A Risk-Free Path to Power System Design - Building Power Design Success

Introduction

Today, high-performance power systems have come a long way, and designers are using multiple input voltages to drive multiple voltage rails for a wide variety of applications. The need to ensure that the PoL regulator is as close to the load as possible requires designers to cram a lot of power conversion functions into a very small area. At the same time, corporate resources are tending to expand to the point where engineers are expected to multi-task, and it is often generalists rather than power experts who are responsible for designing power systems. As a result, today's complex power requirements can be a headache for designers: how to use different resources to provide high-performance power to diverse loads, ensuring that all parts of the architecture operate within their power and thermal ranges, while optimizing efficiency and cost targets.

New applications present further challenges. For example, with the migration to cheaper, cleaner, more efficient energy generation, and government-driven applications, companies are looking at how they can meet legislation and cost targets by moving to high-voltage direct current (HVDC) distribution. This requires a different approach to designing power systems for the electronic devices used.

In this article, we explore the issues, techniques, methodologies, tools, and building blocks that may aid in the successful completion of complex power system projects.

The old approach has been surpassed by system requirements

Looking back over the past 30 years, most systems’ power needs have been provided by centralized power supplies. However, a trend in electronics is for devices to be smaller, lighter, and more powerful. The development of communications equipment has helped to enable the Internet revolution, which requires data centers filled with servers. Systems will become smaller, provide higher performance, and their power requirements will become more and more demanding.

It was clear that something else needed to evolve - the way we power electronics. Large centralized power supplies simply weren't up to the task of powering the next generation of products, so power conversion needed to happen closer to the load. Gone were the days of being able to power devices from a centralized box.

Power conversion needs to meet the different requirements of new loads: one voltage cannot fit all loads. Today's loads require multiple voltage rails, strict regulation requirements, and fast transient response. Therefore, power conversion must become an integral part of the system and needs to be designed in and out of the device.

As devices become smaller and smaller as the need to move closer to the point of load drives us, we have seen a trend of increasing heat dissipation density over time. The drive to develop smaller power components has outstripped the corresponding efficiency improvements. For many years, devices such as power bricks have formed the backbone of power systems, and their power density is being limited. It should be clear that in addition to being more efficient, power components also need to be better at dissipating heat and support flexible thermal management.

New power components enable new approaches

Without fundamental advances in packaging technology, power density cannot continue to increase. New power components that can be used in modern system building blocks are being developed, which provide higher power density, better thermal performance, larger step-down ratios, and integrated magnetic structures. These components also contribute to the emergence of new power distribution designs, including factorized power architecture (Factorized Power Architecture, FPA), and support new applications such as high voltage direct current (HVDC), which helps to further improve efficiency and use alternative energy sources.

So how does an engineer take full advantage of the high-performance building blocks now available, think carefully and embark on a design, responsible for designing an optimized power system for a project? It is indeed a difficult choice - especially when this is not your area of ​​expertise. What is needed here is a risk-free approach with a lot of support, where there is no room for error; respins are expensive and may result in missed opportunities.

Vicor, a company at the forefront of innovation in power design, has pioneered a power component design methodology that allows engineers to predictably and cost-effectively configure high-performance power systems using industry-proven components using a proven methodology.

Power components are dedicated modules that have been fully optimized by professional power engineers for efficiency, power density, transient response and EMI. With this approach, instead of developing a power chain using discrete components, all of these key parameters have been optimized and designers are ready with a most suitable solution for any power design project. In addition, the structure of these modules is fully suitable for reuse in future designs, saving time and effort.

When combined with the array of tools and resources available, this approach will enable a faster and simpler design cycle with less risk to complete projects and bring products to market.

The power component design methodology has three steps: identify, build and implement.

Step 1 - Identify

This is a "big picture" view of the power requirements for a project, defining the number of rails, voltage and current requirements, while taking into account the timeline of the project. At this stage, a list of these requirements is made and initial consideration is given to the types of products that can be used to meet them.

Figure 1: The first step is to list the power requirements for the project. For our example, we assume there are 11 voltage rails, listed in descending power levels. For convenience, we call them Main Rails (MR) and Auxiliary Rails (AR). Any special requirements are included in the Comments column.

What products will meet the requirements? There are many sources of this information. For example, Vicor offers a solution selector tool that searches a database of available components and recommends solutions that meet the customer's input and output requirements. With an intelligent tool such as the Vicor solution selector, the time required to generate a shortlist of possible components can be reduced to almost zero, and the best component for a particular design can be easily selected based on the criteria that are most important to the application. Most engineers simply do not have the luxury of "learning time" to complete this important task manually.

Figure 2: Use Vicor’s PowerBench tool to simplify the component selection process.

What are the typical power components available:

First up is power delivery. Here, power components must take high voltage DC or AC power and convert it to a safety extra low voltage (SELV). In many high-performance applications, engineers are using high voltage and high current to deliver power to their systems. Because of the heat dissipation from the device, it is critical to select thermally adaptable components. These components will need to be placed in multiple locations within the system. This includes power systems mounted in a chassis or motherboard, and the corresponding cooling of each component needs to be considered.

Next comes the delivery of power from the SELV to the point of load. Engineers need to carefully select the appropriate voltage rail for their application. Too many conversion stages will reduce the efficiency of the application. In recent years, power supply designs have begun to move from 12V rails to 48V rails that provide higher system efficiency. The challenge is to select the best components that can provide the right performance at the highest efficiency. Tools like Vicor's Whiteboard can help engineers evaluate the performance of their designs using different SELVs.

Finally, there is the choice of point-of-load components. Based on the SELV selected, engineers need to select the components needed to achieve the PoL requirements so that less than 1V can be achieved at high currents. Isolation and regulation are required, and DC-DC converters such as Vicor DC Converter Modules (DCMs) can be used. Designers can also use components designed for factorized power architectures, where the regulation and voltage conversion/isolation functions are separated. Choosing the latter helps designers achieve high power density, which is equivalent to having the ability to convert a large amount of power in a small space.

Step 2 - Build

The first step in building a system is to create a block diagram of the power system, starting at the output and working backward toward the input. It works best to start at the lowest power level and work from there so that the power component classes can be reviewed and changes made if necessary as power levels increase.

It is important to choose the right component category for the appropriate power level. For example, at low power, a system-in-package (SiP) product such as the Vicor ZVS buck regulator is the best solution. At higher power levels, a better approach may be to use Vicor's ChiP products (Converter housed in Package). Depending on the complexity of the number of voltage rails required to drive the load, a combination of SiP and ChiP can be used in the application.

This will help achieve maximum power density and cost-effectiveness within the system and keep every device in the system operating at high efficiency.

Looking back at Figure 1, it is clear that the first three rails (MR#1, 2, and 3) are the rails that require the highest power level devices, while the last five rails (MR#7 through AR#2) are the lowest power level devices. The rest (MR#4 through MR#6) are somewhere in between. This is where the designer will need to use their judgment to make device choices. Once the outputs are done, we can start to build a picture of the power modules and power levels we need in the system block diagram category.

Step 2 - Build - Differentiate by class

Figure 3: From the power rail requirements analysis, we can determine the most appropriate power component category.

Step 2 - Build - Block Diagram - Work back to the input side (2)

Step 2 - Build - Block Diagram - Optimize and Evaluate if Required

Figure 4: Continuing with our previous work, we can determine the component categories needed to provide the power level for each voltage rail. At this stage, we should always keep in mind the power level required to ensure we balance the load and utilize the power capabilities of each device. Here we see an optimization of our original estimate.

Step 2 - Build - Final Block Diagram

Figure 5: Here we see that the ACFE of the driving voltage rails is now introduced. It is very important here to judge the load on each voltage rail and ensure that the loads are operating close to the maximum with appropriate safety margins.

Step 3 - Implementation

Once the modules are complete, designers need to match part numbers to these modules while taking care to implement all the dedicated circuitry for functionality and emulation of the respective power conversion component chains. Other circuits that need to be developed may include filters, hold-up circuits, and power sequencing. At this stage of the design, engineers should also consider thermal, termination, and packaging considerations.

In our case, there are some special requirements for the power supply: there is a delay on MR#3 before the auxiliary voltage rail ramps up, and tight regulation of MR#3 will require the use of a remote sense loop. It also makes sense to consider configuring the PRM for precise load current limiting and other parameters required to precisely match the voltage rail and load.

For engineers who need to use a PRM to tune their designs, Vicor offers a PowerBench simulation tool that can help further understand the performance of the system.

Figure 6: PowerBench PRM simulation tool.

Design and development tools

In the past, engineers made component selections and analyzed power system efficiency (and overall system performance) at each stage by referring to calculations in device datasheets.

Check the efficacy from the data sheet

Figure 7: Obtaining performance information can be time-consuming and laborious.

While completely satisfying, this approach can become a bit tedious. To simplify the design process and save time, Vicor recently launched the PowerBench whiteboard tool (whiteboard). The whiteboard tool is an online tool for designing and analyzing power systems using a suitable set of Vicor power conversion components. Instead of having to look up the operating and efficiency parameters contained in the data sheet, the engineer can simply draw a circuit block diagram using the online tool, and all calculations can be completed in a few milliseconds.

More accurate and realistic conversion efficiency of 93.17% (automatically generated in milliseconds) generated by the Powerbench whiteboard tool

Figure 8: Whiteboard tools automatically analyze designs in milliseconds and provide performance data, saving time and effort.

By retaining the system's familiar sketch symbols and adding automatic parameter lookup and calculation, the Whiteboard Tool can further reduce the time to complete a design using the Power Component Design Methodology.

In addition, Vicor's solution selection tool can be tightly integrated with the whiteboard tool. Therefore, the design recommended by the solution selection tool can automatically import the design into the whiteboard tool so that engineers do not need to draw the system themselves. At this point, engineers can adjust the design to further meet their needs and quickly understand the efficiency of the design.

in conclusion

Power components have become a key factor in helping engineers design complex, high-performance power systems for today's electronic systems. Because power design experts have optimized efficiency, power density, transient response, EMI, and cost-effectiveness, almost any electronic engineer can use these devices to develop a power system to meet challenging high-performance requirements.

Driven by the need for better thermal performance, many power component innovations have emerged recently. The ChiP platform provides a solution for strong thermal dissipation using double-sided cooling, which is a good example of board power. In the future, other innovations will further simplify the tasks of power system designers, especially on the front end of the power supply.

This article shows that the Power Component Design Methodology provides a simple three-step approach that enables engineers, even those who are not power experts, to build complex power chains that can provide high efficiency and high power density. This approach is further simplified through the use of online tools. However, unlike many design solutions, the Power Component Design Methodology removes the pain and risk from the design process without requiring engineers to spend time learning the technology. Engineers can use this method without special training, shortening development time while ensuring that their next power chain is optimized to provide the required performance.