LaCir: A multilayered laser-cuttable material to co-fabricate circuitry and structural components.

Early prototyping research used digital fabrication to create passive shells for components, but a growing body of work makes use of digital fabrication’s ability to create custom geometries, both internally and externally, to enable object interaction. Geometry-based interactive devices like those LaCir can create rely on a relationship between intricate designs and material characteristics to sense users’ activities.

Additive methods like fused-filament fabrication (FFF) allow for detailed, fully-3D internal and external shapes. Sauron [24] uses this capability to create custom internal geometry for computer vision-based tracking of user interactions, while Digital Mechanical Metamaterials [7] prints structures that guide force around a structure for logical operations. This technique can also be used with multiple thermoplastic materials:./trilaterate [27] creates predictable 3D capacitive patterns. These are but a few of many explorations: 3D geometry, while difficult to design on a 2D screen, allows incredible design flexibility. However, 3D printers themselves are slow, so researchers have turned to laser cutting as a faster method for prototyping structure-based user interaction.

Interactive structures like living hinges, gears, or sliders can be made on laser cutters [16], and research has explored how to port these mechanisms across cutters (e.g., kerf-cancelling mechanisms [22], SpringFit joints [23]) and create 3D [1] or pseudo-3D [15] objects integrating them, thereby increasing laser cutters’ flexibility in object manufacture. JigFab [13] and MatchSticks [30] use digital fabrication to create woodworking joints; we are inspired by these works as we consider joinery the LaCir substrate. These works focus on a single sheet of material at a time, but stacking is also possible: LamiFold alternates adding layers of material and cutting or removing parts in the laser to fabricate interactive devices. LaserStacker [31], on the other hand, selectively bonds multiple layers together by vaporizing and melting acrylic. We rely on techniques from Kerf-cancelling Mechanisms along with LaserStacker to work with our material substrate, but its capabilities go beyond the structure explored in those works to include electrical functionality.

Other efforts exploit intrinsic properties of materials or design their own substrates to enable interactive capabilities in fabricated objects.

Foldem [3] designs a composite material with differing flexibilities per layer: one rigid, one bendable, and one flexible. By selectively cutting more-rigid layers and leaving only more-flexible ones, the meta-material’s local malleability changes. LaCir focuses instead on conductivity.

Several others introduce substrates for cutting which focus on conductivity. iWood [34] creates plywood material with embedded triboelectric sensors to identify vibrations, while Olberding, et al.,’s cuttable multitouch sensor [19] uses patterned, printed electrodes to enhance robustness to cuts. Instead of focusing on avoiding damage to our substrate through cutting, we design it for selective, intentional cutting by the user, more in line with VoodooIO’s flexible conductive substrate [32] or copperclad board. LASEC [5] and Fibercuit [35] introduce circuit-focused material similar to ours in which one layer is conductive and the other is not, but they target flat, stretchable, foldable, and wearable devices, and do not explore the structural requirements of 3D, assembleable, rigid, or jointed structures. Wessely, et al.,’s Shape-Aware Material [33] is designed so that cutting it is the functionality; instead of enabling digitization of physical work, LaCir uses a digital-first process for fabrication. In general, we share these works’ vision for a mass-manufacturable substrate offering new properties in digital fabrication.

As mentioned, 3D printers can extrude conductive thermoplastics in arbitrary 3D patterns [11, 27], but here we focus on rapidly-prototyping 2D and 2.5D circuitry more similar to what LaCir generates; future tools could help users understand speed/functionality tradeoffs like this.

One fast method of creating 2D circuits leverages an inkjet printer (or pen [6]) and conductive ink on paper [2, 4, 10, 20]. This enables flat shapes, foldable origami [20], or flat circuits that can be transferred to other developable substrates [4]. LaCir explores structural circuits with higher strength and durability than these interfaces.

Others make circuitry using the laser cutter. CircWood [8] carbonizes wood to create conductive traces on its outer layer. However, these traces are fragile, and the resulting conductivity depends on factors like humidity; LaCir’s material is metal-based, which reduces these issues. We take inspiration from CircWood’s use of traditional fasteners (e.g., screws) as conductive components in the preparation of our example applications. LaserFactory [17] uses silver ink (fused by the laser head) on acrylic sheets to create traces, but requires a heavily modified laser cutter and does not explore 3D joinery with this technique. In contrast with this, we design a material that removes the need for machine modification and which can include joints.

LaCir thus fills several gaps by uniting structural interaction with a designed material in which users can create electrical traces. Our technique and material require no modification to the prototyping machine, support creating jointed, structurally-sound objects, and accommodate traditional fasteners.

LaCir devices can be modeled in any DXF/SVG-generating CAD tool, like Autodesk Fusion 360, Inkscape, Adobe Illustrator, or Kyub [1]. We have developed a set of design primitives that work with our substrate, and which represent physical and electrical connections of various kinds along with the requisite cuts required to fabricate them.

To fabricate their design, creators must choose a substrate. We explored wood, Delrin, and acrylic as structural layers, and various tapes, leafs, meshes, and paints as conductive layers. Each individual material and combination imparts particular characteristics to the final product, which we further discuss in Section 4.

After selection, the substrate is placed inside the laser cutter for fabrication. For thermoplastic-based substrates, we do not use a pre-sealed stack: instead, we place structural and conductive layers into the cutter separately in an alternating fashion. The various cuts made weld the substrate together in a small-scale version of LaserStacker [31]. Wood-based substrate stacks are placed in the laser already glued. After fabrication, connectors are added, pieces are joined, and the device can be connected to power. In the interest of replicability, we provide cutter settings we used to create our structural circuits (see Table 1). While setups vary, designers can follow a protocol similar to Foldem’s to tune these [3].

For our substrate’s conductive layer, several features are desirable: high conductivity, ease of cutting, capacitive touch capability, and the ability to conduct through laser-cut joints without conductive adhesive. We sandwiched between two clear acrylic structural layers the following materials: metallic leafs (silver¹, copper²), metallic tape (copper³), metallic mesh (aluminum⁴), conductive paints (silver⁵, copper⁶, carbon⁷), and ITO-coated PET⁸ (see Figure 6).

As our substrate is handmade, the thickness of the electric material varies. The leafing material can tear or overlap, the brushed- and airbrushed paint can vary in thickness: this may affect joinery. To test the connectors and joints we created three fixed-geometry objects with each material (where possible) and measured resistance across them (length: 31 mm) three times. We then assembled the objects and measured resistance created across one and two joints (see Figure 7), also three times. Last, we cut a final test piece with tracing and revealing cuts, then attached it to an ItsyBitsy via alligator clips to determine if it was usable as a capacitive device.

Details are in Table 2. In general, the paints were not effective in joints, due to the fact that leafs do not ablate cleanly but instead fold over the edge of the structural layer, improving contact area versus paint. The paints also did not dry well when sandwiched between the acrylic layers, leading to their still being wet and therefore non-conductive even several days after sample was prepared. The aluminum mesh was not cuttable off the roll, but when we darkened it with black water-based paint and ran our laser on minimum speed we were able to cut it with about 95 % effectiveness (19 of 20 wires in our sample separated); this could be mitigated with multiple passes. The metal-based samples also came out of the laser cutter very hot, due to their increased thermal mass as compared to pure acrylic: we had to wait longer than normal (minutes instead of seconds) for them to cool on the bed for our inter-layer welds to solidify. Many samples delaminated and had to be reproduced for the conductor experiment, a failure we explore further in the next experiment (see Figure 8).

Additionally, we were interested in identifying the minimum separation needed between two tracing cuts, as when a single tracing cut is made the “folding” of the non-cleanly-ablated conductor (see Section 4.1) can create a short across the narrow gap. A too-narrow separation fails to solve the problem, and can also affect the structural features of the substrate. To uncover this metric, we created paired tracing cuts on a layered substrate made up of acrylic and silver leaf, separated from 0.1–0.4 mm, in 0.1 mm increments (see Figure 9). We then evaluated the conductivity between the two created layers of the structural circuit using a multimeter.

Our results show that the closest two tracing cuts can be in order to affect conductivity is 0.1 mm, and tracing cuts with this separation did not adversely affect the bottom structural layer.

Different structural materials bring different characteristics and benefits to our layered substrate material (i.e., wood brings flexibility, while acrylic brings meltability) that lend themselves to different applications. To uncover which materials were most suitable for which applications and primitives, we carried out a series of exploratory experiments where we vary the structural layers.

These experiments were carried out using acrylic (both 1.5 mm and 3 mm thicknesses), Delrin (2 mm), and wood (3 mm). As all these materials are rigid by nature, our explorations focused on testing their capabilities to deform while maintaining their structural and conductive properties. To this end we created living hinges with these materials as the structural layers and silver leaf or carbon paint as the conductive layer, and also attempted to build springfit joints (see Figure 3, Figure 10).

The added thickness of a layered material does make soft, bendable living hinges difficult to realize, but all structural materials could do it. Significant bending can cause relative motion and delamination between the layers [28]. To compensate, designers can use thinner, more bendable material like Delrin. Additionally, we confirmed that combining acrylic parts using SpringFit [23] joints is not possible, which was expected as this technique was designed for wood. We note that all tested materials can create satisfactory LaCir substrates: there is no one best material. There is, however, a “best material for a purpose.” For example, our experiments revealed that, due to increased flexibility, wood dielectric layers are suited to applications requiring mobility, while acrylic is better for rapid prototyping with melting and welding by the laser. Delrin is a balance: more flexible than acrylic, but slightly less meltable, which led us to focus more on acrylic in our tests due to our use of LaserStacking techniques.

With knowledge of different conductors, we also needed to know how they could combine with our various structural substrates. We created several combinations to explore lamination and cuttability characteristics: acrylic with all cuttable conductive layers, wood with metal leafs and glue, wood with conductive paints, and Delrin with silver paint.

We found that the thickness of a thermoplastic determines a lot about the strength of the bond; thinner plastic layers (or less-meltable plastic layers, like Delrin) create less welding material during the through cuts, leading to worse outcomes. This can be mitigated through applying additional materials post-fabrication, for example conductive glues, though this of course increases user labor and fabrication time. Some pairs of materials did not provide acceptable results: in our wood and copper stack we saw burning due to the heat mass of the copper.

Having a rocket lamp is the dream of every kid. In our LaCir lamp, we connect power at two of the rocket’s “feet” to light an LED at its nose (see Figure 11). We use our special laser-cut joints between body parts, magnets to attach the LED, and screw inserts to enable alligator-clip connection to power. The entire rocket body is conductive; we did not use any tracing cuts to guide power on particular paths, except that the third leg is designed not to short the circuit between the other two by using a non-conductive traditional finger joint. The third leg hosts a screw insert; its entire surface is a capacitive touch sensor to activate the LED. We fabricated this lamp in acrylic and silver paint, as its coloring gives an other-worldly feel to the design.

We also implemented an electric ‘Wheel of Fortune’ with our primitives (see Figure 12). This design enables spinning the wheel to create contact with a random base segment; the wheel is grounded and each base segment has its own power source—separated by tracing cuts—thus spinning completes one of three circuits. We use press-fit ball bearings and magnets to connect bottom side of the wheel to the base segments both physically and electrically.

To highlight the inter-layer joint capabilities and underscore the utility of tracing cuts, we prepared a PCB which features two vertical vias that together create a flyover trace (also called a bridge) (see Figure 13). Our PCB has 4 coplanar connection points, which are electrically connected in diagonal pairs. To allow for such a jump without a short circuit, we use a 2nd layer.

We leverage a wide variety of existing techniques from lasercutting litterature, but future work should explore our substrate’s compatibility with others, such as LaserOrigami [15] or Fibercuit [35]: anecdotally, we have found that the heat dissipation features of tested conductive layers complicate uniformly, simultaneously deforming both structural layers in a coordinated way, but future strategies may mitigate this.

LaCir’s substrates, in theory, are scalable to arbitrary sizes. In our exploration, we were limited by the sheet size of conductive layers we could purchase, as placing multiple sheets adjacent to each other resulted either in a missing electrical connection or a slight overlap which created unpredictable cutting results. Naturally, other structural considerations come into play at larger sizes (warping, reduced effect of edge-welding, etc.): a topic for future research and development. On the small end, we briefly explored integrating surface-mount components (SMDs) with our tracing cuts, but found that the relative sizes of executed tracing cuts (≈ .3 mm) compared to SMD pitch along with both variability in our hand-assembled layers (e.g., torn leaf) and challenge in exposing our structural layers to solder heat made this difficult. An industrially-manufactured substrate and higher-precision lasercutter could mitigate this.

When using traditional tools that isolate the structural and circuitry design processes, the two sets of designs can be largely independent of each other (though generally circuitry is intended to fit inside of an object’s structure). With LaCir, while it is possible to create multilayer circuitboards through alignment of multiple stacks of cut substrate, due to the nature of structural circuitry the two designs are somewhat more entangled.

We explored a variety of commonly-available materials to serve as structural and conductive elements in our substrate stacks. Some proved difficult to cut—a challenge which could be mitigated through using fiber lasers (see Figure 14), CNC mills, waterjet cutters, or other types of machines. These technologies would remove the possibility of applying LaserStacker techniques [31], but may offer other opportunities for unique substrate manipulation and may be less sensitive to small thickness changes in the conductive layer. Thicker metallic conductive layers may also make it possible to design objects with meaningful thermal transfer capabilities in addition to the electrical ones we explore [14]; the substrates our CO₂ laser can cut are too thin for this to be of much effect.

A conductive layer which can be selectively melted (similar to the behaviour of acrylic) by the laser would also provide interesting opportunities. Low-melting-temperature metals such as gallium could enable better circuit connections at piece boundaries and through joints.

Use of LaCir’s material stackups requires slowing down the laser’s cutting beam to get all the way through, as the material is thicker and the electrically-conductive layers also tend to have heat-conducting properties. This leads to a slowdown of up to an order of magnitude, as seen in our settings table. However, this minor fabrication-time slowdown is dwarfed by the amount of design time spent in integrating tracing, joints, and other hardware cuts into the design. A future design tool could integrate features of e.g., Eagle⁹ into Kyub [1] with flood-filling techniques that map circuitry onto laser-cuttable geometry, but this was outside the scope of our exploration into the possibilities of a structural substrate.