How to find the right power supply
In today’s networked world, where connectivity is increasing at an unprecedented rate, a growing number of links occur between systems – a development which was not predicted to that extent. In some cases, the design expectations of these systems are even exceeded, resulting in interference due to daisy chaining and ground loops. This is especially true for power supplies, since these are often neglected up to the last stage of a system or product development, when, fundamental changes to the application are often undesirable or even impossible. Which in turn means that a power supply unit in many cases must provide filter functions beyond the originally envisaged specifications.
In contrast, the increasing energy efficiency requirements have been the driving force to apply faster switching technologies in power supply designs to optimize efficiency, but at the same time bring about a rise in electromagnetic emissions. As a result, the filter components of many power supplies occupy more than 30 % of its total size, making the power supply not only larger but also more expensive compared to soft switching technologies.
Soft switching technologies have existed for decades, with LLC converters processing the most commonly used technology. However, LLC converters have a very limited basic regulation range, which is the reason why they are often used as a second stage with an Active Power Factor Correction (PFC) front-end, which regulates the input voltage of the LLC converter to a specified value so that a better output voltage regulation can be achieved.
LEI’s Chi-Bi LLC converters have been designed as single-stage AC-DC converters using a resonant LLC circuit with a high Q-factor, which enables highly efficient power conversion with low electromagnetic emissions. In addition, they are highly competitive products offering a relatively high power density.
Single-stage power supply versus universal power supply
Since the advent of the switch mode the output voltage of power supply units may optionally be regulated with a greater input voltage deviation, finally allowing a PSU with universal input voltage range (90 ~ 264 V) to be used worldwide. Even though this proves advantageous with regard to the product specification and reduces the variety of part numbers, universal product development always creates (unnecessary) hidden costs. It is obvious that mobile applications such as mobile phones, tablets, and laptops require a greater range, as users are expected to carry their devices with them when they travel. However, stationary applications such as network infrastructure, printers, TV sets, set-top or OTT boxes are not used in different locations and are therefore suitable for narrow-range power supplies.
Each design window (the range between minimum and maximum values) achieves optimal mid-range performance with proper balance, with both extremes showing reduced performance. This also applies to the input voltage range. When designing a universal power supply unit, the average operating range is approx. 180 VAC – a voltage that is not used anywhere in the world, so that the power supply unit always operates at an operating point that is not optimal.
When designing a single-stage power supply, however, the minimum and maximum values deviate by about 10 % from the center of the design window, which is the voltage at which the power supply is actually used. This leads to a significant increase in performance compared to universal products.
Considering the performance of the components in terms of both price and electrical performance, differences can also be observed, especially with regard to the components of the primary side (AC side) of the power supply. For example, 400 V types are required as buffer capacitors for high-voltage operation (230 V), while a larger capacity is required for low-voltage operation (120 V) to meet requirements such as bridging times. Table 1 shows the result of the requirement profiles for input buffer capacitors, which illustrates that a universal power supply design requires a larger capacitor that is usually more expensive.
The equivalent series resistance (ESR), the loss factor (tan∂) and the price are not listed here, but it is not difficult to estimate the outcome. This results in a potential performance increase for single-stage products in terms of efficiency, electromagnetic emissions, size and cost.
What are the advantages of LLC converters?
Nowadays, the vast majority of energy-efficient products (< 75 W) are built with flyback converters. A flyback converter is a simple converter that uses the transformer as a storage medium before the energy is supplied to the output. This two-stage converter type allows a very wide control range with a relatively simple control mechanism. The input energy is stored in the transformer during the switch-on time of the converter and is transferred to the output during the switch-off time. The isolation of the energy flow is the key to the wide operating range (control range) of flyback converters. The disadvantage of this technology, however, is that the transformer must be able to store all this energy and must therefore be of adequate size.
To reduce the size of a power converter, the transformer can be used in two quadrants instead of one, as is the case with a flyback converter. In this way, the efficiency of a given transformer size can be doubled. The efficiency can be increased even more if the transformer is used as a direct converter component and not as a storage component, i. e. as a direct power supply to the output. By increasing the frequency, the number of turns can be reduced without the risk of transformer saturation. In order to increase the switching frequency, switching losses must be reduced, which can be achieved by a resonant or soft-switching topology.
A half bridge, as shown in Figure 2, allows both to increase operation on two quadrants (the current can flow in both directions) and to use the transformer as a direct conversion component, i. e. the load is supplied directly from the primary side without the need for energy storage. Lp in Figure 2 represents the primary winding of the transformer. When the Q1 switch is closed, current flows from left to right through Lp, charging capacitor C2 and discharge capacitor C1. After opening switch Q1, the residual energy in the transformer will cause a natural transformation. After that, the Q2 switch can be closed with low losses and the current flows in the opposite direction. They can be replaced by a single capacitor, resulting in an asymmetric half bridge with a negligible effect on symmetrical balance and electromagnetic emissions.
If the frequency of the half-bridge converter is adjusted near the resonance frequency of the primary LC circuit (Lp + C1/C2), a reduction of switching losses can be achieved. The opening and closing of the primary switches (Q1 and Q2) can then take place in the vicinity of zero voltage or zero current (ZVS or ZCS), which leads to minimal switching losses. The transformer still recognizes a square wave voltage, but the current flowing through the transformer and feeding the output is sinusoidal. This sinusoidal current is the main driving force for increasing efficiency and reducing electromagnetic emissions, since the harmonics of the switching frequency contain less energy compared to hard-switching technologies.
Flyback converter versus Chibi LLC converter
Figure 3 shows a 60 W ChiBi LLC converter for wall mounting and a normal 60 W flyback converter for desktop use. In addition to the fact that the ChiBi converter is around 30% smaller, its reduced weight also allows wall mounting, which further reduces the total cost as no AC cable is required.
The standard ChiBi converters are cost-optimized and therefore feature only minor improvements in terms of efficiency and electromagnetic emissions, but are very competitive with a size reduction of 30 % and a price reduction of more than 10 %.
The basic concept, on the other hand, ensures greater efficiency and lower electromagnetic emissions. For example, design at a cost comparable to that of standard flyback converters could result in an improvement of 10 dB in electromagnetic emissions or a 5 % increase in efficiency, making this concept ideal for higher-value applications.
One of the benefits of reduced electromagnetic emissions is that the ChiBi converters can be designed without a Y-capacitor. Y-capacitors are usually used to reduce common-mode emissions by placing such a capacitor between the primary and secondary side. This usually results in a leakage current that can disturb capacitive touch panels or cause a line frequency noise (humming) in audio applications.
About the authors:
René Koch is CTO, Leader Electronics, Inc.
Tobias Herrmann is Field Application Engineer, Finepower