Designing sense electrodes for non-uniform and curved 3D touchpad surfaces -

Designing sense electrodes for non-uniform and curved 3D touchpad surfaces

Over the last few years, touchpads have become a standard input method for next-generation user interfaces for everyday applications, from white goods to smart phones, computer peripherals, and remote controls. Touchpad design requires a thorough understanding of capacitive sensing, from electrical to mechanical stack, as the entire device and the user interaction with it need to be understood, especially for hand-held or battery-operated devices. It can be challenging to add touchpads to designs that are not flat or are made with three-dimensional shapes and multiple parts.

Even vs. uneven surfaces
A touchpad is a collection of individual touch buttons arranged in rows and columns, called channels. As the user moves his or her finger over these channels, a delta is sensed in the measured count values and can be used to determine the XY touch co-ordinate.

Traditionally, touchpads are flat with a uniform overlay thickness. One of the main reasons for this is that, with a uniform pattern and a uniform thickness, the touch strength when the user’s finger touches the surface will be uniform over the entire touchpad. If, however, the overlay is not of uniform thickness, the mutual capacitance and therefore the touch deltas will no longer be uniform.
This article explains the most important design choices that need to be made when designing a touchpad for a 3D surface.

Selecting the substrate
The choice of substrate is no longer limited only to flexible circuit boards (FPCs). Depending on the overlay shape and the interconnects to the main board, the designer also has the option of an FR4 PCB. The high sensitivity of the Azoteq Trackpad ICs allows the designer to place the touchpad or slider on the main PCB. This would lower the cost of the solution significantly, as the additional FPC and connector would no longer be needed. However, there are a few factors that must be considered by the designer when selecting the substrate, and these are described in the following sections.

Table 1
below is a list of dielectric constants/permittivity for materials that are commonly used for touchpads.

Table 1: Material properties (estimated for reference purposes only – actual values to be inserted by the designer)

Equation 1 below shows the parallel plate capacitance 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.

Equation 1:    

Overlay considerations. The overlay, which is on top of the sensor pattern, is the user interaction area. It is a critical component of touchpad design and determines the design direction. Overlays can be of various shapes such as round, square, rectangular, etc., and can have thicknesses ranging from 0.2mm to 5mm.

Touchpad and overlay types include:

  • Multi-tier touchpads and touchpads with finger guides
  • Multi-tier touchpads with either one or both tiers moving
  • Curve in one direction – cylindrical or cone-shaped designs
  • Curves in both x and y for a 3D shape – concave or convex designs

Manufacturing limits
Mechanical manufacturing limits need to be taken into consideration for overlay design. In the case of plastic injection molding, there is a limit to the thickness that can be manufactured without resulting in uneven shrinkage. Warping is caused by uneven shrinkage due to the variations in thickness of the part, resulting in cooling at different rates. This warping could lead to an uneven surface between the touchpad and the underside of the overlay, resulting in air gaps between the two layers. Air, with its low permittivity, would reduce the sensitivity significantly in these areas. Determining shrinkage ratios for specific plastics is beyond the scope of this article.

Channel pitch. The pitch [distance between channels or the channel size] of the pattern used must be larger than the overlay thickness, otherwise the E-fields would tend to blur and it would not be possible to determine the touch peaks accurately. This would lead to an increased linearity error.


Mutual and parasitic capacitance (CM and CP)

The mutual capacitance between the sensing (Rx) and driving (Tx) electrode is the sum of all the individual components (Equation 2 ). When the user touches the touchpad, the CM_touch component is lowered and approaches 0 when saturated by the touch.

Equation 2

A simplified single channel is shown in Figures 1 and 2 .

Figure 1: Side view of a channel

Figure 2: Top view of a channel

A simplified equivalent circuit is shown in Figure 3 , with the transmitter (Tx) driving the mutual capacitance and a receiver (Rx) sensing the mutual capacitance.

Figure 3: Simplified equivalent circuit of a single channel

Trace routing considerations. Trace routing for touchpads on Flex PCB (Figure 4 )is critical for minimizing parasitic capacitance and maximizing thetouch delta. For rigid FR4 PCBs, this limitation can be relaxed,however, as the two copper layers are sufficiently far apart and thecapacitance is small enough in relation to the mutual capacitance.

Figure 4: An Rx trace running underneath a Tx pad can add a significant amount of unwanted parasitic capacitance to the channel

Atheoretical comparison of the parallel plate capacitance of two4mm-long tracks of 0.1mm and 0.2mm thicknesses with different basematerial thicknesses is shown in Figure 5 .

The capacitanceincreases greatly for thin base material thicknesses, and each channelthe track runs under adds additional unwanted parasitic capacitance(CP).

It also is necessary to take the RC time constant intoaccount for the driving electrode; this is especially important for ITOand PEDOT touchpads, as the resistance on the traces is large.

Figure 5: How the capacitance increases when the base material thickness decreases

Interconnections. Hot bar soldering and ACF (anisotropic conductive film) bonding aremature technologies used to connect FPC to FR4 PCBs, while ZIFconnectors are also used for the same purpose. If two separate touchpadareas are used and driven with the same IC, then in order to lower theparasitic capacitance the connecting cable must be kept as short aspossible, with Tx and Rx tracks separated with a ground track. Multiplecables are not advised, as the probability of a defective solder jointor connection increases with each interconnect.

Long-term average changes
Whenthe user brings his or her finger close to the touchpad, the mutualcapacitance decreases, and if the touch threshold is chosen correctly, atouch event will be triggered when the user touches the overlay. Thechange in these counts above the long-term average (LTA) is referred toas the ‘delta’.
The deltas of all the touched channels form a touchimage, and the weighted sum is taken to calculate the XY touchco-ordinate (the sum of all the deltas would give the touch strength).If, however, the sensitivity of each channel were not equal, such whenone channel has a lower sensitivity than its adjacent channel, the XYtouch point would be distorted and have a position error relative to theamount of distortion.

At the extreme, if the sensitivitydifferential is large enough, a split touch will be detected (a touch oneither side of the channel that was actually touched) when only oneshould be detected. This could be compensated for to some extent withpost-processing and track pad characterization.

However, this isnot always ideal as it will be dependent on finger size. Anotheralternative would be to use the ProxSense engine compensation settingsto increase the sensitivity of the problem channel. Different overlaythicknesses would also produce a different touch delta in the thin andthick areas.

Figure 6 is a representative image of theE-field lines that can couple with the user’s finger, with the touchdelta being larger in the areas with a thin overlay thickness, ?1>?2.

Figure6: FR4 touchpad PCB mounted behind a curve overlay of non-uniformthickness, showing the E-fields available for the user to interact with

Overlay thickness changes. As the overlay thickness decreases, the mutual capacitance increasesand so does the capacitance available to be removed by a touch,resulting in higher touch strength. For a variable thickness overlay,this touch strength or delta will need to be scaled so that thesame-sized touch on each area will produce the same touch strength. Thisis referred to as mutual capacitance balancing. (Azoteq designers willnormally assist designers with this step.)

Touchpad sensitivity
Theamount of mutual and parasitic capacitance can vary greatly over thetouchpad area as a result of the variable thickness overlay, dielectricpermittivity, track routing, and pattern design. The difference inparasitic and mutual capacitance directly influences the sensitivity anduniformity of touch detection by the touchpad.

The sensitivitycan be compensated for either with routing or capacitive loading in thehardware design, or with tuning of compensation settings with firmware.Variables such as parasitic cancelation (PCC) and ATI target can be usedfor each channel to achieve uniform sensitivity to tune the target andbaseline as needed.

Equation 3

Theabove relationship shows that the areas with a lower baseline wouldhave higher parasitic capacitance. In this case the sensitivity of thosechannels would need to be made higher by increasing the amount ofcompensation. Adjusting the sensitivity in the thick areas to producethe same touch delta can be achieved by increasing the PCC or loweringthe ATI base. As the capacitance over the channel decreases, the countvalues increase. Thus the mutual capacitance is inversely proportionalto the capacitive count values.

Equation 4

whereCxeff is the effective capacitance and CSx is the count values for thatchannel. Increasing the compensation also improves the SNR.

ForFPC touchpads, the parasitic capacitance from routing traces has ahigher impact on the channel’s sensitivity, whereas with FR4 touchpadsthese layout constraints on trace routing can be relaxed. However, withoverlays of variable thickness, the amount of mutual capacitance thatcan be removed during a touch would be different between the thin andthick areas, as described earlier.

The same overlay profile is shown in Figure 7 , but with a uniform thickness and an FPC touchpad design.

Figure 7: FPC touchpad stuck to the underside of the curved overlay of uniform thickness

Theoretically,the above profile would have a similar sensitivity range over thetouchpad area and result in similar touch deltas. However, with thisprofile the FPC would need to conform to the 3D shape and, in order toachieve this, relief cuts would need to be placed in certain areas. Therelief cuts are needed to prevent any surface mismatch to ensure that noair bubbles occur between the FPC and the underside of the overlay.These cuts should also be long enough so that no delimitation occursover the lifetime of the touchpad.

Remember to keep the minimum bend radius of the FPC in mind to avoid folding and breaking the tracks.

Therelief cuts should be placed between channels, as this is already thearea of lower sensitivity, while maintaining the sensitive channelcenter. The tracks that need to be routed around the cut need to beshielded to prevent unwanted user coupling; this can be a ground trackplaced on the top layer of the FPC. Cuts can also be placed in thechannel intersections, but this would result in the peak beingflattened, creating an area of lower sensitivity.

Johnathan Loy holds a bachelor’s degree in Electrical and Electronic Engineering withComputer Science from the University of Stellenbosch in South Africa.Johnathan joined Azoteq in 2012 and specializes in touchpads for consumer electronics.

1. AZD068 Trackpad Design Guide Application Note
2. IQS5xx-A000 Datasheet
3. IQS333 Datasheet
4. US Patent WO2010045662A9: Parasitic capacitance cancellation in capacitive measurement applications 

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.