Better Touch Better Business
Contact Sales at Sophia Louise.
Unlike some batteries, which require only a simple current source to be recharged, lithium-ion batteries have special requirements, and semiconductor manufacturers have launched a wide variety of ICs to meet them. The semiconductor industry classifies these products as "power management" or "battery management." Although the battery may look like a simple two-ended device, users must be careful how they use it to extract the most power from it.
In the field of battery management ICs, each manufacturer takes a different approach to transferring power from the power source to the load, and they provide a wealth of application information for engineers to design battery management circuits. This article will focus on the issue of charging, focusing on how to manage the power supply from several sources and measure the energy in the battery.
The following are the basic features that designers would like to see in a battery management IC:
The designer wishes to program the charging current through an external component and the charger IC or through a simple interface (such as the I2C) to the master microprocessor or microcontroller. A standalone charger doesn't have dynamic programming capabilities, but if the designer plans to let the processor control the charging operation, they may want to set the current dynamically in real time.
The designer selects a charging end current so that the battery management IC knows when the battery has been fully charged. In lithium-ion batteries, charging stops when the current drops to about 10 percent of its normal value. But the designer can choose, for example, to drop the current to 15% of the normal value as the limit for stopping the charge. They can adjust this set point by adjusting the parameter values of external components connected to the IC pins or by interface connections to a processor. Designers also hope to be able to choose different terminal voltages, such as 4.1V or 4.2V, to accommodate different chemical requirements or to cope with batteries from different suppliers.
One or multiple charging status outputs can indicate how far the charging cycle is proceeding. This information is usually displayed by the LED, and the LED can also indicate some other status of the charger, such as pre-charge, constant current, constant voltage or end of charge, etc. These outputs can also alert the main controller or other circuits as appropriate. The available status indicator function depends on the IC model chosen by the designer.
The designer must also consider safety, as the charger IC should be able to terminate the charge or disconnect the connection to the battery if it overheats. BrunoKranzen, director of product development for portable power systems at NationalSemiconductor, recommends incorporating a watchdog timer into the design. "Circuits should not attempt to charge a defective battery indefinitely because doing so could damage the battery and the system it supplies."
Most lithium-ion battery assemblies have a thermistor built into them so that the corresponding management IC can interrupt the charging process if the battery overheats. For example, overheating occurs when a circuit tries to overcharge a battery or pumps too much current into a severely depleted battery. A depleted battery may have to be precharged at a lower current (usually less than 100mA) before entering a normal charging cycle.
(a) The new LM3658 introduced by NationalSemiconductor
Figure 1 Both devices perform a similar function - obtaining power from an AC adapter or a USB connection to charge a lithium-ion battery. An external signal activates these charger ICs and sets the USB current
Figure 2 The lithium-ion battery and external power supply are connected to the device via the LinearTechnologylTC4055 chip. The chip controls the direction in which electricity is sent -- to the device, the battery or both. The two states of the chip glow with LEDs to indicate how the circuit is working
Fever can cause damage
"There is another reason for the fever." Says Peter Fundaro, product marketing manager for power management at TexasInstruments. A linear charger that controls the FET will heat up, so will the device it is powering, and so will the battery when it is being charged. The heat generated will raise the temperature and turn off the charger. But because the device continues to run on an external power source, its user will think the battery is fully charged, when it is not.
If you want to keep the machine running while charging, then the heat will cause some hazards. Designers who have not previously encountered such problems often struggle to figure out what is wrong and how to fix it.
Recognizing this basic need, IC vendors LinearTechnology, NationalSemiconductor, TexasInstruments, and others have designed a variety of devices that allow designers to choose how to control the battery power in a portable device (Figure 1). The range of options covers a wide range of products from dedicated, stand-alone battery management ICs to processor-controlled devices.
But many designers do not want to have as much control over the charging process, and companies offer them a variety of standalone ICs that offer a variety of functions. "We supply ICs that automatically select the power supply -- first the battery, then the battery or the wall power adapter -- to ensure that the power supply to the system is not interrupted," Gurries says. "We've built that 'integrated capability' into the hardware."
Engineers who want to manage batteries through microprocessors or microcontrollers must first make sure that the processor can handle its important tasks. Designers must therefore balance the resources needed for these tasks with those needed to manage the circuitry of the battery. For full battery control, the processor may supply analog and digital I/O pins, and in some cases may perform timer and pulse width modulator functions for battery management circuits. "Designers must also write code to oversee the task of battery management, to monitor temperature, check battery charge levels, and so on." TI's Mr Fundaro insists.
Connecting several kinds of power supplies
Whether the designer uses a stand-alone or processor-based battery management circuit, it must be connected to an external power supply to charge the battery or provide auxiliary power. The diversity of power supplies presents other challenges. "When portable devices contain a battery and an AC adapter or wall adapter, it is best to operate on adapter power." "But if the user disconnects the AC adapter, the power system must switch to the battery and the device cannot be turned off, which is another function that these ICs can perform."
The battery management circuit in a device that can rely on a variety of power sources must be determined:
"It's a difficult problem to solve," says TonyArmstrong, product marketing engineer for LinearTechnology. When a user connects a typical battery-powered system to an AC adapter, all power must first flow through a linear battery charger to the battery, which then drives the system. This arrangement reduces efficiency by 30 per cent because a lot of electricity is wasted on chargers. If the battery in the system is exhausted, the battery charger loses most of the power it receives, and this power does not reach the portable device or be recharged into the battery. "We call it a charger-fed system," says Armstrong, "because the battery is always powering the system, so when you disconnect, your battery isn't fully charged."
Also, will people use USB ports for power? Of course. The USB port can emit up to 100mA of current, and the device connected to it can draw up to 500mA in consultation with the USB master. By battery-charging standards, this doesn't seem like a high level of current, but if a USB device carries a battery inside it, it can be charged from any current source supplied externally. So if someone connects a mobile phone or digital camera to a USB port, they can easily charge the battery for as long as the device stays connected.
NationalSemiconductor's Kranzen adds, "People like to use a USB 'fuel feed. They can use a laptop to charge a mobile phone, PDA or any portable device they carry via a USB port, eliminating the need to carry a separate charger when traveling."
Maintain appropriate electrical energy
No matter how people charge a lithium-ion battery or use its power source, they want to know how much power is left in the battery. The battery "fuel gauge" will supply this information, often in the form of a bar chart or other display indication. Accurate displays of residual power can extend the time users spend using their portable devices as much as possible. The unique advantage of this meter is that the precise measurement of residual power allows the device to be shut down in a proper way, so it can store data to disk, back up system information, and so on before the device completely runs out of power. "If a device doesn't accurately indicate power levels, people don't get a good return on the money they spend on expensive, high-capacity batteries," says TI's Fundaro. "Without understanding how close they are to running out, they charge the battery too often, or not enough."
Unlike a sensor built into a gasoline tank that directly measures the fuel level, a battery meter measures the flow of electrons into and out of the battery. This kind of indirect measurement has its own problems. "Designers need to keep in mind that in an application with, say, a 1A-H battery, you can't supply that much power all the time," says Fundaro. "A rapid discharge of a battery reduces the amount of power it can absorb, but a mild discharge can release more power than its rated capacity. The meter circuit must determine the capacity of the battery based on how the device draws power from the battery."
Like most gauges, the oil gauge is calibrated periodically. These recalibrations or resets must be carried out under specific conditions and planned by the system designer. When the battery management IC determines that the battery can no longer produce power, the IC sets the "oil level to zero", indicating that there is no charge left in the battery. The meter reset can be accomplished using data stored in the battery management IC's EEPROM or Flash memory or a component connected to the management IC pin.
However, if a battery goes through several low-energy charge and discharge cycles, it will never reach the "empty" state that allows the meter to return to zero. In this case, the battery management circuit can notify the device to discharge the battery completely so that the reset operation can be completed, or the battery management IC can reset its charge meter data to zero based on environmental conditions and battery life cycles, etc. Some battery management ICs monitor usage patterns and adjust the metering process accordingly (see Figure 3).
TI is working on a new algorithm that uses the impedance of the battery to calculate the remaining capacity, Fundaro explained. "As time passes, the internal impedance will increase, so the meter can use this information to figure out how much the capacity of the battery has changed. "Next-generation impedance tracking technology will reduce the accumulation of meter errors in some applications."
Don't be put off by the programmable nature of these meters; the chip vendor provides ample information on the implementation of the meter circuit and any firmware operation.