Introduction: Brass Casting With Sand Cores

About: I enjoy working in wood and metal, doing overnight bushwalks, playing music, solving problems etc.

I have a friend who is in a vintage motorcycle club, and he sometimes asks me to cast parts for him which are no longer obtainable. This is the description of the casting an intake elbow, which had cracked and broken. It is a hollow casting, and I thought it might be instructive to show how I went about casting it.

The casting was done in 8 steps:

Prepare the pattern by adding thickness to the part

Make a core pattern

Make a core mould from the core pattern

Make a baked sand core

Mould the item in green sand

Place the core in the mould

Pour the metal

Cut off the gates and sprues, tidy up the casting


Furnace capable of 1200 degrees C

Domestic oven

Casting tools (crucible, tongs, etc)

Green Sand

Casting boxes (flasks)

Fine sand for the core

Corn flour

acetyl cure tile and bathroom sealant

vaseline or grease


Mould for the plaster

3 Kg brass

1 Kg Aluminium

50g Zinc

Butane torch



Step 1: The Pattern

The part was to be cast in brass, rather than in the original alloy. One of the problems was that the walls were thin (2 - 3mm). It is hard to cast brass into thin sections, as the metal often sets before the mould is filled.

To increase the wall thickness I fitted 1mm thick plastic rings in each opening and moulded wax up to the level of the rings, making the walls about 5 mm. This allowed more space for the metal to flow, and also provided some machining allowance.

The final part will also have threads and steps machined into it.

(See photos)

Step 2: The Core Pattern

To cast a hollow item, a core is needed, which means that the shape of the core must be precisely known. This particular elbow tapers from 25mm at one end to 32 mm at the other.

With the wall thickness increased, cardboard extension tubes were put on each end of the elbow (more on this later) the inside was coated with vaseline and silicon putty was used to get an impression of the shape.

The putty was made by mixing equal volumes of corn flour and Acetyl cure tile and bathroom sealant.

It is absolutely essential when doing this to have good ventilation and wear disposable gloves. The putty is very sticky and messy and the acetic acid fumes are intense. It's best to do it outside.

The flour is mixed in until the sealant and the flour are a homogeneous malleable mass, which is then quickly pressed into the part, trying to minimise holes and bubbles. This must be done quickly, because the putty becomes unworkable in about 10 minutes. It sets completely in about 1 hour and can be pushed out of the mould because it is still flexible.

(see photo)

This silicone rubber impression of the inside of the elbow will be the shape of the sand core that will be in the casting. When it is removed from the pattern it is coated with vaseline.

Step 3: The Core Mould:

An impression of the core shape is then made. I used plaster. I found a small plastic container which I modified it with cardboard so that the casting used the least possible material, but still held together (more on this later)

Small metal angles supported the silicone shape in the container.

To calculate the amount of plaster needed,

Calculate the volume needed in cubic centimetres and add 10% for wastage. Multiply this by 0.6. That is how much water you will need in millilitres.

Multiply the water figure by 1.42. That is how much plaster you will need, in grams. It works every time.

The plaster, was mixed, minimizing bubbles by vibrating and vacuuming and then poured into the mould until it came half way up the silicone pattern. After it set, the surface was trimmed, any bubbles were flattened out, then some registration marks were made so that after the mould is split apart, it will go back together in the correct way,

(See photos)

The mould was coated in vaseline, and the second half poured. Once set, the plaster was taken out of the plastic moulding box, the 2 halves carefully split, and the pattern removed.

If you were making hard sand cores using sodium silicate, it might be possible to mould the core in this plaster mould, but I use baked sand cores, and the baking tends to destroy the plaster, and the sand is often difficult to get out of the plaster moulds, so I use the plaster as a pattern for an aluminium mould, which overcomes these difficulties. (see photo) It means having to mould them in sand and cast in aluminium, but it isn't a huge task.

This is why I try to minimise the amount of plaster in the mould, Plaster wastage isn't an issue, but when it's cast in aluminium, large areas of metal with different thicknesses can cause shrinkage, which is undesirable.

Step 4: The Sand Core

To make sand cores,

I use a mixture of 20 parts of fine sand to 1 part (by volume) of corn flour, moistened with water containing 15% molasses by volume. The inside of the aluminium core mould was coated with grease and the 2 halves clamped together with a G clamp, ensuring that the inside edges line up properly.

Core mix was then mixed with molasses water to make it damp and crumbly but not wet, then rammed into the core mould. After this a piece of 4mm steel rod, was used to make a ventilation hole down the middle of each leg of the core, being careful not to let the rod touch the inside edge of the core mould. Without this hole there is a risk that the gases in the core will cause bubbles inside the casting. Optimally, each hole should meet inside the corner of the mould. (See Photo)

The G clamp was removed and the core mould placed into an oven at 125 degrees C for 1 hour.

When it cooled it was removed from the mould by tapping it gently while supporting it. It is quite fragile at this stage. Any flashing was lightly sanded off to create a smooth, even core.

Why is this core so large?

All casting cores must be able to be positioned accurately inside the shape of the pattern, and cores almost always extend past the pattern so that they can be held in the sand while the metal flows around them. In the case of this elbow, to prevent the core falling in to the bottom of the mould, it had to have one very long leg to counter-balance the other end. - hence the long tubes when the initial core pattern was made.

Step 5: Preparing for Casting:

The silicone core pattern needs to be placed back into the elbow so that when the sand is rammed around the pattern, the ends of the core pattern stick out and form a “core print” for the sand core to rest on.

To avoid stress fractures, I ran a bead of silicon sealant around each edge where the core pattern stuck out of the elbow, and to save ramming and preparation time, I used some silicone putty to mould a cube-shaped in-gate/riser against the core where I wanted the metal to enter the pattern. (See photos) If the gates and risers are all rammed up with the pattern, there is less loose sand and less mould collapse than if you cut the gates and risers later.

There are a number of ways that you can mould up a pattern like this: but because I was doing a few of these castings, I made bottom board which speeds up the ramming process.

(See Photo)

Step 6: The Casting:

Once the casting boxes (flasks) were rammed up, while the furnace was heating up, I put the cores back in the oven for about 30 minutes at 125 degrees C to drive out any excessive moisture.

When the metal was almost ready to pour, the cores were put into the moulds and a butane blowtorch was used to char the part of the core that would be inside of the actual pattern. This is just another precaution to expel as much moisture as possible and avoid “core blow” - bubbles and imperfections in the pattern. It is also important at the ramming stage to ensure that there are good vents where each end of the core will be, to allow escaping gases to vent out to the atmosphere. (see photo of gas venting)

After the cores are inserted in the flasks, the boxes are clamped together, the metal tested for liquidity and temperature, and about 50 grams of zinc is added to each 3 Kg of brass, to allow for zinc losses caused by vaporisation and to clean the metal. The metal is then poured steadily into the moulds.

Tip: it's important to have an ingot mould at the ready to take the extra metal in the crucible, as you cannot allow the metal to solidify in the crucible. (it will crack)

Step 7: Finishing:

Once the metal has cooled, it comes out of the sand, and goes into a bucket of water. After the core has been scraped out of the pattern, the sprues and any flashings are removed, and the pattern is wire brushed to remove as much sand as possible, - sand blunts cutting tools in an instant.

Metal Contest

Participated in the
Metal Contest