by Bill Schweber for Mouser Electronics
Just a few years ago, “digital power” was mostly a concept with some prototypes under long-term evaluation, but few actual installations. Fast-forward to 2016, and you will see these supplies are now standard and essential in power-intensive applications such as data centers. Without the attributes and virtues they offer, it would be very hard to provide the hundreds of amps at a variety of DC rails given the space available, the efficiency mandates and thermal constraints, and the sophisticated supply demands of these installations.
The reasons for the widespread adoption of digital power supplies in these power-intensive applications include:
Their high efficiency yields lower operating cost; there is less heat to dissipate; and they make it easier to meet environmental-related regulatory requirements.
They can implement the challenging and sophisticated technical requirements of powering processors and FPGAs.
Their flexibility supports dynamic changes in strategies during operation, and they can handle complicated power-up and power-down sequencing scenarios.
Power-supply designers (and many users) are generally a cautious group, as they must be when dealing with high current, voltage and power levels, and the consequences to equipment and people of supply malfunction or failure. It is a cautious user base that prefers products with a track record and long, viable product life spanning a decade, two decades, or more, and that does not want to subscribe to a trend just for sake of being leading edge.
For these and other reasons, there was some early reluctance to embrace the firmware-based approach, but the situation has changed. However, due to the positive track record of high-end digital power confirmed by solid data, other application areas, such as industrial systems, are seeing “trickle down” availability at lower levels. The gains include improved efficiency from low load to full load, which saves energy, reduces thermal stress on components, simplifies cooling challenges and increases MTBF (mean time between failures).
What is “Digital Power?”
The objective of a power supply or converter is simple to state: provide a stable, regulated DC output at the desired voltage value despite changes in input voltage or load conditions. This requires some form of closed-loop control within the DC/DC converter, based on measurement of the actual output voltage, comparison with the setpoint value, and implementing feedback-based corrections to force the output back to the setpoint and keep it there.
This regulation has traditionally been implemented using a closed-loop negative-feedback with analog circuitry in a switching regulator, Figure 1. (The alternative, a low-dropout regulator, or LDO, is also an option, but only viable at fairly low power levels.) There are many standard architectures for these switchers, with a long list of additional enhancements to increase efficiency across the entire load range, boost performance and ensure consistent operation. These enhancements can become quite complicated and clever, and have impressive names such as SEPIC (single-ended primary-inductor converter).
These variations can become fairly complicated and sophisticated, but all have one drawback: they lack flexibility for real-time setting of operational parameters. For example, the Intel/Xilinx VR13 standard requires the ability to direct the supply to change its nominal output voltage from 1.2 to 0.9 V and back “on the fly,” which an all-analog supply cannot do. This adaptive voltage scaling (AVS) adjusts the supply output voltage to the minimum required by the processor, depending on its clock speed and workload, and also compensates automatically for process and temperature variations within the processor. To do all this requires a fully programmable, sophisticated, firmware-controlled converter.
It is possible to implement some of the desired changes via an I/O port on the supply coupled with digital parameter-setting circuitry. This results in a hybrid supply that has an inner analog-control loop but overall digital supervision and some reporting of supply status, Figure 2.
The all-digital supply uses a very different internal architecture. Rather than implement the control loop using analog circuitry, even with some digital oversight, the digital supply uses analog/digital (A/D) converters to digitize critical internal voltages and currents. The converted values are used by a dedicated, embedded processor (DSP, FPGA) that executes code for closed-loop algorithms. Finally, the algorithms’ outcomes are converted back to analog signals via a digital/analog (D/A) converter, adjusting the voltages and currents as needed, Figure 3.
The control algorithm is firmware-based rather than built as a hardwired analog circuit, so the control strategy can be fairly complicated and sophisticated. Even better, a single processor (if powerful enough) can control two or more independent output rails, and coordinate these rails to manage factors such as output levels, ramp rates and relative power on/off timing between these rails. It can also provide detailed reports and historical data on the supply’s status, conditions and changes, so likely failures can be anticipated rather than just reported after they occur.
Two examples will show how digital supplies are now able to serve applications at lower current and power needs than those of data center racks. The NDM2Z-50 from CUI, Inc. (Figure 4) is an all-digital DC/DC point-of-load (PoL) converter for a 4.5- to 14-V input range and 0.6- to 3.3-V programmable output, providing up to 50 A (165 W maximum). It includes an SMBus interface and is PMBus™ compatible. Despite its small package (30.85 x 20.0 x 8.2mm for the horizontal-mount version), it provides features such as voltage tracking, voltage margining, active current sharing, parametric capture, voltage/current/temperature monitoring, and programmable soft start and soft stop. Its data sheet (Reference 1) includes dozen of graphs showing all aspects of static and dynamic performance.
Figure 4: This all-digital DC/DC PoL converter from CUI delivers up to 50A and is part of a family of both larger and smaller efficient, flexible, compatible and feature-rich DC/DC converters for larger application needs. (Image: CUI Inc.)
The power needs of many of today’s electronic systems can no longer be satisfied by even leading-edge analog supplies, but instead require a new form of power-supply architecture for control. The fully digital power-supply implementation has significant and tangible benefits with its flexibility, performance and adaptability. While it is radically different in concept and execution from the traditional analog-based supply, the digital design is mature and is rapidly expanding to other applications.
CUI Inc. NDM2Z-50 Auto Compensated, Digital DC-DC POL Converter. Close Article
Bill Schweber is a contributing writer for Mouser Electronics and an electronics engineer who has written three textbooks on electronic communications systems, as well as hundreds of technical articles, opinion columns, and product features. In past roles, he worked as a technical web-site manager for multiple topic-specific sites for EE Times, as well as both the Executive Editor and Analog Editor at EDN.
At Analog Devices, Inc. (a leading vendor of analog and mixed-signal ICs), Bill was in marketing communications (public relations); as a result, he has been on both sides of the technical PR function, presenting company products, stories, and messages to the media and also as the recipient of these.
Prior to the MarCom role at Analog, Bill was associate editor of their respected technical journal, and also worked in their product marketing and applications engineering groups. Before those roles, Bill was at Instron Corp., doing hands-on analog- and power-circuit design and systems integration for materials-testing machine controls.
He has an MSEE (Univ. of Mass) and BSEE (Columbia Univ.), is a Registered Professional Engineer, and holds an Advanced Class amateur radio license. Bill has also planned, written, and presented on-line courses on a variety of engineering topics, including MOSFET basics, ADC selection, and driving LEDs.