Having been invited to do a demonstration smelt for the 6th International Conference on Contemporary Cast Iron Art, considerable thought went into designing a furnace that would be suitably authentic and also likely to give a good result. After consideration, a thin-walled furnace was chosen (for quick drying), and Lee Sauder's advice sought on the best technique to use (since this isn't the sort of furnace we have used before). The ceramic used to construct the furnace was based on Lee's "recipe", with adjustment to UK materials. The mix comprised 6 batches, each made up from 25kg of moulding sand, an equal volume of B&Q Farmyard Manure (horse manure was planned, but fell through at the last minute...), 5 shovels of "China Clay No 1" powder and 1 shovel of "Ball Clay Hyplas 71" powder (both clays obtained from "Clayman Supplies"), mixed with as little water as possible.

For an event where most participants would not have seen a bloomery smelt before, it was clearly desirable for the furnace to be manually-blown. The large single action bellows have proved extremely difficult (heavy!) to use in practice - so using the larger double-action smiths bellows seemed best. The best estimate we had from earlier use of these bellows was that they supplied 600-700 litres/minute. This suggested that a furnace of 250mm diameter ought to be within Lee and Skip's magic numbers of 1.2 to 1.5 l/min/cm2 cross sectional area.

The furnace was built around a former comprising the sawn-off end of a pallet surrounded by around 200 4-foot garden canes, all bound up with a little twine. The furnace was constructed inside a ring of brick to provide a little lateral support. As the shaft was built its walls were constructed at roughly 75mm thick, but sagging of the wall led to a thickening at the base, which added to stability. Twine was used to bind the furnace every 100mm or so in order to reduce sagging.

Drying of the furnace was encouraged by pointing both our air blower and one belong to Owen Bush at the outside of the furnace to create a breeze. The air inside the building was rather still and humid otherwise. The furnace was left in this state overnight, with just a few canes removed to allow for a little shrinkage.

The main tasks for today were to dry the furnace and to assemble the bellows frame.

The drying was accomplished by first burning wood around the structure to get the outside firm enough to allow the tap arch to be cut without sagging. Then a fire was lit inside to give a gentle heat. This went on for many hours while the furnace gently steamed. During this process the various canes and woodwork of the former were gradually removed. Towards the end of the day the heat was increased, initally by blowing the electric blower through the tap arch. Once the charge was allowed to become deeper, the bellows could be tested.

The tap arch was cut out with a trowel and then the inside of the opening pared back, so that although the cut block could seal the hole easily, it was not pressed so tight on the edges that it would be difficult to extract in use. The floor of the furnace was added at this point - and was simply a layer of sand. The blowhole was cut by boring through with the piece of scaffold tube that was to be used to line the blowhole and connect to the bellows. As with all our previous experiments the tube did not extend into the furnace - so it functions as a simple blowhole, not a tuyère. In use, the plastic pipe from the bellows just rests inside the tube; the joint is not sealed.

The furnace was left with the remains of the fire glowing well, so will hopefully continue drying overnight.

Smelt 30 was undertaken as part of today's programme. The tap arch was sealed, the furnace filled with charcoal to blowhole level, a fire lit outside the furnace was shovelled in and then blowing was started at 09:25. The initial fill of the furnace took about 9kg of charcoal, with the preheat consuming a further 12.5kg of charcoal, before the charging of ore started at 10:55. Ore was added in 1.25kg batches, with a 1:1 ore:charcoal mix, timed such that a batch of ore was charged with each half-bucket of charcoal (a full bucket of charcoal being assessed as 2.5kg). The time between batches was a steady 8 minutes.

The tap arch was opened briefly at 12:20 (after ore batch #11), but the only slag that had moved comprised a few flow lobes attached to the tap arch blocking towards the blowing side. Charging then continued until 13:10. A further 3 buckets of charcoal were charged (7.5kg), with the last at 13:40. The burndown then progressed very fast, but the process had to be stopped to wait for completion of the smithing hearth (and the "acquisition" of an anvil from the site smithy).

The tap arch block was removed easily and the lower slag mass detached with a crowbar and then pulled clear. The bloom was knocked down from above and detached from both the furnace and the slag cleanly (the top of the slag block contained a little sintered iron from the lower part of the bloom).

Mike Blue and some helpers then took charge of working the bloom. They gave it around 8 reheats and it took a strong hammering without breaking from the start. The bloom was left in an intermediate state, still requiring the teasing together of various component pieces into a coherent whole - a job for another time. As left, it weights just over 2.5kg. Mike spark-tested the bloom and found the outside at least was getting towards a medium carbon steel.

For a "one-off" furnace, worked with a very unfussy regime, this was a very successful result. The tasks of demonstrating the process of bloomery smelting and enabling the demonstration of bloomsmithing were achieved and much interest generated amongst the conference participants. We can work on improving yield and on improving the slag performance in future iterations of a similar size.

Back to Kidwelly again - this time to clear up. Dismantling the furnace revealed some interesting colouring associated with the degradation of thee organic component of the material. The furnace came apart very easily - successive layers of clay had not bonded together very well, so that is something to work on next time. The furnace, apart from a few samples, was transported piece by piece to the skip; a sad end to a nicely performing structure. Within a couple of hours the furnace site had become the spectator area for the production cupola pour...

There are many things to think about from this week. The choice of furnace diameter to balance an easy blow rate from the bellows seems to have been about right. The burn rate was slightly higher than expected (approximately 9kg/h) but within the range of the prediction - but this gave a usefully-sized bloom in a good run-time. The slag was apparently quite viscous - so although the attempt to tap was probably a bit early, its likely that slag would not have flown at a later time. The wall shows quite a lot of attack at bloom level - so its possible that much of the slag was generated here, not higher in the furnace, where wall damage was quite low. Close inspection of the wall samples may allow this to be modelled. The lack of slag above the bloom may possibly be associated with the high C content?

Thanks to all at the conference for making this an enjoyable week, but particularly thanks to Wayne Potratz and his team (especially Mike Blue and Owen Bush) for all the practical help and discussion, to Matthew Tomalin for facilitating materials and many other practicalities and to Bob Booth for getting me involved with the world of cast iron art.