Kranthi Kiran's

What is a capacitor and how does it work? When you go to a showroom and watch some plasma panels maybe you do not know that you are watching some capacitors. Yes, a plasma panel can be considered a capacitor.
The capacitor is a device able to store electric energy. Practically every time two conductor materials (called "plates") will be near and separate from a non conductor material we will have a capacitor. In a PDP (Plasma Display Panel) the plates are the two glasses (front and rear panel) and the non conductor material is the dielectric material that is between them. If we apply a voltage to a capacitor, it will charge at the same power supply potential. In a capacitor the process of storing energy is called "charging" and it involves electric charges of equal magnitude but opposite polarity. Initially, at flat capacitor, the plates are electrically neutral since we have the same numbers of electrons and protons on them.

If we connect the capacitor to a continuous voltage generator, it happens that the generator positive pole catches the electrons from the connected plate while the other plate catches electrons from the generator negative pole (see below). Little by little the capacitor increases its charge, the potential of the plate connected to the positive pole increases and comes close to the generator potential. So the potential difference, at the two resistance ends, decreases and the current intensity decreases also. Therefore the capacitor charges quickly initially, then more and more slowly. Once the capacitor is charged (at the same generator voltage value) we will have a positive potential on the plate A (positive charges predominance) and a negative potential on the plate B (negative charges predominance). The voltage (charge) remains even if we disconnect the capacitor from the generator (no losses, in the case of ideal capacitor of course).
Below there are the voltage and current progress graph and their formulas. The product RC will be explained later.
 
If we connect the plates through a resistance we will have the capacitor discharge, that is we will have an electric charge equilibration and the voltage decreases to zero value (the energy stored in the capacitor dissipates in the resistance). During the discharge process the current direction (that is the electrons movement direction) is opposite to the charge process one.
For the charge, during the discharge process, we will have the following equation:
Q = Q₀ e^–t/RC
where Q is the capacitor charge (Coulomb), Q0 is the charge at the start, "e" is the exponential number (Euler's number =2.718..), t is the time (Seconds), C is the capacitance (Farad), R the resistance (Ohm).
For voltage and current the equation becomes:
Below there is the current progress graph and its equation.
Let's explain briefly the product RC.
The capacitor charge or discharge happens in a time depending from the resistance value (in a series to the capacitor) and from the capacitance value of the capacitor. Laboratory tests have shown that the needed time to charge the capacitor at 63% of the applied voltage is equal to the product result between resistance and capacitance. The product result is called time constant (t), so
t = R * C,
where t is expressed in Seconds, R in Ohm and C in Farad.
Moreover it has been demonstrated that the capacitor is charged in a time T = 5 t because after the first t it charges 63% of the applied voltage and after every other t it charges a further 63%, but of the remaining difference.
The aptitude at the electric energy storing is called capacitance: it is directly proportional to the one plate surface (A) and inversely proportional to their distance (d) and depends, in directly proportional manner, from the relative static permittivity value of the used insulator εr. The formula is
C = ε_r * ε₀ * A/d
where ε0 is the vacuum permittivity, the measure unit is the Farad (F).
The insulator placed between the plates is called dielectric and it can be liquid, solid or gaseous. The dielectric type allows a first capacitors classification. The most used capacitors, in the electronic area, are the ones with air or solid dielectric. The most used types of solid dielectric are: mica, ceramic, plastic film, paper. The capacitance value of a capacitor is clearly showed on the capacitor body (for the big ones) or codified by different codes (colours or alphanumeric). Now let's have a look to some capacitor types, at their features and application areas.
Electrolytic capacitors
The electrolytic capacitors are formed from two metallic sheets, cylindrically wrapped, that are separated by a thin oxide layer (got through an electrolytic process). The very thin layer thickness (approx. 0,001 µm) and its relative static permittivity value, relatively high, allow to get huge capacitances values (until 1.000.000 of µFarad in the aluminium electrolytic capacitors) even if they can suffer a potential difference of a few ten volts only. Due to their structure they are polarized, that is they must observe a polarity verse: one plate must be always positive, the other one must be always negative. Changing the polarity direction is very dangerous: the capacitor could explode.
As we said above they have big capacities, so they can accumulate a large energy quantity. For this reason they are used, mainly, in the power supply units, for the voltage levelling and for the ripple reduction.


Ceramic capacitors
The ceramic capacitors are constituted from a sandwich of conductor sheets alternated with ceramic material. In these capacitors the dielectric material is a ceramic agglomerate whose relative static permittivity value can be changed from 10 to 10.000 by dedicated compositions. The ceramic capacitors, with low relative static permittivity value, have a stable capacitive value and very low losses, so they are preferred in the floating and high precision circuits. The ones with high relative static permittivity value allow to get high capacities occupying a small space. Generally the ceramic capacitors have small dimensions and they are preferred in the high frequencies area. The most used ceramic capacitor shape is the disc one, that is a little ceramic disc metalized on both sides and with the extremities welded on them. Typically they have very small capacities, from some pF to some nF, and they can suffer big potential differences.


Paper capacitors
In the paper capacitors the dielectric material is constituted from a special paper saturated with a fluid or viscous substance. To increase the insulation, in these capacitors, often two or more layers are coupled. The finished envelopment is again saturated under vacuum with insulating oil or is dipped in the resin. Generally they are used as filter capacitors.

Plastic film capacitors
The membranes in plastic film can be produced with lower thickness than the saturated paper and are more uniform. So there are capacitors that use these membranes as dielectric material (a few µm of thickness only) and they can suffer high voltages. The plastic film capacitors are mainly used in the transistor circuits. In the polyester capacitors a metallic sheet is used as electro-conductor layer or the metal can be deposited directly on the film by under vacuum vaporization, with a layer thickness of 0,02 - 0,05 µm. The capacitance of these capacitor can reach some µF. They are used in the low frequency circuits mainly.

Tantalum capacitors
The tantalum capacitors, as the electrolytic ones, are polarized, but they have the tantalum pentoxide as dielectric material. Compared to the electrolytic ones, they are better both the temperature stability and high frequencies, but they cannot suffer over-voltage peaks and can be damaged, sometimes exploding with violence. On the other hand they are more expensive and they have much lower capacity.

Niobium capacitors
The tantalum capacitors have two drawbacks: the tantalum cost due to this material rarity and its susceptibility to certain low level ppm of thermal runaway failures. Because of the increasing demand for tantalum capacitors a new technology has been developed and the niobium capacitors have been launched into the market. With at least 100 times more deposits than tantalum, the niobium guarantees good availability and lower price. So the niobium capacitors are very similar to the tantalum ones, but they have low cost, surge robustness and it is raising the conviction they can have better performances in other fields like voltage range, ESR and miniaturisation.

GNSS (Global Navigation Satellite System) is a common name for all of the satellite based positioning systems, which are GPS (Global Positioning System) from US, Galileo from EU, GLONASS from Russia, and CNSS (Compass Navigation Satellite System) from China. GPS is the first and most popular one among these systems.
In US, FCC ordered all the network operators to follow E911 act(Enhanced 911) to offer the physical address of calling party who is calling 911 emergency call. More and more cellular phones have armed with GPS feature. In the same time, the navigation devices are popular due to fast growing of private cars and mobile phones in the emerging countries. Thanks to Google, more and more consumers can easily connect their GPS devices to the Google web services for navigation, virtual sight viewing or satisfying their curiosity. All of these services are available free of charge.
Google's Inspiration
Google is a great web innovator. Everybody knows about Google Map and Google Earth. And its competitors like Microsoft and Yahoo have to catch up. However Google is not the inventor of the web GIS. Actually web GIS has been available for a long time. Never the less, Google promotes the web GIS with its great influence in the Internet, and furthermore deliveries the free services in a quick and elegant way (AJAX). More and more companies and developers have identified the business opportunities by integrating the existing navigation technologies and web GIS. The new successful stories spread the world and gain the attentions of the venture capitals. As a result, the GPS ecosystem becomes highly competitive and exciting.
Competitive Market
A successful GPS application is made up of GPS terminals, map data services and service centers. That means the GPS applications are blending businesses involve Internet, mobile terminal, mobile network, automotives, and consumer electronics. More and more industries are looking for the new business opportunities in the navigation and location aware services via merging and acquisitions. It is a clear trend that the map data and services are the key factors of a business success. As usual, the silicon suppliers and device manufacturers have to fight for the market share and making devices cheaper. The startup companies must release products with unique features. Some suppliers offer dual mode or tri-mode satellite positioning chipsets for GPS, Galileo and CNSS. Some independent RFIC vendors team up with the software suppliers to promote the software GPS solutions in reduced BOM cost. Some other vendors are promoting the one chip RFIC for all RF features including Bluetooth, FM radio and GPS.
GPS Receiver Architecture
GPS works by making one way range measurements from the receiver to the satellites. In order to arrive at a position fix we must know precisely where the satellites are and how far we are from them. These data are available to the receiver by reading the data message from each satellite which provides a precise description of the satellite orbit and timing information which is used to determine when the signal was transmitted by the satellite. Each satellite transmits on 2 frequencies in the L band (L1=1575.42 MHz and L2=1227.6 MHz). Each satellite transmits a unique CDMA (Code Division Multiple Access) code on these frequencies. On top of this the signal is modulated with a 50 Hz data message which provides precise timing information and orbital parameters. Since the receiver knows which sequence is assigned to each satellite it knows what satellite the data is coming from. The receiver creates a copy of the sequence and correlates or integrates the received signal multiplied by this copy over a period of time (in our case 1 ms). The particular sequence transmitted by each satellite has been chosen to reduce the chance that a receiver will track a satellite transmitting a different PRN sequence. For more detail on correlators see the Zarlink chipset documentation or some of the other references.
The attached figure is a traditional GPS receiver architect (from Zarlink). It is made up of antenna, RF/IF section and a base band processing unit, which usually has correlators and an embedded processor. The host processor talks to the embedded processor in an industrial standard protocol called NMEA (National Marine Electronics Association) or optional proprietary protocols. The physical links between the processors might be a standard UART, USB or Bluetooth. The communication over USB and Bluetooth has to simulate a virtual serial port to talk with high level application software. The default baud rate of NMEA is 4800bps, the higher rate doesn't make sense.
Antenna
Because of miniaturization and multi-functional requirements, the designer faces to more and more challenges in antenna design. These factors include human interference, noises from embedded processors and external interferences. It is better to copy the reference design from the application notes for an inexperienced engineer. Never the less, GPS antenna design is still easier than the mobile phone antenna design. We know, the latest mobile phone has to work on 800MHz, 900MHz and 1800MHz with PA and faces to the big noises inside.
The most commonly used antennas in GPS are the Helix and the patch antenna. Patch antenna has strong direction selectivity, which is used in most of the external GPS mice. The Helix antenna is much suitable for handheld GPS, which offers broader antenna angle, and it works better than patch antenna when it is close to human.
There are some off-the-shelf antennas available in the market. Most of them are external antennas, which offer better performance.
Sarantel offers GPS antenna in full Balun design, which offers 360 degree antenna reception and highly frequency selectivity, and most of the noises can be eliminated. This company also offers the bulk ceramic antenna as the smallest antenna in the world.
Mr. Mark Kesauer offers an inexpensive external GPS antenna design on Circuit Cellar. The PDF document is available on here. This design uses commonly available components and materials.
RFIC
The RF parts of a GPS from different suppliers are slightly different but most of these ICs are sharing same concept. The RF section includes LNA, filter, PLL and BPSK demodulator. Maxim’s MAX2769 demonstrates the general RF IC for GPS receiver.
The RF front-end of a GPS receiver first amplifies the weak incoming signal with a low-noise amplifier (LNA), and then downconverts the signal to a low intermediate frequency (IF) of approximately 4MHz. This downconversion is accomplished by mixing the input RF signal with the local oscillator signal using one or two mixers. The resulting analog IF signal is converted to a digital IF signal by the analog-to-digital converter (ADC).
The MAX2769 integrates all these functions (LNA, mixer, and ADC), thus significantly reducing the development time for applications. The device offers a choice of two LNAs: one LNA features a very-low, 0.9dB noise figure, 19dB of gain, and -1dBm IP3, for use with passive antennas; the other LNA has a 1.5dB noise figure with slightly lower gain and power consumption, and a slightly higher IP3, for use with an active antenna.
There is a provision for external filtering at RF after the amplifier. The signal is then downconverted directly using the integrated 20-bit, sigma-delta, fractional-N frequency synthesizer together with a 15-bit integer divider to achieve virtually any desired IF between zero and 12MHz. A wide selection of possible IF filtering choices accommodates different schemes, such as those of Galileo.
The overall gain from RF input to IF output can be tuned or automatically controlled over a 60dB to 115dB range. The output can be chosen as analog, CMOS, or limited differential. The internal ADC has a selectable output of one to three bits. The integrated reference oscillator enables operation with either a crystal or a temperature-compensated crystal oscillator (TCXO), and any input reference frequency from 8MHz to 44MHz can be used.
Correlators
The correlators of GPS are the essential parts of the whole system of bit synchronization and decoding. The correlators will feed the raw digit output to the embedded processor to acquire, confirm, pull-in, track the satellites, and translate into NMEA protocol, which the host controller can understand.
The correlators can be implemented in hardware and software. Recently, the number of correlators increases dramatically. The early product from Zarlink has 12 channel correlators. The newer SiRF-II has 1920 correlators inside, and the latest SiRF-III has over equivalent 200K correlators to reduce the TTFF. MediaTek (MTK) Taiwan also released a low cost GPS chip, which has 32 channel correlators inside. I use a MTK based GPS for my own testing purpose. It works fine, although I still expect more accuracy from the device. However, its performance is good enough as a consumer class GPS receiver.
Some open source projects released the FPGA based correlators. On the other hand, the correlators can be implemented in a FFT based software algorithm, which is referred as software GPS and cited in an application note from Maxim.
Embedded Processor
The silicon suppliers are trying to promote their own platforms in GPS. The task for the embedded processor is calculation and tracking the different satellites and interfacing with host processor in NMEA. If you check the attached NMEA document, you will realize that the embedded processor has to deal with so many parameters in detail. The requirements for the embedded processors are big enough memory address space and sufficient processing power for intensive calculation. The ARM7TDMI is a 32bit core, which offers sufficient memory space and processing power. The peripherals included UART, USB and Bluetooth have been available for ARM for a long time. As a result, the latest GPS chips from different suppliers have the identical trend to select ARM7TDMI as the embedded processor.
There are some key KPIs for the GPS, which might be related to the correlators and software in the embedded processor.

1. Cold Start: A cold start results when there is no valid Almanac or Ephemeris information available for the satellite constellation in SRAM, or when the time and/or position information is NOT known (i.e. starts at 0 in both cases). Also a cold start will be initiated if an Almanac is valid, but a fix cannot be achieved within 10 minutes of power-up. This could occur if the receiver position has moved significantly since it was last powered-up, but the position change and time are NOT initialized by the user
2. Warm Start: A warm start results when there is a valid Almanac, and the initial time and position are known in SRAM, but the ephemeris is NOT valid (i.e. more than 4 - 6 hours old).
3. Hot Start: A hot start results when there is a valid Almanac, valid ephemeris (i.e. less than 4 - 6 hours old), and when accurate time and position information are also known in SRAM (position error less than 100km, time error less than 5 minutes).

Software GPS
If the hardware can be implemented in software, the total cost of a GPS can be reduced. Different suppliers have different approaches. One solution is merging the embedded processor to the host processor, so the job is done in the host. In this solution, the basic hardware blocks such as correlators are still kept. Some companies call this approach as accelerated software GPS. The other solution is taking the digitalized signal from IF, and implements the correlators and decoders in software. We can call it as full software GPS. The software GPS is only available in commercial licenses on specific chip in linkable library. Sometime the software GPS license might be more expensive than a low cost IC.
SiRF has acquired the Centrality Atlas, who offers SoC for GPS. Its Atlas is a software GPS product, running on a 300MHz ARM microprocessor and 200MHz DSP. This chip is the best seller in automotive navigator. It offers the comparable performance with SiRF-III, with lower price and media player features.
NXP software also offers Spot GPS software for the host application processors. The Spot GPS Software is a commercial software package in the form of ANSI C. It is easy to be deployed since most of the latest smart phones have a 200MHz, even 500MHz processor inside.
There is also an open source GPS project called GPS world, which uses ATMEGA32 for back-end processing. But this project is not a complete software GPS solution, because it was built upon a hardware correlators IC from Zarlink. However if you are developing the firmware of a GPS receiver, it could be the base for your development. In the reference of open source projects, you can find other software GPS designs.
Modules or DIY
Because GPS is quite sensitive to the environment, inexperienced designs will ruin the whole project. According to the field report, the GPS module has quite high failure rate in production site. Some mobile phone manufacturers tried to design the GPS by themselves, finally they found their GPS phone design is totally a fiasco.
The module includes everything and assembly in a can module. The size is as small as a coin. There are many professional GPS module suppliers. LeadTek, Holux and other suppliers are offering SiRF and MTK based solution worldwide.
Application Software
From the point of system software development, the software engineer can consider GPS unit as a standard serial port. No matter which OS is selected, the serial port is always available, either in a real RS232 or simulated serial port on USB and Bluetooth. Any person who has experience of developing UART can develop the GPS application software in NMEA protocol as well.
Finally, the high level application software will combine the map data and coordinates from GPS and present to the users.
There is an open source project called Openmoko, which is sponsored by FIC Taiwan. This project is basically an ARM920T (S3C2440) microprocessor based mobile phone. The GSM/GPRS module and GPS module are connected to the serial ports for ARM920. The phone is working like a desktop PC with GPRS modem and GPS receiver. Anyway, it is a good project which you can start up your own GPS terminals.
The connected GPS project Dash Express is a derivated project from Openmoko.
There are also many open source projects available. You always can find the projects in your favorite languages, Java, Python, C++, .NET and even in web programming languages.

The touch screen technology is widely used in PDA, smart phone, PMP, ATM, information kiosk and many other types of equipment in industrial, medical and commercial environment. Actually the technology enabling these devices is not new, since it was invented by Dr. Samuel C. Hurst in 1971. But it becomes hotter after the release of popular iPhone and iPod touch. With new patents filed for the touch screen technology, Apple brings a new wave to this mature segment and more companies are involved in this revolution with improved interactive UI, ICs, assembly modules and software components.
Conventional Touch Screen
The implementation of a touch screen includes resistive, capacitive, infrared, surface acoustic wave, optical imaging, acoustic pulse, and other emerging technologies. The kit usually is made up of a touch module, and a controller, which measures the touch events in frequency, voltage or current, encodes and transmits to the drivers running in host controller. Different touch screen technologies are suitable for different environments. The selection criteria are:

1. - Light Transmission
2. - Response Time
3. - Touch Accuracy
4. - Environment Requirement
5. - Lifecycle
6. - Surface Hardness
7. - Resolution
8. - Input Mode (Bare or gloved finger, styles, pen)
9. - Display Size (Small, Middle or Large Screen)
10. - Multi-touch

Currently most of the portable devices are using resistive touch screen modules, because this technology has perfect balance between low cost and required performance. Both stylus and finger operation are supported. With the resistive touch screen, most of the technical innovations take place on the GUI system in the host. For example, HTC S1 released a new UI called touchFLO. This new technology works efficiently with single-touch screen. For example, the clockwise and counterclockwise finger event on specific screen area will zoom in/out this part of picture or web page. Sometimes it is more convenient because the user just need one hand to hold, and the thumb of same hand to operate. The iPhone requires both hands to operate on-the-go anyway. Yes, even the user can use two fingers of one hand to operate, but he/she should use the other hand to hold it, unless the device can be installed somewhere. Besides, HTC S1 supports handwriting and virtual keyboard, iPhone uses virtual keyboard. Typing European characters on the virtual keyboard might be a good idea, but not for Asian languages. Some other human machine interactivity research organizations also invented another new symbolic operation UI system with single-touch screen. For example, by handwriting an "h" to represent home page, the computer will show up the index for whole system. This method is deployed in an Automotive PC platform (VIA x86 design), it is very effective and attractive. The driver can handwrite on the screen while looking at the road without staring at the screen and touching some buttons. Therefore this UI improved the safety on the road.
I still remember the replacement of B/W LCD on mobile phones. The color LCD is more expensive and wasting more power, the engineers are arguing about if the idea of using color LCD in mobile phone is stupid or not. The sales of mobile phones with B/W LCD is dropping so quickly, now over 90% mobile phones are equipped with color LCD, in bigger size and dual LCD configuration. This story tells us that consumer electronics' product design is marker driven, instead of technology driven. So my conclusion is, even if single-touch screen is fully capable of every operation with some advantages, the multi-touch screen is a major trend of new touch-screen devices, because the consumers love it. Let us check it out and find resources to implement it by ourselves.


If anyone is just trying to clone Apple's design, he will be disappointed. The multi-touch is a system design, not just a small improvement on touch screen itself. Apple's initiative can not be cloned and it is protected by more than 200 patents. I found that Mezu was going to launch its Mini-One on CeBit 2008, but finally it was enforced to be close down on that exhibition because of another MP3 infringement issue. And the Mezu Mini-One does not support multi-touch as well. I am surprised that Apple didn't comment on the design issue for Meizu's MiniOne officially, even they seem like twins. (Please check left picture. ) Ironically Meizu issued a design patent in China. A clone design asks to be protected from being cloned again? I really hope they can use their resource on other proper projects.
Multitouch
Traditional resistive and capacitive can only support one touch hot area, i.e. one time to track one touch event. If two fingers are put on touch screen or touch pad, the result returns to host is the position of the last touch or somewhere between these fingers. In order to setup multiple touch hot areas, the designer must upgrade all of the three parts of the touch screen sub-system. That means innovation should take place on panel input modules, panel controller (ASIC or MCU) and device drivers in host, of course the application running the host should support multi-touch as well. The capability of touch screen controller limits the tracking number of touch events. I am not so sure about the application software support. The application software should track multi-touch events by themselves in some operations, for example, multiple objects should be tracked individually in a multi-user game. In another case, if the multi-touch can be translated into high-level zoom command, or rotate command, the regular software should be able to support multi-touch operation already. In that case, the device driver should support multi-touch event and translate into high-level commands. It is up to system architect to design the whole software stack.
IBM and Logitech have already issued some patents for multi-touch technology. Apple's iPhone is first consumer products deployed multi-touch. So far Apple has filed two patents for multi-touch, one is self-capacitance, and the other is mutual capacitance. These technologies have a new name as projection capacitive. The multi-touch has already deployed in iPhone, iPod, Mac Book Air. The implementation is described on some patent search sites (Please check reference). More competitors like LG, Samsung, and Microsoft are going to release new models and new operation systems with improved touch screen modules and drivers. Even MTK, a well-known supplier for OEM mobile phones, is going to improve their reference design of touch screen and try to mimic the operation of iPhone. Besides projection capacitive touch, there are resistive and surface capacitive implementation for multi-touch.
Input Modules

 
 
 
 
 
 
 
 
 
 
 
 
To allow people to operate with multiple fingers, the iPhone uses a new arrangement of existing technology. It includes a layer of capacitive material, just like many other touch screens. However, the iPhone's capacitors are arranged according to a coordinate system. The coordinate system does not require very high precision, because it is finger based operation. But I have not idea about the detail accuracy of it. Anyway its accuracy can not support handwriting. Its circuitry can sense changes at each point along the grid. In other words, every point on the grid generates its own signal when touched and relays that signal to the iPhone's processor. This allows it to determine the location and movement on the capacitive material. It won't work if you use styles or wear non-conductive gloves.
In mutual capacitance, the capacitive circuitry requires two distinct layers of material. One houses driving lines, which carry current and other houses sensing lines, which detect the current at nodes. Self capacitance uses one layer of individual electrodes connected with capacitance sensing circuitry. Both of these possible setups send touch data as electrical impulses.
After all, Apple's projection capacitive touch screen works like a big grid of keyboards. Each key is driven by row and column pluses, and encoded with its location and address. That working method and programming skills are very common in microcontrollers. The improvement are materials, these keys are almost transparent (Apple calls them transparent electrodes) and on a single film (maybe ITO film or ITO glass) can be produced by advanced process. Because the surface could be glass, so you will find iPhone's glass is quite hard and anti-scratch. According to Unwired View, iPhone also use force-sensing mechanism to filter the touch events by accident.

Actually Apple's iPhone only works dual touch event, it can support more events if Apple improves the capability of the controller, driver and application software.

Because its nature similarity with LCD drivers, some LCD panel manufacturers are integrating these feature into LCD modules. AUO released in-cell multi-touch LCD modules (480*272) on Oct, 2007. More and more LCD with single-/multi-couth screen will be available. I personally think AUO’s approach maybe more competitive because of the overall BOM cost and performance. Neither AUO in-cell and Apple projection capacitive multi-touch screen have large size LCD deployment, because capacitive is very sensitive to EMC noise, while large size HDTV LCD itself is a big noise generator!
Controllers
Because capacitive touch screen is very sensitive with environment, even with the innovation on input module, multi-touch requires more improvement on controller to work properly. Maybe that is another reason why Apple keeps the method back cover on iPhone/iPod?
According to report from howtostuffs.com, the iPhone's processor and software are central to correctly interpreting input from the touch-screen. There are many processors in iPhone, Infineon GSM processor, Samsung ARM11 as the application processor, and a dedicated screen controller BCM5974 from Broadcom. I can not find any information from Broadcom site, I guess it is a custom chip for Apple. So I checked other products, I guess this chip is a DAC as touch screen digitizer. If so, then the DAC is used to sample the grid instead of switching method. It also means ARM11 acts as the processing for multi-touch. The processor uses software to interpret raw data as commands and gestures.

1. Signals travel from the touch screen to the processor as electrical impulses.
2. The processor uses software to analyze the data and determine the features of each touch. This includes size, shape and location of the affected area on the screen. If necessary, the processor arranges touches with similar features into groups. If you move your finger, the processor calculates the difference between the starting point and ending point of your touch.
3. The processor uses its gesture-interpretation software to determine which gesture you made. It combines your physical movement with information about which application you were using and what the application was doing when you touched the screen.
4. The processor relays your instructions to the program in use. If necessary, it also sends commands to the iPhone's screen and other hardware. If the raw data doesn't match any applicable gestures or commands, the iPhone disregards it as an extraneous touch.

There are some other suppliers for capacitive touch technology. The implementation of controller includes ASIC and MCU based. ASIC is very low cost for mass-production, while MCU has the advantage of custom design but its price is higher. The following vendors are delivering touch solutions for markets. Some of them have already supported multi-touch. More vendors are coming to develop new parts to support. Mentioned vendors have their own patents on touch technology.
Synaptics is a leading worldwide developer of custom-designed user interface solutions for mobile computing, communications and entertainment devices. Synaptics is focusing on touch pad solutions, and offer ASIC with I2C/SMBus.
Taiwan ELAN Microelectoncis Company (EMC) also a patent with trademark of eFinger. It claims that patent is competitive on touch screen application. But it requires registration on specification and has an open lawsuit with Synaptics.
Cypress offers CapSense in many applications. For example, the keypad part of V3/V8 from Motorola, LG chocolate mobile phones. Cypress CapSense is based upon its PSoC mixed signal array with embedded RISC M8 microcontroller. It is very easy to use. The new PSoC CapSense touch screen solution also offers designers the ability to implement multiple additional functions beyond touch screens. The same devices can implement capacitive buttons and sliders simultaneously, replacing their mechanical counterparts, as well as proximity sensing. Engineers can also take advantage of the PSoC mixed signal array to implement functions beyond CapSense. Such functions include, driving LEDs, backlight control, motor control, power management, I/O expansion, accelerometers and ambient light sensors. These functions, in conjunction with flexible communication (I2C and SPI), allow for unparalleled system integration. Cypress's CapSense touch screen solution is available using projected capacitance and surface capacitance sensing techniques.
Leadis is dedicated to creating compelling touch solutions focused on a strategy of Innovation and Integration. Leadis' line of PureTouch (TM) capacitive touch controller solutions will begin sampling in the first half of 2008.
ATLab is a Korean company, co-operates with ST to offer touch screen touch ICs for capacitive technology.
Quantum Research Group, offers QProx(TM) touch control and sensor ICs. As same as Cypress, its solution offers more features for QSlide and QWheel. Atmel acquired Quantum already.
Thanks to Apple, more and more players are trying to compete on this market.
Other Suppliers

ST has two lines for touch pads, one is licensed from Quantum (MCU based), the other is licensed from ATLab (ASIC based). TI is trying to promote its MSP430 in this market. NXP also has its companion chip for PDA with touch screen DAC inside. Some other vendors such as Microchip and Maxim have similar product lines.
Driver and Application Software
I am not quite sure about how Apple implement these, I would rather to use software to refer to both driver and application software part. Apple also filed a complementary patent called "multi-finger gesture" for presentation of the finger movement. But that part requires more know-how on system software, I am not going to discuss here. Anyone who is interested in please visit some professional sites of MIT or NYU.
Latest Development for Alternative UI
Microsoft is developing surface computer, which can track up to 52 touches.
Besides multi-touch, some more topics regarding touch screen is rising on horizontal. They are photo sensor in pixel, polymer waveguide, distributed light, strain gauge, dual-force touch, laser-point activated touch and 3D touch.
Patent Issues
Most the multi-touch related patents are held in 3M, Nitto Denko, Oike-Kogyo, Dupont, Apple, IBM for materials and panels, but more Taiwanese, Korean industries are trying to file their own patents on the touch panels. As a system developer, we just make sure our products will not infringe these patents. It is tricky for those mobile phones manufacturers who are trying to clone Apple design on their own products.