Where does the battery of mobile devices consume

Battery life is an important factor in the design of mobile devices. As many mobile devices add more features, these added features quickly reduce operating time. Engineers must use complex power management schemes to get the longest battery life.

Engineers need to use battery consumption analysis to estimate battery run time, which requires characterizing the device, firmware/software, and its subcircuits both individually and integrated into the system. Analysis techniques include characterizing battery current consumption and how it is affected by various operating modes and usage profiles. With this analysis, engineers can make power management design trade-offs to maximize battery life.

Most power management systems conserve battery energy by putting subsystems that are not actively in use to sleep on a sub-millisecond time scale. As a result, devices have rapidly changing currents during on/off events that occur in less than 1 second. For example, a GSM phone can have a 560μs, 2A pulse when transmitting, and then when in standby mode, the current level may drop to milliamps during the sleep cycle.

Verify battery time

One way to verify battery run time is to use a voltage drop test, using a fully charged battery to power the device under test (DUT) to be verified in operating mode until the battery is discharged. This test can be relatively time consuming, as it requires the entire process to be run through to determine the voltage shutdown point to determine the run time. Again, the results are dependent on the initial state of the battery, which can vary widely.

Another approach is to perform current consumption measurements, which can provide higher confidence in the operating time measurement. The DUT is placed in the operating mode to be evaluated for a short period of time, and the current consumption in this specific operating mode is measured. The operating time is then calculated by dividing the nominal battery capacity by the measured current consumption. Using this method, the designer can determine the operating time without waiting for the battery to fully discharge.

Components of an ideal system

In an ideal system for performing battery drain analysis (as shown in Figure 1), the first element required is a method to put the DUT into the appropriate operating mode for the target test (DUT stimulus). For mobile phones, a base station emulator is typically used.

Figure 1. Several components that exist in a general ideal system for battery current drain measurement and analysis.

Secondly, a proper method of powering the DUT is required, either with a battery or a power supply. The purpose of a power supply is to test the DUT independently of the battery to ensure test consistency, or to quickly replicate various battery states without waiting for the battery to reach those states (fully charged, partially discharged, fully discharged/end of life).

Other important system components are current converters for measuring current, digitizers for recording voltage and current signals, and software for analyzing and storing the test data, which can be very large, amounting to several gigabytes, to complete long-term tests.

Measurement Considerations

The power supply used in battery drain analysis must characterize the DUT independently of the battery. The power supply must have a fast response to minimize transient voltage drops caused by fast slewing current pulses that the DUT has when switching modes or transmitting pulses.

Many general-purpose power supplies can drop as much as 1V transiently under these conditions, so a dedicated power supply that can tolerate these conditions without voltage drop (sometimes called a battery emulation power supply) should be used.

The rapidly changing current waveform flowing from the battery to the mobile device presents two measurement challenges: range and speed. First, the dynamic range of current can exceed 1000:1 or even 1 000 000:1. Full power active current is on the order of 1 to 3A, while low sleep mode levels are on the order of tens of microamps, so the range of currents to be measured presents a challenge in the selection of current converters.

Current-sensing resistors or shunts can be used here, but choosing the right shunt size can be tricky. If the shunt is sized for the smallest currents, a large voltage drop will appear across the shunt during a high-current event, which will place an unacceptable voltage burden on the circuit. If the shunt is sized for large currents, there is a high probability that there will not be enough voltage to measure when microamperes flow. Engineers can solve the signal level problem by having several shunts for different current levels, but then switching shunts means interrupting the measurement.

In terms of measurement speed, digitizers used to measure current shunt voltages and mobile device bias voltages should have sampling rates of 50kHz or faster to capture submillisecond pulses that are characteristic of complex power management schemes.

Simplify complex analysis

Communication systems such as 3G use complex modulation formats characterized by high-order amplitude modulation required to transmit higher data rates. Viewed in the time domain, the resulting current consumption waveform is complex and random.

The current consumption vs. time graph of the RF power amplifier of a cdma2000 handset transmitting using three data channels (see Figure 2) becomes complex and unpredictable when running for extended periods of time and with different operations. This is common for battery run-time testing, and the effects of changing the current consumption design are difficult to observe.

Figure 2. Viewed in the time domain (left), the current consumption waveform of the RF power amplifier of a cdma2000 mobile phone is complex and unpredictable. Viewing the same current waveform as a CCDF graph (b), designers can easily see how often the device is in each current state.

A better way to visualize and analyze complex current consumption patterns is to use a complementary cumulative distribution function (CCDF) plot to view its statistical distribution. In a CCDF plot, the x-axis represents current and the y-axis represents its cumulative occurrence percentage (as shown in Figure 2b).

By looking at the statistical distribution of the amount of current being drawn, designers can quickly see how often the device operates in each current state. Comparing these CCDF graphs for different designs shows when the device consumes more energy (that is, an increasing proportion of its time in the high current state) or when it consumes less energy (that is, an increasing proportion of its time in the low current state). As a result, engineers can assess when a design is better (requiring less energy) or identify design flaws (unexpectedly requiring more energy).

Ready-made solutions

Several test equipment vendors produce products that address different parts of the target test system. Some vendors offer power supplies that can provide a stable, battery-like output while drawing fast current pulses.

The entry-level solution is the Mobile Communications DC Source 66300 Series from Agilent Technologies. This series is specifically tailored to power mobile devices and simultaneously measure their current consumption. It combines a battery emulation power supply with a high-speed digitizing measurement system similar to an oscilloscope to provide accurate current measurements for the device's active, standby, and off modes.

This DC source and its accompanying turnkey software allow users to see their current waveforms in an oscilloscope-like view, data logger view and on CCDF charts without any programming. If higher accuracy and sampling rate are required, other solutions provided by the company can also be selected.