Dan,

I'm following what you're saying, I think. I would like to propose this thought. When finish grinding,after heat treat, we are very careful not to ever get the cutting edge hot. One indicator that we are too hot is watching to see if any part of the edge turns color. As you know, this only takes a fraction of a second for it to turn brown or blue. Does that make the edge soft? I would assume so to some degree. In that short time that it takes to turn colors, I would also assume that I messed up my blade. Not much time involved there. Of course, the material is very thin and I think the color indication would be correct and an accurate assumption.

Does the thickness of the material play a part? I think it does. From what I have read about industry standards, they establish the draw back temperatures based on a given thickness, lets say one inch thick. I say that because I believe that is the case. On a 1 inch thick piece, the outside would reach a target temp faster and time would allow the interior to even out. We're not using 1 inch thick blades but the principle applies never the less.

To me the draw back temperature and procedure, as is generally done in an oven, is primarily for the edge of the blade. Other important steps would be necessary, whether before or after hardening, to address the spine and other parts of the blade. For instance, if you edge quench, you've address the spine area by simply not hardening it. If you full quench, you address the spine area by FURTHER drawing the spine with a torch or some other means. There are a number of ways that could be employed to get a similar (notice I did not say exact) end result. The point being, that the internal molecular state will require a temperature to be reached and maintained until the change occurs. The timing of that might (I think so anyway) be affected by the thickness and how the heat is applied and the intensity of the heat, as long as the heat is in the steel's safe working range for that operation. Within the safe working range, a low heat will take longer. A higher heat, while still in that range, will not take as long. I will say, that I believe any change or variable will make minute differences in the character of the blade's performance. That's my view. I sincerely hope that if I'm wrong, someone will show me the error of my ways.

I mentioned a safe working range. Most of the steels we forge have a working range. A range for forging, a range for critical molecular tranformation, and a range for drawing. It will tell you when you've gotten out of that range. Something will fail. Some failures are not visibly apparent and will take testing before they are manifest. A blade that suffers a catastrophic failure is telling us that it was worked outside a safe working range or an important stage was skipped in it's heat treatment.

Dan, what is the structure of the steel on the spine in the first place?

It would be unnecessary to "draw" a spine that was NOT martensite.

That's the way I think about it.

You are not going to make pearlite tougher by drawing a pearlitic spine. You're just wasting O/A.

Guess what I'm saying is that I've never seen any benefit to drawing a spine on a knife that was not fully martensitic.

And we should all know that tempered martensite is far, far tougher than any un-hardened pearlite.

I'm just tossing this in here to further refine and isolate the conversation.

Personally, I've only drawn one knife in my 16 year career of knife making. And when I think about it, it was differentially hardened 5160, which may - MAY! - have achieved some air hardening on the spine, if, in fact, I did get the spine austenized in the first place.

Lots of variables, and, like a few people here know, my mind is FULL of variables. <img src=' http://www.americanbladesmith.com/ipboard/public/style_emoticons//blink.gi f' class='bbc_emoticon' alt=':blink:' />

(I just posted all this before reading Lin's response.)

Dan, This question has me fascinated. While knowing that you are an expert and adept at the vagrancies of the steel,I understand that when it gets down to these subtle nuances of heat treatment, it bears discussion. I am so glad to see it. I need this very much myself. In fact, very often I talk to and discuss the test findings of some of the makers here on the forum. We usually talk privately by phone. We'll talk about specific steels and their potential problems, symptoms, etc. I find this welcome and very helpfull when it comes to cataloging all of the terminology as well as the unique issues with heat treatment. Karl is one of those who shares in this.

I dont mean to be stating anything in my above post as 100% fact, but it is my view based on what I see. I want to post a photo or two of the recent shop vist at Jimmy Crowell's. I think they bear on this.

Due to the time element, we had to heat treat this blade the down and dirty way. Here Jim is drawing the temper. Notice I did not say drawing the spine. He's using a small torch applying as even a heat as possible and only so hot. He "painted" the blade to a dark straw color. The entire blade. He applied the heat to the spine because of it's mass, allowing the heat to travel or "run" to the edge. He divided this step from tne "drawing the spine" step to emphasize that they are indeed different steps. This step, although applied to the spine is actually to draw or temper the edge. He applied just enough heat so that there was no necessity to dip the blade to stop the running of the colors even though that is ok and the bucket was ready.

In this photo Jim is drawing the spine softer, but to the dgree he thinks this blade would need to perform its intended purpose. Here is where the confusion often is and the opinions might differ.

I might add that this blade was used very successfully in a cutting competition later that day.

There was a lot of good info exchanged there and I appreciate Jim & Lin sharing their expertice with us. I think that we all learned a lot while there. <img src=' http://www.americanbladesmith.com/ipboard/public/style_emoticons//wink.gi f' class='bbc_emoticon' alt=';)' /> I know that my next competition knife will be better because of it.

Gary

It all comes back to the basic rule of time = temperature, however what is missing from that rule is the amounts. Temperature will always trump time since a whole lot of time will equal just a little bit of temperature. At austenitizing temperatures, time is rather short because carbon atoms are very mobile at this heat and it takes very little time for them to move to where we need them. Normal tempering temperatures are around three and a half times less than austenitizing, not only do those atoms naturally move that much slower, but you also have the issue of dealing with an entirely different atomic stacking arrangement at any temperature below the A1 range (from around 1,000F to 1335F.). This lower temperature stacking is not very receptive at all to carbon movement, this is why hardened steel stays hard and why annealed steel is soft. All of this adds up to extremely sluggish carbon diffusion, movement requiring time on the order of hours rather than minutes or seconds, but rather small increments of temperature will still drastically reduce this time.

There is something to the idea of how long it takes for the size of steel to assume the tempering temperature, but that is actually accounted for in most recommendations and tempering times, just as soak times are assumed to be at the target temperature. What does come into play is the driving force of thermal levels. It takes steel many times longer to come to temp in a 400F oven than under a 6000F torch flame. However, if your heat source is 400F, once you are at 400F. it will still take a minimum amount of time for carbon atoms to diffuse through 400F steel, so the two hour suggestion applies if the steel is ¼” or 4”. That carbon needs to move only the smallest of fractions of microns anywhere in the steel so while at temp the steel dimensions are irrelevant.

Dan touched on something to keep in mind- that at 700F and above larger carbides are formed until structures are eventually spheroidized. Most of the steel that we work with will not truly “air harden” as often is believed, but will instead form very fine pearlite and the dreaded upper bainite. Upper bainite is a weak structure to be avoided. Pearlite is tough but nowhere near as ductile as a spheroidal structure, and while true thorough spheroidizing does need more time than a pass with a torch, the carbide lamellae of pearlite will break up into spheres when heated to 900F and above to look like strings of black pearls. So if heated high enough to induce some spheroidization most structures should gain some ductility.

But here we can thank our lucky stars that we deal with steels that are not prone to secondary precipitations which would result in tempering embrittlement in this range, otherwise all bets would be off <img src=' http://www.americanbladesmith.com/ipboard/public/style_emoticons//wink.gi f' class='bbc_emoticon' alt=';)' /> .

All,

Thanks to any who have responded. Why do we do what we do? What do we really know? I think we all know that if you have a martensite blade (thanks Karl), then drawing the back or spine of the blade will strengthen (increase ductility) the blade. One person a few weeks ago that I respect indicated that the reason we draw the back twice was to maker sure we got an even draw across the length of the blade. I argued that time was also a factor. As Kevin pointed out it is both time and temperature. Carbon movement and spherodization in particular take some time and temperature. I personally want to know what is happening when I draw the spine inside the steel. Knowing the variables I can change my technique to get to that intended outcome for that knife. In other words, I want to make the best and strongest knife possible given the knife's intended use. So I was hoping we would get and idea of time and temperature variables on this common procedure that most of teach and do. Kevin added some great information as usual. IF I was really focused on ductility in a sword or a knife, then how hot I get the back of the blade and how long I can maintain that heat without affecting the edge is very important.

None of us would expect a fully hardened blade to bend 90 degrees if the blade were not tempered. At the same time none of us would teach a new bladesmith to oven temper the blade and not do a separate tempering draw across the spine and then have the blade bend 90 degrees. (I am not suggesting all knives should be able to bend to 90 degrees. I am just using it as a criteria in this example). I just want to know best procedures based on a metallurgical knowledge base. What I would love to see is a time at temperature graph showing how hot and how long to a certain ductile outcome. For example, if I heat the spine to 600 degrees how long till certain changes occur in the spine versus I heat the spine to 800 degrees. I would strongly argue that most of what I do know is that the time factor is different and the length of time is likely longer than we thought. I can present some data on this in terms of temperature but the functional time factors are alluding me since most of the articles deal with temperature.

I can relate some practical experiences. As an example of what I am talking about I was attempting to draw the tang portion of a blade where I would be pinning bolsters in place. On this knife the cutting edge (plunge line) is only about 3/16ths to a 1/4 inch from the plunge line. Since I would be drilling in this area and it was hardened, it was important to draw the blade back as close as possible to the plunge line. Protecting the blade with a heat sink in this area is not difficult. I drew the area a couple of time and still found that drilling was problematic. I went back and held the heat in that area of the tang (as well as the rest of the tang) for much longer period of time. After that process drilling was much easier. Time was a factor as much as heat. This has happened frequently in different ways over the years with drilling or filing as a gross measure. My understanding of the tempering literature seemed to suggest I was correct in my assumption that I needed to have the section or area of the blade not only at a certain temperature but also maintain that heat for a period of time. But I am by no means all that knowledgeable and this seemed to be a good discussion point.

Dan

Great discussion!

Russell

I cant tell you how valuable this thread has been. Although I think I have an understanding of the heat treat process, I really can learn by peeling back the layers of this onion. The details and fine point of any subject is where its at. Thank you Dan for bringing up the subject. I'm anxious to see more.

As an example of your point about TIME: In the above photo of Jim drawing the spine back with the torch and the edge in water you'll notice that the full tang blade has the holes drilled already. Just before putting the scales on, we decided to enlarge the holes and use tubing for the pins which would require reaming. We had to use a carbide bit to do it. Even though the tang was drawn back, it still was somewhat hard. That means the required time at the necessary temperature was not allowed for it to be soft enough for our liking. Even though we were short on time, trying to get the knife finished as a demo, we knew enough to go back to the torch and further draw the ricasso area blending it back into the handle.

In a way, our trying to drill and finding it too hard was a test. It tested too hard. Similar to how we test with a file. Since we cannot visibly see whats in the steel, short of breaking it, we can only know by testing. How a bit or a file cuts or dont cut it, how well it holds an edge, does it show to be brittle or soft, etc. We often only can know after the procedure and by testing. In a way, we have to reverse engineer, removing or isolating the variables to the point that some simblance of quantifying can be reliable.

Thanks Dan!

These are the kind of discussions that we as knifemakers like to chew on.

Wish steel was as simple now as it was when I started on this stuff back; well, a long time ago. It's scary when you finally realize what you don't know.

If I draw the spine, I push it till it goes to dark grey. Three times. I just finished a class of new guys who did it that way; no ckacked or broken blades.

Lots of procedures work well for a knifemaker. But; the final word. TEST!!!!! Bend and break a lot of blades. You will find out if your process works. Or not.

J R,Dan, Jerry, Lin, Brion, Kevin, myself; we all approach it in a slightly different manner. Each of us is pretty confident that what we are doing works.

Because....we break it and look at his little innards.

Can I do it better; I hope so. But nothing changes in my process until my testing proves that I am getting better results from a change.

Bend em, break em, drill holes, whatever. But if you don't wreck it; you won't know.

Drawing the spine works for it's intended purpose, several hundred ABS test knives attest to this.

Does it work for you? Only your blades can tell you.

M

I thought I might share some data from a reliable source (Some data from ASTM Tempering http://www.asminternational.org/pdf/spotlights/tempering.pdf). If you go to this site you can find a rather long pdf file that has lots of information on tempering. Of particular interest is the graph on time and temperature for tempering 1082 steel. While the graph is more informative, I thought I would share and translate some of the data into degrees F as opposed to C. From the graph it appears following hardening the steel is about 66/67 on a RC. The following are the data given and hopefully I did not screw up the data as I translated it from C to F and from seconds to minutes.

10 seconds 1 minute 10 min 30 minutes

401 F 65 64 63 62

599 F 62 59 58 57

797 F 56 55 50 48

These data are not derived from a piece of steel with the characteristics of a blade (long, flat with a thickness or 1/4 inch or less). I am assuming standard 1 inch bar. But if you follow the trend you see that at 401 F for 10 seconds the hardness falls from 67 to 65, and that from in one minute it falls from 67 to 64, and that after 30minutes it drops from 67 to 63. As the temperature goes up the drop in hardness occurs faster. Example, at 599 or 600 F at 1 minute the hardness has dropped from 67 to 59 but it takes another 9 minutes to drop another RC point to 58. At roughly 800 F the drop is faster at 10 seconds, 1 minute and 10min and 30minutses. So I am guessing that when I draw the back of a blade I am getting it up to around 600 to 800 degrees F. This graph alone or the data would suggest that I want to hold that temp on the spine for more than just a few minutes if I want the hardness (only one measure of ductility here) to drop to the low fifties in hardness. If all I am getting is a blue color (roughly 550 to 600 degrees) then I need to hold that temperature longer. Also, these data may not reflect actual hardness levels achieved in a shop out of a forge and what percentage pearlite or bainite is retained in the steel. Hopefully, there is not much retained austenite after oven treatments.

Again, this is only one measure of ductility and it is up to the smith to decide the desired outcome. What it made me realize is that this is not just a get it blue and move on with the torch thing. Both temperature and time are important. It also made me wonder about using an oxy-acet torch rather than a propane torch to increase my heat source to keep the spine at temperature longer.

Mike described doing it 3x and likely exceeding 600 degrees based on the color description. That might be closer to time and temperature conditions needed.

However, I think there is much more to this story. Would longer times at 600 F actually achieve the same thing as shorter times at 800 F. In terms of hardness that would be the case but not necessarily in terms of other micro structural changes (spherodization).

This makes me want to really attempt to get some pragmatic measures in the shop where all dimensions of the steel are the same but the time and temperature of the draw is varied.

Dan

Sorry the chart got squished when I posted this. Still, I think you can follow it with the description.

Great thread topic Dan! Most all of my carbon steel blades are just edge hardened. But, I don’t care to see a temper line on most of my damascus, so I often do a full quench and draw the spines on those blades.

I've always had the impression that time at temperature must be allowed for when drawing the spine. Like Mike, I draw at least three times. I also push the color along the top of the spine to gray. I use an acetylene torch for this work. The acetylene torch isn't used with the intention to bring the temperature up quickly, but to be able to push the heat down close to the water. I take my time bringing the blade up to temperature. Once I have the blade up to the temperature that I want, I let it set in the water to cool for about 10 minutes, dipping the point occasionally to keep it from overheating. Then I clean the colors from the blade and do it again. If I don't feel like I have done a good job of "painting" colors on the blade on one of those three draws, I'll add another draw to the process. You can't do this too many times, as long as you don't overheat the blade.

I have been following this thread and it is excellent Dan. I also do the same thing Mike and Steve do. It is nice to see data for this. I use a propane torch and it usually takes me at least 20 minutes per cycle, usually more. One question, since a propane or O/A torch get pretty hot, at least 2000 degrees, for the time we are doing this are we doing anything to the grain structure? As you note Dan, is there any thing going on at the atomic level depending on the temp and time? Like you say we need more study and testing. Great thread.

Brion