MENU

Preventing the nasty surprise of an unexpected system shutdown

Preventing the nasty surprise of an unexpected system shutdown

Technology News |
By eeNews Europe



The irritation and inconvenience brought about by unexpected system shutdowns could soon be a distant memory as new technology for battery fuel gauging offers the potential for a far more accurate picture of how much charge is truly left.

Mobile devices in their many forms underpin much of modern life. Whether it is using a laptop in a public place, a smartphone while out of the office, or even watching a film on a tablet while travelling, productivity, connectivity and entertainment all depend upon the remaining battery life of the device.

However, when the remaining charge begins to drop low, the percentage shown on the device can offer little more than guesswork and could let the user down at a vital moment. This can be particularly notable for laptop users who might lose vital, unsaved work under the assumption that their device has not yet reached a critical point.

Given the importance to the user of knowing accurately how much energy is remaining in the battery at any time, accurate battery ‘fuel-gauging’ is highly important. However, existing methods are not only inaccurate, leading to a greater chance of an unexpected shutdown, but can also be subject to errors caused by temperature and will consume additional battery energy at a point when it is most desperately needed.

The traditional approach
Coulomb-counting has for a long time been the most common method used in mobile devices to inform the user of just how much charge remains within a device. This method uses a precision current-sensing resistor to continuously monitor the battery output current. The current is integrated over time, and the result is compared with the known maximum battery charge to calculate the remaining charge available.

The biggest issue with Coulomb counting is its inherent inaccuracy. An inability to detect battery self-discharge events is owed to the self-discharge current not passing through the coulomb counter’s sense resistor. Further disadvantages of coulomb counting include the relatively high cost of the precision sense resistor, as well as the precious battery energy dissipated by this resistor as the sense current passes through continuously.

The level of accuracy offered by a coulomb counter is likely to be accurate to circa 8%. Hence if the indicator suggests 10% of charge is remaining, the real value may be as little as 2%. This is not acceptable for most users and could lead to substantial loss of work when using a laptop or tablet and for a phone user could expedite the loss of what are classed as non-essential features such as GPS or Bluetooth, and eventually the ability to make calls. As such even dropping below 20% battery could concern the user as they might have only a limited time to use the features they require.

 
Figure 1: With a lack of accuracy in determining how much battery life is remaining, system shutdown or a loss of features can occur suddenly

Improving accuracy in any setting or environment
To address many of these issues, ON Semiconductor has developed a proprietary method, called ‘HG-CVR’ (Hybrid Gauging by Current-Voltage tracking) to calculate the remaining battery energy based on the voltage measured across the battery using a precision analogue-to-digital converter (ADC).

By storing a reference table which more accurately represents the voltage-vs-capacity characteristic of the battery technology being monitored, it is possible to make a more representative estimate of the conditions. By then comparing the measured voltage with values stored in the table, the remaining battery capacity can be calculated. Figure 2 illustrates the principle. Quite simply, if the measured voltage is 4.0V, comparison with the reference table suggests 75% of battery charge is remaining.

 

Figure 2. The remaining battery capacity is inferred from the measured voltage by referring to the battery reference table.

Accuracy can then be further ensured over a period of time by taking repeated voltage measurements at known time intervals. The battery temperature is also monitored. The time remaining before the battery will be completely depleted can then be calculated, based on the voltage and temperature measurements and the changes in the voltage recorded at the known time intervals. More frequent readings are taken at lower battery voltages, to ensure accurate predictions as the remaining battery life becomes shorter.

By measuring the voltage across the battery pack, this approach is able to take account of battery self-discharge events. Moreover, the battery does not need to be fully charged for calibration.  The remaining battery life can be calculated accurately even if the battery is only charged to 50%.

A major advantage of taking these measurements at pre-set intervals is the lack of need for the monitoring circuitry to be continuously active. This allows the fuel gauging circuitry to enter energy-saving sleep mode in between measurements. The active power consumption can also be reduced, in comparison with a conventional coulomb counter, since no sense resistor is needed.

Li-ion batteries also have an unfortunate tendency to be susceptible to environmental conditions and changes in ambient temperature. In particular, the battery impedance changes as the temperature falls to 0°C or lower, resulting in increased battery voltage drop when discharge current flows.

ON Semiconductor has added a unique correction algorithm in its LC709203F battery-voltage sensing fuel gauge IC. This algorithm helps ensure fuel-gauging accuracy remains within 2.8% over a wide ambient temperature range and at all battery voltages.

The effectiveness of this fuel gauge was comprehensively tested by ON Semiconductor. With the performance of a coulomb counting circuit, a smartphone with a new battery was adapted to allow the positive and negative battery connections and the thermistor output of the battery pack to be connected to a LC709203F while still allowing the smartphone’s built-in fuel gauge to continue operating. Data loggers were used to record the output of the built-in fuel gauge, which was monitored via the smartphone I2C bus, and the output of the LC709203F. The smartphone was placed in a constant-temperature tank at 0°C, and was then run in aircraft-safe mode with the backlight on. The layout can be seen below in Figure 3.

 
Figure 3. Comparison of built-in smartphone battery fuel gauge and voltage measurement technique using LC709203F.

The results demonstrated the accuracy of the LC709203F as being greater than 2.8% for the whole duration of the test, and that it was better than 2% at the lowest levels of remaining battery energy.

In comparison, the coulomb counting standard fuel-gauge system operated with varying levels of error and reached its highest level of more than 6% as charge fell closer to zero. From the perspective of the user this is the most crucial time for accuracy as they require a clear picture of how long they have to use their device before encountering a full system shut down.

Low power, small size
In comparison to alternative devices, which may require as many as 14 additional components, the LC709203F allows effective fuel gauging using only one external component. This delivers a valuable saving in costs and design time, and also increases reliability.

Crucially, with a package which measures 1.76mm x 1.6mm it is about 45% smaller than alternative devices. Combined with the reduced component count, this allows the overall PCB footprint of the fuel gauge circuit to be reduced by around 77%. In devices such as smartphones and tablet, where internal space is inherently limited but ever more features are demanded by users, this can be a considerable advantage.

With an operating current of 15µA, which is approximately 1/10th that of the closest competing device (118µA), the LC709203F offers noticeable lower power consumption. In addition to this improvement of over 87% in active consumption, the LC709203F draws up to 60% lower current in sleep mode.

Conclusion
By more accurately sensing the remaining battery life of a device the user can be confident that when approaching low battery states they will not find themselves inconveniently cut-off.

This new technique that makes use of precision battery voltage sensing, with error correction and temperature compensation built in, provides users with a more accurate, cost-effective and energy-efficient solution. 

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News

Share:

Linked Articles
10s