Brake Duct Molds (DRAFT)

There have been a lot of great threads on making molds and body parts by hand. The other Scott’s ability to quickly design and fabricate new fiberglass parts is truly impressive. I don’t have those types of skills so I decided to go a high-tech route. There are a bewildering number of software packages and ways to accomplish this and I’ve worked with Kevin at ??? to design some significant modifications, but I wanted to complete the process for something simple before moving on to more complex parts. The smallest and easiest part was the brake ducts. Here’s the process.

Phase 1: SketchING

Initial sketches are done by hand. In this case, we were copying a design that Allan has done on multiple cars so we skipped this phase.

Phase 2: 2D Rendering

Once a sketch is chosen, the concept is rendered in 2D using Photoshop.

Phase 2: Scanning

Mason’s scans were great for doing the initial renderings, but they didn’t have any reference points to align them with other scans. For example, to design the radiator outlet I needed to align the scan of the nose with the scan of the radiator compartment, so Andrew applied a bunch of reference point stickers, covered everything with CANTESCO D101-A Dye Penetrant Standard Developer Spray and scanned the inside and outside of the nose with a Peel 2 3D scanner at a 0.50 mm resolution. Once the scan was completed he used the Peel software to remove unwanted data.

Phase 3: Rendering

The renderings were done in Modo which a sub-D modeling tool which is used in movie and gaming industry for concept modeling. As such, it’s great to iterate concepts for a car. He sucked Mason’s point cloud scanning as a bunch of polygons. Reverse engineer and surface it. For example, while you can specify the overall dimensions of an object you don’t specify sub dimensions… artist’s view of what looks proportional.

I decided to go with the A2 variant (shown above) of the brake duct discussed in a prior post. Once that decision was made Kevin imported ???? poly sub D (sub divisional)…??? Blah

Phase 4: Designing

He Bring the modo model (guide) and the original scan data into Alias then reversed engineered the scan data to create smooth surfaces and then use surfacing tools to quickly model what you want.

NURBS surfaces

Alias, a Computer-Aided Industrial Design (CAID) package predominantly used in automotive and industrial design for generating precisely sculpted curves and surfaces. Unlike CAD packages such as SolidWorks, it focuses on styling and aesthetics of external surfaces rather than the underlying mechanical details. Mason’s scan data was imported ???? blah blah mesh…blAH

Focused on dimensions, draft angles, etc.

The design has been refined in several subtle ways. Most notably in the rendering the top of duct was parallel to the ground whereas the final design, as can be seen in the image below, the top outside corner slopes up to match the curvature in the nose. This would be difficult to achieve via hand sculpting which makes this part a good candidate for a mold.

When designing a mold you must ensure that you can actually get the part out of the mold. This is accomplished by ensuring that there’s an appropriate amount of positive draft in the direction that the finished part will be pulled. Draft is the amount of taper, usually measured in degrees, perpendicular to the parting line. There are lots of factors (e.g., material shrinkage, depth of part, wall thickness, etc.) that go into determining a mold’s draft, but the larger the draft the easier it is to remove the part. A general rule-of-thumb is that female molds should have at least five degrees of draft and male molds should have at least 3 degrees.

Unlike the Superlite Aero’s buck which was CNC’d from foam, the SL-C’s buck was hand-carved from clay. While the shape is good, it’s not perfect. The above image is a draft analysis performed in Alias showing the amount of draft from front to back (i.e., the direction that the part will be pulled from the mold). Since we’re designing a male mold, the analysis was based on 3 degrees:

• Blue = good (at least 3° of draft)

• Red = marginal (less than 3° of draft)

• Green = bad (undercut)

These imperfections are trivial to fix with a thin layer of body filler when prepping the car for paint, but it’s the type of imperfection that can make it difficult or impossible to pull a part from its mold. For this reason, we decided to go with a male rather than the more common female mold.

Kevin exported the part from Alias as a ??? surface.

Phase 5: MOLD Negative

Andrew imported the surface into Autodesk Fusion 360 and converted the surface into a solid and added flanges. He then imported into Autodesk Inventor, used their mold injection tool (using it the “wrong” way). Many CAD packages have tool that facilitate making molds, but he hasn’t found anything that’s tailored to composites {SPECIFICALLY WHAT???]. This produced???

Andrew then imported into Fusion 360 to validate the dimensions and then he exported it as a STL for 3D printing.

After the parts were 3D printed, the supports were cut off and any remaining bumps were removed with a combination of 80 and 120-grit sandpaper. The molds were then skimmed with body filler and sanded with 120 grit at which point the process was repeated a second time. While the top face of the mold is the duct opening and therefore no used in the final part, it was finished in case I decide to vacuum bag the parts.

The molds were then sprayed with Duratec Gray Surfacing Primer. It’s polyester based and needs to be mixed with MEKP as a catalyst. According to their website:

Duratec® Surfacing Primer offers minimal shrinkage and a porosity-free surface without pinholes or defects. It features a super-fine levelling and filling system that sands easily, builds rapidly up to 40 mils on composite plugs, and can be polished to a Class 'A' mirror finish.

The Duaratec is sprayed as thin and evenly as possible so as to not affect the dimensions of the final part and to reduce post processing. The molds were then lightly with 220 grit and buffed with [???ANDREW].

Phase ?: Making the Part

The mold was then waxed with ??? to keep the part from sticking… on fiberglast website… is a release agent also used??? Parting paste by rexco

PVA but not for a part this simple