The distributed digital manufacturing of free and open-source scientific hardware (FOSH) used for scientific experiments [1
] has been shown to in general reduce the costs of scientific hardware by 90–99% [2
]. These impressive cost savings have proven resilient across both standard [3
] and custom equipment [4
]. This has supported the rapid growth of an engineering subfield to develop FOSH for science, which is represented by the annual Gathering for Open Science Hardware
] as well as two new academic journals, the Journal of Open Hardware
. There are numerous examples of FOSH scientific equipment in all fields, ranging from syringe pumps [6
] to self-assembling robots [7
]. Examples exist in the field of biology [8
], optics [13
], and microfluidics [14
]. Many open tools exist for physics and materials, including radial stretching systems with force sensors [16
], a robot-assisted mass spectrometry assay platform [17
], a large stage four-point probe [18
], and automated microscopes [19
]. Simple yet essential devices for health and medical treatment in the developing world include a mobile water quality tester [20
] and a sample rotator mixer [21
]. There are open-source ventures into Internet of things (IOT) energy monitors for buildings [22
], energy-efficient homes and subsystems [23
], and even smart cities [24
]. With the development of building-block technologies, it is less time-consuming than it was in the past to share and collaborate on open-source scientific instruments [25
One of the primary enabling innovations that provide the opportunity for distributed manufacturing of open-source hardware-based [26
] scientific equipment is the 3D printing capabilities of the self-replicating rapid prototyper (RepRap) project [27
]. RepRap 3D printers have been used to provide high-quality educational experiences for students in a wide range of disciplines in the classroom [28
] and have become scientific platforms themselves [30
]. A maturing network of peer-production [31
] and 3D printing file repositories [32
] provides both time and cost savings within scientific labs [33
]. Combining 3D printing with off-the-shelf components and open-source electronics (e.g., the Arduino prototyping platform) has enabled the automation of scientific equipment. As the fabrication of scientific equipment moves away from a central paradigm of purchasing proprietary equipment to one in which scientists themselves download open-source designs, fabricate components with digital manufacturing technology, and then assemble the equipment themselves, there is a need for a standard procedure that designers can follow when targeting this audience. This procedure is made up of design steps, which are activities that have to be performed to come to a fully defined product [34
], and follow a set of design principles, which are the general rules leading the cognitive activity of design in the appropriate direction [36
This study provides such a generalized design procedure for the development of free and open-source hardware for scientific applications. After laying out and explaining each of the five steps in the procedure encompassing six design principles, a case study is provided for an open-source slide dryer. The case study is discussed as a practical example of the benefits and drawbacks of this approach.
4. Case Study: Slide Dryer
This generalized procedure for design was developed through experience and relied heavily on the Open Source Lab
] and the best practice guidelines from OSHWA [53
]. In order to demonstrate the steps in the creation of FOSH hardware for science, a case study is presented on the development of an open-source slide dryer. Slide dryers are designed to gently warm glass microscope slides to decrease the drying time for experiments after cleaning steps. Slide dryers allow users to increase their productivity. Slide dryers are available commercially for $
]. Commercial slide dryers come in many different shapes and sizes, and with different capabilities [56
]. As a generalization, all slide dryers provide a rack structure and a heat source.
In this case study the target is to design a FOSH slide dryer with an acceptable capacity (30 slides) and a fast drying rate (10 min or less). The numbers chosen are somewhat arbitrary but, due to the parametric design of the system, design constraints may be altered to better fit the requirements of a specific laboratory. Note that the two target features (capacity and dry time) cannot both be optimized using the current design—as drying time decreases, the slide count must also decrease for a given power consumption.
In the first step, the existing literature is surveyed for slide dryer designs. There have been some efforts to patent the concept of slide drying [57
]. One attempt [57
] uses an electric current to generate heat; however, it has since expired. Another design [58
] patented in Russia uses forced air. Yet another design [59
] uses gas forced through a tube in order to create heat and has also expired. Next, a search for open-source solutions is carried out. There is one design available on the Internet, “Glass Slide Dryer” [60
]. Though this design is functional and less costly than commercial systems, it has a few apparent issues:
These issues have prevented its widespread adoption.
Finally, commercialized slide dryers are reviewed. The most expensive option (over $
] is able to heat 57 slides (unless an additional shelf is purchased for $
284) at 70 °C. Many other options are available [56
], but all products are expensive considering their function. The cheapest design that almost fits the target specifications (its slide capacity is too small) comes in at $
] and most slider dryers or warmers were $
300–1000. A more detailed techno-economic comparison is made in Section 3.5
Upon review of the existing options, it is found that the FOSH community is in need of a well-documented, customizable, and effective slide dryer. Concepts are generated, tested, and simplified and refined until an optimal design is found. The simple proof of concept (step 1) that led to this final design was simply aluminum wire wrapped around a box hooked to a variable power supply. The chosen design, which was designed to be parametric in OpenSCAD (steps 2a and 2e), involves 3D printing a base with a peg structure on an open-source 3D printer, (steps 2a and 2d). The 3D printable parts are designed to minimize filament consumption (step 2c). Then readily available wire (step 2f) can be woven across the base. When voltage is applied, electrical energy will be converted to heat due to the resistance of the wire [66
] in a simple design (step 2b).
Twenty AWG copper magnet wire is selected for its low cost and resistance to corrosion [67
]. The resistance is measured by measuring out a long length of wire, in this case 10 m. Then, using a fluke multimeter, the resistance of the length can be found. Simply by dividing the measured resistance by the length, the resistivity can be found. For the specific wire used [67
], a resistivity of 0.000220 Ohm/mm is found. This value is required to find the minimum length of wire to match the selected power supply.
An off-the-shelf (step 2f) 12V 5A power supply is selected [68
] due to both low cost and high availability. Additionally, most off-the-shelf supplies like the one selected have thermal overloads built in to prevent damage due to short circuits. Using Ohm’s law, the necessary length L
can be found, given resistivity ρ
, current I
, and voltage V
The wattage, P
, consumed is simply defined by:
It should be noted that it is not wise to run a power supply continuously at full capacity [68
]. Therefore, it is advised to use a fraction of the available I
. In this case study, 90% of I
is utilized in the design.
has been determined, it is only a matter of distributing the wire among the rack system. The rack is developed in OpenSCAD (step 2a). This allows for the design to be entirely parametric (step 2e), as well as transferable to customizers [37
]. Key parameters that the model depends on are:
Wire Resistance: The measured resistivity of the heating element (in Ohm/mm).
Wire Diameter: The diameter of the heating element (in mm).
Supply V: The voltage of the power supply (in V).
Supply I: Maximum allowable current from the power supply (in A).
Slide count: The desired number of slides to dry (number).
Slide Dimensions: Width and length of the slides (in mm)
Printer Dimensions: The 3D print bed surface area X and Y size of the 3D printer to be used.
There are many lesser dimensional parameters, which specify features such as winding pegs and rack height, which can be adjusted by the user to make a slide dryer ideal for their application. The SCAD model will optimize the design to fit the user’s 3D printer, while minimizing part counts (step 2b). Each rack can be connected using snap-fit connectors also generated by the model. As this is a parametric design, it allows for similar results to be achieved via different means. For example, a smaller printer can be used by printing off a larger number of shelves to accommodate the same number of slides as a larger printer can do with fewer shelves but a greater area. If only a 24-V supply is available, simply by changing the parameters, the design can still facilitate the user’s desired number of slides. The intention of this design is not necessarily for users to replicate exactly what was used in this case study, but rather to empower them to use the materials and tools readily available in their lab or workplace to easily generate a useful and reliable slide dryer for themselves.
The example design based on the desired slide count generated seven shelves for a Lulzbot Taz 5 printer [69
]. The design in the OpenSCAD environment can be viewed in Figure 1
Following guidelines for appropriate documentation (step 4), the bill of materials along with item, number, price, and source are shown in Table 1
. As can be seen in Table 1
, the cost of the materials to build the open-source slider dryer for 30 slides is $
The manufacturing of the device is fairly simple. First, the user must print all necessary components. Then weave wire around the pegs (there should be one strand of wire per set of pegs). Once one shelf is completed, the user inserts the pegs, attaches the next shelf, and wraps the wire once around the peg to tension the lower shelf. This process is repeated for all shelves. Once complete, the user strips both ends of the wire with a razor blade and cuts and places 10-mm pieces of shrink tube over the wire (do not shrink them yet). Then the wire is soldered to the middle tab and the back tab of the barrel jack (the wire is not polarized, so it does not matter which wire is soldered to which tab). Finally, shrink the shrink tube over the solder joints, as well as the unconnected barrel jack tab (as in Figure 2
The slide dryer is sliced using open-source Cura Lulzbot Edition using the high-quality default print settings; 120 g of high impact polystyrene (HIPS) filament and 10.6 m of magnet wire is used. A 5.5 mm barrel jack is soldered to the wire ends in order to easily interface with the power supply. The assembled open-source slide dryer can be seen in Figure 3
As validation (step 3), 30 slides are washed in water and rinsed in ethanol, and then placed on the open-source dryer. The dryer is then powered on, and the time-to-dry is measured while the temperature is being monitored with an open-source thermocouple-based data logger (T400, Pax Instruments). The warming kinetics experiment is repeated three times. A FLIR thermal distribution on a single rack is viewed with a thermal camera to demonstrate uniformity of heating. Lastly, dry time data for commercial solutions were collected via simulated devices using a heated plate and, in the case of the forced air variant, a fan. The plate is brought up to the maximum indicated temperature by the devices’ respective data sheet and then 10 trials are performed and averaged to find the drying times.