A frigging knobhead¶
This week I thougth I should make a handle for a pinball machine launcher.
Modeling¶
I created the following variables:
| Variable | Description |
|---|---|
| shaftDiameter | Diameter of the shaft that is attached to the handle. Leaves space for it to be inserted into the mold. |
| shaftSlotLength | How deep the shaft goes into the mold. |
| totalLength | The length of the handle. |
| knobDiameter | The diameter of the big round part of the handle. |
| knobCenterDistance | |
| waistDiameter | The diameter of the “waist”, the thin part of the handle. |
| waistDistance | The distance of the “waist” from the knob. |
| bottomDiameter | The diameter of the end of the handle, where the shaft is attached. |
Drew a rectandle with shaftDiameter /2 x shaftSlotDistance.
Created a center line for the handle by drawing upwards from origin with length of totalDistance.
Drew a line right from origin with length of bottomDiameter / 2.
Drew a center point arc on the center line with radius of knobDiameter / 2. Set dimesion between the center of the arc and origin to knobCenterDistance.
Drew a construction line right from somewhere along the center line with length of waistDiameter / 2. Set dimension from origin to waistDistance.
Drew a fit point spline connecting to rightmost points of bottomline, waist line and the know arc. It automatically created a midpoint constraing between the knob arc and the new line. While Autodesk Fusion instructed me to end editing by connecting the spline to the starting point, but doing so would have created a circular shape, and I did not want that. In odred to keep it aa a open ended line, I needed to click green tick mark next to the last point to end the line there.
The sketch was finished.
After that, I revolved the finished thing by 180 degrees.
Drew a new sketch at the bottom (the flat edge) of the revolved object to create a mold edges. Created four new variables: moldWidth, moldDepth, moldSlotWidth and moldSlotDepth.
Drew a square around the object with side length moldWidth. Drew another smaller square inside it, with the sides dimensions set to moldSlotWidth distance from each side of the bigger square. Centered the rectangles around the object by drawing a construction line from two opposite corners of the bigger square, and another construction line from the top of the handle to the bottom of the handle. The set midpoint constraint to those construction lines.
Hmm…
At this point I noticed that the shape is not millable as is. The slot for the shaft is a cavity under a hood, and thus unmillable. I needed to rething the design.
After some consideration, I decided to change the plan. Instead of molding the object, and then attaching it to the shaft, I would cast the object around the shaft. This would mean that the shaft would need to be inserted into the mold during casting. I needed to put a large enough hole to pour the cast that would also fit the shaft.
To represent that change in the model. I needed to edit the original sketch by adding a new rectangle at the base, where the shaft should be. I made it with dimensions of shatfDiameter / 2 x 15mm. 15 mm was chosen as a random value that works for now. I also added the rectangle to the previously completed revolve command.
Extruded the walls.
Copied the object with cmd + m, selected copy object and dragged the object a bit to the right.
Extruded the slot surface with cut from the clone by -moldSlotDepth * 2 to create a mirror of the slot, so that they fit together nicely.
To insert the shaft into the mold, the shaft channel needs to go through the outer wall of the mold. I extended the shaft through the lip by drawing a sketch on the surface, and drew lines from the end points of the half circles to the bottom, and extruding that to the outer wall.
I also added small channels for air to escape, and probably to insert castable material to the negative mold.
To make sure that the corners fit correctly, I rounded the corners in the design. This makes it so that the milling machine intentionally creates these shaped, and not just accidentally created by the roundness of the milling bit.
Milling¶
Milling wasn’t that different to other processes we did. Just export .stl file from Autodesk Fusion and pop it into another computer. There I opened it with Modela Player and created tool tool paths, which were run with VPanel for Roland
I created three different tool paths: surfacing, roughing and finishing paths. As my object was fairly large, 10cmx6cmx3cm, the milling times for it were long. Surfacing took almost 30 minutes, roughing two hours and finishing another hour.
The largeness also created another issue, where I orient the 3D model sideways in order for it to fit into the milling bed of our Roland milling machine.
We did not have a large enough block of wax to mill the mold from, so we needed to saw another block in half and melt it into another block to create a big enough piece of wax.
All tool paths were done with 3.12mm (1/8 inch) flat end milling bit. Roughing was done with 0.6mm milling depth.
Then the molds were done.
After that I collected all the milled was into small container, so that they could be melted back into wax blocks to be used again in other molds.
Casting negative¶
I cast them with Smooth-Sil™ 936. I calculated that I needed 185ml of material per mold (so around 370ml in total). I used that number exactly as is, and mixed in 333g of A and 38g of B. As the label recommends a mixture of 100A:10B, I added a bit too much B, but that ended up not mattering. 371 grams of material was less in volume that I needed, and I could not completely fill the molds.
Before pouring the cast, I put it into a vacuum chamber to remove internal air bubbles, but I had mixed the cast in a too small a vessel, and the once the liquid expanded in the vacuum, it almost poured everywhere. I had to stop the process early, and thus the vacuum chamber might have actually increased the amount of air in the mixture.
The outcome was good, even though they only cured for 19 hours, and the cure time should be 24 hours according to the label.
The newly created negative mold was a bit thin. There was probably less than 1mm of material in the thinnest parts.
Casting final¶
There are not a lot of pictures for this process. I cast this with Smoothcast 300, which is toxic stuff that should not be touced with base skin or even by our clothes, so there was a snowballs chance in hell that I would have my phone anywhere near that stuff.
I measured the needed volume to be about 40ml per side, so 80ml in total. As the Smoothcast needs to be mixed with 100 parts of A to 90 parts of B per weight, I calculated that as 50g of A and 45g of B. I rounded it up, because I was not sure what was the weight-by-volume of the material. 50:45 was also the closest possible values that add up to over 80 that use only integer numbers. The other options would have been 40:36 amd 60:54.
I added the halves of the mold together and constrained them with a rubber band. Then I inserted the metal rod inside the mold, and poured the mixed material to the mold through one of the air holes. The air hole was so small that I needed to use a syringe.
After casting, the final result had an annoying air bubble in it, that caused an ugly piece to be missing. I would have recast that part, as the Smoothcast should cure to stick to itself, but we ran out of syringes.
So I decided to fix it. I plonked the handle back into the mold, mixed a bit more Smoothcast 300, and injected it through the hole.
This time the results were not that much better. An air bubble still appeared, and this time I injected a bit too much material and the mold bulged a bit. It resulted in a small sheet of material everywhere along the mold’s joint line, and increased the volume of the upper edge of the handle.
More!¶
I decided to do a couple more, this time with colors. To color them in a multitude of colors, I first mixed the Smoothcast as normal, then poured the mixture into three cups, and added a different color for each of them. For the first one, I added black to 1/2 of the mixture, purple to 1/4 and red to 1/4. For the other one, I added purple to 1/2, and red 1/4 and yellow 1/4.
I added the color by dipping an acrylic stick into the color, and then mixing the smoothcast with the dipped stick, using what ever color stuck onto the stick as the color.
Then I intermittedly injected the material to the mold from each color. For the first one, I started with a huge amount of black, and then injected other colors intermittedly. This caused the bottom to be completely black (or gray). For the other one, to make sure that the one side did not look as boring, I changed the base color to purple and started by pouring red. Also I tried to mix the purple handle a bit more after injecting all the material by tilting the mold a couple of times upside down and back. It might have had a small effect, but I did it after the pot time of Smoothcast was over, so the mixture was already quite unrunny, so I am not sure if it mattered at all.
It was always fun to check what we found under the mold.
I needed to file the injection tube out from the cast handles after I got them from the mold.
I never really fixed the air bubble problem. In the gray one, I tried to intentionally inject the last syringe through the other hole (which was otherwise used as an exhaust hole for the injected air), which solved most of the problems. But the required more material during casting, and I forgot to mix too much material during the purple one.
Interestingly, both colored ones have an “eye” of another color on the backside of where the material was injected. The color probably got there due to the pressure from the injection, only spread a little around it.
Things to take into account in the future
Two sided silicon molds are hard to fit together perfectly. As they are stretchy and sticky, once they are placed on top of each other, they easily deform a little, creating a lip where the molds meet.