Smartphone Batteries
From CSL Wiki
Contents |
Experimental Study of Thermal and Recovery Effects in Smartphone Batteries
Despite the popular belief among mobile application developers, batteries are not ideal sources of energy. The following factors affect the actual capacity of rechargeable batteries 2 that are used in today’s smartphones:
- Rate-capacity effects: The actual capacity of a battery is lower when it is discharged at a higher rate.
- Recovery effects: If a battery is allowed to rest for a while it will recover some of its lost capacity.
- Thermal effects: Batteries do not operate well in very high or very low temperatures.
- Capacity fading: Batteries lose some of their capacity with each full charge-discharge cycle.
These effects make battery models much more complicated than the simple linear discharge models assumed by most developers. Several models have been proposed for batteries ranging from very complex models with hundreds of parameters to simpler ones with a few parameters. I will review existing models in the related work section.
Batteries in Real Life
Almost all the existing battery models are concerned with synthetic workload generated by embedded electronics. Smartphones are novel embedded systems because they are used by human users. Users interactions with their phones are nondeterministic and complex. These interactions, when combined with the non-linear behavior of batteries, result in complicated patterns.
Figure 1 plots battery level, temperature, and voltage of a real user’s G1 smartphone. The data is from traces collected during daily usage. In this figure, the top plot has a point every time the user turns the screen on (a sign of interaction with the phone) These plots show that although the battery level is reported to be constantly decreasing, the voltage and temperature of the battery are fluctuating in close correlation with user interaction bursts.
Another factor that I believe contributes to fluctuations in voltage and temperature of the battery is the confined form factor of mobile phones. While in many embedded systems that are deployed in the field, the heat can easily escape, the material and small form factor of mobile phones prevent effective heat dissipation. As a result the temperature of the phone battery quickly increases, when the phone is used. We do not clearly know if this rise in temperature works against the battery capacity or not.
Weekly Progress Reports
The following are weekly progress reports.
Oct 19 - Oct 25
Done:
- Checked out a G1 smartphone from CENS for the study. I will charge the battery inside the phone (for safety purposes). But I will then remove the battery from the phone and discharge it separately.
- Evaluated the existing equipment to control temperature. I concluded that I will not be able to accurately control temperature. Instead I will conduct the experiments in an environment with high thermal capacity so that the heat generated by the experiment itself will have non-significant effect on the temperature. I will constantly monitor the ambient temperature during each test to make sure it does not vary too much.
- I was briefly introduced to the voltage and current measurement setup at NESL. As per Thomas's suggestion I am trying to start using the measurement equipment of CENS. There is a National Instrument data acquisition unit at CENS. I am trying to learn how to use it and set it up (it has not been in use for a while).
- I cut a battery holder in styrofoam to ease connecting wires to the battery.
- Reading related work.
Next week plans:
- Setting up and learning how to use the data acquisition unit at CENS
- Initial experiments with discharging the battery using resistors. I hope I can find variable resistors otherwise I will use different combinations of fixed resistors.
- Performing discharge experiment while measuring voltage, current and temperature in room temperature as the first experiment.
Oct 26 - Nov 1
Done:
- Took apart an old and broken Nokia smartphone and used its battery pins in the Styrofoam battery holder.
- Consulted CENS senior students and Richard Guy to see if there is any voltage and current measurement at CENS. I found that the equipment is old and the overhead of setting them up would be too high. Therefore I decided to use the measurement set-up at NESL.
- Could not start the experiments because members of the NESL lab who know how to use the the voltage and current measurement instruments were busy and are now at SenSys.
- Richard Guy suggested that I might be able to use a controlled temperature chamber that he knows of in another department at UCLA. We are still investigating the possibility.
Next week plans:
- Setting up and learning how to use the data acquisition unit at NESL
- Initial experiments with discharging the battery using resistors. I hope I can find variable resistors otherwise I will use different combinations of fixed resistors.
- Performing discharge experiment while measuring voltage, current and temperature in room temperature as the first experiment.
- Further investigating the possibility of using a controlled temperature chamber at UCLA.
Nov 2 - Nov 8
Done:
- Found that I cannot use my Styrofoam battery holder during controlled temperature experiments, because it is too good an insulator.
- I was oriented with temperature chamber and got the permission to use it from next week.
- Used fixed resistors to discharge the battery at a fixed rate.
Next week plans:
- Learn how to use the voltage/current measurement setup at NESL
- Run the temperature experiments
Nov 8 - Nov 15
Done
- Learned how to use the voltage/current measurement setup at NESL
- Ran the temperature experiments. I still have to run more experiments.
Next week plans:
- complete temperature experiments
- Run recovery effect experiments
