Manufacturing Hammered Silver Sheet

Jamie Hall 13 years ago

This article was originally posted on the Primitive Method Blog.

One of the key products for any metalworker is sheet metal. It can be used for forming, raising and spinning; it can be decorated by chasing and repoussé, or by etching, engraving and piercing. It can also be converted into other products - further thinning can be done to turn it into foil (see below right) or leaf, or the sheet can be cut into strips for cloisonné enamelling, or for strip-drawn and strip-twisted wire. Without access to sheet, the jeweller would be creatively limited to wire-work and cast pieces.

Hammered Silver Sheet

The technical requirements to make sheet are fairly simple. The material is probably the most important factor; most metals can be made into sheet, but silver is malleable, and gold even more so. With these two metals, the fineness of the alloy plays a significant part - adding copper to silver, for example, will harden it, and require the metal to be annealed more often while sheet is produced; every time the metal is heated, quenched and pickled, the artisan must stop the real work of hammering the piece. When fine silver or gold is used, the metal must be annealed less often, and of course does not need to be pickled, so significant time is saved - particularly if we are talking about the early medieval period, before the advent of strong fast acids like sulphuric or phosphoric acid.


The jeweller must have a hammer and anvil - in prehistoric times, the work might even be done with two smooth, hard rocks and a hot fire (taking many days, of course). The only barrier to producing with a hammer and anvil is time; with the modern rolling mill, many hours of hammering are replaced by a few minutes of rolling, and the skill of hammering sheet is all but unknown in the western world (although it still persists in many of the developing countries, such as India). Although water and wind power were available throughout the middle ages, there is no evidence that there were powered rolling mills until the end of the medieval period; it's hardly an academic citation, but Wikipedia suggests that the first powered mill in England was built in 1590, in the form of a slitting mill. And while it is possible that smaller hand-powered mills were built before then, hammering seems to be the only real option for producing medieval sheet metal, and I have seen no reference to rolling mills in the books of Theophilus, Cellini, Agricola and Biringuccio, despite the detail they provide on other workshop machinery.


Despite the impracticality of producing sheet this way, it's a worthwhile exercise for developing your hammer skills, and it's also critical for researching other methods that depend on that sheet, like producing strip-drawn wire; if we substitute modern rolled sheet or strip for the original hammered bullion, then we can't really see the flaws in archaic production methods. Using strip-drawn wire as an example, so far I have used strip rolled out from modern drawn wire for experimental purposes, but this has very straight edges, and of course has perfectly regular width and thickness, making the quality of the resulting wire far better than it would be if produced without a mill.

Ingot ready for hammering

The experiment was done using about 60g of fine silver, which was melted in a crucible and poured into an oiled and heated steel mould - the adjustable sort in which the metal is cast as a vertical sheet. The resulting sheet was around 3mm thick, but the cast was less than ideal; the "top" edge of the casting was irregular, and despite scraping, hammering and filing it, the edge split very easily, so my sheet ingot had to be cut in half and the edges sheared off, reducing the ingot to about 20g. In future experiments, I'll try casting a thicker round ingot, either using an open mould in sand or charcoal, or using the pressed method that one reader suggested in a previous blog entry. The more regular the edges, the better, and a round ingot would be easier to work with, and probably closer to the finished shape, assuming that a bowl or chalice was being produced.


In his Treatise..., written in the 16th Century, Cellini suggests that it was common to use a large scraper to carefully neaten the surface of the cast ingot; this was when he was working in Italy, but in Paris, he says that the work was done simply by hammering. In both places, the ingot would be cleaned to remove blemishes - if any are left in, they cause problems later on. If the problems are around the edge, then they can be sheared off, but a flaw in the middle of the sheet could ruin the entire piece.


As I was working with a small ingot, there was a considerable issue with "bounce". When the ingot was intially hammered on a large steel block (top photo), it was almost impossible to keep hold of it, and I injured myself with the hammer. After that, I decided to try a small flat stake - this made it possible to control the point of impact, and reduced the "bounce". With practice, it was possible to strike the ingot continuously without it flying out of my hand. During this early stage, a hammer with a slightly domed head was used, effectively planishing the ingot. At this thickness, I wasn't worried about denting or over-thinning the silver; as the purpose of the sheet was purely experimental, my intention was to make it as thin as possible. If if you are intending to produce a sheet of a specific thickness, you would have to take more care, as a domed hammer could thin the sheet too much.


I hammered the sheet on both sides, until it was large enough to hold easily - about 5cm square, and then transferred it to a slightly domed stake; it became more difficult to use the first, flat stake once the ingot was larger than the top of the stake - this problem was resolved by using the domed stake (right), which was only in contact directly underneath the strike of the hammer. The flat flat face of the hammer had to be used with this stake, or the point of impact would have been too small. About an hour of hammering followed.


Once the sheet was less than 1.5mm thick, it was becoming much larger, and was cut in half for ease of handling; the other benefit of cutting it periodically was that it provided pieces of sheet in various thicknesses, for use in later experiments. At this stage, the sheet was moved back to the steel block that I had originally tried - the increased size made it safer to hold the silver, and the issue of bounce had disappeared as the metal became thinner. Although the steel block had recently been reground, it did not have a polished surface, and consequently it wasn't really clean enough for the work. 16112010416I did an hour or so of hammering on this, but the silver had to be cleaned regularly, and the grinding marks in the steel left an unpleasant texture to the sheet (right). Had the block been highly polished, all of the work could have been completed on it. As it was, another stake had to be found - fortunately, stakes are plentiful in the Loughborough raising room, and I was able to find a flat and highly polished stake, not quite as large as the sheet, but large enough that there was no danger of denting the underside of the silver on the edges of the stake. This yielded a much nicer texture (below left). Note that the frosted texture is due to annealing the silver; it would have been better to have photographed the sheet before it was annealed.


By this point, one issue was becoming very clear - there was no room for error when using a hammer to produce sheet; as the metal became thinner, the risk of denting the metal became greater, and I spent a lot of time learning to hit straight. I had always assumed that I could hit straight, but I was apparently wrong. Of all the stakes, this one was used the longest. As much as four hours was spent on this stake, taking the sheet below 1mm, and eventually stopping at around 0.1mm, at which point the sheet technically became foil (further experiments will make this foil thinner and thinner, but I'll write about another time). There are several considerations when trying to hit straight - the hammer itself may be slightly uneven, particularly if it has been repolished. The users wrist and arm may also favour the left or right, and this needs to be corrected for. In the 1912 book "Metalworking and Enamelling", Herbert Maryon talks about the use of planishing hammers: "The secret of it's successful use lies in a certain stiffness of the wrist", and he goes on to suggest that the edges of the hammer face should be rounded for the purposes of practice (chapter XI, "Raising").


In addition to the flat polished stake, some work was also done using a piece of pigskin and the domed face of the hammer on the steel block - the reason for this was the risk of uneven form which resulted from my lack of practice producing sheet this thin. What was almost perfectly flat at 1mm became contoured once it dropped below 0.5mm thick (right; those are contours, not dents, I promise you). The pigskin was used before each anneal, in the hope of leveling out the sheet and regularising it's thickness. This was therefore done many times, and more experience would reduce or even remove the need for this step; if anything, it brought problems of it's own, as the sheet became slightly synclastic (like a bowl), which was hard to hammer out while keeping the thickness regular.

There was perhaps 8 to 10 hours of work on this piece, and the time/cost efficiency could be improved in several ways. First, practice would make the work faster, with less dithering and worry. Second, more confidence in the ductillity of the silver, which could have probably been worked for longer without annealling; if gold was used, the piece would work-harden only very occassionally. Third, a larger hammer and a larger ingot might have been preferable; it would probably only take as long to do the work, but more sheet could be produced at once. Lastly, doing the work next to a hearth would have saved time; the sheet could be annealed regularly and often without stopping the work - although that makes assumptions about having an assistant to keep the hearth hot, and having an assistant will speed up most processes.