Virtually all battery-based power management designs depend on the associated battery, so design starts by picking the specific battery type. The battery may be the non-rechargeable primary type or the rechargeable secondary type. (For more, see “Batteries 101: From Nickel To Lithium And Beyond” at www.electronicdesign.com, ED Online 20737.) The most widely used rechargeable battery-based systems may employ:
- Nickel cadmium (NiCd), which finds use where long life, high discharge rate, and economical price are important
- Nickel-metal hydride (NiMH), which has a higher energy density than NiCd, at the expense of reduced cycle life
- Lithium-ion (Li-ion), which is used where high-energy density and light weight are of prime importance
- Li-ion polymer, with chemistry similar to the Li-ion in terms of energy density
The performance and longevity of rechargeable batteries depend mainly on the quality of the chargers. One type of charger (used only for NiCd) applies a fixed charge rate of about 0.1C. A faster charger can take three to six hours with a charge rate of about 0.3C.
A charger for NiMH batteries could also accommodate NiCds, but not vice versa because a NiCd charger could overcharge a NiMH battery. Lithium-based chargers require tighter charge algorithms and voltages. Avoid a charge rate over 1C for lithium packs because high currents can affect the lithium. With most lithium packs, a charge above 1C isn’t possible because the protection circuit limits the amount of current the battery can accept.
Precise full-charge detection of nickel-based batteries requires special ICs that monitor battery voltage and terminate the charge when a certain voltage signature occurs. A drop in voltage signifies that the battery has reached full charge, known as Negative Delta V (NDV).
After full charge, you can trickle charge into a NiCd battery to compensate for its self-discharge characteristics. The trickle charge for a NiCd battery ranges between 0.05C and 0.1C. To reduce memory effects, there’s now a trend toward lower trickle-charge currents.
NiMH battery chargers use a combination of NDV, voltage plateau, rate-oftemperature- increase (dT/dt), temperature threshold, and timeout timers. The charger utilizes whatever comes first to terminate the fast charge. NiMH batteries that use NDV or the thermal cutoff control tend to deliver higher capacities than those charged by less aggressive methods.
Li-ion chargers use a voltage-limiting device. However, Li-ion batteries have a higher voltage per cell, tighter voltage tolerance, and the absence of trickle or float charge when reaching full charge. Charge time for Li-ion batteries charged at a 1C initial current is about three hours. Full charge occurs after reaching the upper voltage threshold and the current drops and levels off at about 3% of the nominal charge current.
Increasing Li-ion charge current has little effect on shortening the charge time. Although it reaches the voltage peak faster with higher current, the topping charge will take longer. Because Li-ion batteries can’t absorb overcharge, these batteries should not use a trickle charge. Overcharging can cause the cell to overheat.
Li-ion batteries have good cold and hot temperature charging performance. Some cells allow charging at 1C from 0°C to 45°C. Most Li-ion cells prefer a lower charge current when the temperature gets down to 5°C or colder. Avoid charging below freezing.
Several charger ICs for Li-ion batteries make it possible to charge the battery via a USB port or an ac adapter. For USB operation, the user can plug the USB cable into a desktop or laptop computer and use the 5-V output to charge the battery pack in a cell phone or PDA.
Texas Instruments’ bq24150 battery-charger IC is a compact, flexible, high efficiency, USB-friendly, switch-mode charge-management IC for single-cell Li-ion and Li-polymer batteries (Fig. 1). The I2C interface allows the programming of the charge parameters and reports charge status to the host. It integrates a synchronous pulse-width modulation (PWM) controller, power MOSFETs, input current sensing, high-accuracy current and voltage regulation, and charge termination, into a small wafer-level chip-scale package (WCSP).
Continued on page 2
The bq24150 charges the battery in three phases: conditioning, constant current, and constant voltage. The input current is automatically limited to the value set by the host. Charge is terminated based on a user-selectable minimum current level.
During normal operation, the bq24150 automatically restarts the charge cycle if the battery voltage falls below an internal threshold. When the input supply is removed, it automatically enters sleep mode or high-impedance mode.Battery “Gas-Gauge” ICs Battery-based systems are sensitive to the amount of usable life left in the battery. This is particularly important for computers where a loss of power could mean a loss of stored data. In addition, most batterybased systems are portable, so their operating environment can vary over a wide range of temperatures, which affect a battery’s efficiency, rate of charge and discharge, and therefore battery life.
One solution to this battery-sensitive situation is to include a means for providing a real-time indication of remaining battery life. Borrowing from the automotive vernacular, this function within battery-based systems has been called a battery “gas gauge.” The gas-gauge IC calculates the available charge of the battery and also compensates for battery temperature because the actual available charge is reduced at lower temperatures.
For example, if the gas-gauge IC indicates that the battery is 60% full at 25°C, then the IC indicates 40% full when cooled to 0°C, which is the predicted available charge at that temperature. When the temperature returns to 25°C, the displayed capacity returns to 60%. This ensures that the indicated capacity is always conservatively representative of the charge available for use under the given conditions.
Depending on the battery type, the gas-gauge IC also adjusts the available charge for the approximate internal self-discharge of the battery. It adjusts self-discharge based on the selected rate, elapsed time, battery-charge level, and temperature. This adjustment provides a conservative estimate of self-discharge that occurs naturally, and that’s a significant source of discharge in systems that aren’t charged often or are stored at elevated temperatures.
The gas-gauge IC is usually packaged in the battery pack. Because specific inputs on the gas-gauge IC connect directly to the battery, those inputs must consume very little power. Otherwise, battery life will be reduced during long storage periods.
The battery gas gauge continuously compensates for both temperature and charge/discharge rate. Typically, it displays the available charge on LEDs and also can send the charge data to an external processor via an I/O port. The LED presentation usually consists of five or six segments of a “thermometer” display. To conserve battery power, the display is only activated at the user’s discretion.
Battery gas-gauge ICs employ mixed-signal, analog, and digital circuits. One technique uses analog circuits to monitor battery current by measuring the voltage drop across a low-value resistor (typically 20 to 100 mO) in series with the battery. This provides the charge input to the battery and the charge subsequently removed from the battery.
Integrated over time, the scaled voltage drives internal digital counters and registers. The counters and registers track the amount of charge available from the battery, the amount of charge removed from the battery since it was last full, and the most recent count value representing “battery full.”
Initially, the battery must be fully charged and the counters and registers set to states consistent with a fully charged battery. As discharge occurs, the gas-gauge IC tracks the amount of charge removed from the battery. At full charge, all the LED segments are lit. As the battery is depleted, the gas-gauge IC extinguishes successive segments on the thermometer display.
Maxim’s DS2788 operates directly from 2.5 to 5.5 V and supports single-cell Li+ battery packs. It accommodates multi-cell applications by adding a trim register for calibration of an external voltage-divider for VIN (Fig. 2). Non-volatile (NV) storage provides for cell compensation and application parameters.
Host-side development of fuel-gauging algorithms is eliminated. On-chip algorithms and convenient status reporting of operating conditions reduce the serial polling required of the host processor. Moreover, 16 bytes of EEPROM are made available for the exclusive use of the host system and/or pack manufacturer. The additional EEPROM can be used to facilitate battery lot and date tracking as well as NV storage of system or battery usage statistics.
A one-wire interface provides serial communication at the standard 16-kbit/s or overdrive 140-kbit/s speeds, allowing access to data registers, control registers, and user memory. A unique, factoryprogrammed, 64-bit registration number (8-bit family code + 48-bit serial number + 8-bit cyclic redundancy check, or CRC) ensures that no two parts are alike and enables absolute traceability.
Continued on page 3Battery Monitors Battery-monitoring ICs accurately measure the charge and discharge currents in rechargeable battery packs. Intended for pack integration, these devices include all of the necessary functions to form the basis of a comprehensive battery-capacity management system in cellular phones, PDAs, or other portable products. Battery monitors work with the host controller in the portable system to implement the battery gas-gauging and management system.
Battery monitors are actually data-acquisition systems that accumulate data related to battery parameters and then transmit it to the host processor. These mixed-signal ICs, which incorporate both analog and digital circuits, also include one or more types of digital memory and special registers to hold battery data. Analog circuits include temperature sensors and amplifiers, as well as some interface circuits.
The host controller is responsible for interpreting the battery monitor data and communicating meaningful battery data to the end-user or power-management system. To measure current, the monitors usually include either an internal or external current sense resistor. Voltage and current measurements are usually via an on-chip analog-to-digital converter (ADC).
Among the monitored battery parameters are overcharge (overvoltage), overdischarge (undervoltage), and excessive charge and discharge currents (overcurrent, short circuit), information of particular importance in Li-ion battery systems. In some ways, a battery monitor assumes some of the functions of a protection circuit by protecting the battery from harmful overcharging and overcurrent conditions.
The difference between battery fuel gauges and battery monitors is that the fuel gauge provides a self-contained measurement of battery life, whereas the monitor employs a host processor for its calculations. Therefore, the monitor can provide a more accurate reading for battery life and usually a better display of battery capacity.
Linear Technology’s LTC6802-1 is a complete battery-monitoring IC that includes a 12-bit ADC, precision voltage reference, a high-voltage input multiplexer, and a serial interface (Fig. 3). Each LTC6802-1 can measure up to 12 series connected battery cells with an input common- mode voltage up to 60 V. Using a unique level-shifting serial interface, multiple LTC6802-1 devices can be connected in series, without optocouplers or isolators, allowing for the monitoring of every cell in a long string of series-connected batteries.
Multiple LTC6802-1 ICs connected in series can operate simultaneously, permitting all cell voltages in the stack to be measured within 13 ms. There’s also a standby mode. Each cell input includes an associated MOSFET switch for discharging overcharged cells.Power-Supply Controller ICs Virtually all battery-based systems are intended for portable operation, so their power supplies have unique requirements:
- Operate from one- or two-cell Li-ion batteries or three- or four-cell NiCd or NiMH packs
- Provide the appropriate voltage and current for load
- Provide high efficiency for maximum battery run time
- Allow light weight and small physical size supplies to minimize overall size
- Be thermally efficient to prevent overheating
- Minimize assembly and component cost for consumer-based systems
- Provide high-reliability, carefree service
These requirements dictate the associated power-supply controller IC configurations. This also means the controller ICs should require very few external components, and any that are used should be lowcost types. To minimize size and weight, the IC comes in some form of small-outline package. The application will determine whether the controller should provide stepup, step-down, or some other topology.
One tradeoff in selecting a controller IC is whether it employs external or on-chip power MOSFET switches. On-chip ICs minimize external components, but have the potential to increase the junction temperature and degrade thermal performance. Depending on the package employed, this could also shrink the current-carrying capacity of the IC.
One consideration is to reduce power dissipated by the power supply, which in turn increases battery run time. Most controller ICs have a shutdown pin that disables the power supply, cutting battery drain. This can be done in many systems that have a normal “sleep” mode. When the IC comes out of the shutdown mode, it must do so without upsetting the system.
Also available in most battery-based controller ICs is undervoltage lockout (UVLO), which shuts down the power supply if the input voltage drops below a specific threshold. Therefore, if the battery output voltage drops too far, the power supply will shut down.
Another characteristic of these ICs is protection against overcurrent, which protects both the controller IC and the system components. This is accomplished by sensing current to the load and cutting power for an overload condition.
For all switching power supplies, the layout is an important step in the design, especially at high peak currents and high switching frequencies. If the layout isn’t carefully done, the regulator could show stability as well as electromagnetic-interference (EMI) problems. Therefore, use wide and short traces for the main current path and the power ground tracks.
Continued on page 4
The input capacitor, the output capacitor, and the inductor should be placed as close as possible to the IC. Use a common ground node for power ground and a different one for control ground to minimize ground noise effects. Connect these ground nodes at any place close to one of the IC’s ground pins.
The feedback divider should be placed as close as possible to the IC’s control ground pin. To lay out the control ground, use short traces, separated from the power ground traces. This prevents ground shift problems, which can occur due to the superimposition of power-ground current and control-ground current.
The Analog Devices ADP2121 highfrequency, low-quiescent-current, stepdown, dc-dc converter is optimized for portable applications severely constrained by board area and battery life (Fig. 4). Its synchronous rectification improves efficiency and results in fewer external components.
At high load currents, the device uses a voltage-regulating PWM mode that maintains a constant frequency with excellent stability and transient response. At light load conditions, the ADP2121 can automatically enter a power-saving mode. This mode utilizes pulse-frequency modulation (PFM) to reduce the effective switching frequency and ensure the longest battery life in portable applications. During logic-controlled shutdown (EN = 0), the input is disconnected from the output and draws less than 0.28 µA (typical) from the source.
The ADP2121’s input voltage range of 2.3 to 5.5 V allows the use of a single Li+/ Li-polymer cell, multiple alkaline/NiMH cells, PCMCIA, and other standard sources. The converter is internally compensated to minimize external components, and its 1.80-, 1.82-, 1.85-, or 1.875-V fixed output can source up to 500 mA. Other features include UVLO to prevent deep-battery discharge and soft start to prevent input current overshoot at startup.
Battery-based power-supply ICs are also available with multiple outputs. On Semiconductor’s MC34700 is a quad-output, high-efficiency power supply with on-chip power MOSFETs. It features three stepdown switching regulators and one lowdropout linear regulator. The switching regulators employ voltage-mode control with external compensation, which provides the flexibility to optimize a given application’s performance.
The MC34700 is intended for space-constrained applications that require multiple power rails as well as simplicity in power-supply design and implementation. Overvoltage, undervoltage, overcurrent, and overtemperature protection features ensure robust and reliable operation.
Fixed switching frequency, internal soft start, and internal power MOSFETs enable rapid power-supply design and development. It’s well suited for power-supply designs set-top boxes, cable modems, laser printers, fax machines, point-of-sale terminals, small appliances, telecom line cards, DVD players, and other applications.AC ADAPTERS Finally, many battery-based systems employ an ac adapter to supply power for charging the battery. Astec’s AD7216N2L ac adapter, rated at 72 W, provides 16 V at 4.5 A (maximum) at 5% regulation and 85% typical efficiency (Fig. 5).
For more, see “Li-Ion Battery Protection ICs” at ED Online 20738.”