Over the last few years, touch has become a standard input method for next-generation user interfaces for everyday applications, from white goods to smart phones and tablets. However, adding touch to modern designs, which are not flat or are made with more shapes and multiple parts, can be challenging.
This article will show how to accomplish touch detection on non-uniform surfaces, such as discrete keys with overlays of varying thickness and curved surfaces. A later follow-up article will discuss the design and implementation of 2D trackpads with non-uniform surfaces.
Fill the air gap
When a rigid-sensor PCB is fitted inside a curved surface, an air gap is created between the overlay and the capacitive sensor. Air gaps between the PCB and the overlay material reduce the performance of capacitive touch sensors. The reduction in sensitivity is due to the fact that air has a much lower dielectric constant than most overlay materials. A list of commonly used overlay materials with their dielectric constants compared to air is shown in Table 1 . This article will discuss two main options to remove the air gap: spring contacts and flexible PCBs.
Use of spring contacts
Springs and conductive rubber or carbon contacts work well to remove the air gaps between the PCB and the overlay. In applications where it is not possible to fix the electrode PCB to the overlay, conductive rubber or carbon contacts work well to remove small air gaps, whereas larger air gaps can be removed with springs.
Figure 2: Simple helical springs soldered to a PCB pad can remove larger air gaps between the sense electrode and the overlay.
The disadvantage of using springs to connect the electrode to the overlay is that the surface area of the touch electrode is limited to the size of the spring used. According to the parallel plate equation shown in Equation 1 , small surface areas are less sensitive. This limits the possible thickness of the overlay or the relative permittivity of the overlay for a set thickness.
To improve sensitivity, conductive tape (metal foil) can be used. The tape can be stuck onto the overlay to give a larger surface area when the spring presses up against it. In applications where the overlay is required to be transparent (for example to accommodate backlighting) transparent conductive films such as ITO (indium-doped tin oxide) could be considered. Alternatively a metal plate could be used at the top of the spring, as commonly used in battery terminals.
Equation 1 shows the parallel plate capacitor equation where A is the area of the pad, ε0 is the permittivity of the air, εr is the relative permittivity of the overlay material and d is the thickness of the overlay.
Springs could also be connected to a surface of conductive paint. Conductive paint could be used around the touch key area to provide a larger surface area (A) to increase the proximity detection distance. Conductive paint can be applied to any non-uniform surface.
Molding pillars in the overlay
The latest generation of touch controllers from Azoteq provides another alternative for removing air-gaps. New controllers, such as IQS333 and IQS360, allow the designer to remove any air gaps with non-conductive materials that are less expensive. An example would be adding pillars at the touch positions when creating a molded part as overlay, as shown in Figure 3 .
In some applications, the spacing of the touch keys require the sense electrodes to be placed close together. In these applications, spring or rubber contacts are not the best option. The designer might need to use the more directional approach of projected capacitive sensing. Some applications require shielding between keys, which requires ground to be added between the self-capacitive electrodes to reduce the risk of multi-touching.
In these types of applications, the designer can remove the air gap by matching the curved surface. This is accomplished by fixing a flexible PCB (FPC) to the curved surface with double-sided adhesive tape. The designer may choose to create a complete FPC module, which includes the touch controller as well as the sense electrodes. Alternatively, only the sense electrodes could be placed on the FPC, which can connect to the controller through a cable connector or FPC tail.
Care should be taken when designing for 3D surfaces (curves in two or more directions – see Figure 5 ). Fitting FPC to 3D curved surfaces may require cutouts to fit smoothly without creating air gaps (more details to follow in the second article). In cases where there are large curves and a risk of air gaps appearing, the simplest option would be to use a fixed PCB with spring contacts.
An alternative is to keep the PCB area flat (to enable the use of normal flat PCBs with no air gap to the overlay), and only have the curve on the other side, where the user will interact. This could also ease the manufacturing process, especially for more difficult materials such as wood overlays. This does create a non-uniform overlay thickness, and can be compensated for in electrode design. Different electrode designs are discussed later in this article.
The designer would do well to remember that thin FR4 PCBs (0.1 mm to 0.2 mm) provide a more cost-effective solution to FPC in applications where the required curve in the PCB is not too large.
When placing components on flexible PCBs, Care must be taken to place them in areas that will not be bent. This will reduce the risk of component failure such as capacitor cracking.
A variation of standard FPC is to use ITO or PEDOT films for transparent overlays requiring transparent electrodes for backlighting purposes.
How to compensate for curved surfaces
In some applications, it is not always possible to keep a constant thickness in the overlay material. This may be due to injection molding of plastic parts to conform to a designed shape and stiffness, or the requirement for uniform backlighting to accommodate light diffusion layers.
Figure 6: Fitting a flat PCB to an overlay with a curved exterior creates a non-uniform thickness. Electrode layout can compensate for this.
Electrode shapes for thicker overlays
Although capacitivetouch sensors from Azoteq are fully customizable for sensitivitysettings on each channel, as well as having adjustable detectionthresholds for set sensitivity selections, the designer could save timein fine tuning the sensitivity of each electrode by compensating forvariations in the thickness of the overlay material in the electrodedesign. For projected capacitance, bigger gaps between the Tx and Rxelectrodes will give more sensitivity (but reduce stability), or usingthicker receivers in the electrode will allow for thicker overlays (butreduce conductive noise immunity). Figure 7 shows differentelectrode variations of the same design (for projected capacitivesensing) for different overlay thicknesses (in order of increasingsensitivity).
Figure7: Projected button layouts in order of increasing sensitivity. Redrepresents the Tx electrodes, yellow represents the Rx electrodes andwhite represents the PCB cut-outs. The dimensions of all three buttonsillustrated are the same.
For self-capacitive sensing,bigger electrodes will give more sensitivity, and avoiding sharpcorners that form concentrations in field lines will allow for thickeroverlays.
Figure 8 shows different electrode variationsof the same design (for self-capacitive sensing) for different overlaythicknesses (in order of increasing sensitivity).
Figure8: Self-capacitive buttons illustrated in order of increasingsensitivity by increasing the size of the electrode and removing sharpedges that form concentrations in field lines.
Insome applications the overlay is not fully fixed. This is possible inapplications with thin overlays, where a small air gap is possible. Thiscould result in movement of the overlay when excessive touch force isapplied. In these instances it is important to choose the correct touchcontroller. For example, using a device from Azoteq with Dycalcapability (for example IQS228) can avoid stuck conditions byrecalibrating the sense electrode upon touch release events (when theuser releases the overlay).
In some applications, the designer maywish the overlay to be moveable. An example of could be and applicationin which a second level of touch activation is required – the firstbeing a zero force touch trigger, with the second activation occurringwhen the button is pushed to depress a metal dome.
In this situation the designer can opt to place a second PCB for the touch electrode in the moving part (Figure 9 ).This electrode PCB could connect to the main PCB with an FPC tail,which is flexible enough to accommodate the movement. The second optionis to design the touch electrode to sense around the metal dome, asillustrated in Figure 10, instead of sitting on top of it.
Figure 11 showselectrode layouts with enough sensitivity to detect the zero forcetouch event before the overlay is depressed. This type of layout canalso be used with a thick rubber overlay, or with overlays that have thesnap dome molded into them.
Figure11: Electrode designs to accommodate touch sensing around a metal dome.(a) Projected capacitive electrode layout and (b) Self-capacitivelayout.
Some sensors, such as the IQS360, have asnap detection, enabling the designer to place small Tx and Rx extractsfrom the button (Figure 12 ) underneath the metal dome. Touch isdetected when the user touches the overlay, and the snap output isgenerated upon depressing the metal dome.
Figure12: Round projected button layout with Tx and Rx tails to providestability. Enough spacing between Tx and Rx allows sensing through thickoverlays and space on the inside for a metal dome to enable snap/clickfunctionality.
Connecting FPC/FFC to PCB
Inapplications where FPC is used for touch and proximity electrodes, thedesigner has multiple options to connect the FPC to the main PCB.
FFCcables can connect between the FPC electrode and the main PCB. Variousflex connectors are available, with different pin counts and pitches.Although convenient to use, flex connectors can be costly.
Amore cost-effective approach is to use ACF (anisotropic conductivefilm) thermal compression bonding. ACF is a mature assembly technologythat entails a process in which the FPC tail is soldered directly to thePCB with a combination of heat and pressure during the final assembly.
Alwyn Botha joined Azoteq in 2009, and is currently working as part of the applications team. Heholds a bachelors degree in Mechatronic Engineering and a Master’sdegree in Electronic Engineering from the University of Stellenbosch inSouth Africa.