dont specifically do a controlled anneal on the blade. I was just confused about quenching the blade twice. I would love to go to more hammer ins and classes. Maybee in the future.

When only using a forge for heating, my post-forging sequence consists of this:

1. Normalize at a higher heat. (i.e., in excess of Ac3 or Accm) at least 200°F higher than decalescence to ensure all phases are dissolved.

2. Reheat to a lower heat, just at decalescence, perhaps more than once, to refine grain size.

3. Reheat to same temperature and quench to approximately 750°F-800°F, and air cool.

4. Reheat to dull red, around 1300°F several times and air cool for spheroidal anneal.

I no longer teach, and actually discourage the use of vermiculite and the slow cool type of annealing. This method is what is known as lamellar annealing, and while it is OK for simple steel with less than .83% carbon, it can be very bad for steel with more carbon. Lamellar structures (pearlite) are not as soft as other options, and will be harder on mills and drills.

Annealing is an invaluable process that will give many benefits to the material, and your later operations, but there is more than one type of annealing. The confusion comes in when lamellar annealing is referred to as “full annealing” by industry. Spheroidizing is also very much an annealing operation, and certainly no less of an annealing than lamellar, so it is too bad that industry decided to use that term.

Spheroidal annealing is probably a more thorough anneal than lamellar annealing, it will probably result in MUCH more easy machining and greater ductility. And forget about lamellar annealing deep hardening steels, as they are designed not to make pearlite, thus it is virtually impossible to lamellar anneal something like O-1 or L6. Spheroidal annealing is a better way all around, it just wasn’t understood until modern metallurgy found that the lower temperatures could give us a better anneal.

Rather than lamellar, or sheet type, carbide structures, spheroidizing creates round globular carbides within a soft iron matrix. It is done at temperatures below that of recrystallization and so it leaves all of your grain refinement in place, while lamellar annealing will wipe it out and reset it. Lamellar annealing takes hours, while spheroidizing, at least how I teach it, takes minutes before you can get back to work.

I know most reading this are still waiting for an explanation of that whacky quenching business, but I wanted to make you wait for it. If you normalize something like 1084, you will make pearlite. Pearlite spheroidizes like crap, creating very unevenly distributed carbides that look like strings of black pearls in your steel. The vast majority of bladesmiths are fixated on grain size while totally oblivious to something much more important to cutting performance, carbide condition. I would like to change that, and thus I incorporate a lot of information about carbide conditioning when I teach. Fine, evenly distributed carbides are what you want, and pearlite won’t give that to you.

When I quench after normalizing, I am not going for hardening the blade, I do not want to make martensite. The high temperature interrupt is to avoid both pearlite and martensite. This will leave the carbide in a much finer, and more evenly distributed state. Then, when I apply the spheroidizing heats, the spheroidal carbides will form much more easily, much more finely and much more evenly. This is not a sales pitch, but in my DVD “Kevin Cashen’s Guide to 1080/1084” I explain this in much greater detail but, more importantly, I give you step by step micrographs to show you exactly what I am talking about. I am not a fan of asking anybody to go with it just because I said so, I insist on solid evidence to back up my methods and it is all there in the lab work I have done.

This is a great conversation. I have been doing a lot of work with 1095 as of late. I still prefer my O-1, but being that I use 1095 in my Damascus, I decided to spend a bunch of time and effort with this steel. My current process (it has evolved a bit) is now three normalizations, one at roughly 1850, a second at around 1500-1550, and the final somewhere around 1250. For me, it's a simple matter of turning my gas valve down a notch and letting the forge cool down while the steel takes a black heat. I do quench between normalizations 1 & 2, but not typically after #3.

I do all my rough shaping, (profile, bevels, etc.) and drilling from this point. Then I coat the blade in a descale that is equivalent to white, water-based primer, and do a full oven spheroid anneal prior to quench. Then I clean the blades and temper.

Now my question: I started doing the anneal on a whim. I was having difficulty achieving the hardness Kevin has in his papers for as-quenched 1095 & W-2. I decided one day on a test blade, to anneal the blade prior to quenching. My perception is increased hardness in the quench. Am I imagining this? It really seems like ever since I started annealing prior to quenching, I routinely skate my 64HRC chisel on the edge and spine. When I was just normalizing, rough grinding, and going for the quench, that same chisel would cut the "hardened" blade. The 62HRC would skate, but the 64HRC would scrape and cut.

|quoted:

This is a great conversation. I have been doing a lot of work with 1095 as of late. I still prefer my O-1, but being that I use 1095 in my Damascus, I decided to spend a bunch of time and effort with this steel. My current process (it has evolved a bit) is now three normalizations, one at roughly 1850, a second at around 1500-1550, and the final somewhere around 1250. For me, it's a simple matter of turning my gas valve down a notch and letting the forge cool down while the steel takes a black heat. I do quench between normalizations 1 & 2, but not typically after #3.

I do all my rough shaping, (profile, bevels, etc.) and drilling from this point. Then I coat the blade in a descale that is equivalent to white, water-based primer, and do a full oven spheroid anneal prior to quench. Then I clean the blades and temper.

Now my question: I started doing the anneal on a whim. I was having difficulty achieving the hardness Kevin has in his papers for as-quenched 1095 & W-2. I decided one day on a test blade, to anneal the blade prior to quenching. My perception is increased hardness in the quench. Am I imagining this? It really seems like ever since I started annealing prior to quenching, I routinely skate my 64HRC chisel on the edge and spine. When I was just normalizing, rough grinding, and going for the quench, that same chisel would cut the "hardened" blade. The 62HRC would skate, but the 64HRC would scrape and cut.

Joshua, I think something is amiss in your heat treating lineup that we are not seeing here. Annealing is out of synch and should pretty much be of little value when done after machining and just before hardening. Properly normalized steel should harden more easily than annealed steel. This is inherent in the process and definitions of those operations. The purpose of proper annealing would be to soften the material and remove all non-uniform strain energy effects, this is accomplished by taking carbon out of solution to allow the iron portion of steel to behave as much like iron as possible. Martensitic hardening is the exact opposite of annealing, it is intentionally trapping carbon in solution to make the iron matrix distorted and un-moving. Normalizing isn’t really about either, as much as distributing phases and strain energies as evenly as possible. When properly done, normalizing will create a fine pearlite in 1095, which will very readily go back into solution when reheated for hardening.

There are always some odd exceptions but as a general rule ease of dissolution into austenite would generally start with something like martensite, it is already in solution so it is a no brainer. Next would come upper or lower bainites. Then comes fine pearlite. The lamellar spacing is so small that the carbon once dissolved has very little distance to travel for even solution. After that, comes very finely spheroidized material or course pearlite. The most stubborn would be heavily spheroidized material, which has very large spheroids paced at greater distances in a carbon free iron matrix. Spheres a rather stable structures and the carbon has a long way to go once it is free. It is possible to over spheroidize and render the steel incapable of maximum hardness. I have done it more than once when exploring new normalizing regimens in conjunction with industrial spheroidizing. I call this condition carbon locked and it has happened to several supplies of alloy steel that has circulated among knifemaking supplies in the last decade, I know because I was often called in to help solve the problem; and my solution was to normalize before hardening.

Now what could be the issue with your hardening troubles is hard to say from a distance, but there are a couple possibilities that I see. 1850°F is rather high for normalizing 1095, I personally would cap it at 1700°F and even then, would rarely go that high, with something around 1650°F being my comfort zone. Cycling for refinement can be much lower, like in the mid 1400’s range, but not excessive; excessive cycling at lower temperatures tends to bunch carbide up, rather and even it out. I would never leave 1095 in a forge for any amount of time during a cooling cycle. I bring the forge to a low burn and cycle to dull red for spheroidizing. If, on the final heat, that forge is a dull red (1200°F or less) then I may allow the steel to cool in it. From a pearlitic state, or very fine spheroidal, 1095 should be able to reach 65HRC with no soak anywhere from 1440°F to 1475°F.

|quoted:

Joshua, I think something is amiss in your heat treating lineup that we are not seeing here. Annealing is out of synch and should pretty much be of little value when done after machining and just before hardening.

I would never leave 1095 in a forge for any amount of time during a cooling cycle. I bring the forge to a low burn and cycle to dull red for spheroidizing. If, on the final heat, that forge is a dull red (1200°F or less) then I may allow the steel to cool in it. From a pearlitic state, or very fine spheroidal, 1095 should be able to reach 65HRC with no soak anywhere from 1440°F to 1475°F.

Thanks for the detailed response. I thought it was something of an anomaly. For the record, I don't ever leave the 1095 in the forge to cool. Been there, done that, lost the band saw blades. We had that conversation a couple years back when I had trouble cutting my billets. You straightened me out on the annealing process and I have never looked back.

The blades come out of the forge and go onto the rack to air cool. Once black, I quench in the slack tub. I have one of those handy little infrared thermometer guns. Not terribly accurate, but below 800 is when I go to quench and reheat.

The 1850 is roughly where the forge is already, so it seemed like a good place to start. I can lower it down and see if that changes anything, but I already have my doubts that it will. I quench when the paragon says 1475. It might loose a degree or so while I pull it out and quench, but the transfer time is not enough to get below 1440. Why The blades didn't fully harden is still perplexing. Maybe adding another normalizing heat pre-quench would change things? At what temp do you do the pre-quench normalization?

For increasing responsiveness to austenization for hardening, you do not need to go as high unless you have serious carbon lock from heavy spheroidization or segregation. A normal cycle in the 1450°F-1500°F range should be good. The idea is to pre-dissolve carbide so that it is ready for solution on the next heat.

Kevin, since everything I know of must be in some sort of "condition", if you have avoided both pearlite and martensite in this process, what condition did we achieve?

"When I quench after normalizing, I am not going for hardening the blade, I do not want to make martensite. The high temperature interrupt is to avoid both pearlite and martensite."

|quoted:

Kevin, since everything I know of must be in some sort of "condition", if you have avoided both pearlite and martensite in this process, what condition did we achieve?

"When I quench after normalizing, I am not going for hardening the blade, I do not want to make martensite. The high temperature interrupt is to avoid both pearlite and martensite."

The "B" word, that I don't often talk about. It is one of the reasons that I often refer to Bainite as the "unholy lovechild of pearlite and martensite." Upper bainite, while having almost every undesirable property that I wouldn't want in my knife, does have its uses. First- it isn't pearlite, second- it isn't martensite. What it brings to the table is a quasi carbon-trapped-in-solution state without all the unnecessary stress of martensite. It is formed relatively quickly at around 750°F-800°F, as compered to the much larger hold times for lower bainite. You could just full quench to martensite, but I have found that, after all my careful setup with normalizing, I get less distortion issues, now and later on, if I avoid full hardening a scaly and pitted forging.

|quoted:

First- it isn't pearlite, second- it isn't martensite. What it brings to the table is a quasi carbon-trapped-in-solution state without all the unnecessary stress of martensite.

That does it for me!!

I haven't been interrupting at that high of a heat, but thanks for explaining the benefits of doing so.

|quoted:

I often refer to Bainite as the "unholy lovechild of pearlite and martensite." Upper bainite, while having almost every undesirable property that I wouldn't want in my knife, does have its uses. First- it isn't pearlite, second- it isn't martensite.

When does it turn into Dunrite? <img src=' http://www.americanbladesmith.com/ipboard/public/style_emoticons//tongue.gi f' class='bbc_emoticon' alt=':P' />

Seriously though, you're starting to make sense, and I can actually follow what you are saying at least somewhat enough to understand how to teak my process.

Thanks Kevin.