#2 - Celtic Iron Age Bellows

Continuing from the last post ...
What I've built so far:

'Semi Drum' Bellows - top view, extended

So at best I have to consider this a hypothetical design.
Once again (as with my long discussions here on Viking Age twin chamber bellows), there is no archaeology to guide in type or design. In this case there are not even illustrations, much less surviving artifacts (or fragments).

As I suggested in the previous post, I remain unconvinced that the simple open top 'bag bellows' will supply enough air for an * effective * bloomery furnace operation (1).( When I look at other experiments, I most usually see very poor penetration of air into the furnace interior, exhibited by the temperature gradients visible. )

Admittedly, the raw size of the bellows above was almost solely dictated by the leather materials I had on hand. I cut the largest pieces possible from what was about a half full hide leather skin. (This also has used up the last available piece of leather I had of a suitable thickness for bellows sides.)

What you see above is made from one single rectangular piece, folded in half and stitched up the two sides - the same basic construction method as the bag bellows type.  This gave a measurement on the open (top) end of 2 x 25 1/2 inches (51" / 130 cm total as circumference).

The oval top plate was then cut from available 12 x 1 inch rough pine, again utilizing the best width possible from that material. This created an oval 19" / 49 cm x 11 1/4 / 29 cm in dimensions

Top Plate - inlet holes to hand size

Given the oval overall shape thus generated, it was clear that two hand operation would be available. I made a decision not to include interior input valves. Instead, the operator's hands would create the seal - as they pushed down on the plate to collapse the bellows. As with the 'one goat skin' interpretation, a loose leather strap would serve to lift / expand the bellows. I cut the input holes to fit my (smallish) hands - to 3 " / 7.5 cm diameter.

Now I did get a bit more elaborate on the internal structure:

Interior view - showing securing the base straps - and the (modern) output valve.

I had considered how to hold the unit down securely. I decided the easiest way to accomplish this would be using heavy leather lacing across the four lower corners of the bag. These could be tied to wooden pegs hammered into the ground. To ensure there would not be damage to the leather where the crossed thongs exited the bag, I cut a smaller wooden plate for the interior. The thongs are knotted in the centre, then secured in place with a large leather square, itself tacked down to the wood based. (This of course uses unlikely metal fittings. I can imagine the laces running through simple holes in the wooden base as a more historical method.)

Past experience has shown that unskilled operators will often end up sucking air back into a bellows from the output end. As this unit is intended to be attached to a smelting furnace, this unwanted reversal would be sucking back gasses at a temperature range of 1100 - 1250 C. Certain to burn up the leather!
So to prevent this, I did add a very modern 'cheat' to the design. A standard plastic sump pump one way valve was inserted and tied into the output leather tube. This does reduce the available output diameter slightly - to 1 1/8" / 3 cm ID.

The short leather output tube was sized to easily allow the insert of a standard plastic sump pump hose - at 1 1/2" / 4 cm OD. Of course in use this tube could be mated directly to the tuyere. (2)

View of the expanded bellows from the side.

At full expansion, the height of the unit is 10 " / 26 cm. When fully collapsed, the effective height is 4" / 10 cm. This gives a rough 'loft' of about 6" / 16 cm.
Computing the potential volume produced using the formula for a regular oval tank (Pi x major axis x minor axis x length / 4) suggests 17 litres per stroke.
Given expansion of the leather sides, and that the bottom is not another flat oval, this calculated volume is likely high. Experience with other bellows units built in the past certainly suggests the practical working volumes are often closer to 50 % of the calculated theoretical.

Lifting the bellows on the fill stroke (only one hand - the other was on the camera!)

Even so, this all does suggest that this version of a semi drum bellows might easily produced as much as 500 litres per minute, based on 60 strokes per minute. Over the length of a smelt, this is more likely to drop considerably, but hopefully a volume of 350 LpM should prove achievable.
If the smelting furnace intended for the Crannog Centre demonstration is built to the smaller 22 - 25 cm ID, this amount of air should prove workable, if on the lower end of the effective range.

One of the other considerations here is equipment transport (!). This construction lays fairly flat, and is not so large as not to fit into a suitcase. It also has almost no metal parts - often a major concern in these days of airport paranoia.


1) Weasel Words Here:
Certainly you can get * some * iron from a furnace using low volume air. The resulting yield will be extremely low. The density and quality of the iron created will be extremely low. So if your objective is balancing over effort and expenditures of materials over the entire ore to working bar production cycle, low air is just not effective. Other researchers using low volumes have reported 'ore to bar' ratios of 10% (or even less). The loss at 'bloom to bar' is especially high.

2) Our standard practice is to use a Y tube between air source and tuyere. Fitted with a simple wooden plug (in historic context) to the third branch. This allows both easy observation down the tuyere, and clearing obstructions with a long thin 'Radner' tool.