Iron Smelt - Wareham ON - June 12, 2010
Date:June 12, 2010
See also: Darrell Markewitz's ironsmelting site
Team: Ken Cook, Sam Falzone, Pierre La Fountaine
- 1000 years ago the norse visitors at L'anse aux meadows attempted a smelt. We plan to reproduce that smelt.
- This marked the fourth of four smelts in a sequence designed to test several variables currently outstanding in our knowledge of the smelt.
- The smelting area was laid out as suggested by the archaeology of L'Anse aux Meadows, more photos of the furnace hut, and a discussion of the vinland furnace.
|Type||Boxed Short Shaft, clay/sand mix, tap arch|
|Tuyure Type||tapered steel pipe|
|Tuyere height from base||19 cm|
|Tuyere penetration into furnace||5 cm|
|Tuyere angle||22.5 deg|
|Base design||natural set base cm|
|Bellow Type||double bag (smleting) bellows|
|Avg Air rate||700 litres per minute|
|Total Charcoal Mass||unknown Kg|
|Avg Burn rate||unknown min per 2 Kg|
|Total Mass of Ore||20 Kg|
|Burn Duration||3.5 Hours|
|Mass of Bloom||1.6 Kg|
|Bloom Type||lacy and crumbly|
||Reports of all of our iron smelting efforts along with more articles and information are available on the "Iron Smelting in the Viking Age" CD from the Wareham Forge. Copies of the CD can be purchased here.
Full dress rehershal for L'Anse aux Meadows, including a lack of time pieces.
In retrospect, the use of unknown clay, a choice caused by simple time pressures and supply problems, appears to be the main cause of the construction problems with this furnace itself. This statement based on the fact that this there were three other furnaces constructed earlier on the same layout, and the methods employed were those fairly standard for the building of many others. With the new clay the only major variable, this certainly appears to be the source. My plan for the actual LAM demonstration smelt is to have enough of of standard clay on hand for the build. There remains the possibility of acquiring enough of the local clay to build the furnace of. This however would then add a second level to the experiment itself.
Some better understanding needs to be acquired on the effects of various mixtures of sand and clay. I had originally thought that mixing 50 % sand was a generous amount, but the advice given based on the direct experiences of others suggests this is not the case. Without definite knowledge of the exact mixtures utilized at LAM originally, this all remains guess work at best. The furnace used for Vinland 4 did experience hardly any interior wall erosion, normally a significant concern on at least the first firing of such small clay wall furnaces. A number of others have explained that increasing sand or adding grog normally helps *prevent* cracking. Generally this was linked to problems during the actual firing cycle of the furnace, which was not the case here. Since the cracking developed during the initial drying, some modification of that method should be considered.
The exact composition of the ore used was not recorded. It was however a mixture of our DD1 (red) plus some amount of granular hematite. The straight DD1 (at 45% Fe) has proven to produce lower yields (in the range of 15%) in the past, so perhaps the low yield on this experiment should have not been so unexpected. The addition of any amount of hematite (at 68% Fe) could have only served to enrich this ore.
There has been some indication that it may be possible to utilize locally gathered primary bog ore for the LAM demonstration smelt. My plan is to bring enough of the previously tested DD2 (black) analog. Switching to an unknown and untested ore is most likely to modify the entire theme of the LAM smelt from 'practical demonstration' into 'experimental archaeology'. Consultation with Parks Canada would certainly be called for before switching the focus.
The other large variable most likely to impact bloom formation is the air system. As has been indicated, the working volume of the Norse type smelting bellows may not have been adequate to produce enough air, considering the expansion of the slumping furnace from the planned 22 cm to 28 cm internal diameter. Low air almost always equals low yields and less dense blooms, both of which occurred in this experiment. The other use of the same human powered system (Vinland 3) also resulted in lower yields than expected, despite the use of a significantly higher iron content ore. It is hoped that with more practice (and a solid working team) fluctuations in delivery (thus temperatures) can be leveled out.