3D printing is a relatively new technology which has been hugely beneficial to almost all engineering and product development activities, for example in space, in the automotive industry, and in giving people superpowers. When stereolithography first began to widely used in industry in the mid-1990s, fragile, small parts could cost thousands of dollars, but there is now a wide variety of 3D-printing technologies and materials, and the cost has reduced dramatically, such that 3D printing is now accessible to almost anyone.
Our tricorder CAD was a constantly evolving mechanical design for well over a year, as we pulled together all the elements that we wanted to fit inside the skin of this iconic prop. Several of the early mechanical concepts required physical testing before being developed into a design that could be manufactured and 3D printing was a key part of this process. 3D printing allows us to quickly produce real parts in practically any shape we like, which could be difficult or even impossible to injection mould. This is helpful, as we can get a concept made quickly which allows the parts to be tested in mechanisms before being tweaked and developed into models that are designed as production parts. Once the mechanical concept is tested, then the parts have to be designed so that they will be manufacturable in an injection mould tool with the associated limitations and essential draft angles.
The tricorder CAD needed refinement right up to the point where the design was frozen so that injection moulding tools could be cut. Injection mould tooling is tremendously expensive, and once the tooling is being made, further changes are difficult and costly to make. Of course, if there are issues that crop up in the realisation of the design, sometimes there is no choice but to revise the design after the tooling has been started or even completed.
To minimise the potential for changes this late on in the process and also, from the earliest point possible, to find out if the CAD design would work in the real world, each of the components have been 3D printed, often multiple times. 3D printing the tricorder has been critical in realising the design and understanding fully how it will fit together, be built on the production line and actually work in the hands of a user. Seeing the geometry turned into real 3D printed components assembled, all mechanically slotting into each other, truly brings the tricorder to life. After many months of working through the CAD, it’s exciting to see the model actually working and makes the product very real in every sense.
But exactly how did we get from a virtual tricorder design to a real-world printed mechanical prototype model?
Initially, the tricorder component files were created and stored as a parasolid geometric modeling kernel; a format used by SolidWorks. In order to 3D print a given tricorder component, the file is output as an STL file which can be imported into the 3D printer’s native software where the orientation is chosen and supports (scaffold structures to hold overhangs) are created. This semi-automated process generates a position and support scaffold but enables the user to move or add supports and alter the orientation of the part with respect to the print surface in order to get the best finish on the most visible or most mechanically critical area of the part. Often the best orientation is not what you might expect. For the tricorder’s screen cover, with its thin low-profile section, from a speed and simplicity perspective, it would have seemed obvious to lay it flat on the printer bed. However, to minimise the number of supports attached to the middle of the screen, and avoid contour-stepping on the gently curved surfaces, it was turned to stand upright, thus producing a much better-looking and functioning part. Once the part to be printed looks as though it will print without errors, the 3D printer converts the file to a series of 0.025 mm slices which are then output as a layer-by-layer set of g-code commands for the 3D printer.
We have a Formlabs SLA (stereolithography) machine which we used to print the tricorder prototypes due to its superior resolution. A tray of liquid resin is cured, one 0.025 mm slice at a time, by an ultraviolet (UV) laser, allowing the model to grow out of the resin reservoir layer by layer. Based on the mechanical properties we need, we have a selection of resins with different properties to choose from depending on which tricorder part is being printed, although our printer has a 145 x 145 x 170 mm print volume which makes it only suitable for smaller, more detailed parts. For the display glass and LED light pipes, we used a clear resin that can be polished to make it transparent. For the first mechanical trials of the disc ejector mechanism we used a blue resin called ‘Tough’ which has more durable and resistant properties, similar to a production plastic.
Despite the accuracy, each SLA part requires a great deal of finishing. Newly printed parts must be carefully washed and then cured using a UV light chamber.
Once cured, the part must have its scaffold support structure removed. As the support structure is the same material as the model itself, if a blemish-free surface is necessary, then this stage can be very time-consuming.
We also have various FDM (fused deposition modelling) machines to use for larger, less critical parts, or if we need to choose a specific 3D printing material to mimic the tough, flexible properties of the injection moulding plastic like ABS for example. FDM machines are the most common 3D printers because they are relatively simple, cheap, clean and easy to use. A reel of plastic filament is melted through a hot nozzle and applied layer by layer onto a platform. Components printed by the FDM process are generally good for block models and other more basic parts because they are quick, but the finish is not smooth, as the layer thickness is usually around 0.1 mm which shows as visible layer lines.
For some of the larger tricorder parts where we wanted a tough, resilient, but also accurate, print, we had to use an external service. For these parts we used HP’s Multi Jet Fusion (MJF) process to produce the highly accurate, thin, stiff and nicely finished panels needed for the tricorder’s doors and sides. Parts emerge from a fine powder bed as a modified inkjet printer first prints layers of a binding agent, then rapidly solidifies the composite with heat. The powder bed supports the part during printing so no support scaffold is needed, greatly simplifying the amount of component surface finishing required when the printing is complete.
It’s not only us that need to see and feel the tricorder: our factory also makes use of rapid prototyping and Stratasys PolyJet 3D printing machines to get a better understanding of what they are going to have to manufacture. A fully working prototype is essential for the factory to test the fit and build sequence and make sure that the CAD really works before tool cutting is started. There are many aspects of the design that are very difficult to gauge from just looking at CAD on a screen or by testing functions with simulation software – for example, how bundles of cables are routed, what the detents on the hood opening action feel like in operation, or how the range of internal state sensors perform together. It is usual for these factory models to be expensive and time-consuming to build and, more often than not, the process is repeated a number of times whilst the design is refined as issues are discovered and ironed out in subsequent revisions, which then also need testing. They are, however, a very necessary step towards full manufacture.
To make a fully working tricorder prototype, 3D printing can only get you so far. The tricorder is constructed from a range of different materials, and in addition to 68 injection moulded parts which can be 3D printed, there are 22 metal parts, a number of different sized screws, bolts, springs, and magnets, a hand-stitched leatherette strap, and more than 350 electronic components assembled on to six printed circuit boards.
Where it is easier or more accurate to do so, many of the metal parts are created using traditional CNC machining, laser or waterjet cutting. For the tricorder’s side frame plates, we had access to a laser cutter which allowed us to initially cut acrylic sheet to shape to create quick prototype models in order to test our initial design concepts. As soon as the design had been refined, the side plates were sent out to an external fabricator to be laser cut from aluminium. For some of the parts it was quicker to use some of our more traditional prop making skills and, when some angled handles were needed in short order for the prototype model, the basic shape was milled from stock, in-house, and finished by hand with a file.
In fact, a file, sandpaper, a Dremel, scalpels and just about every other small model-making tool are not yet redundant and are employed throughout the 3D printing process as each part requires a significant amount of hand finishing.
So, while on the face of it, 3D printing direct from CAD files appears as though it might reduce the need for the craftsmanship that gave us some of our most cherished and iconic props from yesteryear, those traditional prop-finishing skills have not been entirely outmoded by the 3D printer. And, for now at least, there is still a huge degree of hand finishing required to build complex models such as the tricorder from 3D-printed components. Nevertheless, 3D printing is enabling us to get ever closer to creating prop replicas that truly embody the original intent of those gorgeous fully hand-made props from decades ago.
Coming next time
Functional blocks – working out how it will work
Fans who register with us will be the first to read our news and the progress of this exciting Tricorder development and, later, where and when to purchase it. You can catch up with the story so far here on this blog, but if you haven’t already done so, why not register your interest in the Tricorder – you’ll then receive a personalised registration certificate and early access to these updates (before we publish them on our website).