top of page
Wooden Hut
  • Writer's pictureDylan Crismani

Making a Sonambient Sculpture


In mid 2023 I was approached by the curator of the Samstag Museum of Art, Erica Green to deliver a music performance for an exhibition in 2024 titled Mulka Yata (The Knowledge of Place). The concept of the exhibition was for a number of artists to respond to the work of Adnyamathanha artist Kristian Coulthard, and his family’s cultural practice of mulka wood carving, and his connection with and knowledge of Adnyamathanha country. The concept of the exhibition was excellent, and Kristian’s work (and that of his family) outstanding, as was the work of the other artists involved in the exhibition.

Erica had not approached me looking for a musical sculpture, but this had been something playing on my mind since I had returned from America earlier in the year, having encountered the work of Harry Bertoia whislst travelling on a Churchill Fellowship. Essentially Erica was looking for some music for the opening night of the exhibition, and I pitched a few different ideas that I thought she and her colleagues may be interested in. In the end, she decided that the best proposal I had put forward was to produce a sonambient sculpture in the style of Harry Bertoia, with my father Christopher Crismani who is a visual artist, which could be displayed in the exhibition and which I could deliver a performance on for the opening night. Following this initial meeting to discuss the potential to work with the Samstag Museum of Art, my father and I set about seeking funding from Arts SA to support the project. Later that year Arts SA generously agreed to support the project. Without the support of Arts SA we would not have been able to realise the project, and particularly we would not have been able to realise it at the scale which we did.


Failed attempts

The very first attempt I made at putting any kind of sonambient type sculpture together was actually around six months or so before Erica Green contacted me about the Mulka Yata Exhibition. In my naivety I was trying to attach steel rods into a slab of wood – shortly after I had first encountered the work of Harry Bertoia. The wood was too thick and too soft and the holes I had drilled were immensely too wide. The experiment was indeed a complete and utter failure.


The second attempt involved threaded aluminium bar and plate. Slightly better than the wood but still poor for various reasons (the plate too thin, the resonance of material average, the bars wayward in their orientation, and so forth). Despite these initial attempts being complete failures, each refinement, each change, each new attempt, each critical evaluation of failed process and method brought Barraabarraa closer to reality. Tangentially, since Barraabarraa has been completed, I have spoken to a few people (boilermakers included) about how they would go about solving the problem of attaching the bars to the plate and interestingly multiple people have suggested threading the bars. It would be interesting to see if anyone could make the instrument work using this method. I also suspect, but do not have the specialised engineering knowledge to confirm, that the quality of the welded joint has a direct impact on the resonant quality of bar. If the joint is good, then the bar will ring. If the joint is poor, then the bar will go thud. I suppose the logic being that the plate acts like a soundboard, and if there is inadequate connection between ‘string’ or ‘bell’ and soundboard, then there will be little sound to be amplified.  


First brass delivery and first jig


Initially we got a smaller order of brass for doing some initial proof of concept testing. Just a small offcut plate, and a few bars of different diameters. During these initial tests we spent some time investigating how Bertoia got his pieces so perfectly straight. During one of my discussions of the piece with my colleague Luke Harrald, he suggested using a jig and upon further investigation we of course discovered that this was indeed how Harry Bertoia achieved such perfectly straight lines in his work. The first jig (one of many) that we used was just some plywood clamped to some cabinet shelves.

First delivery of brass (6mm thick plate and rods of various thicknesses):

First jig (a few piece of loose plywood which were later clamped to the shelf). Even this rudimentary set up results in very straight positioning of bars:

The first successful test piece helped establish that the brass bars all had certain lengths beyond which they would begin collapsing under their own weight, in a completely wayward manner. This helped determine that the thickest bars would be the longest, while the medium and thin bars would be slightly shorter to keep the bars appearing as straight as possible.

The first test piece also helped establish that the brass plate – which was only around 6mm thick - would need to be considerably thicker at around 12mm in order for the plate to have enough weight and strength to bear around 30 – 40 kilograms of brass bars.


After we conducted the initial tests, we calculated roughly how much brass we would need and placed a couple of orders which ended weighing in at around half a tonne.


First big delivery of brass bars (a few hundred kilos across two of these boxes):

The thicker (12mm thick) and much more suitable brass plates water jet cut to size for absolute precision:

The eight plates were 455 x 455 x 12mm, and the bars, of varying diameters were 3,600mm. To keep the plates all together in the one place at around waist height we needed to buy a piece of MDF and reinforce it with some 2x4s to bear the weight.


There were then three main tasks that lie ahead – cutting and finishing the brass bars, drilling and finishing the plates, and joining the bars to the plates. I wish it were as simple as it sounds, but a number of these tasks expanded out into much larger and more complex battles. Luckily we were adequately resourced to manage such a position.


Due to the large amount of materials that were needed and the considerable expense of the materials, we needed to do some calculations on exactly how many brass bars were needed to achieve the lengths we wanted. The final sculpture has two rows of bars that are 1,700mm, two rows that are 1,600mm, two rows that are 1,500mm, all the way down in 100mm increments to the final two rows that are 600mm. The brass bars only come in lengths of 3,600mm, and we were working with different diameter bars also. So we spent some time calculating how we could divide up the 3,600mm bars with maximum efficiency and minimum waste.


When the bars arrived, we then needed to implement this plan weeks later, revisit and double check our calculations (which thankfully were correct!), and then cut the bars. Cutting the bars and deburring both ends of the bars took around two days in total. It doesn’t seem like a particularly intensive task, but the scale at which we were working caused this to take rather a long time. We then finished the bars by cleaning with mineral turpentine, and sanding with an orbital sander to create a finish that looks like the bars are wrapped in brass foil. Apologies, no photos of the bars close up with beautifully deburred ends and foil finish.


The next step was to mark up the points on the plates where the holes would be drilled. We used a 1:1 scale piece of cardboard to help us determine the grid size, and also to determine where the snake would be placed (the snake appears in the negative space of the final piece – i.e. there are no rods placed where the snake runs).

The brass plates with the grid cut, and the paper snake cut out to determine where to not place any holes:

The tail of the snake, at this point centre punch has been used to mark points where the lines meet:

To mark up the grid we used a regular woodworking marking gauge, but eventually needed to make a custom marking gauge as the standard, and almost exclusive, length of these tools is around the 15cm mark. The custom gauge allowed for marking lines up to 30cm from the edge of the plates - about twice the length that a standard gauge would allow. Once the grid was scored, the points where all the lines met were punched with a centre punch.


Pilot holes of around 2-3mm were drilled initially to ensure accurate placement. Once the pilot holes were made, then the holes were drilled through to the three diameters of the brass bars.

Testing the flap disc finish on the plates:

Again working from what information we could gather about Bertoia’s work, we decided to make countersunk holes so that we could make little pools of silver solder when attaching the brass bars to the plates. The countersunk holes were all 13mm across, and about 3-4mm deep.

All the drilling ended up taking quite a while. In total the sculpture has roughly 1,200 holes (give or take a few). So in total I think I ended up drilling around 3,500 holes – (1) pilot hole, (2) hole at final diameter, (3) countersink. There were even more holes drilled later on when I realised the tolerance was too fine between brass bar diameter and hole size…. more on that later.

Ordinarily when drilling metal I set the drill press to the slowest speed, so most of the time my press is set to the slowest speed. When working on this project however, I did end up breaking the rules a little. Because so much drilling was required and drilling at the slowest speed was so time consuming, I ended up cheating a little and using a slightly faster setting. I only ended up breaking two or three drill bits in total, and I rotated between around 15 – 20 drill bits as I blunted them rather quickly.

Once all the holes were drilled we needed to finish the plates prior to brazing as once the bars were attached to the plates we wouldn’t really be able to do too much finished on the plates. We trialled a few different finishes including mirror finished polish, orbital sanding, and flap disc (on angle grinder). In the end we settled on doing a few different passes with the flap disc to create a kind of 3D finish.

Arriving at the final finish for the brass plates (multiple passes with the flap disc):

Once all the holes were drilled we needed to shift our attention to making the jig(s). We started off with a standard shelving unit from the hardware store. Clamping the various levels of shelving to the plates we drilled holes in the exact orientation of the plates. The only problem with this approach was that standard shelving units are pretty flimsy in their design and although they may be rated to hold 200kg per shelf, they’re not designed to be tipped over repeatedly which is what we needed the jig to be capable of in order to remove each piece of the sculpture.

After almost completely destroying the sides of the shelving unit, we re-built to the point where little was left of the original shelving unit. Solid steel angles were used in place of the flimsy bits of 0.75mm metal angle that are normally used for these units. Shelves were screwed in place, and the whole unit was cross braced.

This is the shelving unit jig in more or less original state, note the extremely flimsy sides, and no cross bracing. Shelved are not securely attached (other than by gravity):

Revised jig, little of the original uni remains. New steel angles used instead of flimsy tin full of holes. Multiple cross braces used, shelving sides now screwed in place. Dad and removing the MDF sheets:

In addition to the main jig, we needed to create two additional components to get the whole thing working (and we also ended up making a third component in the end). The first component are the series of little staircases that appear on each level of the shelving unit. These mini staircases set the height of the brass bars – sculpture is oriented upside down in the jig, so that silver soldering can be done on the underside of the plates. This additional component was only necessary because we had chosen to have brass bars of various lengths. If we followed Bertoia’s path of all bars being the same length, this wouldn’t have been needed.

The second additional component that we needed to make was the little wooden box on top of the jig. This was needed so that we could orientate the brass plate above the main jig in the correct position before threading through the brass bars. I lost track of how many times I threaded the brass bars through before figuring out that this was needed.

This little box also needed to be fastened to the main jig - for when the whole structured is tipped over to thread the rods, and then to remove the sculpture. It also needed to be easily removed, as during the brazing process this box structure needed to be replaced with heat bricks to create a mini hearth which would capture a little bit of heat, and also prevent the MDF sheet from spontaneously combusting due to the high temperature of the brass plate.


Older jig (before it got completely ruined by being tipped over repeatedly), with brick hearth:

Completely revised jig. Top shelf reinforced to bear weight of the brass plate. Removable box securely fastened on top of top shelf, will later be replaced with hearth prior to brazing. Little staircase in position:

Once all the jigs were constructed, we were almost ready for brazing. The plates were sat in the little box on top of the jig, and the brass bars threaded through. At this stage we realised that the holes – being drilled to exactly the same size as the bars – were a bit too tight. So we did a little more drilling.

More drilling:

The difference between metric and imperial proved to be a good difference to bore the holes slightly wider but not too wide. Once all the bars were threaded, it was then inevitable that we needed to prop some of the rows up slightly (on the little staircase) and then trim some other rows down a little with the angle grinder. The final result we were aiming for were bars that were almost perfectly flush with the plate.  

This plate is pretty much ready for brazing, a few bars need to be raised up a little so they are flush with the bottom of the plate:

Firing up the torch:

Learning how to braze was fun, but the material and the size of the material proved to be a significant challenge. Ideally we could have baked the brass plates prior to brazing, but the nature of the design meant that even if we could have baked the plates, they would have been cold by the time we’d finished threading the brass bars through.


In hindsight, a second oxy acetylene torch would have been ideal – one torch for continual heating of the work piece, and a second torch for doing the actual brazing. The main challenge was that with a single oxy acetylene torch it was difficult to get enough heat into the plate for the silver solder to take very well.


The second reflection I would make here is one to do with flux. I understand that many sources advise to use flux when doing this kind of work. I personally found that I got the best and cleanest joins without using any flux at all. I found this product to be more of a hindrance than anything and if using it in the future I would use it in only the smallest quantities possible.


Hot off the press, weighing in at around 60kg each, it was very difficult to move the piece of the sculpture around so we initially used stock standard dollies that you can buy from the hardware store:

While useful for moving the pieces of the sculpture around the workshop, these dollies would have been a precarious way to transport the sculpture on to and off of the moving truck. As such, we ended up needing to produce a custom dolly for each piece of the sculpture that acted as both a trolley for transportation as well as a semi-protective box around each piece which could be strapped to the truck for transport.

On the home stretch, six finished pieces in the transport units:

It was not my finest carpentry, but the custom dollies were functional, and they served the task that we needed them to serve.


From here, the pieces of the sculpture were covered in bubble wrap, wheeled onto a truck and strapped using ratchets to the side of the truck. When I arrived at the Samstag Museum of Art, Erica had arranged for a team of young, strong, fit, and enthusiastic assistants to set the sculpture in place. Erica had decided to place the sculpture on a large custom built plinth (around a metre tall). It was time consuming getting the sculpture in place on the plinth but the museum had the tools and resources needed to achieve it and in the end well worth the effort when you see the final result and the sculpture towers over the viewer.  

Barraabarraa set up in the Samstag Museum of Art:

Thank you to Arts SA for their generous support of this project

73 views0 comments

Recent Posts

See All

2023 Electric Cristal

Follow the link to see a brief youtube video about the Arts SA funded 2023 iteration of the electric cristal.

Cristal Resources

The following post does not contain any knowledge of my own, but is simply a collection of resources relating to Cristal instruments. The list will grow in time:


bottom of page