Analog and digital methods for power management

The long-standing debate between digital and analog has recently expanded into the power supply field, causing attention and response in the analog field. The numbers are still popular, but the real world is analog. And the power supply is no different from any other real-world system: the power output is analog ( voltage, current, power, etc. ) , which is analog output with digital TV, digital camera and digital cellular phone ( mobile video, still image, sound) Etc. ) have the same way. In addition to the obvious communication aspects that have been highlighted for quite some time, digital power supplies are simply related to some of the processing units at the heart of the power conversion system and are now being improved for digitization. However, there is no doubt that the interface to the analog world – the outer layer remains analog.

An analog circuit is a device that represents continuously variable and can measure physical quantities ( such as length, width, voltage, and pressure ) . However, in the analog market segmentation, we see data conversion and interface products as devices with digital data and circuit components.

In fact, in the analog world, we see a wide range of mixed-signal applications ranging from pure analog to digital and from power to signal. In fact, simulation is a rich and basic element in generating new circuits, new structures and new solutions in every technological journey. As shown in the market intelligence company's data report, simulation has become a large mixed-signal field that spans all analog circuits from pure analog to digital.

The simulation continues to grow, regardless of every actual phenomenon of digitization. This is because the reasons and simulations described above remain an unparalleled area of ​​innovation: 10 years ago without a capacitive regulator ( charge pump ) , 5 years ago without LED drivers, this innovation will continue. As a result, the digitization of pure analog circuits maintains home growth in a broadly defined analog market, confirming the leading innovation capabilities of the simulation. To be sure, no simulation can be digitalized without simulation.

Digital limitations

Figure 1 shows a block diagram of a general power conversion and management system. The elements in the block diagram ( power reference, D/A , driver, filter ) are the features that connect the "external" analog world, and for the above reasons they will remain analog. The communication unit in the block diagram is digital ( serial or parallel communication bus ) . The control unit, traditionally implemented using analog methods, has changed to digital implementation in the last five years.

Today's industry trends indicate that digital control structures for power conversion ( servo control algorithms ) and power management ( new serial or parallel bus protocol communications, sequencing circuits, etc. ) are maturing. In the next few years, these structures are predicted to replace the analog counterparts.

The power supply continues to be in a tough environment, putting a lot of pressure on semiconductor devices. The inductor in the switching regulator or the coil in the motor periodically energizes the electronic circuit with a voltage spike greater than the VCC supply and below it. Such overvoltage and undervoltage offsets are bound to wake up parasitic transistors in semiconductor devices, which can have a detrimental effect on the system. How to make these harmful effects not connected to the outside world is not the scope of digital electronics. This is a fairly difficult problem to solve, even for the most experienced analog designers. In fact, parasitic parametric problems turn power / analog design into an art, not a science. There are no SPICE simulators that can simulate the three-dimensional effects of parasitic transistors, and as long as this continues, the simulation will continue to be the "black magic" in the hands of a few skilled designers.

Analog and digital structure

Figure 2 shows a block diagram of a typical analog control implementation of a voltage regulator that builds a pulse width modulation (PWM) switching regulator around the modulator . The analog modulator consists of a comparator (the modulation waveform is shown in Figure 3) , and the comparator input is a periodic piecewise linear ( triangular or sawtooth ) modulation waveform VST of period T and the other input is the error signal V ε. If the quasi-steady-state error signal V ε is between the minimum and maximum of the modulation waveform, the intersection of the two waveforms determines the period Ton of the ' on ' pulse . Therefore, the comparator output produces a square wave Vsw whose average value is the same as the DC output voltage Vo . In this method, the PID ( proportional - integral - derivative ) unit can be implemented with an op amp and external passive components ( compensation resistor Rc and capacitor Cc) or with a single chip integrated with Rc , C2 compensation networks.

Figure 4 shows a digital control structure, wherein the input error signal (Vfb-Vref) by the analog / digital converter (ADC) into a digital signal, and thereafter PID compensator digital modulation (DPWM).

At the heart of the digital power conversion control loop is a digital modulator. Figure 5 shows a digital modulator method implemented with a ring oscillator, which is a simple and efficient method. In this example, the ring oscillator operates at 1MHz (T = 1ms) , which is also the clock frequency of the digital PWM system. The ring oscillator consists of 255 circuits ( in the simplest implementation of ring oscillation, the number of gates Must be odd ) to correspond to 8 -bit resolution. Each gate output is delayed by 1/255 clock cycles from the previous gate , approximately 4ns .

By properly selecting the time delay between the gates, an ' on ' pulse at the output of the digital modulator can be generated, which can be made by a digital selector driven by the digital error signal voltage DV ε.

Selection control algorithm

If the system being regulated is truly linear, which means that the mode of operation is continuous or stationary, then the simulation is usually the method used. This is the case with desktop CPU voltage regulators, where the regulator output must be continuously controlled using the same algorithm from no load to full load. If the system is not stable, which means that the working mode is discontinuous and changing, then it is better to choose the digital method.

For example, in a notebook or mobile phone voltage regulator application, it is better to choose a digital method. Because power should be saved at light loads, mode changes are required at this time. This usually happens from the PWM algorithm to the PFM ( Pulse Frequency Modulation ) . PFM is a mode that adjusts the frequency with the load, thus producing a lower frequency at lighter loads and thus lower switching losses.

In analog systems, such mode changes require a sudden transition from one control loop ( such as PWM) to another ( such as PFM) , where the load is changing. This algorithm is not continuous and must result in a temporary decrease in output stability.

Instead, digital controls inherently configure to handle discontinuities. Therefore, digital control has the ability to handle mode changes in a single control algorithm.

Power management and conversion applications

The obvious advantages of digital power management are ease of communication, programming, status reporting, and more. A typical application example of such digital control is a smart battery system, a smart battery charger that powers a notebook computer. This system includes a smart charger, a smart battery and a main microcontroller. In this system, the smart charger slave receives commands from the "master" controller through the system management bus (SMBus) . The smart charger then adjusts its parameters to provide the required current, voltage, and power to the smart battery and then reports its value to the microcontroller.

In power-conversion applications, microcontroller-based digital structures have many useful uses, especially in applications that require more than just programmability and current and voltage shaping.

Current shaping application

Current shaping is required in light-duty ballast applications, and the intensity and period of the current and the three operating phases ( preheat, ignition and dimming ) can be flexibly set for different lamps . Current shaping is also required in PFC applications, and the current must have the same shape as the line voltage.