WIP: Pattern-welded sword from bloomery steel

[Andrew W]

Here goes nothing--let's see if I can make a sword from dirt.

 

[Jan 2021]

 

The Plan

 

I'm aiming to reproduce a sword that you might encounter in late 6th- or early 7th-century lowland Britain (ie, an "Anglo-Saxon" community).

 

I want to make something typical of archaeological finds from the period, an "average" sword rather than a reproduction of one specific find.

 

I'm going to make it from ore, with maybe a bit of nineteenth-century wrought iron mixed in for fun and contrast.

 

Sources

 

The best survey of swords from this period is Paul Mortimer and Matt Bunker's new book, The Sword in Anglo-Saxon England: from the 5th to 7th century (2019). Paul and Matt do living history, and they wrote this book for an audience of makers and re-enactors as well as academics. It's a great survey. In the book, Paul shows that herringbone patterns in either 2 or 3 bars were the most common on blades from this period. Matt wrote a great chapter on hilts, most of which were made from unadorned cattle horn.

 

I also re-read all of Brian Gilmour's metallographic analyses of sword blades (Tylecote and Gilmoure 1986, and his briefer studies on blades from Castledyke South, Park Lane, and Saltwood).

 

Sketching out a pattern

 

So, what I want to do: 2 twisted bars on each face of the blade forming a herringbone pattern. These twisted bars will sandwich a core of plain iron. For the edges, I'm thinking a simple stacked bar of bloomery steel, but with the welds turned 90* so as to give the edges a subtle striped pattern (something Gilmour noted on several of the blades from Saltwood which I think would look cool).

 

Like so:

 

 

So I'll need to make 4 twisted bars, a core, and 2 edges.

 

The metal

 

I've been smelting for 2 years, and I've got a few bloom pieces lying around that should work well.

 

This bloom came out high carbon (hypoeutectic) steel, and it forges very well. I made it on the winter solstice 2019 (photos of the smelt here). I forged half into a dozen knives last summer, but I held onto the other half:

 

 

This one, smelted two months ago on the 2020 solstice, is also hypoeutectic steel:

 

 

And this one I made last month is lower carbon mild steel (basically just iron):

 

 

I've also got this bucket of bloom scrap that I can melt down into high-carbon hearth steel for the blade edges:

 

 

All these blooms were made from a low-phosphorus hematite ore, however, and they don't give great contrast for pattern welding.

 

So I grabbed some mid-P wrought iron to mix in with the twisted bars.


 

Edited February 18, 2021 by Andrew W

[Andrew W]

Making Hearth Steel

 

[Feb 6, 2021]

 

For the blades of the sword, I wanted high-carbon steel. I decided I'd make that by melting my scrap into a charcoal hearth.

 

I recorded a video of the process:

 

 

2lb (.9kg) of scrap, plus a copious serving of high-iron bloomery slag, gave me nearly 2lb of high carbon steel.

 

Hopefully that'll be enough for my blade edges!

[Alan Longmire]

No pressure, but I'm pinning this one. 

[Andrew W]

The core bars

 

[Feb 2021]

 

For the twisted core bars, I used more hearth steel (carburized bloom) + some nineteenth-century wrought iron that has a moderate phosphorus content. The last time I used this combination I got a great color contrast, so: fingers crossed?

 

1.5lb (700g) of hearth refined bloom steel + about as much phosphoric iron, ready to weld:

 

 

Heating it up...

 

 

Success!

 

 

And drawing it out into a 1/4" (6mm) square bar:

 

 

This netted me enough for 2 bars.

 

So I did it again, getting me 4 x 25" (64cm) long bars.

 

Altogether, to make the 4 core bars and the edges, I used at least 10lb (4.5kg) of unrefined bloom + 2lb (900g) of refined wrought iron. I'll use at least 2 (900g) more lb of bloom to make the center core bar--all to make a sword blade that will weigh less than 2lb / 1kg!

 

---

 

And that's as far as I am right now. Next step: lots of careful twisting!

[Joshua States]

Nice! Following this.

That looks like my Tom Clark Hammer.....

[Ryan Hobbs]

Oooooooo, this looks like fun!

[Andrew W]

Twisting, twisting, twisting...

 

The bar on top isn't twisted tight enough yet--I took this photo between heats, then kept twisting tighter.

 

The 1/4" bars are so easy to snap while I'm twisting them! I may have had to scarf weld a few back together.

 

I've twisted 3 of the 4 bars. I'm hoping to do the last after work tonight--more updates coming!

Edited February 22, 2021 by Andrew W

[Emiliano Carrillo]

Looking good! I've had a lot of issues twisting bloomery pattern welded material, and the vehmaa sword I did has lap welds aaaaaaall over the place. 

[Clifford Brewer]

[Francis Gastellu]

This is awesome, can't wait to see where it goes! 

[Kevin Colwell]

1. damn.

2. I am impressed with your smelting/hearth melting ability. very impressed.

3. Even with modern steels, I start with 8 or so pounds to get a 2 pound sword. That is because of the complex pattern welding. It always seems a shame, doesn't it? Like the old Bugs Bunny cartoon where big lathes were taking trees to make toothpicks. One tree = one toothpick.

4. I know that phosphorus adds contrast to wrought iron or bloomery steel. I seem to remember (sort of) that it can also cause some other properties, like hardness and brittleness. Are there any reasons / how do you work the material with phosphorous differently that you work the non-phosphorous material?

5. Thank you for documenting this. 

 

Help me understand (anybody)

.76% and lower - hypoeutectoid steel < .77% carbon - eutectoid steel - .85% carbon < .86% and higher - hypereutectoid steel 

Edited June 21, 2021 by Kevin Colwell

[Jerrod Miller]

4 hours ago, Kevin Colwell said:

Help me understand (anybody)

.76% and lower - hypoeutectoid steel < .77% carbon - eutectoid steel - .85% carbon < .86% and higher - hypereutectoid steel 

The eutectoid point is where the A1/A2, A3, and ACM lines meet (the valley where austenite formation happens the coldest).  Depending on the phase diagram you're looking at, that could be about 0.76% to about 0.84% (hence why 1080 and1084 are made and are basically the same thing).  The thing to keep in mind is that the Fe-C phase diagram attempts to consider only Fe and C.  Other elements will push A1/A2, A3, and ACM around a bit, so the eutectoid point moves.  From there, 'hypo' means below and 'hyper' means above/beyond.  

 

Also, eutectic is is similar except it isn't with the austenite formation point, but the liquidus (i.e. the lowest melting point).  Working in a foundry, that is the one I am typically dealing with the most, and so often say eutectic when I mean eutectoid.  I had to go back and correct this post as I was typing it.  Stupid habits.  

[Kevin Colwell]

thanks Jerrod - I sent you a pm.

 

 

[KenH]

Hey it's been several months now since the last update.  We're all waiting anxiously for the next post.  This is GOOD - I've sent links to several friends to follow along.

 

Jerrod, THANK YOU so much for starting this thread - it's tagged for me to follow.

 

Ken H>

[Jerrod Miller]

On 10/9/2021 at 10:17 AM, KenH said:

Jerrod, THANK YOU so much for starting this thread - it's tagged for me to follow.

Not me.  This is @Andrew W's thread.  

[KenH]

Andrew W:  Sorry I gave Jerrod credit for the thread, not sure at all why I didn't get it right.  Andrew, I'm more than impressed with your work, please continue the thread.

 

Jerrod, sorry about that, and I'm glad to corrected me so I can correct it.

 

Ken H><

[Jaro Petrina]

More! More!

[Patrick Brown]

On 10/13/2021 at 7:11 AM, Jaro Petrina said:

More! More!

Seconded!

[Andrew W]

Sorry y'all, I got very distracted by life!

 

Let's resume where we left off--with me twisting lots of tiny little bars.

 

I eventually finished enough unbroken bars to make the central core (I kept all the broken fragments to use for knives).

 

I decided to weld the faces on one at a time.

 

Here's two of the twisted bars welded together, strapped on to the central iron bar (to which I've already welded the other pair of twisted bars):

 

 

They welded together pretty well!

 

 

But I did make my first mistake here. When I welded the second pair of bars to the core, I got things a little off center and part of the bar came out rhomboid.

 

No big deal, I thought--I can just forge it square again! (lol, nope)

 

I forged the bar square again, but I stretched my twists unevenly and the pattern shifted. Lesson learned.

 

Next I added the steel edges:

 

 

And, done!

 

 

 

Here I've made my second mistake: I underestimated how much steel I needed in the edges, because I misjudged how much I would lose to hammer scale. (That's foreshadowing!)

[Andrew W]

That, above, is where I stopped working on this project last spring. Life happened, I got COVID, etc--and this got pushed back into the pile of half-finished things.

 

Last weekend, I found it again and decided to finish

 

First, I cut a notch into the tip of the blade and welded it shut:

 

 

Then I forged in the distal taper. I followed Peter Johnsson's advice in this thread, which was extremely helpful:

 

 

Next, I forged out the blade. The thin layers of pattern welding (basically just a veneer over the iron core) made this unusually challenging. I learned that I usually rely too much on my grinder to clean things up--something that just doesn't work with a blade like this.

 

 

And: done!

 

 

I noticed while I was forging the blade a few weld flaws in the pattern-welded veneer. Hopefully they're shallow enough to grind out (without cutting too deep into the veneer and marring the pattern).

 

Grinding, grinding, grinding--and I discovered one of my two mistakes. The blade's thickness was correct, but there was no way to clean up the edge profile while preserving the width I'd originally intended. Turns out, I'd severely underestimated how much material I would lose to scale while I forged the blade. I should have used twice as much steel on each edge to achieve the width I'd wanted. Damn. Too late now!

 

Once I'd ground the blade to 120 grit, I did a quick etch to see what I was working with.

 

 

Hooray!

 

This is the point where I realized my other mistake. Because I had forged (and badly corrected) that rhomboid section in the blade, the pattern wandered off to the side. It did this on each face, and there was nothing I could do to correct it. Oof.

 

I also found that I'd cut too deep into the veneer on the sword's tip, changing the pattern. Given the thickness of the blade here (about 1.5mm) and the veneer (<.5mm), this didn't surprise me. I honestly might have lost the herringbone to scale while forging it out, it was so thin.

 

I think I may, on the next one, end the iron core about 60% down the blade's length so the veneer at the tip is thicker? Let me know if you've encountered this and solved it.

 

 

That's as far as I got Thursday night.

[Andrew W]

I got off work early on Friday, and fired up the forge to harden the sword.

 

And that's when everything went horribly

 

Fine. It was just fine. Bloomery steel is super easy to harden.

 

I did have trouble with the blade sagging under its own weight while I was normalizing it. I had to forge out a bit of a sabre bend at one point. I think I may need to build an oven to hang these vertically if I make many more of them, because it was very annoying!

 

I ended up putting a piece of 2" angle iron in my forge and using that as a track to keep the blade straight, and this worked pretty well.

 

It hardly warped at all in the water quench, thank goodness.

 

 

Success!

 

Next up: deep fryer tempering.

 

I heated it to 450*F in a tube full of canola oil.

 

 

 

After letting it soak for about an hour, I pulled it out and straightened the remaining kinks by bending it in my hands while wearing welding gloves. I'd bend it for 5-10 seconds, inspect the results, and put it back into the oil to heat it up again. It took a few dozen heats to get it straight.

 

There were a few places where the blade was twisted, and I had to clamp it in my vice and counter-twist it with a pair of tongs. This scared the shit out of me, but the blade held up just fine.

 

And that was it for Friday!

[Andrew W]

Saturday, my wife and I took the dog for a walk in the swamp.

 

 

Then I got back to work on the sword

 

I ground it up to 400 grit on my 2x72, then switched to hand sanding up to 1500.

 

I didn't finish until Sunday.

 

Once it was polished, I gave it a light etch in ferric chloride. I etched it 4x 3 min.

 

 

 

You can see the pattern wander from where I got it off center, if you look closely.

 

 

The blade has lots of small flaws on the surface. These are normal on all the real blades I've handled on museums, though mine has more of these flaws than the originals. It's definitely a first try.

 

 

Where the pattern changed toward the tip.

 

Overall, I'm both pleased and disappointed with this blade.

 

Disappointments first: The finished blade came out significantly narrower than I intended. I was aiming for between 45mm, but the finished blade is only 35mm wide. I severely underestimated how much steel I needed on the blade's edges. Lesson learned for next time. I also wish I'd been more careful forging out the central core to keep the pattern from wandering, and the blade's surface has more welding flaws than I'd like for cosmetics.

 

But there's a lot that I love: I made a blade from ore! Despite the cosmetic blemishes, the structural welds are sound. The blade is also very dynamic--Peter Johnsson's advice on how to forge the distal taper was unsurprisingly excellent, and I'm happy with the result.

 

----

 

At this point, I was so close to the end that I decided to stay up all night making the hilt.

 

I wanted to keep this sword simple, so I would make the hilt entirely from horn with no metal fittings, copying a cemetery find from a 6/7th-century boat burial at Snape (UK).

 

(Image from Filmer-Sankey, W., and Pestell, T. (2001). Snape Anglo-Saxon Cemetery: Excavations and Surveys 1824-1992. East Anglian Archaeology, 95, page 150.)

 

The original hilt was made from cattle horn, so I used the same: a piece of mixed-color horn for the grip, and two pieces of dark brown / black horn for the pommel and guard.

 

Because my blade came out narrower than the historical examples I was copying, I made the grip a little narrower to match. The Snape grip is 45mm across (edge to edge); mine is about 35mm, just like my blade. I kept my grip, guard, and pommel lengths the same as the early medieval original.

 

Here are the pieces test-fit on the blade. I finished fitting them in the wee hours before dawn, so I'm waiting until tomorrow to rivet it all together.

 

 

I gave the blade a final polish before I went to bed (not pictured--yet!)

 

Overall, I'm pleased. This project was both easier and harder than I anticipated. Easier in that, despite every step taking an immense amount of time to complete, the process itself is pretty straightforward. Lots of forge welding, but that applies to everything made from bloomery steel. Harder because the fine details matter so much in these blades. This thing proudly displays every single error I made--there's so much less margin to "fix" and fudge mistakes. I have a much better understanding now of how much further down this road I'll have to walk to make a blade that might earn the old masters' approval.

 

That's going to be a fun journey to walk

Edited March 21, 2022 by Andrew W

[Alan Longmire]

Thanks for coming back to this!  Excellent work. 

[Andrew W]

I glued the hilt together last night. I cheated with epoxy (the originals probably used pitch).

 

Today, I trimmed the tang down to 1mm and peened it over the pommel.

 

 

 

I considered using a washer here, but most of the finds I’ve seen are just peened like this. It worked well and feels secure!

 

 

All that remains is to finish the hilt with linseed oil and build a scabbard.

[Andrew W]

I forged up some leftover bits of twisted bar into two little knives: