Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

In general, the higher the switching frequency, the smaller the size of the output filter elements L and CO. Therefore, the size of the power supply can be reduced and the cost thereof can be reduced. Higher bandwidth also improves load transient response. However, higher switching frequencies also mean higher AC-related power losses, which require more board space or heatsinks to limit thermal stress. Currently, most step-down power supplies operate from 100kHz to 1MHz ~ 2MHz for output current applications ≥10A.for

By Henry J. Zhang, Analog Devices

Switching frequency optimization

In general, the higher the switching frequency, the higher the output filter elements L and COthe smaller the size. Therefore, the size of the power supply can be reduced and the cost thereof can be reduced. Higher bandwidth also improves load transient response. However, higher switching frequencies also mean higher AC-related power losses, which require more board space or heatsinks to limit thermal stress. Currently, most step-down power supplies operate from 100kHz to 1MHz ~ 2MHz for output current applications ≥10A.for

Output Inductor Selection

In a synchronous buck converter, the inductor peak-to-peak ripple current can be calculated as:

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

At a given switching frequency, the low inductance provides a large ripple current and produces a large output ripple Voltage. Large ripple currents also increase MOSFET RMS current and conduction losses. On the other hand, high inductance means that the inductor size is large, and the inductance DCR and conduction losses may also be high. Typically, when selecting an inductor, choose a peak-to-peak ripple current that exceeds the maximum DC current ratio by 10% to 60%. Inductor suppliers typically specify DCR, RMS (heating) current, and saturation current ratings. It is important to design the inductor’s maximum DC and peak currents within the supplier’s maximum ratings.

Power MOSFET Selection

When choosing a MOSFET for a buck converter, first ensure that its maximum VDSrated above the supply V with sufficient headroomIN(MAX). However, don’t choose a FET with an overly high voltage rating. For example, for 16VIN(MAX)Power supplies, FETs rated at 25V or 30V are ideal. A 60V rated FET is overvoltage because the on-resistance of a FET generally increases with voltage rating. Next, the on-resistance R of the FETDS(ON)and gate charge QG(or QGD) are the two most important parameters. usually requires a gate charge QGand the on-resistance RDS(ON)trade-offs between. In general, FETs with small silicon die size have low QG, High on-resistance RDS(ON), while the large silicon die size FET has a low RDS(ON)and big QG. In a buck converter, the top MOSFET Q1 absorbs both conduction losses and AC switching losses. Q1 usually requires low QG FETs, especially in applications with low output voltages and small duty cycles. The AC losses of the low-side synchronous FET Q2 are smaller because it is usually at VDSTurns on or off when the voltage is close to zero. In this case, for synchronous FET Q2, a low RDS(ON)than QGmore important. Multiple MOSFETs can be used in parallel if a single FET cannot handle the total power.

Input and Output Capacitor Selection

First, capacitors with adequate voltage derating should be selected.

The input capacitance of a buck converter has pulsating switching currents and large ripple currents. Therefore, an input capacitor with sufficient RMS ripple current rating should be selected to ensure lifetime. Aluminum electrolytic capacitors and low-ESR ceramic capacitors are usually used in parallel at the input.

The output capacitance determines not only the output voltage ripple, but also the load transient performance. The output voltage ripple can be calculated by Equation (15). For high performance applications, to minimize output ripple voltage and optimize load transient response, both ESR and total capacitance are important. Generally, low-ESR tantalum capacitors, low-ESR polymer capacitors, and multilayer ceramic capacitors (MLCCs) are good choices.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Close the feedback regulation loop

Switch-mode power supplies also have an important design stage—closing the regulation loop through a negative feedback control scheme. This task is usually more challenging than using LR or LDO. It requires a good understanding of loop behavior and compensation design to optimize dynamic performance by stabilizing the loop.

Small Signal Model of a Buck Converter

As mentioned earlier, switching converters change operating modes as the switches are turned on or off. It is a discrete nonlinear system.To analyze feedback loops using linear control methods, linear small-signal modeling is required[1][ 3]. Due to the output LC filter, the duty cycle D to the output VOThe linear small-signal transfer function of is actually a second-order system with two poles and one zero, as shown in Equation (16). There are two poles at the resonant frequency of the output inductor and capacitor. There is a zero determined by the output capacitor and capacitor ESR.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

in,Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components,

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Voltage Mode Control vs Current Mode Control

The output voltage can be regulated by a closed-loop system, as shown in Figure 11. For example, when the output voltage increases, the feedback voltage VFBincreases while the output of the negative feedback error amplifier decreases. Therefore, the duty cycle is reduced. The output voltage is pulled back so that VFB = VREF.The compensation network for the error op amp may be a Type I, Type II, or Type III feedback amplifier network[3] [ 4]. There is only one control loop to regulate the output. This scheme is called voltage mode control. The ADI LTC3775 and LTC3861 are typical voltage mode buck controllers.

Figure 12 shows a 5V to 26V input, 1.2V/15A output synchronous step-down power supply using the LTC3775 voltage-mode buck controller. Because of the LTC3775’s advanced PWM modulation architecture and extremely low (30ns) minimum on-time, this power supply is suitable for converting high-voltage automotive or industrial power supplies to the low-voltage 1.2V required by today’s microprocessors and programmable logic chips. application. High power applications require multiphase buck converters with current sharing. With voltage mode control, an additional current sharing loop is required to balance the currents in the parallel buck channels. A typical current sharing method for voltage mode control is the master-slave method. The LTC3861 is such a PolyPhase®voltage mode controller. Its ultra-low current-sense offset voltage of ±1.25mV enables precise current sharing between parallel phases to balance thermal stress.[10]

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 11. Block Diagram of a Voltage Mode Controlled Buck Converter

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 12. LTC3775 Voltage Mode Synchronous Buck Supply Provides High Step-Down Ratio

Current-mode control uses two feedback loops: an outer voltage loop similar to the control loop of a voltage-mode control converter, and an inner current loop that feeds the current signal back into the control loop. Figure 13 shows a conceptual block diagram of a peak current mode controlled buck converter that directly senses the output inductor current. When using current-mode control, the inductor current depends on the output voltage of the error op amp. The inductor becomes a current source. Therefore, from the op-amp output VCto the power supply output voltage VOThe conversion function becomes a unipolar system. This makes loop compensation much simpler. Control loop compensation is less dependent on the output capacitor ESR zero, so all ceramic output capacitors can be used.

There are many other advantages of current mode control. As shown in Figure 13, since the peak inductor current is affected by the op amp VCcycle-by-cycle limit, so the current-mode control system will limit the current more accurately and quickly during overload conditions. Inrush inductor current is also well controlled during startup. In addition, the inductor current does not change rapidly when the input voltage changes, so the power supply has good line transient performance. By using current mode control, it is also easy to achieve current sharing between supplies when paralleling multiple converters, which is critical for reliable high current applications using PolyPhase buck converters. All in all, current mode controlled converters are more reliable than voltage mode controlled converters.


Current-mode control schemes require accurate current sensing. The current sense signal is usually a small signal at the level of tens of millivolts sensitive to switching noise. Therefore, the PCB layout needs to be designed correctly and carefully. The current loop can be closed by sensing the inductor current through a sense resistor, inductor DCR drop, or MOSFET conduction drop. Typical current-mode controllers include Analog Devices’ LTC3851A, LTC3855, LTC3774, and LTC3875.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 13. Block Diagram of a Current Mode Controlled Buck Converter

Constant frequency and constant on-time control

The typical voltage-mode and current-mode schemes in the “Voltage-Mode Control vs. Current-Mode Control” section have a constant switching frequency generated by the controller’s internal clock. Easy synchronization of these constant switching frequency controllers is an important feature of high current PolyPhase buck controllers. However, if the load boost transient occurs just after the control FET Q1 gate turns off, the converter must wait the entire Q1 off time before responding to the transient until the next cycle. In applications with a small duty cycle, the worst-case delay is close to one switching cycle.

In such low-duty-cycle applications, the current-mode control responds to load boost transients with shorter delays at constant on-time valleys. In steady-state operation, the switching frequency of a constant on-time buck converter is nearly fixed. If a transient occurs, the switching frequency can be changed rapidly to speed up the transient response. As a result, the power supply improves transient performance and reduces output capacitance and associated costs.

However, with constant on-time control, the switching frequency may change with line or load changes. Analog Devices’ LTC3833 is a valley current mode buck controller with a more complex on-time control architecture, which is a variant of the constant on-time control architecture, except that it controls the on-time so that the switching frequency is remains constant under stable line and load conditions. Using this architecture, the LTC3833 controller has a minimum on-time of 20ns and supports 38VINto 0.6VOof step-down applications. The controller can be synchronized to an external clock over a frequency range of 200kHz to 2MHz. Figure 14 shows a typical LTC3833 power supply with 4.5V to 14V input and 1.5V/20A output.[11]Figure 15 shows that the power supply responds quickly to bursts of high slew rate load transients. During load boost transients, the switching frequency is increased to speed up transient response. During the load buck transient, the duty cycle drops to zero. Therefore, only the output inductor limits the current slew rate. In addition to the LTC3833, the LTC3838 and LTC3839 controllers also provide fast transient, polyphase solutions for multiple output or PolyPhase applications.


Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 14. Fast, Controlled On-Time Current Mode Power Supply Using the LTC3833

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 15. LTC3833 Power Supply Provides Fast Response During Fast Load Step Transients

Loop Bandwidth and Stability

A well-designed SMPS should be noise free. This is not the case with undercompensated systems, which tend to be unstable. Undercompensated power supplies are often characterized by noise from magnetic components or ceramic capacitors, jitter in switching waveforms, and oscillations in the output voltage. An overcompensated system is stable and has little noise, but has a slow transient response. Such systems have loop crossover frequencies at very low frequencies (usually below 10kHz). Designs with slow transient response require large output capacitors to meet transient regulation requirements, increasing overall power supply cost and size. The excellent loop compensation design is stable and noise-free without overcompensating, resulting in fast response and minimal output capacitance. ADI’s application note AN149 article details the concepts and methods of power circuit modeling and loop design[3]. Small-signal modeling and loop compensation design can be challenging for inexperienced power supply designers. Analog Devices’ LTpowerCAD™ design tool handles complex formulations, greatly simplifying power supply design, especially loop compensation design[5] [ 6]. LTspice®The simulation tool integrates all ADI device models and provides additional time domain simulation to optimize the design. However, during the prototyping stage, loop stability and transient performance are often benchmarked and verified.

In general, the performance of a closed-loop voltage regulation loop is evaluated by two important values: loop bandwidth and loop stability margin. The loop bandwidth is determined by the crossover frequency fCQuantized, at this frequency, the loop gain T(s) is equal to 1 (0dB). Loop stability margin is usually quantified by phase margin or gain margin. Loop phase margin ΦmDefined as the difference between the total T(s) phase delay and C180° at the crossover frequency. Gain margin is defined as the difference between T(s) gain and 0dB at a frequency where the total T(s) phase is equal to C180°. For buck converters, 45 degrees of phase margin and 10 dB of gain margin are generally considered sufficient. Figure 16 shows the current mode LTC3829 12VINto 1VOTypical Bode plot of loop gain for a /60A 3-phase buck converter. In this example, the crossover frequency is 45kHz and the phase margin is 64 degrees. The gain margin is close to 20dB.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 16. LTpowerCAD Design Tool Easily Optimizes Loop Compensation and Load Transient Response

(Take a 3-phase, single-output LTC3829 buck converter as an example)

PolyPhase Buck Converters for High Current Applications

As data processing systems get larger and faster, their processors and memory cells require more current at lower voltages. At these high currents, the demands on the power supply multiply. In recent years, PolyPhase (polyphase) synchronous buck converters have been widely used in high current, low voltage power supply solutions due to their high efficiency and uniform heat dissipation. In addition, with a multiphase interleaved buck converter, the ripple current at the input and output can be significantly reduced, thereby reducing input and output capacitance and associated board space and cost.

Precision current sensing and current sharing become very important in PolyPhase buck converters. Good current sharing ensures even heat dissipation and high system reliability. Current-mode controlled buck converters are often the first choice due to their inherent current-sharing capabilities at steady state and during transients. Analog Devices’ LTC3856 and LTC3829 are typical PolyPhase step-down controllers with precision current sensing and current sharing. For 2-phase, 3-phase, 4-phase, 6-phase and 12-phase systems with output currents from 20A to over 200A, multiple controllers can be daisy-chained.

Additional Requirements for High-Performance Controllers

A high-performance buck controller requires many other important characteristics. Soft-start is usually required to control inrush current during startup. Overcurrent limit and short-circuit latch protect the power supply when the output is overloaded or shorted. Overvoltage protection protects expensive loading devices in the system. To minimize EMI noise in the system, sometimes the controller must be synchronized to an external clock signal. For low-voltage, high-current applications, remote differential voltage sensing compensates for PCB resistive voltage drops and precisely regulates the output voltage of remote loads. In complex systems with many output voltage rails, timing and tracking between different voltage rails is also required.

PCB layout

Component selection and schematic design are only part of the power supply design process. Proper PCB layout in switching power supply design is always critical. In fact, its importance cannot be overemphasized. Good layout design optimizes power efficiency, mitigates thermal stress, and most importantly, minimizes noise and interactions between traces and components. To do this, designers must understand the current conduction paths and signal flow of switching power supplies. It usually takes a lot of effort to gain the necessary experience. See ADI’s application notes 136 and 139 for a detailed discussion.[7][ 9]

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 17. 3-Phase, Single V using the LTC3829OHigh Current Buck Converters

Choose from a variety of solutions – discrete, monolithic and integrated power

At the integration level, the system engineer can decide whether to choose a discrete, monolithic or fully integrated power module solution. Figure 18 shows an example of a discrete power module solution suitable for a typical point-of-load power application. Discrete solutions use a controller IC, external MOSFETs, and passive components to build the power supply on the system board. One of the main reasons for choosing a discrete solution is the low bill of materials (BOM) of components. However, this requires good power design skills and a relatively long development time. Monolithic solutions use ICs with integrated power MOSFETs, further reducing solution size and component count. The design skills and development time required for this solution are similar to discrete. A fully integrated power module solution can significantly reduce design effort, development time, solution size, and design risk, but the BOM cost of the components is typically higher.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 18. (a) Discrete 12VINto 3.3V/10A LTC3778 power supply;

(b) Fully integrated 16VIN, dual 13A or single 26A LTM4620 µModule®Buck Regulator Example

Other basic non-isolated DC/DC SMPS topologies

This application note briefly illustrates SMPS design considerations using a buck converter as an example. However, there are at least five other basic non-isolated converter topologies (boost, buck-boost, Cuk, SEPIC and Zeta converters) and at least five basic isolated converter topologies (flyback, forward , push-pull, half-bridge, and full-bridge), these topologies are not described in this application note. Each topology has unique characteristics suitable for specific applications. Figure 19 shows a simplified schematic of other non-isolated SMPS topologies.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 19. Other basic non-isolated DC/DC converter topologies

There are also non-isolated SMPS topologies that are composed of basic topologies. For example, Figure 20 shows a high-efficiency, 4-switch synchronous buck-boost converter based on the LTC3789 current-mode controller. It operates with input voltages lower than, equal to or higher than the output voltage. For example, the input voltage can range from 5V to 36V and the output voltage can be regulated 12V. This topology is a combination of a synchronous buck converter and a synchronous boost converter, sharing a single inductor. When VIN > VOUT, switches A and B act as active synchronous buck converters, while switch C is always off and switch D is always on. When VIN At OUT, switches C and D act as active synchronous boost converters, while switch A is always on and switch B is always off. When VINclose to VOUT, all four switches work effectively. As a result, the converter has high efficiency, up to 98% for typical 12V output applications.[12] The LT8705 controller further extends the input voltage range to 80V. To simplify design and increase power density, the LTM4605/4607/4609 further integrate a complex buck/boost converter into an easy-to-use high-density power module.[13] They can be easily paralleled to share the load and are suitable for high power applications.

Application Note 140 – Part 3/3: Design Considerations for Switching Power Supply Components

Figure 20. High Efficiency 4-Switch Buck-Boost Converter Operating with Input Voltages Below, Equal to, or Above the Output Voltage

Summarize

All in all, linear regulators are simple and easy to use. Since the series regulating transistors operate in linear mode, the power supply efficiency is generally lower when the output voltage is significantly lower than the input voltage. Linear regulators (or LDOs) typically have low voltage ripple and fast transient response. SMPS, on the other hand, use transistors as switches and are therefore generally more efficient than linear regulators. However, the design and optimization of SMPS is more challenging and requires more background knowledge and experience. Each solution has advantages and disadvantages for specific applications.

References

[1] V. Vorperian, “A Simplified Analysis of PWM Converters Using PWM Switching Modes: Parts I and II,” IEEE Transactions on Aerospace and Electronic Systems, March 1990, Vol. 26, No. 2.

[2] RB Ridley, BH Cho, FC Lee, “Analysis and Interpretation of the Loop Gain of a Multi-Loop Controlled Switching Regulator,” IEEE Transactions on Power Electronics, pp. 489-498, October 1988.

[3] H. Zhang, “Modeling and Loop Compensation Design for Switch-Mode Power Supplies,” Linear Technology Application Note AN149, 2015.

[4] H. Dean Venable, “Optimal Feedback Amplifier Design for Control Systems,” Venable Technical Literature.

[5] H. Zhang, “Designing a Power Supply in Five Easy Steps Using the LTpowerCAD Design Tool,” Linear Technology Application Note AN158, 2015.

[6] LTpowerCAD™ Design Tool at www.linear.com/LTpowerCAD.

[7] H. Zhang, “PCB Layout Considerations for Non-Isolated Switching Power Supplies,” Linear Technology Application Note 136, 2012.

[8] R. Dobbkin, “Low-Dropout Regulators Can Be Directly Paralleled for Heat Dissipation,” LT Journal of Analog Innovation, October 2007.

[9] C. Kueck, “Power Layout and EMI,” Linear Technology Application Note AN139, 2013.

[10] M. Subramanian, T. Nguyen, and T. Phillips, “Sub-milliohm DCR Current Sensing and Accurate Multiphase Current Sharing for High-Current Supplies,” LT Journal, January 2013.

[11] B. Abesingha, “Fast and Accurate Step-Down DC-DC Controller Directly Converts 24V to 1.8V at 2MHz,” LT Journal, October 2011.

[12] T. Bjorklund, “High Efficiency 4-Switch Buck-Boost Controller Provides Accurate Output Current Limit,” Linear Technology Design Note 499.

[13] J. Sun, S. Young, and H. Zhang, “µModule Regulator Fits 15mm × 15mm × 2.8mm, 4.5V-36Vin to 0.8V-34V VOUT(nearly) complete buck-boost solution”, LT Journal, March 2009.

The Links:   NL10276AC30-03L LM64C031

micohuang