The cellular telephone technology is one of the best success stories of recent years with its capability to keep the user working untethered for an entire day, and requiring only a single overnight recharge. The ultimate vision for this technology is the "smart phone," having the advanced functionality of a hand-held computing device and mobile phone in one convergent device. Reaching such level of functionality, without compromising the usage model of the cellphone, will present enormous challenges as well as opportunities for the electronics industry and in particular for power management.
The Wireless Landscape
The wireless landscape is today----and will remain for many years----very fragmented along both geographical and communications standards lines. Three generations of digital cellular technologies; second (2G), third (3G) and in-between (2.5G), do already coexist (see Table 1).
Table 1. Common cellular standards.
Japan is ahead of the pack with 3G (W-CDMA and CDMA2000 flavors), while U.S. manufacturers are building the infrastructure to provide 2.5G technology, at this very time. Europe and Asia are somewhere in between in development. The Japanese typically do not own home computers and are relying increasingly on their cell phones to exchange text messages, as well as to access email and the Internet.
If this behavior of relying on overlapping technologies takes hold elsewhere, the future of smart phone devices, which combine the computing capabilities of a hand-held device and the communications capabilities of a cellular telephone, is assured. That smart phones have a real possibility to become the next "disrupting" technology seems to be confirmed by the recent entrance in the wireless arena of powerful novices like Intel and Microsoft.
The majority of cellular and hand-held devices is powered today by a single cell Lithium-Ion battery. The wireless semiconductor "smart" ICs in the signal path of this technology follow an established industry-wide trend, and are mostly designed in sub-micron, low voltage and high-density processes. Consequently, the power management ICs are----with a few exceptions----low voltage devices themselves, bridging the gap between the power source voltage range (2.7 to 4.2V) and the operational voltage of the signal IC's (1 to 3.5V). Such low operational voltages, in conjunction with the necessity of low quiescent currents for long stand-by times, have established low voltage CMOS (0.5um minimum feature at his juncture) as the process of choice for wireless voltage regulators. Since space is premium in these applications, these voltage regulators come in very small packages, from leaded to lead-less to "chip scale" flavors (See fig. 1).
Figure 1. Small form factor packages for voltage regulators.
For the next few years the cellular telephone will remain by far the most dominant wireless device, accounting for 95 percent of total units. Hand-held devices and smart phones will account for the remaining 4 percent and 1 percent respectively. Figure 2 shows the typical block diagram of a 2.5G digital cellular telephone, in the class of the recently announced T68 mobile phone by Sony Ericsson.
Figure 2. Block diagram of a 2.5G mobile phone.
In this configuration, each block requires a specialized power supply. The RF section is particularly sensitive to noise and is best served with low noise linear regulators, while other sections will be served by either linear or switching regulators, which is based purely on architectural and cost constrains.
Figure 3 illustrates a possible strategy for the configuration illustrated in Figure 1. The battery can directly power the audio LDO since its output voltage (2.5V) is below the minimum operational battery voltage (2.7V). The rest of the LDO outputs fall somewhere inside the battery range of operation (2.7 to 4.2V) and consequently need a higher supply voltage, in this case provided by the boost converter. The DSP core at 1V will need a dedicated buck converter, while the LCD display contrast at 20V will need a dedicated boost converter.
Figure 3. Power management strategy for a cellular telephone.
Figure 4 shows the block diagram of the power management system. In this case a total of 7 voltage regulators are necessary to power this device.
Figure 4. Mobile phone power management system.
Finally, Table 2 shows a wide selection of chips, classified by function, from which to draw each of the elements in Figure 4.
Table 2. Fairchild semiconductor building blocks for wireless applications.
Wireless and Hand-Held Devices
A lot of activity is going into wireless and hand-held devices thanks to their potential to, at some point, intercept and to take over a share of the cellular market. Figure 5 shows the typical block diagram of a 2G wireless hand-held, in the class of the recently announced Palm i705.
Figure 5. Block diagram of a 2G hand-held computer.
Here, again, this diagram shows how each block requires a specialized power supply; but because of the absence of a DSP and of the SIM card the power management is a bit leaner than for the previous case. With similar considerations done, as in the previous case, Figure 6 shows the strategy chosen for the hand-held power management device. Figure 7 shows the implementation, obtained with a total of 5 regulators. Again, the specific components can be drawn from table 2.
Figure 6. Power management strategy for wireless hand-held.
Figure 7. Hand-held power management system.
Battery charging issues
Other important elements of both cellular phones and hand-held devices are the external AC adapter and the internal charger. Many AC adapters on the market are very simple implementations based on a transformer, bridge rectifier and a resistive current limit. More sophisticated controls can be obtained with integrated controllers, such as those indicated in Table 2.
The Lithium-Ion charger is a constant-current/constant voltage regulator that is either implemented with specialized controllers (see Table 2 for an example) or by PWM of a pass transistor controlled directly by the CPU.
Protection and fuel gauging
Another important consideration in wireless hand-held designs is the in-battery electronics, namely that section of power management residing inside the Li+ cell. The energy density of Lithium-Ion cells make them dangerous elements that need a very precise protocol for charge and handling. Overcharge, as well as undercharge, must be prevented, which leads to reduced energy storage. To this end, protection electronics measures the battery voltage and opens a pass transistor as soon as the charge voltage threshold is crossed.
Fuel gauging is necessary to be able to display the state of charge of the battery and to predict the residual time of operation in battery mode. This is an interesting and challenging feature because residual time of operation matters to the user only toward the end of the battery charge, exactly when the accuracy of the prediction begins to falter. In fact, no matter how precise the measurement system is, eventually the residual time of operation will translate into an amount of residual charge that is of the order of magnitude of the system precision, leading to increasing prediction errors as the battery approaches the empty state. This results in amazing levels of precision in the current measurement, or Coulomb counting, with analog front-end amplifiers resolving micro Volts of voltage drops across small sense resistors, followed by 10 bit or higher order A/D converters.
The actual processing of the row data today----at the 2 to 2.5G juncture----is generally done in the central processing unit. With both smart phones and 3G systems and above we expect that the taxing of the DSP, or its successor, will be such that the fuel gauge data processing will be de-centralized, and this will lead to smart fuel gauge devices incorporating compact 8 bit micro-controllers. Figure 8 shows an example of integrated combo fuel gauge and protection control IC utilizing Fairchild's dual MOSFET FDW2508D as pass transistor for the protection section.
Figure 8. Fuel gauge and protection with Fairchild's FDW2508P as pass transistor.
Convergence of the cellular telephone and the hand-held device
By analyzing the block diagrams in Figures 2 and 5, it becomes obvious how similar these two systems are. Both rely on the same radio technologies and frequency ranges; Bluetooth for device-to-device networking, single Lithium-Ion for power source, etc. In fact it is easier to point at the differences between the two devices. The DSP, present in the cellular phone only, is a key differentiator allowing for voice processing. Other than that, it really comes down to size. The hand-held will typically have a bigger screen and a bigger memory (flash for operating system, phone book and files storage and SRAM for temporary data storage), as well as stereo audio for music player emulation (MP3).
A few examples of smart phones, namely hand-held devices with DSP on board, are already appearing in the market. One example is the Blackberry 5810 from VoiceStream, a handheld PC that can be transformed into cellular phone by means of a hands-free module connected to the device via a 2.5mm Jack.
As pointed out at the beginning of this article, the challenging questions are part technological (will the smart phone be able to deliver the same stand-by and talk times the cellular customers are used to?) and part cultural (will the Japanese model of connectivity illustrated at the beginning prevail?).
At the 3G juncture, the system complexity for cellular and smart phones will be such that one DSP will not be enough to support video and audio compression and an additional DSP or ASIC will be necessary. In turn, this will increase power consumption and reduce battery operation time. Adaptive Computing Machines, or ACM's, are a new class of ICs appearing on the horizon and promising to solve the power dissipation problem by means of a flexible architecture optimizing software and hardware resources.
Power management in wireless devices is a pervasive issue encompassing every element in the signal as well as in the power path. With systems complexity increasing dramatically from one technology generation to the next, the long duration of untethered operation in wireless devices can be preserved only with the introduction of new breakthrough technologies. New architectures, such as the aforementioned ACMs as well as the conversion of large scale IC's from bulk CMOS to Silicon on Insulator (SOI), should go a long way toward reducing the power dissipation of the electronic load. At the other end of the equation, new and more powerful sources of power, perhaps fuel cells, should be able to provide higher power densities inside the same cell form factor. The entire technology arsenal should be able to continue to provide more features without compromising performance
Power conversion technologies have achieved impressive efficiencies already, reaching peaks of 95 percent. Accordingly, they are a critical element of the power management equation but not its bottleneck. The analog building blocks for an effective power management of a wireless device in its present and future incarnation are already in place and it is not uncommon today to find them assembled inside custom combo chips integrating the entire power management function, on board, of a single IC. Voltage regulators----today designed with 0.5um minimum features----will continue to follow the CMOS minimum feature reduction curve, staying only a few technology generations away from the loads they are powering (state of the art 0.13 micron minimum feature). Accordingly they will continue to be able to adequately sustain the power, performance and cost curve that will be required to power future generations of wireless devices.