System effectiveness, status monitoring and energy efficiency require AC-DC power supplies to be monitored and controlled, and though the concept of doing this digitally has been around for some time, as end-designs are becoming more complex, many engineers are just starting to investigate the benefits that incorporating digital control into their system may bring.
A design engineer looking into digital methods of monitoring and controlling an AC-DC power supply may be drawn by the prospect of access to a high level of information compared to an analog system and the ability to control the output digitally. However, these functions come at the cost of system complexity.
Analog vs Digital
Digital control undoubtedly offers higher levels of flexibility. Power supplies can be turned on and off remotely, limits can be set on output voltages and currents and alarms on these levels enabled. Digital power supplies can provide monitoring data, such as temperature and fan performance, to help predict failures. Digital control can also ease calibration – parameters may be programmed instead of tweaking a pot in an analog power supply. On the down side, full digital control of a power supply involves many operating parameters and can be extremely complex. It requires a DSP with sophisticated software, which can be difficult to troubleshoot.
By comparison, analog control, having been used for 50+ years, is a tried and true method. Analog power supplies are stable and less prone to glitches, though they have limited flexibility. Parameters are defined during the design phase and cannot be changed later. While status can be monitored, control is usually limited to one possible reaction per parameter; with a digital power supply, there is usually an extensive range of input scenarios with multiple possible actions based on those scenarios, albeit limited by the speed of the processor.
Finding a balance
Finding a compromise between analog and digital can be tricky. XP Power’s approach to finding a balance between the two domains is to use a tried and true analog PWM controller with an addition of a digital interface, to create communication flexibility. In a typical design, an analog PWM controls the output voltage in real time, and a microprocessor monitors and adjusts that. The digital control board can artificially create multiple operating states by combining the single parameter control and status signals, allowing a higher level of flexibility than with the analog controller alone. This level of control suits the majority of customers who prioritise a cost-effective solution.
The analog PWM controller has the advantage of eliminating certain types of inaccuracies that are inherent to digital designs. For example, sampling signals with an ADC to feed to the DSP will introduce aliasing errors. Using a DSP to control the output will also introduce jitter into the PWM signal – this can cause sub-harmonics in the output leading to EMI problems in the application. Using an analog PWM effectively avoids this. However, since an analog PWM’s switching frequency is set by hardware, this can’t be adjusted. Also, a fixed control loop means this can’t be optimised for different conditions without changing components.
Communications
The most widely-used communications protocol for digital control of power supplies is PMBus, a clearly defined industry standard intended to make power supplies plug-and-play, so that the interface from the end equipment can be designed without even seeing the power supply. While it’s perfectly possible to control a digital power supply using a CAN bus, or Ethernet for example, this requires much more time developing a protocol stack for use in each specific application. PMBus is much simpler. It features a relatively simple interface (two I/O lines) which can help make hardware smaller, and there are only three layers to the PMBus stack, compared to 10 or 12 for a comparable Ethernet implementation. For this reason, designers in a hurry, and those without deep networking expertise or resources often choose PMBus.
Aside from monitoring and control functions, utilising the PMBus interface allows several additional applications. Battery charging is one example – since the output current and output power are monitored, a battery can be charged from the power supply without the need for an external interface. Power sequencing can also be implemented, using a multiple output power supply turned on or off at a particular time or in a particular sequence as determined by the system.
Real applications
As an example, the GFR1K5, a 1U rack-mount 1500-W AC-DC front end from XP Power uses a PM Bus interface with a very simple command structure to give a practical amount of control, as shown in table 1. The output can be turned on and off, and the overcurrent shut down point can be set; hardware based shut down occurs at 110-140 percent of Inom, but a firmware shut down point can be set for 105 to 0 percent. Response to a firmware shut down can be specified as well (latch off or a hiccup mode with an adjustable number of retries or continuous retry).
The PMBus interface also enables monitoring capabilities: signals are available for output voltage, output current and temperature, and the data bits for the alarms can be set easily using the interface. Information can be retrieved about the model number and serial number (useful in a system with many power supplies) and the runtime is also available.


In summary, there are pros and cons to both the analog and digital control of power supplies, but an analog PWM controller with a PMBus interface picks the key benefits of both approaches, achieving a balance between functionality and cost. Different methods for accessing power supply functions are available, but the widely-used PMBus protocol is specially designed for power supplies and is therefore the easiest to implement, while offering full access to all the functionality of a modern AC-DC PSU.