Tuesday, July 31, 2007

Smelters & Archaeology - Some Questions

The following is taken from an (earlier) e-mail conversation with Dr. Kevin Smith (who has been excavating a large Viking Age iron production site in Iceland)

(Dr Smith)
First, interesting that after 4-5 repeated smelts the base of the furnace stack itself (below the tuyere) was still essentially unconsolidated, poorly fired or unfired clay cobb, plastic enough to slump 10 degrees forward. Unlikely that anything this "soft" this would preserve through 800 annual winter/summer freeze/thaw cycles and yet we're always looking for the bottoms of the furnaces, expecting them to be preserved. Makes perfect sense that the best fused/fired parts of the furnace wall are going to be above and around the tuyere but it's important to realize also that the base may well be ephemeral from an archaeological perspective.

Showing the remains of a clay smelter left to weather since June 2004
I would put that line actually bellow the tuyere. The heat effect is consistent with the location of the entire slag bowl mass, forming a gradient from highest temperature effect just above the tuyere to a level marking the bottom of the slag bowl. The way we set our bases, this marks a line roughly at the top of the tap arch. With the angle of the tuyere and size of furnace being pretty constant, the heat is not penetrating much beyond about 20 cm below tuyere level.
The clay cobb bellow that line is just mud. So depending on the construction of the furnace, I would expect to see unaltered earth, then maybe disturbed earth, then the ring of raw cobb. Inside this would be whatever remains from the extraction process.
If you pulled from the top, the slag bowl might just be left in place. That would leave you with the classic shaped bowl. Below that would likely be some combination of ash, unburned or partially burned charcoal and maybe charcoal fines. The looser this lower packing is, the more likely there would be 'icicles' of slag from the bottom of the slag bowl. There might also be the same shapes of bright white cast iron. We have also seen plates of this metal forming between the slag on top of the fines layer we use. (Although I suspect our high air rates and temperatures have something to do with that material).

One question for us is what happens next. Our process has been to knock loose the cold slag bowl to prepare the smelter for its next use.

If you pulled from the bottom, you first dig out all the lower packing material through the tap arch to leave a space. Then normally you would try to puncture the slag bowl to let some of the liquid run into the space created. This serves to help heat and loosen the structure of the slag bowl. Next you either pound on the top or work to free the edges of the slag bowl from the top. With some levering from the bottom, you work to loosen the entire slag bowl and get it to drop down into the space. Then the whole slag mass, with bloom still in place, is pulled out in one large piece. In the past this is then dragged forward a couple of feet, Then most of the loose slag is removed by striking with chisel ended heavy iron rods. (This is how Skip and Lee extract)
So this method pulls the entire core out of the smelter. Nothing would remain of the slag bowl save some fragments still attached around the edges. There would be some mix of slag / ash / burned charcoal, very low in the smelter base. This at the level of the raw cobb. I would expect a much more disturbed artifact - likely one that pretty much results in loose pile of mixed elements as it collapses.

(Dr Smith)
Second, interesting how the patches in the walls finally burned/melted through with subsequent influx of external packing material and consequent production of excess glassy slag. And interesting, as well, to find that 4-5 reuses was the limit for the utility of the patched stack. How long, though, had that stack been standing (i.e. over how many winter/summer cycles?). Would be interesting to know how many times in one summer a single cobb-built stack could be used before it became necessary to tear it down and rebuild it, as the "typical" structure of the Norse smelting sites is based on a series of overlapping furnace bases representing the reconstruction of a furnace on more-or-less the same site over and over again...yet it's not that clear how much time (or how many smelts) each base in the "stack" represents.

Clay cobb smelter - after 5 smelt experiments
A number of things here:
Limits on re-use - we are finding typically that our smelters loose something in the range of 2 - 4 cm of wall thickness each use. This effect most drastic just above the tuyere. (We have found that tuyere angle effects this - remember that this smelt had that angle at only 10 degrees due to winter slumping. The ideal has been found to be 22.5 degrees. The lower angle suggested that excessive melting would in fact take place.)
Also significant for this smelt was the double smelt. In the past we have had a gap between smelts that allowed us to repair and patch up the lost wall thickness.
The repair before this experiment was done using a clay that is known to have a lower firing / melting point (local 'Blue Mountain Red'). This likely was an important factor.
The heat profile in the smelter is a torus. This washes back over the walls above the tuyere (relates to angle again). The distribution should be roughly symmetrical. What happened was that the excessive erosion was primarily on the right side of the tuyere - the area which was inset into the bank and had a layer of ash and sand packing against it. Again I would say this is significant. Free standing clay smelters can radiate heat off their surface and thus avoid excessive heat build up. There is some balance between clay properties and ideal wall thickness.
This particular smelter has gone through one entire calender year, plus two sets of firings (June and November). In our area it is pretty dry from mid June through to mid August there is always a lot of rain In September and October. The smelter had a piece of sheet steel over the top through all this time, but the exposed surface (roughly half) was open to winter snows. When the snow melted the interior base actually had several inches of standing water in it (this at the level of the raw clay cobb.)
So it may be that a single cobb wall smelter, especially if free standing, would withstand even more uses - providing that it was fully patched between smelts. This smelter had started to develop some fairly serious top to bottom cracks - several of which were leaking jets of hot gases. The cobb mix helps to tie together even cracked furnaces. My best guess is that after 5 - 6 uses of the furnace that it might just be too damaged to continue using the same structure. The damage done during the extraction process is actually more significant to the ability to re-use the structure. Most of the patching is done at the base around the tap arch - most especially if a bottom extraction is used. Skip and Lee have never gotten any more than two uses of a Flue Tyle furnace - and there is almost no wall erosion with that system. The base at the tap arch is the problem.
Depending on your access to clay (!!) building a new smelter is not that big deal. At best two days for two people (thats including digging and preparing the clay). Given the physical difficulties of top extraction (heat related), and the vast reduction in charcoal required for a second 'hot swap' smelt, it may just have been more efficient overall just to re-build entirely.

(Dr Smith)
Third, I was wondering whether you plan to dismantle this furnace prior to rebuilding it? If so, it would be interesting to know how oxidized/burnt/modified the soil appears beneath the furnace itself and in the matrix that surrounds the bottom 20-25 cm of the furnace base. Again, Icelandic archaeologists tend to look for burned/oxidized/intensely fused soil beneath and around the furnace base to convince themselves that they have found a furnace base. Yet, if the lower 15-20 cm of the furnace stack itself, below the tuyere, is still unfired clay and the charcoal charge beneath the tuyere remains unburned (and I well remember finding unburned paper in the base of the furnace we ran in 2002), it seems questionable that the sediment/matrix/soil beneath the furnace will always be fused/heat-reddened/burned/fired and perhaps questionable, as well, that the matrix outside the cobb core will always be intensely thermally altered unless there has been a disaster, like having the wall melt through. Does this make sense to you?


Our plans right now are to uncover this furnace and leave it exposed to the elements. With the hole blown in one side and the cracking I don't think we can actually mover the thing without breaking it into pieces. We backfilled the trench we dug at the front to allow the second bottom extraction effort with sand. This mainly to give us a flat surface in the work area. We will build our next smelter to the left side, so we can leave the current smelter in place for some time.
What you mention certainly suggests a useful experiment would be the ability to probe for temperature at various levels of the furnace structure. I have a tool coming that may allow for some attempt at that (I think - Its actually an electric meter - but it comes with a thermocouple attachment. This may take some number crunching to covert electrical measurements to temperature.) As you mention, we have pulled unburned paper out from under the slag bowl, (Mind you that WAS that first attempt here - where we did almost EVERYTHING wrong!)

I would suggest that there is minimal to zero effect on the condition of the underlaying soil. Remember however that when we dug in this smelter, we cut a larger hole then back filled the gap between the cobb cylinder and the undisturbed ground with that ash and sand insulation layer (also provided drainage !) We also raise our smelters up and create a base layer with charcoal fines. We did this initially to provide insulation from the damp soil underneath - which proved not a problem. It also allows the slag bowl to settle slightly below our estimate of its correct level. If you simply used bare earth for your base, then remember the slag bowl would form on top and transfer heat directly. I would think that you might end up with more of a flat bottom to the slag bowl if you worked directly on the soil. The distance of the tuyere from the base earth layer is likely the greatest effect.

All this suggests (yet again) things that we may be doing that are different that what was done in the Viking Age. I will take some more images of the weathering smelter from 2004 (seen above). What is happening there is that the sintered clay is becoming a standing ring surrounded by a wide smear of clay mud. As the ceramic is exposed, it cracks from freezing and largely falls into the hollow where the slag bowl was extracted. Any visible heat effect is limited to the walls of the smelter itself.

We have always dragged off any slag that leaks out of the smelter away from our work area as soon as it cools enough that we can pick it up. This mainly to keep our work area clean and safe to move around it. This last smelt we had a huge amount of just waste glass slag. It mostly had enough iron contained to turn it black - but not much more of than that (not magnetic or we would have re-cycled it). Originally I wanted to also record our slag volumes and weights - but with the extra produced because of the hole in the smelter this information is not helpful to an understanding of ore to bloom conversions. Anyway, I'd guess we have at least enough of this material to fill two milk crates. Time to start our first slag pit?

As always - more questions than answers!


Friday, July 27, 2007

Viking Age Blacksmith's Bellows

(edited with images from a series of posts to NORSEFOLK)

A layout for a Viking Age blacksmith's bellows that I have seen copied numerous times is the one presented on the Regia Anglorum web site. There are two main problems with the Regia design as depicted. The first is the bellows bags are cut incorrectly as simple triangles. The second is that the overall measurements are not based from the actual historic references - but the proportions are based on modern commercially available lumber. This pictured layout has then been duplicated without those people checking back to the original sources.

I'm not going into great detail here. For a detailed commentary and
working plans for a (speculative) Norse style sand table forge set up -
purchase 'Experimental Iron Smelting from the Viking Age' available on the Wareham Forge web site . (There is a
photo essay on bronze casting on the disk as well.)

There are photos of two versions of the reconstructed Norse double bag
bellows on the DARC web site
(scroll down to the sections on bronze casting and smithing work)

There are two available images of bellows from the Viking Age. One is a
wood cut from the stave church at Hylestad, Setesdal, Norway. This
gives us a side view in proportion (??) to a human figure.
The second is the 'cartoon' from the rock carving at Ramsund, Sodermanland,
Sweden. This gives us top down proportions of length to width and relative size of the inlet hole.

reconstruction for 'World of the Norse'
My estmates from the period illustrations give the rough overall size is about 28
inches long by 20 inches wide (including handles). Each of the main
chambers are about 20 inches long by 10 inches wide. In use, a
comfortable lift to the top plate is about 12 inches (at the handle
end). Average air delivery is about 2.2 litres per stroke.

The shape of the bags is actually a segment of a sphere. So the cutting pattern from flat material is a series of three lens shapes roughly 8 inches at the widest point tapering to maybe 3 inches at either end over a total (for the version described yesterday) of about 48 inches. Double stitching the seams allows for the threading of a wire stiffener into the joint. I have used leather for all of my blacksmith sized reconstructions. To make the 'Ubber Bellows' used for the 2005 iron smelting experiments leather was going to be too expensive. Instead I used a heavy canvas which was painted with tar roofing patch from the hardware (to simulate pitch). This worked extremely well in terms durability and being air tight.

I have made maybe about a half dozen of these at this point. My standard design
uses one solid wood block for the head piece - to which is attached the
two separate bottom plates. This makes it a lot easier to apply the
leathers first! Top plates float of course. The best performance results
from the use of a Y shaped bellows tube (three short tubes joined by a
leather box Y. The advantage of the leather is that it is a flexible joint. This reduces any motion from the bellows action towards the bellows stone itself. I set up the smithing equipment for LAM using this same coupling method.

I have used a heavy modern commercial door hinge under a
leather cover to extend the working life of this joint. The leather is required to seal the top end of the bellows anyway. Kevin has made a
couple of bellows using a heavy leather strip which is re-enforced by a
metal plate on either sides of the hinge joint. This is a more historic
method and has proved quite durable.

Mark Pilgrim, the staff interpreter at LAM has certainly more time at this forge than anyone else in North America. He showed me the trick to the action that Bruce mentioned in his last. You have to start each downward stroke with a small snap of your wrist which comes JUST BEFORE the completed downward stroke on the opposite side. This seems a bit weird till you get the hang of it. The purpose is to start the second exhaust stoke before you start the first intake stroke. This motion prevents any kind of suck back from the fire end.

I have used a heavy modern commercial door hinge under a
leather cover to extend the working life of this joint. Kevin has made a
couple of bellows using a heavy leather strip which is re-enforced by a
metal plate on either sides of the hinge joint. This is a more historic
method and has proved quite durable.

Note on the design - by using the correct wider planks, it is possible
to locate the handles down the centre axis of the bellows plate. It has
proved critical that the operator does not torc / twist the plate as it
is moved up and down. ( I have had metal hinges destroyed by unskilled
I have found that a U shaped handle, of a width large enough to place
your handle in the bottom of the U as the most efficient. Both the
historic illustrations show straight bar handles (but also the operator
not actually using that handle).

PS - sorry about the lack of pictures - my terminally slow rural dial up is keeping me from uploading any new images!)

Wednesday, July 11, 2007

What I've Been Doing - Summer '07


You may have noticed that new postings to 'Hammered out Bits' have been a little thin of late. Summer is always my busiest time of the year. I'm also involved in a major commission right now. I have been documenting this railing project (for a home in Toronto) as a photo-essay:


Note that this commentary itself is a work in progress. I will be adding to it as the project moves through its various stages through to completion.


February 15 - May 15, 2012 : Supported by a Crafts Projects - Creation and Development Grant

COPYRIGHT NOTICE - All posted text and images @ Darrell Markewitz.
No duplication, in whole or in part, is permitted without the author's expressed written permission.
For a detailed copyright statement : go HERE