Anyone can bake a cake, but while some chefs bake dry, uninspired bricks of dough, there are other chefs who make cakes we would die for. The ingredients may be the same, but the outcomes are so very different.

This is also the case between average electronic products and world-class, market-changing products. One of the most recent technical sensations is the capacitive touch screen. But what makes some touch screen-based products amazing, while others get such poor reviews?

This article explores key touchscreen performance parameters, critical touchscreen design features, significant design tradeoffs, and key issues product designers must consider when choosing their touchscreen supply chain. Don't get caught making an uninspired product; instead, create something people would "die for" to have!

Perhaps the single most significant technology change to affect the performance of today's touch screens has been the shift from resistive to capacitive touch screens. Industry analyst iSupply forecasts that nearly 25% of the mobile handsets with touch screens will have shifted from resistive to capacitive screens by 2011, while Jeffries and Co. has increased their projections for capacitive touch screens in 2010 from 100 million units (Mu) to 188 Mu. The market is exploding, in large part because of the benefits which capacitive-touch screen technology brings.

While traditional resistive touch panels detect a finger or stylus touch when a flexible top layer of clear material is pressed down to contact a lower conductive layer of material, projected capacitance screens have no moving parts. The projected capacitance sensing hardware consists of a glass top layer, followed by an array of X, Y, and insulating layers of indium tin oxide (ITO) on a glass substrate. (Some sensor suppliers create a single-layer sensor that includes both X and Y sensors in a single layer of ITO with small bridges where they cross.)

As a finger or other conductive object approaches the screen, it creates a capacitor between the sensors and the finger. This capacitor is small relative to the others in the system (about 0.5 pF out of 20 pF), but it is measurable using several techniques. For example, one technique used with Cypress Semiconductor's TrueTouch parts involves rapidly charging the capacitor and measuring the discharge time through a bleed resistor.

This all-glass touch surface gives the user a solid, smooth feel across the entire screen. Glass screens are preferred by customers because glass gives the end product a smooth industrial design and provides a good capacitive signal for measuring touch. In the end, however, it is not only how the touch panel looks, but also how it operates. For designers, gaining excellent performance from a touch screen entails first knowing several key parameters:

Accuracy: Defined as "the maximum position error across a pre-defined area of the touch screen, as measured in units of distance along a straight line between the actual finger position and the reported finger position." Accuracy is measured with a simulated or mechanical finger. The finger is placed at precise locations on the panel and actual finger position is compared with reported finger position.

The importance of accuracy cannot be overstated. Users want the system to properly locate their finger. One of the biggest frustrations with resistive touch screens is their low accuracy and accuracy loss over time. Capacitive touch screens' accuracy enables new applications like virtual keyboards and handwriting without a stylus.

As an example, Figure 1 shows a poorly constructed touch panel's data, which displays finger drift above the true straight line traveled by the simulated finger.

Figure 1: Example showing inaccuracy or error in touch panel tracking.
(Click on image to enlarge)

Finger Separation: Defined as "the minimum center-to-center distance between two fingers placed on the touchscreen while two separate fingers are still reported by the touch screen controller." Finger separation is measured by placing two simulated or mechanical fingers on the panel and moving them towards each other until they are reported as a single finger, Figure 2.

Some touch screen suppliers report finger separation as edge-to-edge, while others report it center to center. A 10 mm finger separation specification for 10 mm mechanical fingers could mean that the fingers are touching or that they are 10 mm apart, depending on your touch controller's method of specification.

Without good finger separation, you cannot implement reliable multi-touch functionality. Finger separation is especially important for virtual keyboards, where two fingers are commonly on the screen in close proximity.

Figure 2: Measuring finger separation
(Click on image to enlarge)

Response Time: Defined as "the time between a finger touchdown event on the touch screen and the touch screen controller generating an interrupt signal." This can be measured by electrically stimulating the touch screen to simulate a finger or by physically moving a simulated finger to the panel.

Response time is particularly important because it directly translates to how fast users can move their fingers on screen for a "swipe" or "flick" or to write with a finger or pen. A touch panel with a slow response time may look choppy or may miss a movement altogether. Touch screen response time is one component of system response time that includes:

• X/Y scanning: Time for the touch controller to scan and measure the change in capacitance on the sensor.
• Finger detection: Comparing the capacitance change on the panel to a predetermined "finger threshold". If the change is over the finger threshold, a finger has been detected.
• Finger location: Interpolating between the results from several sensors to determine the exact position of the finger.
• Finger tracking: When more than one finger is on the sensor, each finger must be identified and assigned a unique identifier.
• Interrupt latency: This is the delay between the interrupt indication and interrupt servicing on the host. In most systems, this delay is less than 100 μsec.
• Communication: Typical systems use I²C at 400 kHz, or SPI at 1 MHz, to communicate to the host.

Parallelism is another key tool. Scanning, finger processing, and communication all use separate hardware and can all run in parallel. Highly optimized algorithms for finger detection, finger location and finger ID all reduce processing time and reduce response time.

Refresh Rate: "The time between two consecutive frames of touchscreen data available in a data buffer while a finger is present on the touch screen". A low refresh rate will result in jerky movement and curves that appear to be made up of line segments rather than smooth curves.

Instead, if a touchpanel has a high refresh rate, it provides many more datapoints for interpretation of a smooth or complete shape or motion. A high refresh rate improves gesture interpretation as well. Smart touch screen controllers like Cypress Semiconductor's TrueTouch products can adjust their refresh rate to match the system's requirements. A drawing or handwriting application needs a fast refresh rate, but a mobile phone dialing keypad only needs to interrupt the host when buttons are pressed or released.

Average Power Consumption: "The average power in a touch system is comprised of time scanning, time processing, time communicating, and time sleeping for the controller IC and host processor receiving and interpreting touch data." Power consumption seems like an obvious performance parameter to capture: Measure the current used by the device, multiply by voltage and you know the power consumption.

In the world of touch panel power consumption, however, a more sophisticated model is needed because power consumption depends on usage. Phone "standby time" will depend on touchscreen "standby" or "deep sleep" current use. Even when the touch screen is active, it can be in several modes such as "wake on touch" (WOT), "touch", and "cheek detect". In a typical 5-minute call, your phone may be in touch mode for 10 seconds as you lookup or type in the phone number, then in WOT or cheek detect for the remainder of the call. Even sending a text (SMS) message is a mixture of WOT mode and actual finger contact, as the controller IC dips in and out of sleep modes while you are typing and thinking.

It's easy to be misled by system power promises if you don't take these power modes into account. In almost every case, the touch screen spends 90 to 99% of the time in "cheek detect" mode and "wake on touch" mode. Some systems allow customization of the ratio of processing time to sleep time even while a finger is on the panel. There is little need for a 200 Hz refresh rate if the system is only reporting "there is still a finger at the same location". To develop a high performing touchscreen, it is important to take advantage of systems with low sleep specifications along with creative sleep and wake modes.

There are several other important parameters that system designers must keep in mind when designing a capacitive touch screen system:

Finger Capacitance: The capacitance measured between a finger and a single sensor element. Finger capacitance is measured using a real finger rather than a metal finger to ensure real-world data. Factors that affect CF include cover lens thickness and the dielectric constant of the cover lens material.

System Noise Floor: The amount of noise measured at the output of the capacitance-to-digital-converter referred to the input (capacitance) of the data converter.

Signal to Noise Ratio (SNR): The finger signal measured on one sensor, divided by the observed measurement noise. While this is an important shorthand for the two measurements above, it must be understood that to create an effective touch panel, the system must be able to accommodate, adapt, and filter parasitic noise in a mobile system. In order to observe high signal counts and very low noise counts, an accurate analog front end should be considered for the touch function.

Figure 3: Signal to noise (SNR) example
(Click on image to enlarge)

Programmable solutions like Cypress Semiconductor's TrueTouch family provide excellent mechanisms for filtering noise. The PSoC programmable analog can be reconfigured to integrate signals over longer time periods to filter noise. Different signaling frequencies, including spread spectrum and pseudo-random frequencies, can be used to avoid EMI.

Standard digital filters can remove one to two bits of signal jitter, or provide a low-pass filter like an IIR (infinite impulse response) filter. Smart digital filters can discard samples that don't "look right" compared to the samples near them on the panel. Smart filters are limited only by the system designer's ingenuity. Figure 3 shows an example pattern of device noise floor and a registered touch. In this case, a SNR of 5 is registered.

Understanding and controlling key touchscreen performance metrics will lead to a significantly better performing touchscreen design. Understanding these metrics also allows you to select design partners who have technology which can accommodate the unknown noise and electrical issues in a mobile consumer product.

The beauty of touch screens is their seemingly simple design. Devoid of clunky buttons, sticky roller balls, or barely-readable screens, touch screens are an entirely new way of creating an enjoyable user experience. The difficulty with touch screens, however, is that to be able to deliver an elegant, simple design, you must use sophisticated hardware, firmware, and manufacturing techniques. Understanding the language of touch screens, key design parameters, key touchscreen performance parameters, and touchscreen design tradeoffs is the first step in building a world-class touch screen product

Related articles of interest
1. Using capacitive sensor user interfaces in next generation mobile and embedded consumer devices, Mariel Van Tatenhove and Andrew Hsu, Synaptics, Inc.
2. Designer's guide to rapid prototyping of capacitive sensors on any surface, Mark Lee, Cypress Semiconductor Corp.
3. Capacitive sensors can replace mechanical switches for touch control, Wayne Palmer, Analog Devices Inc.
4. Building a reliable capacitive-sensor interface, Wayne Palmer, Analog Devices, Inc.
5. The art of capacitive touch sensing, Mark Lee, Cypress Semiconductor Corp.
6. Practical considerations for capacitive touchscreen system design (Part 1 of 2), Yi Hang Wang, Cypress Semiconductor Corp.
7. Basics and implementation of capacitive proximity sensing (Part 2 of 2), Ganesh Raaja, Cypress Semiconductor Corp.
8. Touchscreens 101: Understanding touchscreen technology and design, Steve Kolokowsky and Trevor Davis, Cypress Semiconductor Corp.

About the authors
Steve Kolokowsky is a Member of the Technical Staff in Cypress Semiconductor Corp.'s Consumer and Computation Division (CCD). Steve's focus is capacitive touchscreen products using PSoC™ technology. Steve has a BS in Computer & Systems Engineering from Rensselaer Polytechnic Institute in Troy, NY. He is based in sunny San Diego, California, USA and can be reached at syk@cypress.com.

Trevor Davis is currently the Director of Marketing for Cypress's Consumer and Computation Division (CCD) focused on Universal Serial Bus (USB) in consumer products. Trevor received his undergraduate degree from the United States Air Force Academy and also holds his Masters in Business Administration. He served as an Air Force Officer for five years before joining Cypress, and lives in San Diego, CA and can be reached at tmd@cypress.com

.