Retrofitting the Seek Thermal Compact with a USB Type-C Connector using 3D Printing Nov 10, 2018

Putting all of this together produces the above type-C interface board, with schematic, layout, and assembled photo shown. Two USB footprints are included, depending on whether the flexible printed circuit (FPC) can be reused from the original cable assembly on J3, or a substitute flat flexible cable (FFC) needs to be connected to J2. Since the ID pin is internally connected to GND on the FPC, it needs to be explicitly connected on the schematic for the FFC. Care has been taken to match trace lengths on the D+/D- differential pair to minimize propagation skew, but it is difficult to ensure 90Ω trace impedance on 0.8mm FR4 substrate. With revision v2.0, I corrected the aforementioned mistake with Rp and Rd, and with revision v2.1, I moved Rd and changed its footprint from 0603 to 0402.

Mechanical

Aside from building the type-C interface board, I also needed to design a replacement plastic assembly that could fit it.

Modeling

Initially, I started by measuring various dimensions of the original connector assembly using calipers, then redrawing them to build a replacement in SolidWorks. Typically, this process involves creating 2D sketches that are anchored with respect to various geometric references (e.g. points, planes, edges, etc.), and then extending them into 3D features using built-in operations (e.g. extrude, cut, fillet, etc.). Through incremental composition, this forms a complete 3D model that represents a single part, which can itself be composed with other parts through mating relations to produce an assembly. At any point in time, this construction timeline can be modified, allowing new modifications to be introduced, with subsequent changes propagated on top. This resembles a version control system for source code, where new commits can be introduced at any point in the commit history, and subsequent commits rebased on top. Having never used SolidWorks before, I made many mistakes, and needed to revise my design multiple times.

To verify correctness of the final type-C assembly, I was able to export a model of the interface board from Autodesk EAGLE in IDF 3D format, and then import it as a subassembly into SolidWorks using the CircuitWorks plug-in. Simultaneously, I was able to import vendor models for the FFC and USB type-C connectors into the subassembly. Final renders of the completed assembly are shown above.

3D Printing

I had some difficult creating a design that could be easily printed and assembled, due to the size of the model and its sub-millimeter features. In particular, the interior needed to be hollow to fit the interface board, which complicated the printing process. Most 3D printers utilize fused deposition modeling, which incrementally prints an object from the ground-up using plastic filament. During the printing process, slicing software needs to introduce support material to prevent the model from collapsing due to voids or elevated features, but then this material needs to be removed in post-processing. Typically, this is facilitated by minimizing the amount of contact introduced between the object and support material, and certain printers may even support printing with special dissolvable plastic for the support material.

This is shown in the above image, where the three white-colored prints were produced using removable supports, and the middle black-colored print was produced using dissolvable supports. The rightmost two white-colored prints were produced using ABS plastic on a Stratasys Continuous Build printer, whereas the leftmost white-colored print was produced using VeroWhitePlus on a Stratasys Objet30 printer, and the center black-colored was produced using ABS plastic on a Stratasys Dimension Elite printer. Unfortunately, all of these prints suffered from various problems; the rightmost two white-colored prints were quite brittle, structurally weak, and easily damaged by support material removal. Likewise, despite the water-soluble support material, the middle black-colored print still required manual processing for complete removal, and also suffered from structural damage. Perhaps the best was the leftmost white-colored print, which preserved all four mounting pins, unlike the others, and was also both strongest and most flexible. But, all of these prints were rather imprecise, and suffered from large tolerances; for example, the USB type-C connector didn’t always fit through the front slot, and the interface board occasionally didn’t fit when inserted from the back.

Overall, this entire process involved multiple iterations of modeling, printing, and testing, not only due to printing issues, but also from measurement errors and bad designs. Though these prints were useful for prototyping and validation, I soon realized they weren’t up to par for the final assembly. Although I had access to other 3D printers, such as the NVBots NVPro and Ultimaker 3 Extended, those were not recommended due to even worse accuracy and resolution. Instead, I had the final prints produced externally by commercial vendors, and I tested both Shapeways and Sculpteo. Due to process constraints on minimum wall thickness for this design, both companies could only utilize a HP Multi Jet Fusion printer, which uses a powder-bed fusion printing process. Unsurprisingly, the quality of the final print was essentially identical between both companies, as shown above. In terms of cost for black-colored prints, the closest that Sculpteo could offer was a gray-colored print followed by an expensive dyeing step in post-processing, whereas Shapeways could print directly with black-colored plastic. Shown below is the final assembled unit.