A 3D Printed Toolbox for Opto-Mechanical Components
AA 3D Printed Toolbox for Opto-Mechanical Components
Luis Jos´e Salazar-Serrano , Juan P. Torres , Alejandra Valencia Quantum Optics Laboratory, Universidad de los Andes, AA 4976, Bogot´a, Colombia ICFO-Institut de Ciencies Fotoniques, Mediterranean Technology Park, 08860 Castelldefels(Barcelona), Spain Dep. of Signal Theory and Communications, Universitat Politecnica de Catalunya, 08034 Barcelona,Spain* [email protected]
Abstract
Nowadays is very common to find headlines in the media where it is stated that 3D printing is a technologycalled to change our lives in the near future [1]. For many authors, we are living in times of a thirdindustrial revolution [2, 3]. Howerver, we are currently in a stage of development where the use of 3Dprinting is advantageous over other manufacturing technologies only in rare scenarios. Fortunately,scientific research is one of them [4]. Here we present the development of a set of opto-mechanicalcomponents that can be built easily using a 3D printer based on Fused Filament Fabrication (FFF) andparts that can be found on any hardware store. The components of the set presented here are highlycustomizable, low-cost, require a short time to be fabricated and offer a performance that comparesfavorably with respect to low-end commercial alternatives.
Introduction
3D printing technology is evolving continuously in various directions. The development of scientific toolsis one of the areas that is more rapidly growing since it is opening the possibility of using state-of-the-artscientific equipment at a fraction of the cost with respect to commercial alternatives [5–8].The implementation of opto-mechanical components using 3D printing has a direct impact on thephotonics community [9, 10]. Researchers are no longer constrained to work with available commercialelements and therefore their experimental setups can be more versatile. Thanks to the fact that thefabrication time is minimal, the components can evolve very fast from the researchers experience andthat evolution process can be a creative way to engage young researchers in photonics. Moreover, thecomponents 3D printed can be considered as prototypes and thus large manufacturers can take this ideasand build better equipment.Additionally, since the 3D printers based on Fused Filament Fabrication (FFF) are becoming moreaffordable the opto-mechanical components of our set can be fabricated practically in any location.It is important to highlight that since the components fabricated have a similar performance thanlow-end commercial alternatives in terms of stability and robustness, the approach presented here makesphotonics more accessible to industry and academia both in developed and developing countries, enlargingsignificantly the size of the community interested in performing experiments in photonics and relatedfields [11]. In economically developed countries, the ideas presented here constitutes a way to reduceincreasingly rising costs, and in developing countries, allows to overcome funding restrictions and largelead times due to customs and administrative procedures [9].The toolbox presented in this letter is composed of components highly customizable, low-cost, thatrequire a short time to be fabricated, offer a performance that compares favourably with respect tolow-end commercial alternatives, that are also aimed at complementing other component libraries alreadyavailable on internet [5, 7]. 1/9 a r X i v : . [ phy s i c s . i n s - d e t ] J un aterials and Methods We have designed and fabricated a set of opto-mechanical components that are essential in any opticslaboratory either for research or teaching. For the sake of illustration we have fabricated all the opto-mechanical components to construct a Michelson-Morley interferometer, since it is an experimental setupthat requires a very diverse set of opto-mechanical components for its implementation. In figure 1A wepresent some of the components fabricated, such as kinematic mounts , used to set the tip and tiltof mirrors or lenses, translation stages , used to set the position of a component along a single axiswith high precision, kinematic platforms , used to support and set the position of components such asprisms or beam splitters. In addition we have also built an integrating sphere , used to measure thepower of a light source and other basic components such as post holders and post clamps . All theplastic parts were printed using a Prusa-Tairona [12] 3D printer that cost ≈ A C and is manufacturedin Colombia. (a) (b) (c)
Fig 1. Set of opto-mechanical components printed. (A) A Michelson-Morley interferometerimplemented with opto-mechanical components made in plastic. (B) Fused Filament Fabrication (FFF),a method of rapid prototyping: 1. Nozzle ejecting molten material (plastic), 2. Deposited material(modelled part), 3. Controlled movable table. (C) Printing a component using a Prusa-Tairona printer.The complete scheme of fabrication used to build the opto-mechanical components consists of twoparts. In the first, the main elements are built in plastic using a standard 3D fabrication process thatcan be described as follows: to start, a 3D model of the component is designed using a CAD softwarefor instance OpenSCAD, Blender, Solid Works, or can be downloaded from a digital design repositorylike Thingiverse if the model has been already designed. For our purposes, all the plastic componentswere designed using OpenSCAD, an open-source, script-based software that generates 3D models bycombining (adding or subtracting) primitive shapes such as cylinders, spheres and cubes, that is veryeasy to use. Once the model is available, it is further converted using another software, like slicer orcura, into printing instructions for the 3D printer. In our case, the Prusa Tayrona printer uses theprograms Repetier Host and Slicer to generate the instructions (gcode) and print the piece, respectively.Afterwards, the component is fabricated using an additive manufacturing technology known as FusedFilament Fabrication (FFF) in which a filament of PLA (Polylactic Acid, a non-toxic, biodegradablethermoplastic polymer made from plant-based resources such as corn starch or sugar cane) is heated andthen extruded through a hot nozzle. The hot plastic is deposited layer by layer following a given patternso that each layer binds with the layer below to build a solid object. A moving platform or moving nozzledetermines the position of the hot filament and thus the shape of the solid object printed (figures 1B and1C).Once the main components are printed, the second part of the construction process starts. The plasticelements printed are combined with components like nuts, screws, bolts, washers, springs and rods, easilyfound on a hardware store, to create a fully functional opto-mechanical component. Figures 2A and 2Bdepict a drive screw mechanism implemented by embedding a nut in the plastic. Figure 2C shows theimplementation of a linear bearing using a rod. It is interesting to note that even though the individualcomponents added in the second part are not designed for high precision applications, we have found thatthe opto-mechanical components fabricated with them, provide very similar performance with respect to2/9ts commercial counterpart, as is shown in the results section. (a) (b) (c)
Fig 2. Mechanism Implementation. (A) Nuts of different shapes embedded on the plasticimmediately after fabrication. (B) Drive screw mechanism implemented using a nut embedded in theplastic and a screw. The precision in the displacement is limited by the nut’s thread. (C) Linearbearings can be replaced by holes carefully made in plastic through which passes a metallic rod.
Kinematic Mount / Kinematic Platform
A kinematic mount (KM) is an opto-mechanical component used to adjust precisely the tip and tilt ofa mirror (or lens), while it holds the component securely in place, as shown in figure 3A. Similarly, akinematic platform (KP) may be seen as a rotated kinematic mirror mount that is mainly used to controlthe tip and tilt of a flat surface where other components such as prisms, beam splitters or non-standardoptics are secured. (a) (b) (c)
Fig 3. 3D printed kinematic mount. (A) Finished kinematic mount with mirror mounted. (B) 3Dmodel of plastic components (KM top: leftmost component, KM bottom: rightmost component). (C)Components required to build the kinematic mount.Our proposed implementation of the kinematic mirror mount, follows the widely used cone, groove,and flat constraint scheme [13], and is based on the original design of Doug Marett [14] that can be foundon Thingiverse. The mount is implemented by printing two pieces of plastic (figure 3B) that are jointtogether using a sphere and two springs secured with two rods on each side. The drive screw mechanismis built using two nuts that are embedded into the plastic. Two M4 screws with rounded nuts in one endare used to adjust the tip and tilt respectively (figure 3C). The rounded nuts are used to keep the twoplastic pieces into position and to reduce any unwanted motion. A complete list of the required materialsis presented in table 1.Typically, a mirror mount may cost between 35 A C, for a basic mount, to 150 A C, for a more advancedcomponent. The values indicated correspond to an average cost over very popular manufacturers suchas Newport, Thorlabs and Edmund Optics. Regarding the lead time, it may vary between 2-7 days forEurope or United States to 1 to 6 months for other countries (for instance 6 months for Colombia).3/9 able 1. Bill of Materials for Kinematic MountComponent Comments Quantity Unit cost [EUR]
KM top † Vol 9.663 cm † Vol 10.588 cm φ = 8 mm 1 0.25Hex nut M4 4 0.25Rounded nut M4 2 0.25Screw M4, L = 4 cm 2 0.25Tension spring L = 1 cm 2 0.25Metal rod L ≈ TOTAL 13.35 † Printing cost of 0.5 EUR / cm is assumed.It is clear that with plastic we cannot compete against high-performance components, due to theplastic mechanical properties; however, we can still reproduce the behaviour of low-end opto-mechanicalcomponents at a fraction of the cost and with a significant reduction in lead time to a few hours.The kinematic mount shown in figure 3A was built in three hours: the first two hours were spentprinting the two plastic components, and the last hour was spent building, adjusting and testing thecomponent. The total cost of manufacture is 12 A C, including the components purchased in a hardwarestore. As a result the cost is significantly reduced, but more importantly the lead time is dramaticallyreduced from months, to hours.
Translation stage
A translation stage (TS) is a component typically used to vary the position of an object along a singleaxis. The position of the moving platform is controlled with a drive screw mechanism more precise thatthe one used in the kinematic mount (figure 4A). (a) (b) (c)
Fig 4. 3D printed translation stage. (A) Finished translation stage. (B) 3D model of plasticcomponents (TS top: leftmost component, TS bottom: rightmost component). (C) Components requiredto build a translation stage.The translation stage is composed of the two pieces shown in figure 4B. Unlike a traditional system(for instance see ref [15]), where the rods are fixed and the moving section has two linear bearings; thestage printed has the two rods fixed to the moving section, and the walls of the main platform act asbearings. With this scheme it was possible to obtain a very stable and robust linear motion along theaxis of the rods. The drive screw mechanism is built using a single nut that is embedded in one wall ofthe main platform. A M4 screw with a rounded nut in one end is used to set the position of the movingsection over a range of 1 . A C. This4/9mount corresponds to a considerable investment for a laboratory or company that is on an early stage ofdevelopment. Regarding the lead time, there is no much difference with respect to the kinematic mountpresented above.The component shown in figure 4A was built in four hours. Three hours to print and clean the plasticcomponents and one hour to leave the stage fully operational. In table 2 are indicated the requiredmaterials and the corresponding manufacturing costs.
Table 2. Bill of Materials for Kinematic MountComponent Comments Quantity Unit cost [EUR]
TS top † Vol 14.848 cm † Vol 9.663 cm φ = 4 mm, L ≈ . L = 4 cm 1 0.25Grub screw M6, L = 1 cm 1 0.25Spring L = 1 . TOTAL 33.2 † Printing cost of 0.5 EUR / cm is assumed. Integrating sphere
An integrating sphere is an optical component composed of a hollow spherical cavity that has two smallwindows orientated at 90 ◦ one with respect to the other [16]. The first window corresponds to the inputport whereas the other is the output port where a detector is located. For commercial devices, the interiorof the sphere is covered with a diffuse reflective coating. When light enters to the sphere, it is reflectedequally in all directions due to scattering. As a result, an integrating sphere is a device that can be usedto measure optical power, while the spatial information of the input beam is destroyed (i.e. beam shapeand entrance angle).In Fig. 5 is shown the 3D printed integrating sphere developed. The hollow sphere is printed usingwhite PLA as shown in figure 5B. In our design the output window allows to use any of the followingdetectors: photodiode, webcam, or the camera of a mobile phone. In the case of using a webcam, aprogram written in Python is used to calibrate the integrating sphere with respect to a known referenceand also to perform further power measurements. (a) (b) (c) Fig 5. 3D printed integrating sphere.
Panel (a). Integrating sphere using a webcam as a detector.Panel (b). 3D model of the plastic integrating sphere. The inner sphere is printed as a complete piece.In the figure the sphere is divided in two sections for ilustrative purposes. Panel (c). Integrating sphereoperating with lights off. 5/9or each image taken with the webcam, a corresponding light input power is determined by calculatingthe area behind the grayscale histogram of a masked version of the image. The mask used is a circularmask of radius 100 pixels. In our setup we used a Genius Eye 100 webcam.In table 3 is presented the cost of fabricating the integrating sphere shown in figure 5. The couplercharacteristics depends on the detector used.
Table 3. Bill of Materials for Kinematic MountComponent Comments Quantity Unit cost [EUR]
IS top † Vol 20.475 cm † Vol 1.6 cm TOTAL 11 † Printing cost of 0.5 EUR / cm is assumed. Results and Discussion
To determine the performance of the opto-mechanical components fabricated, a direct comparison withrespect to its commercial counterpart was carried out by using the experimental setups presented in Fig.6. In Fig. 6A it is shown the scheme used to compare the kinematic mounts. The input beam isgenerated using a He-Ne laser with Gaussian spatial profile and a beam waist of ∼ − ◦ , +1 ◦ ) in the horizontal and vertical directions is dividedequally in seven points. Each direction is scanned ten times by rotating either the X and Y knob, inorder to evaluate the hysteresis and repetability of the component.Figures 7A and 7B show the centroid position as a function of the X or Y knob rotation, respectively.In both cases is clearly seen that the 3D printed kinematic mount, exhibits the same behaviour as itscommercial counterpart, in this case a Thorlabs KM100 mount. In fact, when the beam is shifted inone direction, the other is confined within the same small interval as the commercial component. Fromthe characterization, the only difference observed between both components appears in the sensitivityexperienced in the knob rotation, determined by the screw thread.For the translation stage, two set of measurements where performed over the interval (0 mm , +10 mm),where the translation step was set to 2 mm. The results are shown in Fig. 8. From the results is observedfor the commercial TS, the beam drift lies within the interval ( − . ◦ , +0 . ◦ ). On the other hand,the 3D printed TS provides a worse performance, particularly for displacements beyond 6 mm, wherethe rounded screw exerts a significant pressure on the moving platform. This pressure gives rise to theunwanted displacement observed in the y-direction.Fortunately, for small translations, below 6 mm, the device presents an acceptable performance wherethe beam experiences a drift that lies within an interval of tenths of degrees. Notice that this unwantedbeam displacement is imperceptible to the eye. 6/9 a) (b) Fig 6. Experimental setup used to characterize 3D printed components. (A) Setup used tocompare two kinematic mounts; KM kinematic mount under test. (B) Setup used to compare twotranslation stages; TS translation stage under test. (a) (b)
Fig 7. Experimental results kinematic mount.
Centroid position as a function of the X knob (A)and Y knob (B).Regarding the integrating sphere, it is found that it exhibits a non linear response as a function of theinput beam intensity. Figure 9A displays an example of the device response as a function of the inputintensity. From the fitted data, a calibration curve is obtained that is further used to determine theresponse of the power meter. For the sake of example, Fig. 9B presents a single wavelength ( λ = 633 nm)comparison between a commercial power meter and a 3D printed integrating sphere. From the data isobserved a maximum relative error of less than 2%. In the experiment the light intensity is controlled bychanging the angle between two crossed polarizers. 7/9 ig 8. Experimental results translation stage. Centroid position as a function of the mirrordisplacement in millimeters. (a) (b)
Fig 9. Experimental results integrating sphere. (A) device response as a function of the intensity(area behind image histogram). (B) Comparison with a commercial power meter after calibration.
Conclusion
We have developed and characterized a set of opto-mechanical components that can be easily implementedusing a 3D printer based on Fused Filament Fabrication (FFF) and parts that can be found on anyhardware store. In particular we have compared three of the main components required to implementa Michelson-Morley interferometer, namely a kinematic mount, a translation stage and an integratingsphere with respect to commercial alternatives.From our results, we have found that 3D printing provides a suitable alternative to implementexperimental equipment in scenarios where is not required a high precision. Surprisingly, the 3D printedkinematic mount provides a very similar performance with respect to its commercial counterpart. Eventhough the results obtained for the translation stage are not so optimistic, since the beam drift isimperceptible to the eye, the device can be suitable for undergraduate laboratories. Regarding theintegrating sphere, we have developed and demonstrated a simple accessory that can be printed in orderto convert a webcam into a power detector.Importantly, in all cases we have found that a 3D printer is an extremely useful resource in anylaboratory since it opens the possibility to fabricate experimental equipment that is highly customizable,at a low cost with respect to commercial alternatives, and more importantly in a very small period oftime. 8/9 cknowledgments
LJSS and AV acknowledge support from Facultad de Ciencias, U. de Los Andes.