Iron Smelting with Human power

This overview prepared as background to an interview I will be undertaking on October 7. This is in support of a research project by Amy-Eva Nuttall at the University of Sheffield (UK). For the Master’s Dissertation : ‘Investigation into Bronze Age bellows in literature and experiment’.

My original development of blacksmithing skills included working at a 'Settlement Era' (1850's) living history museum in Toronto. The forge used a 'Great Bellows' - a type standard for that time and place. Over the years, I have worked with later period (1860's +) hand driven rotary blacksmith's blowers on forges - and have several of that type here at the Wareham Forge.
It should be noted that my own experimental work has been primarily related to ‘Late Iron Age’, cultures (Norse, Pictish, Celtic), and focused on bloomery iron smelting. Some testing of bronze casting method (Norse) using historic equipment has also been undertaken in bits and pieces over the years. Bellows have been used extensively in the glass bead making experimental series.

Norse Twin Chamber - ‘Blacksmith’ (4 smelts)

Based primarily on the (well known) Hyllestad Church Carving

There are a large number of commentaries related to this piece of equipment available on this blog :
https://warehamforgeblog.blogspot.com/2008/01/bellows-reconstruction-2.html
I have made at least a half dozen individual equipments based on my evaluation of the two available Viking Age illustrations.
A series of delivery volume tests were made :
http://www.warehamforge.ca/ironsmelting/bellowstest.html

- This bellows has been used over the years on a Norse ‘sand table’ type blacksmithing forge, firing charcoal. It has proved quite effective for general forge work, and capable of generating welding temperatures. 

Viking Age forge demo - Haffenreffer Museum, RI - 2006

- The same type of bellows has been used for a number tests and demonstration of small scale bronze casting, heating roughly fist sized crucibles effectively.

Upper Canada Village Medieval - 2019

- The same type of bellows is used for the series of glass bead making experiments by Neil Peterson . In this case it has proven that care needs to be taken to not produce too much air volume for effective glass working.

Neil Peterson, Megan Roberts - 2008

- At least one set of demonstrations of the 'Aristotle' re-melting furnace used the Norse blacksmith's bellows.

Trillium War (SCA) - 2008

The earliest iron smelts I undertook were attempted with this same bellows. Over four experiments, it was certainly demonstrated that this bellows, which produces on average 120 LpM just did not produce enough air to effectively smelt iron, inside a typical ‘short shaft furnace.

Early Iron 1 - 2004

Norse Twin Chamber - Blacksmith’s linked by bladder. (1 smelt)

This was a single ‘concept’ experiment, where a series of three smaller blacksmith’s sized bellows (described above) were linked to a single large air bladder, then from there into the smelting furnace. Although the overall system proved reasonably effective, the labour pool required was extremely large.

SCA 50 Year - 2015

Norse Twin Chamber - ‘Ubber’ Bellows (6 smelts)

In connection with the experimental series related to the archaeologically proven iron smelt in Vinland (at L’Anse aux Meadows NL) a greatly oversized twin chamber style bellows was built. In this case the form of the Norse illustrations was extended, with the intent of creating a piece of equipment that would produce the air volumes indicated for best function in the short shaft furnace.
This bellows, due to its huge size, was found to be far too punishing on the workers, especially over the many hours of constant operation required for a full bloomery smelt.

with Dave Cox, Keven Jarbeau, CanIRON 5 prep - 2005

Norse Twin Chamber - ‘Smelting’ Bellows (5 smelts)

Based on the work above, a mid sized Norse styled bellows was built. This was used for the ‘Vinland’ series, experiments leading up to demonstrations at L’Anse aux Meadows itself. It also has been used for workshops where either historic setting or lack of electricity has required human power.

Ken Cook, Vinland 3 - 2009

‘Celtic’ Drum Bellows (test only)

A short attempt was undertaken to work up a bellows of this type. The results were not the best, primarily a heat failure of the (plastic!) one way valve used.
https://warehamforgeblog.blogspot.com/2017/06/2-celtic-iron-age-bellows.html

’Chinese’ Box Bellows (partial)

This system was quickly dubbed the ‘Franken-Bellows’. A wooden square box bellows was mechanically powered by an electric motor driving a modified bicycle crank set to convert rotary power to back and forth action. This was only used for part of one smelt (Vinland 2, 2009)

'Franken-bellows' in use, Vinland 2 - 2009

Looking back, I have to date undertaken 85 individual iron smelts. (Plus a good number more as observer or participating as work team for others).
Electrical powered systems dominate - for the obvious problems with any human powered systems of labour required, and effective equipment builds.
- ‘Vacuum Cleaner’ - at least four different types (9 smelts)
- ‘Leaf Blower’ - used in Scotland for the Turf to Tools series (7 smelts)
- The standard electric blower used for all the remaining smelts is a high end ‘compressor’ style blower, (US Navy surplus : rated at 1400 LpM - continuous duty).

It should be mentioned that 'Victorian' hand cranked rotary blowers have been found to deliver high volumes - but at such low delivery pressures that effectively air is not forced into smelting furnaces as required.

There may be some effect from the way Double Chamber (and Box, Drum or Bag) Bellows deliver air in repeated pulses, against the constant blast of mechanical blower systems. At this point this line of experimental investigation has not been undertaken by this team. (I suspect that considerable instrumentation might be required to develop any useful data.)

 

Although part of a much larger discussion, repeated duplication of nearly identical smelts (same furnace / furnace build, same ore * ) has clearly demonstrated the 'high volume air' effect first documented by Lee Sauder and Skip Williams (2002). With the change from a high volume, constant air blast, blower, effective yields have been found to drop from the range of 25% down to the 15% range. I consider this significant. It is certainly true that ore can be reduced into metallic iron blooms with the use of less efficient furnace builds, and employing lower are volumes. But the overall yields and importantly the quality in terms of density of blooms created with low volume, human powered bellows does not match the known artifact blooms (from the Viking Age, at least). With high volume air, the blooms we create look most like these artifact samples.

* ) This is most clearly seen in the Vinland series, where effective yields dropped from 28% (electric) to 14% (bellows).

Reproduction, Replica or Interpretation ?

 

Reproduction, Replica or Interpretation ?

On making a Viking Age cauldron

 

This is a shortened version of the article written for the Fall 2020 issue of the Iron Trillium – the newsletter of the Ontario Artist Blacksmith Association. Because of this, formatting may be a bit off.

Note that through this description, there will be a blending of Metric (artifact) and Imperial (shop work) units given.

Completed cauldron : ‘hero’ shot.

  

Part One : Defining a Project

I was approached some time ago by a serious re-enactor (1) about making a version of a specific artifact.

The object in question was the large sheet iron cauldron found in the famous Oseberg Ship Burial. (2) This is a royal status burial from Norway, uncovered in 1904. The grave is thought to be that of Queen Aså, who died in 834 AD. In keeping with the pagan customs of the early Viking Age, this wealthy individual was buried with an extensive collection of often quite elaborate grave goods (3), inside a complete 22 metre long oak plank ship.

Excavation of the Oseberg ship in 1904.  Museum of Cultural History,

University of Oslo / Olaf Væring.

Viking Ship Museum web site

Of special interest to blacksmiths is the group of cookware, including two cauldrons, an elaborate chain hanger and a unique iron tripod. (4)

As with so many objects in museums, there are few images of the cauldron available. Due to its display position, and the general low lighting levels now common in museums, these images tend to be from the same angle, and are poorly exposed, both serving to obscure details.

 

Tripod and Cauldron, on display at the Viking Ship Museum

 Drawing of the archaeological find. From Osebergfundet (original excavation report)

I freely admit that I personally have never been to the Viking Ship Museum in Norway, so have not seen the object myself. I do have a (scanned) copy of the complete 5 volume set of Osebergfundet, but this has limited illustrations - and is in Norwegian.

My normal practice when I am undertaking a commission like this one is to :

a) Gather as much actual hard data about the object as possible (including images)

b) Determine what ‘authenticity level’ the customer requires

c) Check and see what others might be offering in terms of detail, quality (and cost!)

a) I do have a considerable reference library of my own. I am fairly aware of the rough description of the objects from Oseberg (with a number of more popular descriptions as well as the report mentioned above). From that report I was able to get the general measurements, and then working with a scale off photographs, at least estimate other dimensions.

Because I do have a bit of a track record of work with Viking Age objects (5), I directly contacted the curator at the Viking Ship Museum with some specific questions, hoping for better details. (6)

 

Scaled artifact drawing of the cauldron – from the Viking Ship Museum.

I had a number of primary questions for the curator :

i) What was the exact construction of the cauldron (i.e. number of plates).

- The scaled drawing appears to show 3 plates per quarter section (so 12 plates), plus a flat bottom disk with slightly upturned edges. There is a conflicting descriptive note stating “14 side plates”. (7)

ii) What was the thickness of the iron plates used.

- The drawing notes “Plates approximately 1.5 mm thick”

iii) Exactly how was the upper rim constructed. (Where modifications / supports added in the conservation.)

- It is clear from the drawing that there is an upper rim (not a re-enforcing band added in conservation). This rim is formed by forging over the individual plates to create a rolled edge (certainly not a separate strip of metal)

Working from the excavation report, the scaled drawing, measuring from photographs, and some rough math, this is what I came up with for the dimensions:

Diameter at rim = 50 cm

Diameter at base ≠ 25 cm

Height ≠ 21 cm

Plate thickness = 1.5 mm

Volume ≠ 32 litres (8)

b) Just how close to an original artifact a customer is (most importantly) willing to pay for is something you really need to establish. Too often raw cost becomes a (massively) limiting, and dominating, factor. I often hear things like ‘I want it just like the original - but perfectly water tight, (9) and made of stainless steel’.

Obviously, the closer an object comes to the exact details (materials, measurements, fine details) to an original artifact, the more time and skill is required - and so the higher the associated cost. For me, this is a dividing line between ‘Replica’ and ‘Reproduction’. I class a Reproduction as being as close as possible to the details on the original. The object should stand examination when placed in your hands. For a ‘Replica’ should withstand observation from 3 feet distance – and have as few changes from the prototype as possible. In either case the materials may not be identical to the artifact, but should be chosen so as not to be visible.

I absolutely insist on ‘Truth in Advertising’.

c) So let’s take a look at what is available out there (10)…

A general search via Duck Duck Go using Oseberg Cauldron. Top 10 suggestions, only 5 of which were actually related to the topic (2 of 5 my own work).

 

Royal Oak Armoury

“ Royal Oak Armoury is a small, quality-focused company located in Saskatchewan, Canada. At the helm is master armourer Jeffrey Hildebrandt “ (11)

from the web site :

“ This is the largest version we have currently made of the cauldron found buried with the Viking Age ship at Oseberg. Capacity is ~12 l, compared to an estimated capacity of 20 l for the original, which was likely expected to feed an entire ship’s crew. The cauldron is constructed of steel, and is thicker at the bottom to aid in heat distribution. We offer the addition of tinned seams (as shown) for those who would prefer not to waterproof their cauldron with boiling porridge, …”

US$390 (currently $514 CDN)

 

Critical Evaluation

Note that quite importantly that this is described as only a “ version “ of the artifact.

- Size : Not given (!)

(The description above gives “estimated capacity of 20 l for the original” ??)

Note that the description gives only a volume measurement at 12 l. I attempted to run the math backwards (using the reference site indicated) and this appears to work out to a ‘cauldron’ that is 25 cm diameter and roughly 15 cm deep.

- Metal thickness : Not given

- Construction : Not stated, but from images is 13 side plates. Hard to tell from the images the shape of the bottom plate, but it does appear dished.

- Top Rim : Separate L shaped piece

Entirely different than the artifact construction (no idea where Jeff came up with this).

- Attachment lugs : One U shaped piece of forged round rod, rivet at each end

Different shape and attachment than seen on the artifact.

- Handle : square rod with twists - as original

- Sealing : Tin solder

Specifically described as an ‘alternative’ - with no attempt to relate to artifacts

The Practical Viking (Etsy)

Ísgerðr and Kjartan (obviously ‘re-enactor persona’ names)

Chesterland, OH, United States

 
“The Practical Viking strives to create well researched hand crafted items for all levels of re-enactment…”

from the web site :

Viking Cauldron- Oseberg Cooking Pot

This handmade Cauldron is a perfect addition to your camp cooking set up-

Inspired by the original found in the Oseberg Ship this cauldron will help your Viking encampment.

CA$493.96

Critical Evaluation

Note that specifically this is described as “ inspired by ” the artifact

- Size : Not given (!)

Although it is hard to tell from images, this appears to be quite shallow as related to width.

- Metal thickness : Not given

- Construction : Not stated, but from images is 14 side plates. Hard to tell from the images the shape of the bottom plate, but it does appear dished rather than flat.

- Top Rim : Separate L shaped piece

Again, not as artifact. Actually appears to be copied from the Royal Oak version

- Attachment lugs : One U shaped piece of bent round rod, rivet at each end, only vaguely forged. Different shape and attachment that seen on the artifact.

- Handle : round rod - no twists as original

- Sealing : Not given

Weland Smithy

Torvald Sorenson (suspect is ‘re-enactor person’ name)

Benton City, WA

“ Many of the items in this section are reproductions of articles found in the Oseberg Ship, a Royal Viking burial ship, while several others are of general Norse origin. The rest of these items are of more generic period origin. “

price not given

(It should be noted the majority of items on offer are clothing or costume accessories)

from the web site :

Critical Evaluation

Note that there is vague reference to “reproductions”, but specific objects are not described individually.

Other than reduced size, this appears to be closest to the artifact.

- Size : Not given (!)

There is a second image of one of these in use, a best guess is roughly 30 - 35 cm dia.

- Metal thickness : Not given

- Construction : Not stated, but from images is 10 side plates. Hard to tell from the images the shape of the bottom plate, but it does appear dished rather than flat. (Note that this object is closest seen in overall shape to the overall ‘half sphere plus’ shape of the artifact.)

- Top Rim : None

- Attachment lugs : Hard to tell from image, but appears at least close to shape of original

- Handle : square rod with twists as original

- Sealing : Not given

(rant mode off)

  Closer look at the finished Cauldron – note the error in the shape of that first plate.

 

Reproduction / Replica / Interpretation ?

Overall Measurements (to Artifact) :

Diameter at rim = 42.5 cm (50 cm)

Diameter at base ≠ 12.5 cm (≠ 25 cm)

Height = 24 cm (≠ 21 cm)

Plate thickness = 3 mm (1.5 mm)

Volume ≠ 24 litres (≠ 32 litres)

Weight = 7.7 kg (??)

Critical Evaluation

- Size : slightly smaller than original

- Metal thickness : thicker than original

- Construction : 12 side plates, much smaller bottom plate in proportion

- Top Rim : None

- Attachment lugs : as original (maker's mark)

- Handle : square rod with twists, but flattened central section

- Sealing : bronze brasing

First consideration relates back to material choice. For this piece, all the material was standard modern mild steel bar and plate. The artifact was of course made on bloomery smelted, wrought iron. Although it might have been at least theoretically possible for me to have actually created the identical starting materials, even a rough estimate would be adding at least an extra $3000 to the cost!

Reproduction ?

As I discussed above, my standard for classing something a Reproduction is ‘within 10 % of the original’.

Here the size is off by roughly 15 % - so based on this alone, I can not call the piece I made a Reproduction.

Replica ?

My standard for classing something a Replica is ‘with no more that 3 major points of difference’

Here the various differences have been marked by underlining each as they happened:

- overall diameter

- thickness of plates (? 12 )

- size of base to top diameter

- oval shape to bottom plate

- (number of plates ?)

You can see that there are at least three major points of difference. Four if the number of segments is considered (although the exact number of plates is uncertain without direct observation in this case).

Interpretation ?

There were a number of deliberate modifications from the original prototype made :

- no rolled edge

- use of bronze braise on seams

- flat section on handle

I do consider the overall impression the completed object makes on the viewer, at least on first glance, is that when compared to the original artifact, this appears to be a reasonable Replica of the Oseberg Cauldron. The use of heavy, hand forged plates is the major contributing factor here. The other versions shown above are all cold worked from light sheet - and have a level of smooth finishing not exhibited by the artifact (even when it was new).

In all honesty, I will describe this work as an Interpretation. The difference in size, and the three chosen modifications from the original, are the reasons.

Images from other commercial web sites are used without permission – but attributed to the indicated web sites.

Other images as credited in the text, or by the author.

End Notes:

1) “Re-Enactor” itself is a bit of a problematic term, often best applied to those who re-stage specific historic battles. The customer in this case might be better described as a ‘Living History Interpreter’, someone who is interested in replicating aspects of general ‘daily life’ from a specific time period and cultural setting. As might be expected, the degree of detailing (so requirement for accuracy) varies considerably between individuals!

2) A general description of the burial via Wikipedia : https://en.wikipedia.org/wiki/Oseberg_Ship

3) For a fuller look at some of the grave goods : https://www.khm.uio.no/english/visit-us/viking-ship-museum/exhibitions/oseberg/the-women-in-the-oseberg-burial/index.html

4) Although there are other cauldron hanging chains that have been found, the tripod itself is the only one known. Although often desired by re-enactors, it is a problematic object at best. My own interpretation is that this object was never used over a fire - but in fact was more like a highly ornate ‘serving stand’.

See fuller commentaries on the blog :

Oseberg tripod - my interpretation

Oseberg - putting up the pot

5) It was my interest in the Viking Age in general that lead me into blacksmithing, back in the later 1970’s. I would go on to extensive work for Parks Canada at L’Anse aux Meadows NHSC, plus work on a number of major exhibits around ‘Viking Millennium’ in 2000 and beyond.

6) For those truly interested in the details, these are described in a further blog posting :

Oseberg tripod and cauldron - refining the details

7) Just to muddy this further, I had a personal communication from an individual who reports he has seen the artifact several times, and has counted 13 side plates.

8) http://ambrsoft.com/TrigoCalc/Sphere/Cap/SphereCap.htm#segment

9) Actual Viking Age metals are all ‘wrought iron’ (actually bloomery iron). The plates are hammered to best fit, but only riveted together. A pot is sealed by cooking a thick stew or porridge, which leaks through the seams, then is allowed to burn into the cracks - never cleaning the outside surface. Although there are some objects with traces of tin ‘soldering’ this was not common practice.

See a discussion on this : Purchasing a cook pot

10) These direct quotes and images used with no permission - although the web sites are included.

I did also check the offerings at Jelling Dragon, a well known UK supplier to Viking Age re-enactors. There is nothing based on the Oseberg cauldron on offer there :

https://jelldragon.com/collections/hand-forged-cauldrons-cooking-pots/

11) I should mention that I know Jeff, casually, from way, way back. His work as an armourer is exceptional. The degree of finishing on his version is certainly very good.

12) The note on the artifact drawing indicates 1.5 mm. I have a large quantity of 1/8 thick (3 mm) steel plate on hand, which I used for this project. One certain question would be if the reported 1.5 mm represents the thickness of the artifact plates, certainly much corroded, after preservation. Even with the object recovered in as good shape as it appears, it is certain to have had variations in thickness, and much material lost to surface rusting. At 1.5 mm, it would have been extremely difficult for the original Viking Age smith to have produce such thin plate.

Interview - on 'Last to Sea' - ESP

 In an attempt to help promote Elora, through illustrating the Sculpture Project, Wrightman Cable shot a series of short interviews with participating artists.

Well, some of the participating artists.

The edited video was rendered down into something for their own cable channel. 

Only their own cable channel subscribers...

This is the segment that I presented for this :

A couple of notes:

1) Wrightman, when asked, stipulated that the broadcast program NOT be shared on any alternate platform besides their own (very!) limited subscription service.

I have cut off all the content but the specific segment that involves me. Other than the camera work, there is no interviewer or narration other than my own voice. In this I feel I have a perfect right to show * myself *.

2) We were given a potential list of points to express by the Sculpture Project. We were also requested to keep our statement to a length of 4 - 5 minutes. 

As this is certainly not my first time providing a news / promotional video, I carefully kept to these suggestions, especially the time limit. As it later turned out, relatively few of the participating artists showed up for the filming. The 'around town' program Wrightman was producing ran 60 minutes, and in the end they did not have enough interview footage to fill the time. The result of this is a huge variation of on screen time allotted to individual artists. As I was the only one who actually held to the original time frame - my interview segment is the shortest of the entire final program. 

I'll leave it up to the viewer to decide how well I represented myself, this specific work, and the Sculpture Project overall.

 

Some Bloom to Bar

 My iron smelting partner, Neil, came up Monday for a scheduled one day 'bloom to bar' session. What with COVID, and honestly, Neil's limited experience blacksmithing, this turned out to be a combination of observation and trying out. I worked up two pieces, then later Neil did one himself. Although all this was using smaller pieces of bloom, there were good results all around. 

Partially sectioned raw bloom, plus finished 'currency bars' - Norway, Viking Age.

The results should at best be considered as only approximations of the overall conversion of raw ore into final working bars. Much attention has been directed to the first element of this process - that of bloomery smelting itself (ore to bloom). In truth, an understanding of the full cycle also needs to assess the important next step, bloom to bar. The desired finished product was never blooms, but the working bars sold into the hands of blacksmiths.  

There are a couple of major limits that need to be remembered when comparing the experiments here at Wareham, and how them might be applied to a larger understanding of historic iron making.

First is the relatively small size of the blooms produced. Intentionally, the normal ore amount used is in the range of 25 - 30 kg. The standard bog ore analog has a rough iron content in the range of 55 - 60 %. It certainly has been found that smaller ore weight smelts tend to have lower yield results than those using larger amounts. In effect, it takes a certain amount of ore to create an effective 'working base' in the furnace, and later additions of ore simple pack more and more iron on to the developing bloom.

The objective has been to aim for blooms in the 3 - 5 kg weight range. This is large enough to show effective production, yet also leaving individual blooms small enough to make later forging to bar possible at the hands of a single worker (admittedly  assisted by powered equipment!) Yields have been found to vary considerably however, the general impression is as the effect of air supply systems. The expectation is yields around the 20 % mark for these small ore smelts. (1)

A clear second exception is the use of modern forges and forging equipment. The Wareham Forge is a well equipped, if small scale, artisan blacksmith's shop. A two burner propane forge is commonly used for an initial 'heat soak' of the bloom pieces (to ensure full heating of the central core). Then pieces are transferred to a coal forge to bring up to full welding temperatures. As well as hand hammering, compression sequences are undertaken under a 30 ton hydraulic press. Fuller compaction and drawing makes use of a 75 lb air hammer. (2)

Bloom pieces from the 'Bones' smelt : June 2020. 

The piece seen to the left was one of those used (# 3).

A number of roughly fist sized pieces of earlier blooms were considered, most in the 500 gm range. These admittedly small pieces were chosen primarily for the ease of heating them, and also to reduce the amount of raw force that would be needed while hammering.

1) The first piece attempted was a smaller fragment from the 2018 'Gromps' experiment. The texture of the piece certainly suggested a high carbon content. At the first light hammer stroke, the piece completely disintegrated - a sure sign that in fact it was composed of unforgeable *cast* iron!

2) The second piece was part of the bloom from June 2008, one of the early tests of the 'Dark Dirt' bog ore analog. This was a specifically small ore smelt, with only 20 kg used. This seriously impacted the initial yield, although a considered a 'nice compact bloom' the total produced was only 1.9 kg (so 10 % yield). This section weighed 727 gm at the start - and had already had some compacting undertaken. The material worked down into a bar fairly easily, using a single overall sequence of compressing and compacting, with no major flaws developing. The end result was a bar 26 x 2.2 x 1 cm at 435 gm, considered ready to forge into some object. This then represents a roughly 60 % return from bloom to bar. 

Finished bars # 2 / # 3. Inset is the starting piece used for bar #2

3) The third piece selected was from the 'Bones' experiment - June 2020. This smelt had lower yield results than expected, a 25 kg smelt with 2.6 kg / 11 % return (3). The individual bloom pieces had been compressed then cut right after extraction via the hydraulic press. The starting weight was 577 gms. Through a full compression and compaction series, the resulting block still had some cracking and voids on the surface. It was prepared for a second weld series, being cut into two roughly equal pieces. At this point the weight is 382 gm, at roughly 15 x 3 x 1.8 cm, tapering in width to the two ends. This represents a bloom to bar return at 66 %, but it is important to note that some loss can certainly be expected over the remaining welding series. 

4) The last working sequence was carried out by Neil, who has limited forging experience. The piece selected was a fairly dense segment of the October 2005 smelt (very early!). This used a combination of industrial taconite and Virginia limonite, a roughly 19 kg smelt resulting in a 4.3 kg bloom / yield at 22 %. The bloom segment was 492 gm at the start.

The end result was a small bar, down to 209 gm. Measurements at this stage are roughly 8.5 x 2.7 x 1.2 cm (later addition from Neil). There were still some pits remaining on the centre surface, and a major flaw along one edge. The combination would require cutting and another welding series (as seen with bar #3). Neil was justifiably pleased with his 'first time' result, and decided to retain the bar as a lecture sample. The difference in the bloom to bar yield, here 42 % (at this stage) was put down to the relative lack of experience in the worker. 

It is clear that not too much in way of conclusion can be drawn from this limited experience. The starting quality of the blooms will obviously have a major impact of the bloom to bar sequence results. Because the of the intentionally small ore smelts certainly significantly lowers the yields at the ore to bloom stage, it can be expected that any overall ore to bar numbers are clearly distorted. 

Other tests rendering the DARC team blooms down into bars have also resulted in roughly 60% returns at this phase. Although not conclusive, this at least is suggestive of the kind of results than may be expected. 

A big thank you to Neil Peterson - for suggesting undertaking this working day. Overcoming personal inertia is a huge problem for me of late (if not for us all in the world of COVID).

1) This is certainly bore out by comparing smelts with different air systems and ore amounts.

When human powered bellows of various types have been used, yields typically drop dramatically, with roughly 15 % ore to bloom being a rough average.

When ore amounts are increased to 40 - 45 kg, so the yields also dramatically improve, into the range of 35 - 40 % ore to bloom (occasionally as high as 45%!)

2) Even still, all of these equipments also impose certain size limits. The die surface on the press is about 5 x 6 inches. The die on the air hammer is a still smaller 2 x 4 inches. The existing propane forge is limited to about 4 x 8 inches, The coal forge has an effective heat zone at best roughly 6 inches in diameter. The combined result is that any piece much larger than a half grapefruit simply is to large to effectively heat or work. (Another reason to limit smelting production to 3 - 5 kg.)

3) The best guess here is that the additional moisture from the several additions of bones with meat still attached might be the reason ( ?? )