Week 9 - Molding and Casting
Assignment
Individual assignment
design a 3D mold around the stock and tooling that you’ll be using, machine it, and use it to cast parts
Group assignment
review the safety data sheets for each of your molding and casting materials, then make and compare test casts with each of them
Process
Designing the model
For this week assignment, which required from the student’s behalf a good time management approach, I wanted to design something that could remind of psychologist Nathaniel Kleinmnan’s studies on Basic Rest-Activity Cycles, so that … you know … I could make a better use of my time (hopefully).
The final product would be a table ornament with an hendecagonal shape useful to keep track of your top performance time span and suggest some rest when your brain would get tired. To do so, I started designing on OnShape such idea and after having sketched out most of its components I added two cylinders that would make the pouring hole and the breather in the final mold.
At the moment of this first design, I hadn’t had a better idea for splitting the model in two than by substracting a thin plane at its half.
Then I located the two helves over a working surface were I also added four couples of registration keys.
After downloading the resulting STL file, I was ready to prepare the milling file for my mold.
Preparing the milling file
Being new to the world of milling and engraving softwares, before getting to work the softwares available at our laboratory, I did a good skimming of the official tutorials by Vectric. In our lab we had available Aspire and Cut3d, which essentially do the same job but with slightly different workflows, and I ended using Cut3D because Aspire gave some exportation issues with the file requeste by our milling machine.
To prepare my milling files, I had at disposal different milling tips which I rapidly reviewed to see what where the options for my work. To do the roughing of the mold the best choice would be a 1⁄8” flat end mill, while for the finishing I chose a 1⁄16” flat end mill.
A very quick and non-customized processing of my model in Cut3D made emerge a serious issue that I didn’t consider at first.
As it is possible to see in the picture above, most of the details of my model were totally missing the preview of the final result. This happened because when I first designed I wasn’t taking into consideration that sizes and accuracies of the milling tips. In fact, it’s impossible for a 1⁄16” flat end mill to work on spaces smaller that 0.0625 inches and since my model was full of text and small details in the best case scenario the details wouldn’t be carved out; in the worst case scenario the tools could suffer some damage.
Before proceeding, I redesigned from scratch the whole model, this time by directly extruding the model shapes from the working surface and removing excessively small details.
I was then ready to work on all the steps that Cut3D provides to make the final milling files.
The first is about model scaling and orientation. If you design the model form the top as I did, it requires no changes and usually even scaling is not necessary.
The second step asks for a bit more of attention. In the Material size area you have to set the origin position (the lower left corner in my case) and the dimension of the material you’ll carve out (a good rule of thumb is to set dimensions slightly smaller than your material, in order to have a little safety offset). Depth of model below surface is where you set a thickness that in my case would be very useful, since it generates an outer wall in the future cast. To do so, I added a 3mm depth that I also summed to the Thickness value in the Material Size section. Finally, setting the Cut Plane Position in Model to the Bottom value make it so that the machine will engrave the model in all its depths. A mistake in this value can result in an erroneously partial mold or with non-matching registration keys.
The next steps regards the making of the roughing path, meaning the route the 1⁄8” end mill will make to trace an appoximate shape of the final model out of the material. The most important thing about this step is about the parameters the machine will have to follow regarding the drill bit.
Setting parameters with little care can result in several problem for the machine and I’ve learned at the expenses of my peer milling before me that starting up again a work that takes up to 2 or 3 hours can be very, very frustrating. Pressing the Select button in the Roughing Toolpath windows gives access to a wide range of default tips. After choosing the 1⁄8” I carefully considered every parameter before proceeding. In particular, Pass Depth must always be at most the half of the bit diameter, and Feed Rate and Plunge Rateneeds a little more of speculation. There are many ways to calculate precisely these values, but not having at hand the datasheet for the material I was going to mill I opted to a value of 25 for the former and 2 for the latter. I extrapolated them by other tips setting and specifically for the Feed rate, the most insidious parameter, I figured out to have an averagely low value because and erroneously setting it to a value of 40 took my peer Dario to repeat his milling several times.
After pressing the Calculate button I was able to see a first preview of the roughing route.
An less accurate job is required for setting up the parameters for the finishing path, the following step, because the tip usually doesn’t suffer the stress reserved in the previous task. In this case, the only important thing to consider is that if you use a flat end mill (like I did with my 1⁄16” one), to have a more polished result it’s better to check the option Create extra pass at 90 degrees to first. This will make the milling machine to repeat the finishing route and erase certain corners that are inevitably left by a flat bit when routing only in one direction.
After pressing Calculate, it already possible to see an approximate preview of the final result.
A fully-computed preview of the result showed me that the mold would result fine and zooming to certain risky details ensure that everything was going to be processed as required.
Milling
The material I was going to use to cast my mold was wax block that resulted slightly longer that the working surface of our milling machine. To fix this issue I just trimmed a bit a side of the block until it got steady on the machine…
…then I sticked some biadhesive tape below it so that it would move during the milling operation…
…and I was ready to mill it with the roughing toolpath.
This operation can be very time-consuming (it took almost 2 hours for my model, which I considered a simple one) and very stressful for the end mill, so it’s really important to pause the process every so often to remove the residual dust from the machine. If not done, it’s very easy for the machine to get stuck and for you to start the whole process over again.
When the milling was done, I changed the milling tip to 1⁄16” and prompted the finishig path.
When this step ended with no issue, I had my block finished and ready to be casted upon. A quick check showed me a little issue that I could have foreseen easily: having milled a model a little bit too deep, resulted in the repeated collision of the mill bit with the borders of the central part of the model.
To avoid this perhaps I should have considered better the depth of the drill bit, or design more sloped shapes, but in the end it is not a real inconvience and the end result wouldn’t have been so bad if it wasn’t for the thing that I did just after.
Have you seen the picture above? It’s an hot-air gun sometimes it’s used to smooth the surface of wax molds. I thought that it was an easy and useful process, especially for molds milled with flat bits, but I was wrong, very wrong. Using it not only didn’t smooth very well my model, but it ruined it in several parts (one of the most damaged perhaps was the result of prolonged exposure during the shooting the that picture).
So, if I ever can give an advice regarding this assignment that is: do not use the hot-air gun!!!
Your model is just fine even with a grainy surface, and surely better than a Dali-esque contoured mold like mine.
In the end however, I kept it to proceed with the assignment.
Casting
For the first cast of my negative mold I used a polyurethane rubber compound by Smooth-On called PMC-121 with an hardness of 30 shores.
Upon reading its behavior and specifics in its official datasheet, I started mixing its two separate parts with a 1:1 proportion by volume and poured the resulting fluid over the wax block.
The datasheet claims that this compound has a cure time of 16 hours, so I left it rest overnight and kept a sample from the mixing process to evalute the reaction of the cast before pulling it out of the wax block.
When the time elapsed the cast dried just fine and after cleaning it with water and soap I was ready to use it as a mold.
Molding
Somehow, when redesigning my model, I forgot to add the two channels for pouring and breathing, but there are very few things in life that cannot be solved by an precision cutter and so I made them by hand after separating the two halves.
With the help of a some tape, two wooden bars, a clamp and a makeshift funnel, I prepared the proper setting for casting my model.
For this task I used a polyurethanic resin by Prochima called Sintafoam. As described in this compound datasheet, it is a very hard and resistant product that has many applications in the world of modeling and rapid prototyping. In fact, its pot life (the time you’re given to mix A and B components and cast them in the mold) is very short - less than 5 minutes - and its cure time is less than 30 minutes. Having that in mind, with highly calulated, fast and (not so) precise movements, I mixed the components with 1:1 proportion by weight and casted the mix inside the mold.
After almost 30 minutes I could see that the resulting model was flawed on many, many aspects…
First of all I miscalculated the amount of resin to use, resulting in a half-molded cast. Then I think I might have spent a little too much time during the pot life span, resulting in a badly solidified model with the compound clogging halfway to the bottom.
But all these mistaked could have meant very little - it could have been just a bad try, if it wasn’t for the fact that the resin badly abraded many details and sides of the mold definitiely ruining it. That’s because removing the resin from the mold wasn’t itself a very easy task (it got stuck in many parts, especially in the crevices) and not having used any demolding substance like a teflon spray resulted in a permanent damage of the mold.
Not willing to quit with this mold, I tried a different approach to see if it could still be useful, and so I opted to cast some gypsum.
In our lab, we had at our disposal an powder called Ultimate Drystone by USG which, according to its datasheet seemed very easy to prepare and use, allowing much more room for errors in the mixing process and longer pot life.
To help myself in the process, I used one of the funnel I made for the 3D printing week’s assignment and it was handy to pour the gypsum without besmirching the working desk and myself.
As before, I saved the remainder from the mixing process to check when the gypsum had dried. Since I added a bit more of water that prescribed (I wanted a compound that could easily enter in all the crevices without clogging inside the mold) I waited almost 1 hour more than the 25 minutes suggested by the datasheet, but by then I was sure that my cast was done.
Even though the cast came out fine, with not mixing flaws or bubbles, and I was able to remove it from the mold very easily, the result was pretty sloppy, with many letters, details and borders ruined or absent.
This was indeed caused by the ruined mold, so I realized that I had to do another one to get a proper result.
Once again
To prepare a new mold, I opted for a differet siliconic rubber called AL20 by producer Antichità Belsito. According to its datasheet, compared to the PMC this rubber has less shore hardness (approx. 20) and and a longer high cure time, but it granted to be much more nonstick and resistant to laceration.
In fact, once prepared the compound with an A-B weight ratio of 20:1 and having it rest for 24 hours, the cast result was very nice looking and felt elastic and resistant at touch.
I tried to make a new gypsum cast, and to to try something either I added a bit of SO-Strong green tint by Smooth-On (it is intended for polyurethanes and epoxies, but it works just fine even with other substances).
WARNING: the dyes are really powerful and just a tip can fully color a lot of material (and the mold, and the surface you’re working on, and your clothes, and your hands… for days).
Result
In the end the green gypsum cast resulted in a nicely finished model, with just few casting lines easy to remove and few tiny bubbles - surely I mistake from my own mixing - that overall don’t impact the appearance and consistency of the model. Accidentally I broke a piece of the upper border while handling it.
I also tried to craft another resin model with Sintafoam, and this time too I could experience a smooth demolding process without ruining the mold. I also added some purple dye, maybe too much.
In this case, I think I might have mixed badly the compound initially, because it appears to be kind of sticky and some edges are oddly soft, but after a wash the result was fine.
Following are the links to the various file generated to complete this assigment:
Gist & Further development
Overall, the designing and milling process went smooth and fine, and most of the following failures occurred are due to my profound lack of crafting skills. But, if there’s a lesson to be learned, and that Neil warned us about in the first place, is that for casting and molding you should design you ideas around the tools you’re going to use, not the other way, and having had this in my mind since the beginning could have helped me and made me save some time for better results. I’m surely going to apply these techniques for other applications, both for my final project and other purposes, but first I need to study much more the tools and the materials involved.
Tools and software used
- OnShape
- Vectric Cut3D
- Roland SRM-20
- Smooth-On PMC-121 30 Dry
- Prochima Sintafoam
- USG Ultimate Drystone
- Antichità Belsito AL20
- Smooth-On So-Strong
- Various cutting tools, working materials and disposable stuff