My First Hardened Files

It is a strange thing to re-create skills and knowledge in the age of the internet. There is at once too much information and then not enough that is applicable. And then there is the annoyance of blogging about something apparently so obscure that your search queries return your own content.

If you want to make a file and have the limitation of a simple forge for heat treating, prepare for a challenge. But what a great challenge for learning this stuff, let me explain. You’ll need to forge your own file cutting chisels, they’re pretty much the unicorn of metal working chisels at this point. So you get the experience of heat treating an edge, and then testing that edge by repeatedly hammering it into a bit of steel. Right away you know if your work is not good enough, the edge tells you if the grain of the steel or temper is bad from how it chips or folds or tears.










I’ve been testing my chisels on a bit of mild 1/4″ bar stock, which is buttery soft compared to the tool steel I’ve been working on, for a number of interesting reasons. Jason dropped a comment in my last post that was manna from heaven, annealing tool steel is not as straight forward as ninety percent of the information out there might lead one to believe. Having previously tested all of my chisels, seeing how many inches of teeth they could cleanly cut with a given bevel angle, I re-tempered.

Normalize, normalize, normalize. That is what I was missing with my heat treatment for these chisels. It was obvious that even though I was hitting the right temps to harden, the grain growth was really bad, leading to an edge that would blow out large pieces. Moving the chisel through several rounds of normalizing (above non magnetic, and then a bit of soak, and then air cool to no longer glowing) made a tremendous difference in edge holding. I probably also did a better job of quenching at not too high of a temperature. What can I say, you’re kind of fucked trying to judge metal temps even in the shade during daylight.

I’ve been drawing the temper to a basic straw color after hardening, seems to be the right range for engraving tools, which I consider these file cutting chisels.


So now I have a few chisels that I feel have basically decent edge holding, how about my problems cutting the file blanks? I knew that my simple annealing process using wood ash as an insulator wouldn’t be the same as a full kiln annealing, but its so often spoken of as adequate I didn’t consider that there was more going on in the structure of the steel. Certainly heating my file steel up to non-magnetic and allowing to cool in an insulated medium softened the steel, that was clear from how easily it bent. The unintended consequence, best explained by Kevin Cashen here:

“….,the ideas of overnight cooling all work fine with 1084, 1080 and other steels with less carbon, but will give you no end of troubles with a hypereutectoid steel like 1095, and is exactly why you found it less than pleasant trying to drill your holes. Heating to above critical (or even non-magnetic) and allowing the steel to slow cool is called a lamellar anneal and it forms coarse pearlite within the metal. This state is made up of extremely hard carbide segregated out into sheets between areas of soft iron, so such a blade will bend and seem soft but will rip drill bits up! No amount of reheating (tempering type) short of a full normalization will have any effect on these sheets. But of even more concern is that leaving hypereutectoid steel that is in solution to slow cool will allow all the carbon in excess of .8% to fill the grain boundaries and cause a very weak an brittle condition that would be rather bad for a knife.

Hypereutectoid steels, and forgive me for not explaining the term which simply means steel that has more carbon than the eutectoid or .8%, really should be spheroidized if you are going to be doing any machining with them. It is best done with well controlled heat treating equipment but it can be approximated in a simpler shop as well. To accomplish this with just a forge you should heat the steel up and normalize it well to homogenize things and get any carbide our of the grain boundaries that may be there. Next reheat it a couple of times at lower temperatures to refine the grain and on the last cycle heat it to just above non-magnet and then quench it in any oil you may have.

Now for the anneal- reheat the hardened blade to red but never allowing it to lose magnetism, do this several times and it will ball all that extra carbon that you trapped in the quench up into little spheres suspended in a soft iron matrix. The steel will be soft and bendable and when a drill or mill hits those little spheres they will just move out of the way and allow you to keep cutting .

This jives with how the chisel felt cutting the file blank previously, little hard spots that wreak havoc on the chisel edge.


So after grinding and hand shaping a couple of files I went through the above normalization and spheroidization. I didn’t wind up with the buttery consistency of my mild steel test piece, but it was an improvement.


My work holding is way off from historical president, same with anvil shape, but you use what you have and make it work. If I can actually get a file to harden properly I’ll have to think about refinement here.

Only last night did I read back through a book written by the early Nicholson File Co, “A Treatise on Files and Rasps Descriptive and Illustrated“, which of course speaks only favorably about their new at the time “increment” cut file that reproduces some of the valuable irregularity in hand cut files. It does however mention one important detail about sharpening the chisel for hand cutting, that of rounding off the newly honed edge by dragging a few strokes of the stone across the edge, much as one might sharpen a plane blade for working abrasive material like bamboo that might otherwise lead to chipping.

Another small detail, but important not only to edge retention, but gullet formation for the file tooth. In some sense then the sharp edge of a file tooth is formed by upsetting the steel, not merely cutting into the file blank.

It took a hell of a lot of sharpening to cut these three files. I then mixed up a thin paste of flour and water, coating the file and wiping off the excess with a swipe of my fingers. In a mortar and pestle I ground some table salt to a powder, which was then used to liberally coat the sticky files, as much as I could get on there. After drying they kind of looked like incense sticks, I figured a heavy coating of salt was better than too little. I didn’t include any pictures of this yet because I need to work with the technique more. Supposedly I should be able to see the salt subliming as it reaches non-magnetic temps.

In practice I ran into a simple problem. If I’m moving the file around in the fire to check its heat, try to heat it evenly, the salt coating gets worn off. I resorted to laying the prepared file on the coal bed, covering it over, and trusting in a bit of intuition to remove it a the right moment for quenching.


Such is the way of things that my first file warped quite noticeably in the water quench. I’ve read that if they’re pulled out at just the right moment with enough residual heat its possible to straighten. Of course, a tiny file like this cools very, very rapidly, so I pulled it out an promptly snapped it in half trying to straighten. At least I know my hardening was good!

And it gives the chance to look at the grain structure, which doesn’t look too terrible thanks to the normalization.


The above photo shows a couple of interesting things. The salt coating protected the top half of the file quite nicely, but fell off too soon on the bottom and scale developed. Testing for this file proved immediately destructive of the thin edges. Perhaps I should try an oil quench for a forging so thin and delicate of proportion?


This file (with 135 deg. face angles for sharpening chone-gake on maebiki-oga) seemed to have the best annealed state as you can see by how cleanly the teeth were cut, and didn’t warp during quenching. Both of which I’m attributing to the more even cross-section. I still need to clean it up with an acid wash followed by a neutralizing alkali wash.

Finally, finally I get to the point of my first hardened files! I’ve learned more about blacksmithing and heat treating tool steel in the past week than in the previous seven years, seriously.

9 thoughts on “My First Hardened Files”

  1. Awesome!

    That broken file shows a decent grain structure, nice and fine, from a simple backyard forge. As you train your quenching eye, try the “file” experiment, where you cut notches as break points along the length of a file, bring the tip of the file to a high yellow, then oil quench. Break the individual segments to see how steel grain size increases with heat, the yellow end of the file will look like coarse sugar, then progressively decrease towards the tang end, until you can’t even break it. This helps you to judge the optimum quenching color, and if you are observant, you’ll see the point of decalescense (forgive my poor description, haha).

    Fun stuff! We need to get together and figure out the finer points of yakiire and fire blackening. I’d love to just drill down and work this stuff for a couple of months, see what we could learn. Next stop…….forge welding.

    Great progress!

    1. Also…..could you go into more detail regarding the water/flour/salt mix? I’ve tried water/flour once, but had no luck in getting it to stick once the steel came up to temp….fell right off (a’ course that was a knife blade, and it was awfully smooth). I also tried mixing all three together, figuring that the salt might give it some tooth, much as sand does, but the flour just kind of baked away. Is this something from the Murray Carter book?

      The “tsubo”/pot symbol that many of the file manufacturers use is a reference to the miso paste fermentation pot. Miso paste was used to coat the files, to protect the steel during the quench. At least that’s my take on things, something that I’ve seen referenced. I was trying the flour thing, didn’t have any miso, and was just guessing. I didn’t know that it was a real techniques, haha.

  2. So, when can I place my order?

    Seriously, can’t wait to see how far you’ll be in a year. I really wish I could become a toolmaker some days- I might see if the ABVII knows any solutions for blacksmithing for the visually impaired. I really would love to make planes.

  3. The flour paste/table salt thing was from an earlier text that Sebastian had linked to, suggesting ale grounds as the sticking agent. I do a bit of cider brewing and was curious if I could use the lees, but haven’t had the chance to try yet. I think I had the same problem you described with the salt barrier falling off in the forge. I didn’t try mixing the salt with the flour paste, just mixed relatively runny flour/water, brushed onto file, wiped of excess with finger, and liberally sprinkled on powdered salt until it stopped adhering, then tapped off the excess. I could have also gone the route of a bone paste but I think it would smell horrendous.
    I’ve thought of a couple ways to speed up forging yasuri with various dies, but for the next couple of files I’m going to focus on a flat single cut so that I can try some variations in heat treating. I have a nagging suspicion that the high carbon point of this steel combined with the thin cross section of yasuri can cause a problem with cooling too fast normalizing in ambient air. I can definitely see us tool forging together one of these days, that would be most excellent.
    Stephen, you can place your order when I have a computer controlled kiln for annealing, salt bath for hardening, and plenty of good pine charcoal to work with. If I can get the rough forging much quicker and closer to finished dimension with some dies and the right starting stock its just a matter of being able to reliably and consistently heat treat. Half of the fun of this is seeing what I can do with a simple forge, the old timers knew how to do this stuff, but industrialization has made it almost lost knowledge.
    The trickiest part seems to be annealing properly. I can start to understand what might be desirable about using a lower carbon point steel and case hardening if you can’t kiln anneal.

    1. Hi Gabe,

      So cool to read, though my blacksmithing knowledge is close to nill. I was thinking, though, if you (and Jason) have trouble getting the salt/flour mix to stick: Aren’t you lacking the slimy/sticky nature of the ale grounds? Perhaps that why it doesn’t stay on for you? Maybe you could add a bit of active yeast and sugar to a very runny flour dough, let it ferment well and then coat in that very sticky substance. Then there is something for the salt to cling onto.
      Rye flour (or oats) tend to become much more slimy/sticky than processed wheat as well.
      It sounds like a good trick to protect the steel, but no good if it gets baked to a hard crust and knocked off during the operation. I’m guessing there was a reason they used ale grounds specifically. (which would have rich animal feed otherwise) – or think of the fermented nature of miso paste.

      Best of luck, looking forward to seeing more

  4. I propose we all go visiting you next summer and have a mini-production line, so each one gets good at one thing during the “school”, maybe we can produce some nice files. I think I’d like the shaping of the blanks after forging in fact. And you seriously should stop complaining of your set-up, you got one of the nicest fuigo I’ve seen online. (Perhaps you need to add the white ribbon, to make it a sacred place and improve the quenching.) Your 6th picture looks like a really well made shape, I’m sure in a few weeks you will start sending them in the mail for testing purposes 😛

    1. Oh, and something I forgot to ask for clarification last time. Can you make a drawing of your chisel edge should look like? I’ve only seen one and it seemed too blunt for cutting well.

  5. Very nice information! I had never heard of the triple normalizing before. I guess it would not occur to anyone who was not forging then annealing 1095 or similar tool steals. I always read the same old heat to nonmagnetic then cool in ash, had no idea. I wander if this is necessary for chisel making or anything with high carbon tool steal? Also wander if this is necessary using W2 for file, wasn’t that what was used traditionally?
    Thanks for sharing the journey!

    1. Josh, man, I’ve been watching your videos on youtube, really great! From what I’ve read on the various blade smithing forums the triple normalizing is good for grain refinement on any edged tool, though there is the risk of decarb with so many extra heats of the metal. By far the most useful thing for me has been the sub-critical annealing, recently I’ve added an extra stress relief anneal after the major grinding is done on yasuri blanks, seems to soften things up quite nicely, though still harder than annealed mild steel.

Leave a Reply

Your email address will not be published. Required fields are marked *