Anders Häggman

MAS.863 Portfolio

WEEK 6 // Moulding & Casting

My goal was to create an interestingly shaped ice cube mould, and I did not want to spend too much time working in CAD, as that is a skill I feel I already know.

I tried to find a .stl file of an iceberg, but could not find any models that were free and looked anywhere decent. There were, however, several quite nice models of various mountains, so I decided to pick on of my favourite mountains, Matterhorn, and cast two mountains mirrored on top of each other, to create an iceberg.

I looked at several different files before deciding on a specific file, so I am not sure which I ended up using, but I think it was this one: Matterhorn.stl

  • Level of previous experience:              
  • Time taken:               
  • Tools used:
    Roland Modela MDX-20  //  SolidWorks  //  Autodesk Meshmixer  //  Vacuum Chamber  //  Moulding Materials & Equipment
  • Got help from:
    Charles Guan
  • Main learnings:
      Crash course in Autodesk Meshmixer
      Being careful to think about which is the negative/positive
[Image_Missing]
The beginning.
[Image_Missing]
We were told to hot glue the machinable wax onto the modela, which was great advice. It was easy to attach, felt very secure, and was surprisingly easy to remove.
[Image_Missing]
I don't know if there's a more elegant way to do this, but I was worried that the spindle block might impact the material I was milling, so I wanted to measure the length of the protrusion of the end mill. A caliper was too bulky to be able to measure the tool in place. I knew my part was at a maximum 35mm deep, so I took an allen key, and made a mark with some electrical tape, and used that to measure the tool.
[Image_Missing]
Deep enough, with a little bit to spare.
[Image_Missing]
After the disappointing Friday (when I tried to mill the mould the first time), this time the computer and mill actually communicated, and I was on my way to getting a mould. The tool path had some (what I thought were strange artefacts) in the first layer, so I on purpose zeroed the tool slightly above the material. (This would turn out to be a problem later on...)
[Image_Missing]
The rough cut looks kind of neat. The cutting time was 50 min per mould (I have a two part mould). To save time, I was at one point contemplating just using the rough cut version as my mould, but looking at it when it was done, it was clear that I would need to run the finishing pass as well.
[Image_Missing]
After machining the second mold, I realized I had made a mistake... I had forgot to mirror one of the mountains, so when they are put together they match. Since machining took almost an hour per mould, I did not have the opportunity to machine another one right away, but I went in to Meshmixer and created the mirrored file, so that it would be ready to go in case it was needed. In the end I did not make another mirrored mould, since the asymmetry of the mountains actually added quite a nice aesthetic to the iceberg, and made it look a little less planned, and perhaps more natural/genuine. Between my two rough cuts and the final finishing pass, someone moved the tool up, so I lost my z-zero, and as I hadn't zeroed off the top of the part, I was now in a bit of trouble. I tried eye-balling it back to the same level, but as the z-control on the Modela is horrendously bad, I overshot once (hence the white circle). I started to run a finishing pass, but noticed that the mill was not removing any material, so I aborted the run, re-eyeballed the mill, and tried again. It took a few tries to get it where I wanted it to be, but in the end I think it turned out alright.
[Image_Missing]
The wax shavings had a nice aesthetic and looked a bit like snow on the side of a mountain. I used Matterhorn (a mountain on the border of Switzerland and Italy) as the mountain to make my iceberg out of. During the milling process a lot of wax shavings were created, and since I was forced to use the compiled fab module there was no way to 'pause' the job to remove the shavings, so while the machine was running I tried to periodically remove the shavings with the vacuum, which worked OK, but one had to be careful not to get the vacuum nozzle stuck between the machine and the part which could result in the machine aborting the job and resetting itself, and would cause problems with the x/y zeroing as well (I was told this happens every now and then).
[Image_Missing]
As I ended up going with the same 1/8" flat end mill for the finishing cut, to avoid a tool change (and the large increase in machining time), I decided to increase the overlap to 80% instead. That gave the finishing pass a machine time of a bit over 20min per mould. In total, each of the two moulds took 1h10min+ to machine.
[Image_Missing]
It was easy to correct small imperfections with an x-acto knife.
[Image_Missing]
The final moulds. The cylinder in one of the moulds is to allow water to be poured into the two part mould.
[Image_Missing]
In order to make the iceberg-ice cubes food safe, I used the Smooth-On Smooth-sil 940, which was much more viscous than I anticipated. The catalyst had a consistency of roughly water, but the other part was so thick that you could turn the cup upside down for a good bit of time without it falling out. Mixing the two compounds was therefore incredibly challenging, and a lot of bubbles were introduced into the mixture. The ratio of the two compounds to be mixed was 10:1 by weight, but as I could not find a scale I decided to do it by volume. I am planning on making another set of moulds that I would actually use for ice cube production, in which case I will probably try to weigh the compounds to be sure the proportions are correct.
[Image_Missing]
I placed the mixture in a vacuum chamber to remove (hopefully) most of the bubbles. I had placed some sacrificial papers at the bottom of the vacuum chamber in case the mixture overflowed (which turned out to be a good idea). My mixture expanded roughly 2.5x to its original volume, and it only took a few seconds of distraction to have the mixture overflow.
[Image_Missing]
Since the pot life of the Smooth-sil 940 was indicated to be 30min, I left the mixture in the vacuum chamber for 15-20min, cycling between ambient pressure and 30" Hg to try and expand and collapse the bubbles. After the first vacuum treatment the mixture looked quite good and bubble free, but as it was so thick, when pouring it into the moulds, air was again trapped into the mixture. I tried pouring it from a higher distance in a thin stream, but as it hit the mould the mixture folded onto itself, and trapped air in the folds. I think for this thick a mixture, a better alternative could have been to pour it from a very low height slowly. After the mixture was in the moulds, I put them back in the vacuum chamber, and cycled between ambient and 30" Hg for the remainder of the pot life.
[Image_Missing]
Due to the generous draft angles I had created, the silicone mould was relatively easy to extract from the machined wax mould, but I can imagine that with an even slightly more complicate mould (or less draft angle), it would have been nice to have made some sort of handle from the silicone that could be used to extract the silicone mould. I let the silicone cure in the wax mould for about 26h, after which I placed it in a 200°F (93°C) toaster oven (not used for food) for 2h, and then for another 2h at 225°F (107°C). The instructions were somewhat unclear on the smooth-on website. A .pdf on their site instructed you to cure it at 80°C for 2h, and then 1h at 100°C, while the text on the website instructed you to cure the silicone at 100°C for 4h. When I came to pick up my parts, I noticed that the oven had turned off / had been turned off, so I do not know how long the moulds had been at 225°F, but as they were still hot to the touch (and based on the conflicting instructions on the Smooth-on website, it doesn't seem like it is very sensitive to having precisely the correct temperature) I felt they were probably fine. The surface texture also felt like the silicone forms that can be bought in stores.
[Image_Missing]
Both halves of the mould combined, pouring water into the mould. The water entered quite well, but it was hard to know if there were bubbles trapped in the mould or not. I tried squishing it to see if bubbles would come out, and inserted a straw to see if I could suck out any bubbles, but none seemed to be present (and once the water had frozen I could see that infact the mould had filled completely with water). The connection between the mould halves was good, and the form had sufficient weight that I don't think the rubber bands were necessary, but I left them on anyway.
[Image_Missing]
The final product. The moulds were easy to remove, and the iceberg looks nice (the only problem is that it is too small). I had originally wanted to make an ice cube the would be so big that only one cube would fill the glass. Unfortunately the machinable wax blocks I could find were not large enough to do this.
[Image_Missing]
I realize now that I should have taken better picture before putting it in a drink. I fished the cube out and took these pictures, but the rougher edges that I thought made it look nice had already melted away.