The embossed panels
The embossed panels are approximately 60 mm × 55 mm rectangles and an average of 0.004 inch thick; the originals are primarily silver gilt or with a gold enriched surface [Footnote 12]. The reproductions are stamped from 0.004 inch copper shim that has then been polished and electrolytically gold plated once embossed. The archaeological Hoard research literature describes these components as being ‘die impressed sheet fragments’ [Footnote 13].
There are five distinct iconographic images embossed on to the various panels the largest number of which are described as ‘standing warriors’ where these warriors face either to their left or to their right with distinct differences in their dress, armaments, and pose so a simple CAD driven mirroring or reversing of the image was not going to be possible.
The next grouping are the mounted warriors where there was only clearly identifiable physical evidence of the mounted warrior facing in one direction; however, the experts asked us to produce left- and right-facing versions of what they considered the original Helmet probably looked like and which was based on comparisons with other Anglo-Saxon, Scandinavian, and Roman examples, some of which are cited later in this paper. So, in this case, we were able to use our digital technologies to create a mirrored or reversed digital data set for the mounted warrior.
The final grouping consists of just two panels on each helmet and was, to us non-experts, intriguingly described as ‘The Priests’ who are two of the most warlike warrior priests you are ever likely to come across in that between them on each panel, they have six spears and two swords and are wearing fearsome-looking horned helmets!
Our original plan was to create the appropriate CAD STL file and in a castable photopolymeric resin, print the panels on our ProJet 3500 CPX 3D Printer and then cast directly into bronze; early experiments with this process proved disappointing in both achieving the desired thickness and an acceptable level of surface detail that would withstand the polishing process without the loss of important features. In Image 12, you can clearly see the horizontal ‘build lines’ that would eventually prove to be very difficult and time consuming to remove during polishing without compromising the warrior’s more important surface details.
Luckily on the SoJ DDM team working on this project was a colleague who is an apprenticed tool maker and die sinker with many years of experience in die and force making for the jewellery industry who is now the senior CAD/CAM tutor here at the SoJ. He proposed we adopt an innovative and cost-effective method for making the necessary sets of dies and forces in a resilient material that could be CNC milled yet not require the time consuming and costly post-processes, like furnace hardening, that are normally associated with more traditional dies and forces made from expensive tool steels.
This process started with his knowledge of a material called SikaBlock® [Footnote 14] which is a very dense model-board material resource which has a density of 1000 kg/cm3, good milling behaviour, a very low coefficient of thermal expansion coupled with good compressive and edge stability. It is most typically used for foundry pattern making, gauges, and other applications in mould and tool making. In use, we found it to have excellent dimensional stability when CNC milled, despite its comparatively low density, which was good for handling, yet coupled with a high-tensile surface which makes it an excellent choice for being easily mechanically workable which happily for us translated into being very machineable on our ‘lower end’ CNC milling machine which is more typically used to CNC softer materials like woods and wax for our student body.
Stamping tooling development
The first stage of creating the stamping tools involved cutting down on a bandsaw the as supplied fibre boards from approximately 1.5 meters square, and 2 centimetres thick, into useable and suitably sized blocks. Image 13, and the subsequent images in this section (Images 14, 15, and 16), will amply demonstrate how well the SikaBlock® fibre board accepts the rigours of the various cutting, milling and press stamping operations it was subsequently put through.
The edges then needed to be squared and made true on a Lagun FU-TV 125 miller
We then milled in the edge slots to enable safe and secure ‘dogging’ (locking in place) of the tools into the CNC miller and the large hand press that was to be used to stamp the various replica Helmet panels
Die and force cutting
Taking the data captured and created from the process of laser scanning and digitally re-assembling the multiple fragments in the museum’s Hoard collection and filling in the blanks are covered in some detail earlier in this paper. By the end of the digitisation and file creation process, we were able to create the necessary tool paths to allow cutting to begin using ArtCAM software (now discontinued as part of the Autodesk stable of software’s). We used a 2 mm and 1 mm diameter roughing ball drill and a 0.50 mm conical finishing tool. The die and force cutting was done on our Charlyrobot 4U CNC milling machine which uses Grail Pilot control software (Image 17).
Six millimetre diameter holes were drilled into diametrically opposite corners of each die and force and 6 mm dimeter steel dowels tightly hammered into the force to act as location guides during the stamping operations (Image 18).
The offset used to cut these dies and forces is a 2D offset only that subtracts the meatal thickness of the copper shim from the vertical sides of the force (or male) die. ArtCAM is perfectly capable of offsetting the 3D form and purists would say that this is the correct thing to do, to create a gap between the male and female dies that is occupied by the material being stamped. When stamping simple shapes this would probably work quite well.
However, in reality, what we tend to find when stamping complex shapes like these panels that tend to ‘pull’ the metal in multiple directions in a short space, and in a very malleable metal, is that detail on the upper faces tends to look as though it has not been struck hard enough. This is because the dies are held apart by the thickness of the stamping material more on the top surfaces than at the sides. Think in terms of a stack of inverted plastic drinking cups where the top edges will never touch. To allow for this we removed material from the male die on the vertical sides only. In ArtCAM, a boundary vector is created around the figure on the stamping; this is then offset inwards by approximately 80% of the metal thickness, so 0.08 mm in this particular case (Images 19 and 20).
Next step, the same as we would with steel stamping dies, was to test stamp some annealed 0.004 inch copper shim squares to look for splitting at any high spots on the die that would require easing. The stamping was carried out on the SoJ’s largest, No3, hand press with extra weight added to the counterbalance swing arm which gave us around 30 tons of downward force and with a false nose and heavy metal plate between the stamping tools and the ram of the press in order to widely distribute the downward force of each blow as evenly as possible over the stamping tools. This exercise revealed a number of ‘high spots’ on the tools that caused splits in the stampings to be identified that required some ‘easing’ of the tools (Images 21 and 22).
Image 23 (below left) shows the use of a spit stick graver to ease one of the panel stamping tools, use of such a graver was possible due to the softer nature of the SikaBlock® fibre board when compared to working on steel dies and forces, although a pendant motor and fine steel burr were also used quite extensively during these easing exercises. Both the stamped panels and the stamped strips required a pre-raise of the copper sheet and strip followed by an open hearth anneal* and scouring clean to remove oxide before a second, and occasionally third, and final raising blow on the hand press. The sticky tape visible on the right-hand side of Image 23 below provided for a bit of extra pressure in any places where the stamping is considered not quite sharp enough. With a hand press and semi-soft tooling, we could not just endlessly turn up the tonnage, as we probably could have done with a hydraulic press and hardened steel tooling.
*Annealing — when being worked (hammered, rolled, stamped etc.), metals harden due to changes in their crystalline structure and need to be heat treated in order to soften the metal back to a more malleable state. Typically, in this project, the annealing was carried out on an open hearth using a flame generated by the mixing of gas and air through a handheld torch and applied until the metal acquired a deep cherry red colouration [Footnote 15].
Progression or follow-on tooling
The Helmets project also required the development of stamping tooling able to create two different repeating strip patterns, one that was unofficially known by us as ‘the ducks’ which are described by the Anglo-Saxon iconography experts to be a representation of a ‘zoomorphic quadruped’ the second strip was a much more easily identifiable ‘kneeling, or running, warrior’. Luckily, both had an identifiable repeat line that gave us a repeat blow length we felt could be accommodated using a similar stamping tooling method as that used for the panels to produce the progression force and die in SikaBlock® whilst also manually feeding the strip through. In the case of the warriors, it was a dotted vertical line after every fifth warrior and for the ducks, it was a less visible ‘witness mark’ after every sixth duck (Image 24).
In the case of the ducks, we constructed the tool with a seventh duck in only the die which was used to locate the strip into the tool after advancing it forward by six. The seventh duck deliberately did not get re-stamped as this could have led to additional splitting issues, also note that the ducks, or quadrupeds, and warriors are deliberately not identical as per the original strips.
The warriors were easier to index as they were in two groups of five with a dotted dividing line, so they were advanced in their groups of five with ten warriors cut into the die and only five into the force. The stamping of both of these strips was again done on the large No3 hand press and the strips were being torch annealed as required, the strips were also being advanced and located by hand. Early strip stamping trials resulted in a decision to double up the thickness of the bases and tops of these tools compared to when being used for stamping the panels as they were found to flex a little as they were being stamped, this doubling up also helped to distribute the press ‘blow’ more evenly along its length rather than being concentrated through the centre of the tool directly below the headstock. The progression tooling was also doweled to improve the accuracy of each blow of the tool; it was also necessary to develop a strip guide and feed system in order to present the annealed copper strips evenly and aligned straight into the progression tools. The 0.004-inch-thick copper strips had been sheared to a width of 25 mm on a small hand-cranked shear we have here in the SoJ and finally open hearth annealed (this annealing process is described earlier in this chapter) in one of our silversmithing workshops using a gas and air handheld torch, after annealing they were pickled and cleaned with pumice* before rinsing and drying. Again, some splits were observed in the stampings, despite annealing and a pre-raise blow, so the forces or dies were eased as appropriate in much the same way as the panels (Images 25, 26, and 27).
*When using an open flame to anneal dark oxides will form on the surface of the copper these are removed by first immersing in a dilute acid solution (pickling) which loosens the oxide layer which is then removed by rubbing pumice powder over the metal and scouring the surface with a toothbrush under warm running water.