Reconstructing the Trundholm Sun Chariot

Anders Söderberg


The Trundholm sun chariot was found in 1902 in the Trundholm bog at Sjælland, Denmark. In order to celebrate the anniversary of the finding, and to provide Trundholm with a replica of the chariot to be used in childrens history classes, a cooperative project was initiated between craftsmen from different European countries who reconstructed and cast the different parts of the chariot. This was made during the summer 2002, and the parts were assembled at an anniversary festival in Denmark in September 2002.


The chariot, with a horse and an ornamented sun-disc, measures 59 centimetres in length and it dates from the 13th or 14th century BC. Its wheels are made of solid bronze, while the horse and the gilded sun-disc are made around clay cores. The artifact tells us not just about Bronze Age religious customs, but probably as well a lot about ancient designs for full-scale war chariots. It is exhibited in the National Museum, Copenhagen.

The piece I made for this project, was the central pole connecting the chariot to the horse (the dark piece in the picture below). It measures about 44 centimetres in length and is not a very spectacular part of the artifact, but an important one and just enough of a challenge to suite me.




I decided to cast the piece by the lost wax-method in a ceramic mould. It's rather obvious that at least the horse of the original chariot were cast in this way, as it was cast around a clay core onto which the original wax-model most probably was made.

The Danish archaeologist Preben Rønne has introduced convincing evidence that many Bronze Age-objects with ornaments were cast by the lost wax technology. Spiral ornaments on several bronze collars and tutuli, show obvious signs of having been died onto soft wax with dies made from metal wire twisted into spiral-shapes (Rønne 1991).

I started with casting a simple copy without ornaments from a wax model made in 1:1-scale. I made this in order to estimate the shrinkage from original to cast object. After this, I could make the real wax model a little bit larger than wanted, in order to get a bronze-object in perfect 1:1-size in the end. A technical problem when casting copies and replicas, is that ceramic moulds shrinks during firing. The bronze as well will shrink slightly during solidifying and cooling. The shrinkage of mould + metal usually keeps within a range of 2.8 - 5.6 % when making smaller objects.

This first casting became a rather ugly little piece as the mould broke during firing, but enough for calculating size. The lenght-shrinkage showed to be as modest as 2.9 % from original wax to cast object, which I found rather remarkable for such a relatively large and long object.

When I later made the moulds for the final castings, I reduced the percentage of sand in order to get a less fragile mould; which meant that I had to be prepared for a slightly higher shrinkage percent.


Making the wax model

The model was made in bees wax, a material that we can be sure was accessible to the Bronze Age man. If you ask a modern goldsmith, you'd probably be told that bees wax is far too soft and sticky to be used for making models for casting; and I'd say - yes I agree, but you do get used to it!


A wax bar is prepared. The cast piece in the background is the ugly little trial-piece I made.


The piece is cut out of the wax, and polished with a spoon.


I started polishing the wax piece with a spoon, but later I abandoned the spoon in favour of my own fingers; you can attain wonderfully shiny surfaces on bees wax by rubbing it with your finger, wet by some saliva - surfaces reminding a lot of the fantastic smooth surfaces you sometimes find on well preserved Bronze-Age objects.

These surfaces have often been assumed being attained by days and weeks of polishing the cast metal objects, which themselves have been assumed originally being of a fairly low qualities; all as an example of the tremendous sacrifice of surplus-time that was put into these things - surplus-time that the societies possibly didn't really dispose but did put in anyway because of the (religious?) importance of the craft itself.

I think, in cases when the lost wax technology was used (Rønne 1991), these people could attain almost perfect surfaces directly out of the mould; just needing fairly modest polishing afterwards. Experiments have shown that the lost wax method and simple ceramic moulds of ancient type can reproduce metallic surfaces surprisingly true to the wax-originals - almost perfectly reproducing details as small as 0.2 millimeters in size (Meeks, Tulp & Söderberg 2001).

This doesn't mean that large quantities of "surplus" time wasn't put into this craft by the Bronze Age societies. Wax working will take a lot of time as well, and so will the mould making and the casting procedure. And there will as well be need for some afterwork even on a "perfect" lost-wax casting. But I don't think we should exaggerate it - the need for polishing the cast object afterwards, is set by the volume of work you put into the wax model and the mould before casting.


The ornaments have been pressed into the soft wax. Finished!


Right: A waffle patterned die which I cast in connection to the work of Meeks & Tulp. It's not an all-perfect casting, but in my opinion it shows the potential of the lost-wax method and "ancient" ceramic moulds in a convincing way. Width of die 11.8 mm.

See related web-resources at bronze die for making box pattern backing foil in cloisonnÚ cells and at Scientific analysis of the gold disc-on-bow brooch, from the Laboratorium voor Conservatie & Materiaalkennis LCM, Groningen, The Netherlands.


Making the mould

The mould-material I usually use is a mix of Southern Swedish - Scanian - glacial earthenware clay, fine glacial silt I've found in northern Sweden, chamotte from ground moulds and organic temper; in this case sawdust from a yachtyard in Stockholm; and sand.

Making a clay-mix mould around a wax model isn't that easy. You can't, like when you make ceramic shell moulds, simply dip the model in a liquid mould-mix. This means that you have to wrap the rather heavy material around the model, and you must do this carefully as your worst enemy is air-pockets trapped next to the wax - these will be very visible on the surface of the cast object in the end.

First, I usually put a string of clay-mix on the table, which will serve as an under part of the mould, and I put a thin layer of clay/silt-mix on top of it. This mix is fine enough to pick up even the finest details of the model's surface, and I put a layer on the under-surface of the model as well. Then I gently press the model into the clay.



After that, I usually wait a while for the under part to set. By doing this it'll be stable when I make the upper part of the mould, which is an advantage. This one is made like the underpart; a thin layer of silt mix is applied on the wax model and the upper mould part is pressed onto the under part. The parts are thoroughly sealed together along the sides. The mould is now ready for drying, which will take a few days or up to a week - depending on size, shape, temperature and air humidity.



Three moulds were poured into on August 29, resulting in one fractured mould already during firing, one poor chariot pole (surface pitting - gas problems) and one OK enough to pass even my strict and critical control.

The moulds were fired in a charcoal hearth, and the bronze was melted in a graphite crucible in an electric goldsmith's furnace. The bronze used was a 12 % tin-bronze of Swedish standard SS 5465.

After studying the last piece a bit closer, I had to realize that casting a long and slim object like this isn't really piece of cake. Its surface showed patterns of rather large dendrites towards the end - the bottom of the mould where the hottest bronze slowly have solidified. Towards the top, during the pour, the metal in the crucible had became gradually colder; which means that it solidified a lot quicker at the top of the mould. This part of the pole showed a very smooth surface, with no visible larger dendritic pattern - a surface that wouldn't have needed much polishing to get shiny.



Above: True backyard foundry - perhaps a traditional West African founder would recognize the type of mould and the simple environment?

Right: Opening the second mould. The piece's surface is rather heavily oxidized.


The Chariot Pole

Depending on the rather poor result of the pouring, I didn't get the chance to proove my idea of getting a piece without any larger needs for polishing directly out of the mould. Because of extensive dendrites and gas pits on the surface, the object came to need a harder polishing than I'm used to when making smaller objects.

Of course, this is the largest object I've so far have cast in an "ancient" ceramic mould. I've now experienced that when I go up in size, I suddenly have to face different problems than the ones I've got used to when casting small jewellery. Quite naturally.

Nevertheless, it was a great experience trying to make an object of this size, and getting a chance to test the limits of the mould material. It did stand the test, but I think going up further in size would maybe be a bit too much for the lime-rich earthenware clay I used. When I poured the bronze, the top of the mould actually bent a little bit at the top from the heat - like soft bread in the oven.


Part of the pole before finishing, showing a rather uneven dendritic and pitted surface with flashes where air pockets have been trapped next to the wax when making the mould.

Probably the forming of the rather large dendrites was caused by slightly too hot metal poured into a too hot mould.

Below: The polished pole.



Because of the poor surface, I had to polish the piece rather heavily. There still are a few pits visible on the surface, and my opinion is that these actually save the piece a little - I wouldn't have liked it at all if it had been a 100 percent smooth and perfect.

Then, what about the shrinking? I estimated that the shrinking from wax over mould to cast object would be 4 %, and from that estimation I decided the size of the wax original. In the end, the actual shrinking showed to be 3.8 %. A rather close hit!


Was this piece at all cast in a "Bronze Age Way"?

Let's first make clear that archaeology, more than many other disciplines, is a noble art of probabilities and interpretations, as our source materials mainly are very sparse. This means that academic reconstructions of ancient technologies are dependant on qualified guesses, modern ethnographic studies and lots of never ending discussions. Accordingly, a practical reconstruction must as well be considered an interpretation and not a truth.

Second; this was a very anacronistic reconstruction I made, with no wider ambitions to make a "true" reconstruction of the process itself. The goal was to reconstruct a chariot pole, no matter how I'd choose to make this. But I still think there are a few parts of the process, that a Bronze Age founder could have recognized fairly well:


    The making of the wax model is one of those parts - bees wax was of course fully availabe in the Bronze Age.

    The mould material is mainly based on analyses of Iron Age and Medieval moulds. There are fragments of Bronze Age ceramic moulds found, but I haven't studied analyses of them closer. Probably, they wouldn't look very different from the ones I made. The replacement of silt with sand from the pet shop for the outer layers, is my own invention as my silt-source is situated a very long way from where I live - I have to use it sparsely. A Bronze Age mould would, like a medieval one, possibly have been consisting of silt and clay all through. I mainly just use it next to the wax model. The ancient founder would probably have used manure as organic temper, but I stick to saw dust.

    The lost wax mould. As mentioned earlier, Preben Rønne has introduced proofs for use of the lost wax technology in the Bronze Age. It must as well be accepted that objects cast around clay cores, like lures and horns would have been made by the lost wax technology. Besides, there's an old and forgotten (?) indication on lost wax casting described by E. Voce, referring to moulds made of stone and bronze: "...bees wax have been found in the crevices of an ancient mould" (Voce 1951:112). This may support the idea that those moulds could have been used for reproducing wax models. Voce doesn't further describe this mould, where it was found or if any analyses of it are published.

    The firing of the mould was made in a charcoal fire. Charcoal is timeless.

    The alloy used, SS 5464 - CuSn12, a 12% tin bronze with a low Pb-content (less than 0.8%), is rather similar to the alloys in many Scandinavian Bronze Age objects (Oldeberg 1976:120ff).

    The electric Italian goldsmith's furnace that I used for melting the bronze, would have been rather useless 3400 years ago. Not that the distribution net wasn't there for shipping it up to Scandinavia - it most certainly existed - but the guys probably would have had a tough time finding a suitable hole in the wall for connecting the cable ;)


Assembling the chariot

The chariot was assembled on Saturday September 7th 2002, at an anniversary festival close to the Trundholm bog where the original was found in 1902. Unfortunately it couldn't be made fully complete in time for the celebration, as the horse and the wheels were missing because of circumstances beyond the different craftsmen's control.

We had six iron wheels available made by a Danish blacksmith which we used, and the horse was replaced by a wooden model. It actually looked rather neat; and as the chariot wasn't to be displayed until after sunset, the wood and the iron didn't disturb the impression during the anniversary procession in the evening. Later on, the chariot will be made complete with bronze wheels and a cast bronze horse.


The chariot. Not complete, but giving a good hint of what it will look like when finished. The sun-disc is a splendid piece cast by art founder Kurt Hansen, Denmark. The axles were cast by Karl-Heinz Moser-Fiedler, SkŠlsk°r, Denmark. The wheels were made by Danish blacksmith Hans Erik Rasmussen, but new bronze wheels will be cast by archaeologist/bronze founder Ken Ravn Hedegaard, Denmark. The horse will be made by museum's director Frank Erstling, Friedland, Germany. The aim is to finish the chariot before December 21 2002.


After visiting the anniversary show, I went to Copenhagen and studied the original chariot at the National Museum. There are visible differences between the original and the replica we made, and this illustrates the problem of making replicas from drawn patterns and photographs without having access to the original object. Every artist's eye and hand the impression of the chariot passes through, is a filter stating an interpretation - the less such filters the better the replica in the end.

And now? Well, maybe I'll start building a sun chariot of my own - I've already got a mould for the pole, and that's a good start!



Meeks, N. Tulp, C. S÷derberg, A. 2001. Precision lost wax casting. In: Proceedings of the 1st International Workshop. Experimental and Educational aspects on Bronze Metallurgy, Wilhelminaoord 18 - 22 October 1999. Tulp, C. Meeks, N. Paardekooper, R. (eds). Vereniging voor Archeologische Experimenten en Educatie. Leiden.

Oldeberg, A. 1976. Die ńltere metallzeit in Schweden II. KVHAA monografier 53. Stockholm.

Rønne, P. 1991. Forsøgsarkeologi og bronzealderens ornamentik. In: Eksperimentel arkæologi 1991. Madsen, B (ed). Lejre.

Voce, E. 1951. Bronze Castings in Ancient Moulds. In: Notes on the Prehistoric Metallurgy of Copper and Bronze in the Old World. Occasional Papers on Technology, 4. Pitt Rivers Museum. Coghlan, H. H. (ed). Oxford.




Main Page - Photos © A Söderberg & Johan Borg - October 8 2002 - Updated April 14 2007