Skip to content

12. Molding and casting



For this week we need to:

  1. Mill a mold using our available CNC machine.
  2. Cast an object using the milled mold.
  3. Test materials for the group assignment

Things I’ve learn so far:

  • 3D modeling for molding.
  • Process 3D CNC operations.
  • Machining Wax bricks for molds.
  • Calculate volume of material needed.
  • Mix food-safe silicone.
  • Cast silicon.
  • Use the vacuum chamber.


  • This is the rotor for my milk flow sensor.


3D design on Fusion

  • Model inspired on this:


  • And this 6 mm diameter x 50 mm long shaft I’ve found and cut in Agrilab.



  • Sketch.
  • Diameter: 50mm
  • Shaft hole diameter: 5mm
  • Blades width: 6mm
  • Round corners 3mm


  • Cross section of the model extruded.
  • I’ve made this by using revolve on a profile and then applying difference.


This is a video of my entire CAD design process:


  • I’ve placed thi origin in the corner shown in the capture.


  • Rough cut.


  • Smoother cut.



Tool configuration:

  • Diameter: 3 mm
  • Shaft diameter: 3 mm
  • Overall length: 50 mm
  • length bellow holder: 25 mm
  • Shoulder length: 16 mm
  • Flute length: 15 mm
  • Tool number 3. (Just for current agrilab CNC machine tool configuration)

fusion_07.jpg fusion_08.jpg


  • Delete line with G43 code. This adds 16 mm compensation to Z axis.
G43 Z16. H3

Preparation prior to Milling process

The milling process requires the Wax brick to be fixed to the CNC bed. But the WAX brick dimensions do not coincide with the pattern in the bed that holds pieces using the vacuum pump.

So, we had the task to design a holding tool, for that I’ve measure the CNC bed.

  • I had the idea to make also taps for the vacuum holes of the CNC machine.
  • The tap dimensions are 50 mm square and 4 mm radius in every corner.


  • The lenght from rail to rail to use the holder pieces is 380 mm.


  • The depth of the cuts in the CNC bed is 10 mm.


  • The squares of the CNC bed pattern are 40 mm with a cut of 10 mm.


  • These are the holder that are used to hold pieces in using the rails in the CNC bed.


Holder design

I’ve chose to design a holder device for the machinable wax bricks, using 6 mm MDF plates and the laser cutter.


Inkscape captures missing


FTP captures missing

Mold base

  • This is the resulting piece, it holds steady in the CNC bed without the need of the vacuum pump.


  • You can see here the small flexes that help the bricks to fit in place. All the bricks have slight 1-2mm differences in their dimensions.


  • This is the brick placed in the holder.

mold2_07.jpg mold2_05.jpg

  • Since I didn’t have time to finish my entire design for milling. In the milling day I’ve used Theo’s holder for the 75 x 176 x 37 mm Wax brick. So I’m saving my brick and holder for later.




  • 2D pocket cut (Rough cut)
  • 3D parallel cut (Smoothing cut)

Mistakes during the process:

  • At first my process didn’t include a 3D machining process.
  • Generated the GCODE with the tool Number 1.


  • Change the model adding angled spins to my design.
  • Generate the GCODE with tool number 3 which in our CNC was the 3 mm Flat-end milling bit.


  • This is the rough process, the machine just cuts the silhouette as a 2D but it lefts scales and steps the later will be smoothed.

cnc_09.jpg cnc_01.jpg


  • This is the smoothing process, the machine moves in along side the Z and X axis during this process.


  • The resulting piece has a smoother surface.


Group assignment

For the group assignment I had the idea to make a sample table and cast some coins. For that I’ve used:

  • 2 pieces of 6 mm MDF plate.
  • Plastic sheet.
  • Laser cutter.

  • This is the sample table.


  • Putting altogether the parts.


  • I’ve also bring paper towels and globes for my team mates.


  • I’ve cast a coin using Food-safe silicone, as Elina, but with the difference that I’ve used the vacuum pump to extract air bubbles from the mix.


  • This is the result, the very small features were replicated perfectly in the mold.


Food-safe Silicone

  • This is the Food-safe silicone RTV polyaddition 3428 made by Esprit Composite.

  • It’s a two components silicone A 100 and B 10 by mass.

Esprit composite Food-safe Silicone
Datasheet Link
Safety sheet Link
Pot life 90 mins @ 23 C
Demold time 16 hours @ 23 C
Color White
Mixing ratio A:100g, B:10g MASS
Indentation hardness 28 Shore A
Tensile strength 7.5 MPa
Elongation 600%
Tear strength 20 KN/m
Viscosity 25000 mPa
Shrinkage 0.1%
Safety precautions Not hazardous
Solubility None

I’ve found this information on English Stephane Muller’s Fab-Academy page

And in French in Esprit Composites web site.


  • I’ve used a disposable cup and a chopstick to mix the components.


  • This the vacuum chamber and pump.


I’ve found two food-safe demolding agents:

  • Petroleum jelly and,
  • Sunflower seed oil.


Silicone preparation

There are two important steps before preparing the mix.

  • To know the volume need so by that avoid any excessive amount of silicone.
  • To use a demolding agent and so help the silicone come out easily.


I’ve used water and the disposable plastic cup to estimate the volume needed of the mix. It’s important to remark that I still had to measure the mass of the volume to know the amount of the component B.

  • I’ve prepared also Theo’s Silicone for his food-safe silicone assignment. And this is the volume needed for him.

  • The top mark was the initial water level.

  • The following mark is the current water level.


  • The bottom mark is the water level after filling my mold.


  • After over-estimating the residuals I’ve marked the level of MIX in the cup.



  • I’ve weighted the empty cup and then used the Zero function in the balance. So the next grams measured will be only the weight of the components.



The component A and B weighted 97 g together 4 mm above to the level of the volume mark.

  • 88 g Component A
  • 9 g Component B



I’ve used the chopstick to mix the components following the instructions of Professor Neil Gershenfeld to avoid filling the mix with air bubbles.


  • This a sample of the movements I’ve used to mix the components.

Vacuum chamber

I’ve used the vacuum chamber to extract the air bubbles of the silicone Mix.


  • Pressure: -1.5 MPa.
  • Time: 10 mins.


  • After releasing the vacuum.



Demolding agent

  • I’ve chose Sunflower seed oil as demolding agent, following the instructions of the RTV3428 Silicone, it states that any agent should be food-safe as the silicone.

  • Coating the borders of the mold with the Sunflower oil using a coffee filter.


  • Coating the inside of the mold pouring oil into the cavity.


  • After realizing the excessive coat of oil I’ve placed the mold backwards to help ease the excessive oil out of the cavity.



  • Theo’s mold.


  • My initial mold.


  • A big mistake.
  • I’ve chose to put some PLA wires I found out nearby with the idea to help me in the demolding process. Same for the Chopsticks.


  • I didn’t have the chance to record myself pouring the mix into the mold, but here I show out the technique I’ve used.

  • The idea is to make a small stream pouring into the mold to break any bubble in the mix.

  • I’ve show to Theo the same technique.



  • The oil worked very good.
  • Next time I’ll not use any weird attachment, like I did with the chopsticks.

  • I’ve pour the silicone on Monday 11 am and cast off on Tuesday 11 am.


  • The mold after removing the silicone part.


  • The casted part.


  • I’ve used a knife to cut the excess of material.


  • After clearing the excess I’ve inserted the 6 mm shaft as planned. The hole in the silicone is 5 mm so it is tight and stable.



Casting bismuth

I’ve helped Theo’s with his test for soft metal casting.

  • This is the setup of the pizza oven and a improvised table I’ve found in AgriLab.


  • The oven with the special container for molten metals.


  • We used a K-type thermo-pair attached to the container using Kapton tape.


  • Theo coated the mold with a thick layer of baby powder.


  • This is Bismuth and his characteristic light reflective crystal structure.


  • This device is a XTC-Profiler a thermo-pair reader that we use in the Electro-Lab.


  • This was Theo’s first attempt to cast Bismuth. The Bismuth got cold quickly, since the mold was 4 meters away from the oven. So the Bismuth didn’t flow on the mold.

bismuth_01.jpg bismuth_06.jpg

  • This is a video of a second attempt performed by our instructor Luc
  • This is Theo’s third and successful attempt to cast the Bismuth into the mold.




  • Final cast piece.


Data casting Bismuth
Oven max Temp 290 C
Time of practice 46 mins
Time to reach 280 C 26 mins
Attempts until success 3


  • Silicone RTV3428 it’s really easy to use.
  • The viscosity was perfect to replicate all the features of the design.
  • I’d like to use a cap for the top of my design, or a two parts mold.
  • I’d keep using this material for my next sensor parts.


Design changes

  • Design a mold that covers or embeds the shaft to avoid direct contact with the milk and the metal.


Future projects

I’ve the idea to cast the entire body of my flow sensor, but since it’s the first time for me using Food-safe materials I’ll wait until finish my design properly.


NC Rotor Rough Pocket process file.

NC Rotor Smoother Parallel process file.

Inkscape laser mold base file for 6x6 inches Wax brick

Fusion 360 Rotor design and manufacturing file

Last update: April 20, 2021