The Role of Digital Power in Portable Applications
By minimizing the high-speed circuit components and taking advantage of smaller process geometries, digital power solutions have become practical for portable appliances.
Power management is vital to extending runtime for portable applications. Power conversion in these applications has been dominated by analog circuits, while the power management has been performed by embedded processors and simple sequencers. These analog solutions have served the system well. With recent trends in power solutions focusing on digital control, new opportunities emerge for portable power solutions.
This article presents digital power solutions for portable applications that take advantage of digital power’s combination of management and control. When management and control are separated, the interfaces between these circuits limit flexibility and the level of complexity that can be addressed. Digital power solutions integrate these functions and are capable of addressing much higher levels of complexity. Portable applications have many operational modes. Digital power solutions allow each mode to be powered uniquely and, therefore, optimized for best performance. Examples include adaptive compensation and phase management in portable applications.
Digital Aspects of Portable Applications
Figure 1: Power management benefits in sub-micron design
The developments in deep sub-micron processes have greatly benefited portable applications such as multimedia cell phones and notebook computing. However, these smaller geometries come with their challenges. Power has become a major issue. There are very effective solutions to decreasing leakage current and methods for improving performance without increasing the required power. As shown in Figure 1, power management can provide nearly an order of magnitude reduction in quiescent current. This standby and sleep current is reduced by dynamic voltage and frequency scaling (DVFS), reverse body bias (RBB), and power domain management (PDM).
Both RBB and PDM require that application circuit design integrate these functions. PDM is managed by the system application.
The active power can be reduced by employing a technique known as forward body bias (FBB). This method which in integrated into the design reduces the power required for fast clocking. Typically it is managed by the system and used when high speed is required. Coupling this technique to DVFS can reduce active power by 25 percent while also reducing the bias power.
These circuit techniques help improve standby and runtime in portable applications. However, they do require collaboration between the power sub-system and application system. There are certainly analog power solutions that offer a system interface for voltage adjustment as well as power shut-down. However, digital power control provides a greater degree of power control and management that can be used to address higher levels of power management complexity.
Digital Power Solutions
Figure 2: Digital Power Controller
Figure 2 depicts a general digital power controller. Such a controller may use hardwired state-machines or embedded CPU for general processing of communication, configuration, and housekeeping tasks. In today’s mixed signal processes, embedding a CPU requires much less silicon area compared to only a few year ago, so the flexibility of a more general solution does not add appreciably to the total cost. Such CPUs are similar to those used in portable applications for general user interface and system management, and have high MIPS per milliwatt rating. Many digital power solutions also embed non-volatile memory for configuration as well as recording real-time system information.
The digital controller typically has at least one communication interface. An industry standard interface is PMBus. This interface is similar to SMBus which is often found in portable devices for battery management. The PMBus can be used to monitor the various power parameters enabling the system to make choices about operating modes. The PMBus also can be used to configure the digital control prior to operation to set output voltage, sequencing, and response to operating and fault conditions. During runtime, the PMBus can be used to adjust the power supply for the desired output voltage as well as other power dynamic characteristics.
Figure 2 also shows other functions generally provided in a digital power solution such as a general purpose ADC for analog signal monitoring, support circuits such as references and oscillators, and general purpose I/O and timers. These peripherals can be configured to provide the required functions for housekeeping and system control.
The primary differentiator is the digital compensation controller. This circuit is comprised of a high-speed analog-to-digital block, a digital compensator block, and a digital pulse width modulator block. A digital power integrated circuit may have more than one of these compensators. The advantage of multiple compensators/output control blocks is that sequencing and tracking can be very flexible and configured within one IC. Multiple blocks can share common resources such as references, oscillators and system interface hardware.
The digital compensator block generally has programmable loop compensation registers. The registers are used to configure the compensation filter for the proper loop response. The advantage of a configurable filter is that it can be adjusted for various operating modes. For example, in lower power modes the switching frequency may be reduced. In this case, the compensator can be adjusted to provide the proper loop response given this lower power and frequency mode.
Digital Flexibility in Portable Applications
Integrating more and more functions in portable applications has become the norm. Along with this functionality comes a broad range of operating modes and active power. Digital control allows the designer to adapt to these conditions. One example of adaptability is the ability to perform non-linear compensation. A simple approach to non-linear is to express the compensation as PID (proportion, integral, differential) as shown in Equation 1.
Equation 1. PID Compensation:
Kp is the proportion gain
Ki is the integral gain
Kd is the differential gain
The Kp can be made a function of the difference between the desired voltage and the measured voltage, the error term. Using the values from the high-speed data converter in the loop compensation path, Kp can obtain from a table based on the magnitude and direction of the error. Figure 3 shows the case where small error values use linear compensation (Kp held constant). But as the error becomes great due to a transient, the gain is increased starting at errors of ±20mV. This non-linear method can be applied to compensate during operating modes where the load will vary greatly. Another aspect of digital control is the ability to switch between compensators as a function of operational modes. For example, during standby or sleep modes, compensation may be selected for stability but with a narrow linear compensation range so that if the mode changed suddenly, the compensation could maintain the desired output.
Figure 3: 2A to 12A transient, blue is linear and red is non-linear
For applications using multiphase regulators like those for processors and graphics processors, digital provides phase management solutions that allow phase shedding and adding based on load requirements. Although phase management can be done with analog power solutions, digital solutions can be more adaptive. Even in a two-phase solution, there are benefits in selecting the best phase of the two for low power operation. Additionally, temperature inputs to phase management may be important. Simply balancing the current between two-phases may result in one phase continuing to get hotter, given the positive temperature coefficient of the loss components. Hence, in the digital multiphase controller, phase temperature can be added to the phase current balancing algorithm.
Benefits from a digital power solution are derived from the flexibility of the solution. Although digital power solutions have been used for more than a decade, it has only been recently that the solutions have been able to address high-frequency switch-mode operation required in portable applications. However, generating the required duty-cycle resolution at high frequencies must be done with low power operation in mind. By minimizing the high-speed circuit components and taking advantage of smaller process geometries, digital power solutions have become practical for portable appliances.
The combination of non-linear compensation with complex decision configuration, digital solutions can help reduce power losses during conversion as well as control the operational voltage to take advantage of other system level power saving methods. As the power industry continues to develop digital power controllers these devices will become an important part of portable power solutions.
About the Author
David Freeman is an Engineering Manager in System Power Products at Texas Instruments where he manages systems engineering and is a core member of TI’s digital power team. David has 16 years experience in power-related areas with a strong focus in portable power management, is a Texas Instruments Fellow, and has 18 patents related to power and battery management.
Texas Instruments Inc.
This article originally appeared in the July, 2008 issue of Portable Design. Reprinted with permission.