Today’s rapid evolution of capacitive touch screen technology makes digital prototyping more necessary than ever. While simulating prototypes just before the manufacturing process can enhance product quality and significantly reduce cost, the difficulty of incorporating digital prototyping into the production flow is a serious roadblock because simulation software is still quite complicated.
Here are the real pains in touch screen simulation as we’ve known it so far.
Draw or import the model
Most commercially available simulation tools offer the option to draw or import the model that is to be simulated. However, drawing a touch panel from scratch is not a good idea; electrode patterns are too complicated, consisting of many tiny details. For instance, how many hours would you need to design a touch panel of the snowflake pattern shown in the figure below by using the draw option of a simulation tool? And more importantly, how valuable is your engineering time to be spent in such handiwork?
Figure 1. Drawing from scratch a touch panel of such complex patterns (i.e. snowflake) in a simulation tool is not the most practical option. Image: “Electrostatic Simulation Methodology for Capacitive Touch-Screen Panels”, Thesis submitted by Barry Cannon for BSc Applied Physics, Dublin City University, July 2014.
A more workable option to create the model is to import it directly in the simulation tool. Most commonly, touch screen designers create the electrode patterns in 2D drawings in .dxf or .dwg file format. A 2D drawing is also created for the cross-section of the stack up, showing the dimensions of each layer. Therefore, in order to simulate a touch screen, these 2D files must be imported and transformed by the simulation tool into the real 3D touch screen model as shown below. Depending on the process followed by each simulation tool, this demanding task may often result in a flawed model (Fig.3) that either does not accurately represent the real 3D touch screen model, or cannot be accurately solved. In such cases then, an expert is needed. Someone who can “heal” the model and make it ready for digital prototyping.
Figure 2. In order to simulate a touch screen the .dxf or .dwg files of electrode patterns must be transformed into the real 3D touch screen model.
Figure 3. An example of a flawed model; the curves of the crossover are not properly connected.
Define Boundary Conditions and Generate the appropriate Mesh
Once the model is created, the next step is to define the boundary conditions. That is, to select which entities are perfect conductors, which of those are stressed by voltage or are left as floating conductors, and which volumes are specified as dielectric materials; for the latter, the values of dielectric permittivity (εr) are also required. It is clear that such a job can be done efficiently only by an engineer, or a physicist. In other words, the touch screen designer is required to be too much of an expert to set up this kind of simulations.
Figure 4. Detail of triangular mesh in a double-diamond touch screen model.
Given that boundary conditions are defined, the model must be meshed by using triangles, polygons, tetrahedrons or hexahedrons, depending on the simulation method. In particular, the user has to mesh either the surfaces of the geometry (Boundary Element Method, Fig. 4) or all the volumes of the model as well as the space around the model (Finite Element Method). Regardless of the selected method, it is a common truth among simulation engineers that the results are more or less affected by the density of the generated mesh (see Fig. 5). By rule of thumb, the denser the mesh, the more accurate the results. But that also increases simulation time. In addition, although Adaptive Mesh Refinement (“AMR”) may bring you somewhat closer to accurate results, it is certainly not a panacea. Which are the best convergence thresholds and the optimal number of iterations with respect to simulation time?
Figure 5. As the mesh elements increase in number simulation results converge to the correct value.
The competitive environment of the touch screen industry holds both perfect product quality and short time-to-market as major objectives. In the field of digital prototyping, this means getting simulation results of the highest accuracy in a very short time. For the majority of commercially available simulation tools, this twofold demand seems unattainable, as a denser mesh leads to more accurate results on one hand, but increased simulation time on the other. Also taking into account that touch screen designers need to check how the model responds for many different finger (or stylus) positions (i.e. Fig. 6), simulation speed combined with high accuracy is of the uppermost importance.