Blooms to Bars - some Data

Regular readers here will have noticed a bit of a theme over the last posts. Again, this whole thing was sparked by two things :
First an attempt to catch up on long promised cutting and distributing some blooms made during combination demonstration / teaching events.
Second how I can't give an simple answer to a simple question (!!). In this case 'How much stuff can you make from a single bloom'.

Now I had undertaken a specific 'Bloom to Bar' project, back in 2012. (1)
When I looked back at my notes - I realised that although I had compiled a massive overview table of the data - I had never actually published this. (The only table available on the web site documentation is from Spring of 2006 !)

'Conclusions' - Blooms to Bars - c 2018

I have dug around through my notes and various past blog reports. There are likely a few individual attempts missing. You can see that in many cases, the compaction / refining process has not been carried out to its full sequence into a 'perfect' working bar. Only some of those bars had gone on to completed objects. (2)

A much more complete version of this table has been (just) published on the main documentation. This includes links back to each of the individual smelt events, plus links to the objects as created.

Some Observations :

Bloom Yield :
This is based on raw ore to finished bloom.
- Quality of iron ore is an important element in overall yields. The majority of the smelts here used our proven 'DARC Dirt' analog (red iron oxide).
- Most of the smelts were on the lower ore volume end, most typically using 30 kg ore addition. It has been shown that there will be a rapid increase in bloom size, as additional ore is added (The two 'Smeltfest 06' blooms were created using roughly 45 kg ore, for example).
- All the smelts listed are using the 'Sauder & Williams' system of high air volumes, supplied by electric blowers. This method certainly results in higher yields than those typically found with use of various human powered air systems. (3)
- The standard here is the blooms have undergone a single compaction series, hand hammered, from the initial extraction heat.

To / Welds :
My own normal process is :

Bloom / Section compressed to 'Plate'

Plate cut & folded, welded and drawn to a 'Book'

Book has been welded and drawn to a 'Billet'

3rd Weld series, flattened at 90 degrees to last series

Welded / Drawn / Flattened to finished Working Bar

The sequence shown above is for #9 on the table, but does represent a general pattern I undertake.  (This full sequence is shown in detail on a blog post)

'Welds' here refers to one complete weld series, not an individual forge weld. My normal practice is to make my first weld using a lighter hand hammer. This allows me both control and speed, but not as good penetration. The result is more of a 'tack' weld, which will (mostly) seal the seams between pieces. I then follow up with a heavy 'consolodation' weld, using the air hammer. Due to the irregular shape of blooms, especially at the compaction of plate and plate to 'book' There are likely to be additional welding heats taken to forge in the fractured edges. Taken together, there may be 3 or 4 individual welding heat cycles to each of the 'sequences' as recorded here.
This is important, but there is a clear relationship between number of weld sequences and yield. A casual observation is roughly 10 % of the mass is lost for each weld sequence I have undertaken.

As you see, a number of these sequences were only taken to the first step, bloom to plate. This done specifically, as the intent was to create bowl forms from the material. As a raw material turned into artistic forgings, I specifically chose to retain flaws in the 'finished' bars.
Again this would skew any attempt to calculate an average loss.

Total Return :
This may be the more interesting number of the table. The number here is based on the loss of ore to bloom, then again from bloom to bar = ore to bar.
Taken overall, the rough average here is about 30% working bar from ore. I would not present this as definitive (especially given the variables discussed above), but at least suggests some earlier reported figures (as low as 10% return) may not be truly representative.

Obviously, the major flaw in these numbers is the result of the dominance of purely modern methods and especially machine tools :
- use of electric blowers for smelt air
- coal forge for the bloom to bar phase
- use of hydraulic press and air hammer for consolidation and bloom to bar phase

Ideally, this work should be continued, working closer and closer to all historically accurate tools especially. (4)


( 1)

Blooms to Bars : Supported by a Grant from the Ontario Arts Council

(2) It is obvious here that my main stress has been to testing various historic based furnace systems. To that end, each smelt is typically using a dependable, but smaller volume, of ore - the standard is 30 kg.
To date, I have produced about 75 individual blooms. You can see that only 15 of these have been worked down into a 'bloom to bar' sequence.
(This is not actually either the full number of sequences - or the correct total of objects I have created from bloom iron. Looking back, I can see several work sessions never had detailed notes recorded at the time!)

(3) The difference in yields between machine air and human air is dramatic. Typically, our own tests with variations of 'Viking Age' bellows systems have produced yields in the 15 - 20% range. The same furnace / ore / sequence, only with increased air volumes from electric blowers more typically produces yields in the 25 - 30 % range. A clear example is the listed Vinland 2 (electric) @ 27% and Vinland 4 (human) @ 8% - where everything else was virtually identical.

(4) Fortunately, others in the Iron Smelting / Living History community have also been working on the same problems :
Gotz Breitenbucher, a member of the wider Experimental Iron Smelting community (based in Norway)
Črtomir Harald Lorenči, who had expanded his work through academic studies into Early Irish ironworking / smelting
Thijs van de Manakker, one of the originals in this field. Thijs (from the Netherlands) has well documented his work via video on YouTube
Note that these are hardly the only experimenters in the field! These people had made recent comments / contributions to things I have posted here.