You mean ONLY with more air?

 Warning : Often when you e-mail me a question, I tend to ramble on at length in reply. Having spent the time (for me typically an hour or more) on an attempt at a full consideration of that question, I will turn the reply into a blog post as well.

1) so you've shown the early medieval iron bloomery furnaces only produce accurate blooms when you push more air in than you can currently manage using what you think was the bellows technology of the time?

 

 What a ball of wax that is! 

I consider the effect of air a big disconnect, especially with what has happened here in North America as interest in bloomery iron making developed.
I consider myself one of the extremely small group that started the whole thing, in the early 2000's. Lee Sauder and Skip Williams are most certainly the very first to seriously undertake and repeat the process, and finally come up with a system that not only produced significant blooms, but consistently. Their objective was not historic method, but functional results - how to get the best yields at the highest density. Modern electric blowers were employed from the start. One very important difference between North America and Europe is over here work with bloomery furnaces has been primarily in the hands of blacksmiths - not archaeologists or re-enactors. Lee's original point of inspiration was African systems, for which there was some 'traditional' recording still available. He would return to this interest into the later 2010's.

Making bloomery iron is an elaborate, expensive, time consuming, and labour intensive process - with a very steep learning curve to positive results. So the resulting metal has a high 'investment' value. Lee has set the 'selling price' of his bloom pieces at roughly $60 CDN / kg (the few times I have been asked, I quote $100 / kg). This is roughly ten times the cost of modern alloy steel bars. So for blacksmiths, the only object type that can justify this kind of investment in materials to object price is knives. Unfortunately, here in North America, this has lead the entire Early Iron movement to become dominated by bladesmiths. This in turn has lead to what I feel is a quite unrealistic obsession with carbon content, an extremely modern consideration of what is clearly a 'pre industrial age' material (and processes).

The 'Gange o Fer' in 2004 : (L-R) Lee Sauder, Skip Williams, me, Mike McCarthy

My interest started with Viking Age systems, most specifically sparked by the single smelt event at Vinland by the Norse. c 1000 AD. As part of the original 'Gange o Fer', it was so clear that there were many individual variables effecting the dynamic inside a small scale furnace. So the early years were simply testing variable after variable, in the hopes of getting some understanding (and control?) over these both individually and in combination. So my focus has never been either to best possible yields or specifically 'quality'. If anything, my estimate of 'good iron' is based on the ease of compacting the bloom to bar, then the ability of that bar to be easily forged to object. Yes, we do end up with some blooms to bars being higher carbon, and set these aside (as the Norse would have) for cutting edges.
Through almost all of our experimental work, we have quite deliberately aimed to making smaller blooms, in the 3 - 5 kg range. (Yield % climbs sharply with larger ore volume additions!)

In Europe, much of the work with bloomery iron is in the hands of living history sites and hobby re-enactors. What has been so frustrating to me is the lack of recording. (See my piece in EXARC : 'Standardized Reporting...'). Lee has pointed out to me many times the overall difficulty of getting any kind of effective measurements of air volumes, and that the only uniform field reporting can be 'time of consumption'. Even there, it is obvious to me that most people are actually reporting total charcoal consumed / total time of smelt. This is actually only a vague average at best.

One one major problem is the simple lack of historical accuracy I see. If you are using what is at best a Late Medieval double chamber bellows (to chambers stacked on top each other) you are NOT using 'Viking Age' method. Too often I see smelts described as 'Viking', which are using different furnace builds and ore types, than the known Norse archaeology. *

Our 'Econo Norse' teaching furnace : Brick, pipe tuyere, vacuum blower, using taconite. The only thing 'Viking Age' is the furnace diameter?

Don't miss understand - people are most certainly getting iron blooms. 

The times we have used a proven furnace layout and standard ore, yet with variations of a Viking Age type twin chamber (side by side) bellows, consistently our yields drop about 10 % overall, from an expected 20 - 25 % down to closer to 12 - 18 % return. The blooms also tend to be considerably less dense (so harder to work into bars, with more loss at this stage of the overall process). 

Early twin bellows for smelting ? : 'Ubber-Bellows' for CanIRON V prep, 2005

- Obviously, one clear possibility is that the whole furnace layout and overall method we are using is just not effective, and so may be entirely different than historic process. (This might also be a simple as 'we still are screwing up'!)
- We are working with an Fe2O3 based ore analog at typically about 55% Fe content. Natural primary bog iron ore is actually FeO-OH, which potentially could be as much as 63% Fe. Natural ores vary considerably, even from the same location, but the difference between both chemistry and especially iron content may be a significant difference?
- We have certainly found a considerable 'learning curve' with use of human powered air. It may be that we are just not working correctly with this entirely. (One experiment using a secondary collection bladder may be suggestive, but there is nothing from archaeology to suggest this method. Latter Medieval illustrations which do show bellows use, don't show bladders. )

Norse 'blacksmith' size bellows linked to a bladder : SCA 50 event 2015

So key to this whole thing is a more correct statement :
 'We don't get historic blooms when we use Viking Age suggested bellows.'
- There is not much data available on the actual measured density of the existing artifact blooms.
- Others are certainly getting iron - but there is often no clear reporting of actual yield or most importantly the quality of those blooms. 

* What REALLY aggravates me is a most recent trend to individuals who are using 'Viking Age!' as a mere marketing label.