Thanksgiving Smelt - some considerations

Neil and I have been looking at our measurement system with a bit of a critical eye:

First off - there have been inconsistencies in some of our past experimental data. This has most often come from not making correct account of the variation in the unit weights of some of our raw materials.

Charcoal varies a lot in terms of its density, this primarily due to the amount of water that may be incorporated in various sources and individual batches. The material from Black Diamond Charcoal for example is particularly dry. The primary adjustment here is recording totals as volume - rather than as raw weight as has been done in the past. During an actual smelt, we are measuring by a standard bucket volume anyway (then calculating the weights.)
The measured volume of our standard bucket is 10 litres.

Ore is also measured by a standard scoop, again working in volume. The calculations from number of scoops to weight has proved to contribute the largest error in the past. The problem here is the widely differing densities of the various ore types that have been used. The Virginia Rock ore for example has an average weight of 12 ounces per scoop, with the hematite grit closer to 20 ounces for the same volume.
The best suggestion here is to measure the weight of a standard scoop of ore for each individual smelt event.
The weight of a scoop of the hematite grit is 560 gms.

Air is delivered by an electric blower controlled by a marked switch. It is also becoming obvious that the normal vacumn blower is starting to show its heavy use. (Although a rugged piece of equipment, it is several decades old!) The light dimmer switch being used as a volume control is also likely to degrade with time. This became appearent during the last smelt, when the dial mark that used to show the 'just kicking over' position was in fact producing not enough current to start the blower running.
The best suggestion here is to measure the various marked volumes with the anemometer for every experiment.
The just measured air volumes are posted HERE

The relationship between air volume, charcoal consumption, reactive column height, and ore particle size needs to be re-considered for the next experiment in the Icelandic series. Up till now we have been following Sauder & Williams' guide lines for high volume air delivery - and with great success. According to their model of 1.2 - 1.5 litres per square centimeter of smelter area, our working volumes have been the range of 500 - 800 l/min. (against our typical base dimension of 25 - 30 cm).
Williams had reported that Michael Nissen (of Denmark) had been undertaking working smelts at volumes in the range of 300 l/min. This is with a blow hole set up and using a version of the Norse double bellows. (His unit is about 50% larger in physical dimensions than my blacksmith type reconstruction.) One thing to note here is that Nissen's set up has the bellows tube sitting just slightly outside the actual blow hole. This may create an extra air intake, from a possible venturi effect from the blast of air. This effect might be increased because of the way air flow pulses under the physical action of the double chamber set up.

Our own experience (admittedly only a single test) of the blow hole set up suggests that this effect may in fact be limited. Note that the last test used the electric blower, which delivers air in a constant blast (no variation cycle):
- When the charcoal was first added, a significant volume of the applied air did not penetrate into the furnace, but instead heated air splashed back out of the blow hole.
- As the internal charcoal slowly ignited, this black splash effect was visibly reduced. This suggests there was some natural draft effect from the rising hot gasses within the furnace.
- The high temperatures within the furnace fairly quickly started to melt material off the interior wall above the blow hole (stone surface). The slag produced then dripped down - freezing when it hit the relatively cold air a the mouth of the bellows tube. This effect was seen at roughly 30 minutes into the main smelt sequence. After about a hour after the addition of charcoal this melted slag had effectively blocked over the entire surface of the blow hole (save for the actual mouth of the bellows tube).
- In that test smelt, there was no attempt made to remove this solidified material. The normal care was taken to ensure the actual mouth of the tuyere remained clear (by rodgering the opening to remove hardened slag).
One question here would be - In the Nissen smelts, was extra care being taken to keep the whole of the blow hole surface clean of slag? I took a second look at some of the images from one of these smelts, and in fact the bellows tube fits snugly into the hole on the front plate. This suggests minimal air loss and no extra venturi effect from the blast.