Structural

After researching several different types of insulation used on other foundries, including firebricks, rockwool insulation, and fiberglass insulation, I decided on using refractory cement to create an aircrete. This decision was based on cost, how low the thermal conductivity of the material was, and how easy it was to make the foundry.

Aircrete is a cement or concrete that has been fluffed up while mixed to aerate it. Air has a low thermal conductivity, and by aerating the cement it lowers the thermal conductivity of the cement, allowing the foundry to lose less energy to it's surroundings. I chose to use refractory cement for my aircrete because of refractory cement's ability to withstand high temperatures without adverse reactions, such as cracking or crumbling.

In order to make aircrete I also had to design a system for getting air into the cement as I mixed it. There are plenty of resources online on how to create a bubble foam to mix into the cement and aerate it. I created a slightly different apparatus to create the foam to make it easier to build with materials we had around the house and 3d printed parts. This device, as well as foam created by it, is pictured below.

Since the aircrete is made at home, each batch has a different density and thermal conductivity based on how much air is mixed into the cement. In order to approximate what thermal conductivity my foundry's walls would have, I made samples of aircrete and took their densities.

After creating this shaving cream consistency of bubbles, I mixed refractory cement as instructed on the bucket, and then added the bubbles and mixed it. As you can see in the pictures below, it's noticeably a different consistency than regular cement and is very lofty and incredibly light compared to regular cement. In the blue rubber mold we placed samples of different densities depending on how much bubbles we mixed it. It turned out that by mixing in too much bubbles the mixture collapses as it begins to set from the mixture being too wet. The bottom two sample in the picture began to collapse because I mixed too many bubbles into the cement.

By comparing the densities of my aircrete to densities of an aerogel infused concrete in a research paper titled Foam concrete-aerogel composite for thermal insulation in lightweight, I was able to approximate the thermal conductivity of my foundry based on a graph that compares density of their aerogel concrete to thermal conductivity (pictured to the right). As you can see as the mixture becomes less dense and more air is mixed in to the thermal conductivity lowers. The lowest density aircrete I was able to achieve without the mixture collapsing was about 500 kg per m^3, which based on the graph I would expect to have a thermal conductivity of roughly .15 W/m*K.

Having an approximation for my aircrete's thermal conductivity allowed me to calculate what thickness I should make the walls of my foundry in order for it to reach my target maximum temperature of 2200 degrees F. By calculating the R value of the foundry walls at different thicknesses and then applying this value to calculate both the rate of energy transfer leaving the foundry at maximum temperature and how much energy would be needed to melt a mass of copper if it started at room temperature. From these calculations I was able to get some idea of how much energy I would need the heating element to output, about 1.6 kW if I made the walls of my foundry 11 cm thick. This was a nice balance for me between thickness in walls and the amount of energy the wire would have to output.

Electrical

While researching what heating element I wanted to use in my foundry, I found several options, including using a heating element taken out of an old microwave or stove, a pre coiled resistive wire such as nichrome, or buying and coiling my own resistive wire. I settled on the last option due to the ability to easily adjust how much energy was being put into the foundry by how long I made the heating element and out of what gauge wire.

I based my decision for what gauge of wire to buy off of the energy need I calculated based off of the foundry's walls, detailed above in the structural design decisions. Since these values I calculated are incredibly rough approximations, I used a safety factor to make sure I would be able to get up to heat and specced my heating element for 3 kW. Based off of the inside diameter of the foundry, how many times I wanted the heating element to spiral around the inside of the foundry, the size of coils I created in the heating element, and the spacing between these coils, I was able to calculate the total length of the heating element wire and what resistance the wire would have to be to output 3 kW at this length. I decided upon 20 gauge Nichrome 80 resistance wire, coiled .5 inches diameter with 231 coils going around the inside diameter of my foundry 3 times, for a total length of 59 inches. This wire was rated at .6348 ohms of resistance per foot of wire, making the total resistance just under 20 ohms. I ran this circuit off of 240 V, pulling 12.5 A.

I also decided to build my own temperature control for the foundry. By using an arduino uno I was able to control a relay to turn my foundry on and off with simple controls. My arduino code performs the following functions:

• allows the user to change the target temperature that they want the foundry to heat up to using a potentiometer

• turns the foundry on and off based off of a set amount of error that determines how big of a range of temperature the foundry should be allowed to fluctuate in around the target temperature

• uses a LCD display to show the user what temperature the foundry is currently at, what the max temperature it's reached is, what the target temperature is, and whether the foundry is on or off

• records data to an SD card while plugged in (current temperature, target temperature, if the foundry is on, and real time)

The code can be found here on github.

Structural

In order to build the refractory cement aircrete structure of my foundry, I cast the aircrete into a metal bucket that I used for the exterior of my foundry, along with it's lid.

For the lid I created a form to cast the aircrete into with foam board and hot glue. To make sure that the aircrete didn't come apart from the bucket lid, I screwed loops of steel into the bucket lid that the aircrete was then casted around. After the aircrete dried, I removed the foam board.

Electrical

To get the right length of heating element, I followed the specs outlined in the designing the electrical section. SPECS. In order to wind the wire into the correct size coils, I attached a metal rod with the right diameter to a drill and rotated the rod while winding the wire around the rod. It is crucial to keep the wire winding tightly around the rod in coils close together to have the coils of the heating element be evenly spaced and the foundry evenly heated. After winding the wire, I took the wire off of the metal rod. I held each end of the wire and stretched it to the proper length in order to fit the spiral in the wall of the foundry. As long as only the two ends of the coiled wire is pulled, the coils will space themselves out evenly if they were originally spaced evenly on the metal rod.

After creating my heating element I wired the rest of the foundry and temperature control.

The first thing that I tried casting was just soda cans and other aluminum scraps into these muffins out of a steel muffin tin.

Afterwards, I tried melting down aluminum wire into a Moai head, like the ones found on Easter Island, and the aluminum muffins into a velociraptor. To make these molds I started out by 3d printing a Moai head and velociraptor. I glued these down to a piece of scrap wood, and surrounded the 3d printed parts by disposable cups. I inserted tubes from the points that are deepest in the mold to the surface to allow air flow and keep from forming bubbles. I then filled these disposable cups with plaster and let the plaster harden. To get rid of the plastic inside the mold, I heated up the molds inside the foundry and let the plastic melt out of them into a pan below. When the molds were done, all of the moisture had left them from being inside the foundry for so long, and there was no plastic left on the inside of the mold.

The molds are likely to break if subjected to too much thermal shock (the temperature of the mold rising really quickly). This can be made worse by any moisture left in the mold, so before pouring into the mold I left the mold and bucket of sand out in the sun for a few hours to help evaporate any moisture in it. Once the aluminum had melted completely in the foundry, I poured it into the cast while it was sitting in the bucket of sand. Once the metal and mold had cooled, I removed the mold from the sand and gently tapped it with a hammer to break the mold off of the part. After the part cooled enough to be able to be held without gloves, I removed imperfections by sanding the part down. The maui head on the left is after the cleaning process, while the maui head on the right and the velociraptor have not been cleaned yet.