Select the right battery fuel-gauge for smart phones and tablets
Smart phones and tablets are becoming indispensable as they gain more and more functionality. At the same time, the expectation is for these devices to become both thinner and cheaper. This requires balancing the competing objectives of both smaller batteries and increased usable capacity. To solve this dilemma a highly accurate battery fuel-gauge is needed. In this article we discuss the trends in smart phone and tablet design. We also identify how accurate fuel-gauging can increase customer satisfaction, reduce battery costs, and extract the maximum run-time. And, check out the sidebar discussion: Pros and cons of embedded batteries for consumers -- and the environment.
Smart phone applications such as video players or gaming systems require application processors (APs) based on relatively power-hungry two- or even four-core AP chipsets. The smartphone display screen is also getting bigger, from 4.3 inches to 4.7 inches and upwards, which also require more power.
The traditional battery pack in smartphones has been removable. As the devices form factor continues to become thinner, many are shifting to non-removable battery packs, a trend first started by tablets. Embedding battery packs within the device allows for more creative placement options to save space. With space at a premium the system designer needs to use the smallest cells possible to achieve the desired specifications of run-time. Not only are cells with higher energy density used, but also all available cell energy is needed. If an inaccurate fuel-gauge underestimates the available energy, it can lead to early termination. This means that energy is still available, but the end user only sees is a shorter run-time.
One solution is to use an over-sized battery to compensate for the unused capacity while allowing a looser gauging estimation, but this wastes space. A more cost-effective solution is to use a more accurate fuel-gauge to efficiently utilize all of the available energy while still avoiding a surprise shutdown.
To save pc board space and reduce cost, additional features are integrated into the baseband processor, application processor, and power management IC (PMIC). On PMICs, in particular, there is an attempt to integrate battery fuel-gauge functions. However, compared to the complexity of advanced discrete fuel-gauges, the battery fuel-gauge found in PMICs basically is a simple “coulomb counter” that requires the host processor to provide the gauging algorithm. Furthermore, the heat and interference generated by the PMIC switch-mode charger and converters can impact the integrated voltage and current measurement. This greatly reduces the gauging accuracy of this all-in-one approach.
Most tablet designs place the fuel-gauge within the battery pack, and most tablet batteries cannot be removed by the user. This trend of embedded batteries has spread to smartphone designs as phones get thinner. More designs now use pack-side gauges to reduce development and manufacturing cost.
There are some misconceptions about battery fuel-gauges due to the traditional thinking of coulomb counters. A true battery fuel-gauge must do more than just simply count the Coulombs or report the battery voltage. So what exactly is a battery fuel-gauge?
A battery fuel-gauge is an intelligent device that can provide information such as state-of-charge (SOC), available charge (in mAh or mWh), battery run time (in minutes), and battery state-of-health (SOH). One point of confusion is that there are many types of “fuel-gauges” that are no more than a battery monitor. A battery monitor simply reports battery voltage, charge and discharge current, and battery temperature. It usually incorporates a coulomb counting analog-to-digital converter (ADC) to track charge. A battery monitor can’t accurately report the actual battery remaining capacity without a sophisticated gauging algorithm. A battery monitor just reports measurements and requires the host processor to apply compensations for temperature, load rate, and other factors that can affect the usable capacity. In this approach, cost of software development can be higher while the accuracy will be lower when compared to a fully-featured fuel-gauge. Performing gauging algorithm by the host processor will also increase power consumption and still lack the features of most standalone fuel-gauges.
A true battery fuel-gauge accurately measures the battery voltage, charge and discharge current, and battery temperature. It then uses its own sophisticated gauging algorithm to calculate the battery SOC, available capacity, and other useful information. Some advanced battery fuel-gauges can learn the actual battery capacity and impedance as they change over time and usage. For optimal accuracy the learning is critical since battery cells from the same production line do not have identical capacities and resistance profiles. The profile also changes as the cell ages. The gauges with profile statically configured only for a new pack cannot provide accurate results after some time in the field.
A Lithium-Ion (Li-Ion) or Lithium-Polymer (Li-Poly) battery is an electrochemical system with complex behavior. This presents challenges to simplistic concepts of fuel gauging. The battery SOC is no longer a simple function of battery voltage. It is a function of battery voltage, load, temperature and internal impedance Equation 1.
SOC = f (V, I, T, R) Equation 1
A battery’s behavior depends on its chemical solution, anode and cathode materials. However, the open circuit voltage (OCV) curve of the same type of battery is very similar. A fuel-gauge can use the OCV to determine the initial SOC of a battery. A “fuel-gauge” that only uses coulomb-counting cannot determine the initial SOC due to lack of OCV information. These gauges require a full cycle (charge to full then discharge to empty) to get accurate results and are best used when embedded in the battery pack. For host-side coulomb counters, this becomes an issue since once the battery pack is removed.
Traditional voltage-based gauges can no longer be used for smart phone or tablet applications. Figure 1. The only condition under which the voltage can reliably be used as an OCV measurement is when the battery is relaxed.
Figure 1: Even under normal operation the load profiles for high-end smart phones are so dynamic that battery voltage and SOC no longer directly correspond.