Attempting Clairvoyance with Battery Performance Page 1 of 4 Power Management Designline | Attempting Clairvoyance with Battery Performance

The lightweight , long-lasting, high-performance attributes of cellular phones and laptop computers, among other equally impressive portable devices currently in the marketplace, are responsible for igniting the overwhelming growth of the battery-powered electronics industry. The demand for smaller and longer lasting solutions, in fact, is only increasing, and key to this success is the battery, which can range from single-use alkaline and zinc-air to rechargeable nickel-cadmium, nickel-metal hydride, lithium-ion, and lithium-polymer technologies. Unfortunately, however, advancements in circuit and system integration have outpaced energy and power density improvements in the battery. Consequently, as batteries conform to the size constraints of portable applications, capacity and output power are necessarily compromised. Degradation in battery performance over time not only affects functionality but also operational life, proving inadequate the traditional assumption that the battery is an ideal voltage source. Including the effects of the battery on state-of-the-art systems during the design phase is therefore of increasing importance for optimal life and performance. The problem is securing a suitable Cadence-compatible model. Battery Models State-of-the-art electrical models for batteries are either Thevenin-, impedance-, or runtime-based. Thevenin-and impedance-based models, shown in Figures 1(a)-(b), assume both open-circuit voltage and capacity or state-of-charge (SOC) are constant and approximate loading and ac/transient effects with an impedance network of passive devices for transient response and/or curve-fitting impedance blocks (for example, Z AC) derived from electrochemical-impedance spectroscopy experiments for ac response. Both these models concentrate on either transient or ac response, but neither considers temperature, lifetime limits, or steady-state open-circuit voltage variations (that is, DC effects). Reported runtime-based models (Figure 1(c)), on the other hand, predict operational life and steady-state variations of the open-circuit voltage, but at the cost of complexity and therefore increased simulation time, which is why an impedance matching network for predicting transient and/or ac response is usually forfeited, yet this is exactly what is needed to fully comprehend the degrading characteristics of the battery.