Many power supplies, especially offline power supplies, require low standby power. For power levels below 100 W, the most cost-effective isolated topology is the flyback because it requires the fewest components. Flyback converters typically produce multiple secondary outputs, which require relatively precise regulation. This article will describe the challenges of achieving a well-regulated output voltage while still achieving low standby power consumption.
Low power AC/DC flyback power supplies are widely used in industrial applications such as motor drives and appliances because they can achieve good voltage regulation and low standby power consumption. Typical applications of isolated low-power designs often require multiple secondary outputs. Figure 1 shows an example of a flyback topology that generates outputs V OUT1 and V OUT2 from a universal input (85 V AC to 265 V AC). Transformer T1 provides galvanic isolation between the AC power line (mains) and the load. The auxiliary winding AUX powers the primary-side flyback controller.
Figure 1 Simplified schematic of a multi-output flyback that provides galvanic isolation between the AC line and the load.
How to reduce standby power consumption
Let's briefly review known techniques for reducing standby power consumption. Standby power mainly depends on circulating energy, start-up circuit, snubber network and minimum load requirements. Reducing the no-load switching frequency and using an active start-up circuit and a Zener snubber network instead of a resistor-capacitor-diode snubber circuit reduces standby power consumption. Unfortunately, other circuit characteristics also increase standby losses. Therefore, it can be helpful to develop strategies in advance that will help keep standby power consumption low.
One of the main challenges facing power supply designers is that it is impossible to build an ideal circuit, since any practical circuit board has to deal with parasitic capacitance and inductance, as well as noise in the system.
These challenges are exacerbated when two or more isolated outputs are generated, as shown in Figure 1. Typically, the voltage control loop regulates only one output; the coupling of the transformer windings half regulates the other output. Figure 2 shows the regulation of one output. An external error amplifier (U1) is connected to the output V OUT2 through a resistor divider ( R high1 , R low1 ). An optocoupler helps transfer the error signal to the primary side.
Figure 2 schematic diagram of an external error amplifier connected to V OUT2 shows an adjustment of the output
The other output, V OUT1 (3.3 V), is only half-regulated due to the coupling of the transformer windings. But what happens in standby mode under light or no-load conditions? To answer this question, consider Figure 3, which shows the secondary winding voltages (also known as secondary switching nodes) for V OUT1 (3.3 V) and V OUT2 (12 V).
Figure 3 Overshoot of the secondary-side switch node can be a challenge under light or no-load conditions
You can easily identify overshoot and then ring after the on-time is over. Basically, the overshoot of the primary switching node is reflected to the secondary side. This overshoot can be a challenge under light or no-load conditions, especially for unregulated outputs, because it charges the output capacitor through the output diodes D1 and D2, as shown in Figure 1. Overshoot can cause the unregulated output voltage to rise to a very high value.
What are the main causes of unexpected overshoot and ringing? This is a parasitic effect of the power stage and the board, including the leakage inductance of the transformer. Leakage inductance is caused by magnetic flux in one winding of a transformer that is not coupled to the other windings. This energy is dissipated outside the transformer and overshoot occurs. Figure 4 shows the primary switch node voltage, which is basically the drain-source voltage of a metal-oxide-semiconductor field-effect transistor (MOSFET).
The primary switching node of Figure 4 is the drain-source voltage of the MOSFET.
Effect of Transformer Leakage Inductance
Now that you've seen how overshoot can adversely affect light-load cross-regulation, the question arises: why isn't it strongly clamped? Typically, a snubber clamp circuit limits the overshoot voltage to a certain level. The clamping circuit absorbs the energy stored in the transformer leakage inductance and, depending on the value of the clamping voltage, also absorbs a small fraction of the magnetizing energy. As the clamping voltage decreases, the energy loss in the clamping increases rapidly.
Due to the high energy losses, you must allow some switch node voltage overshoot. The minimum overshoot mainly depends on the leakage inductance. Using existing transformers, it is not possible to clamp overshoot to every desired level. Before ordering a custom transformer sample, you must consider the optimized transformer construction. The goal should be to minimize the ratio of leakage to magnetizing inductance.
Leakage inductance strongly depends on the geometry of the physical winding. In general, there are two changes that reduce leakage inductance: reducing the dielectric spacing between the primary and secondary windings and increasing the overlapping surface area between them. Therefore, using a staggered winding configuration and wider spools and moving the layers further together will result in low leakage inductance. Unfortunately, there is a trade-off. These changes usually involve increasing parasitic inter-winding capacitance, which increases common-mode EMI. Therefore, you should work closely with the transformer manufacturer from the outset to find an optimized transformer construction.
Now, let's take another look at the design that produces two outputs: 3.3 V (V OUT1 ) and 12 V (V OUT2 ). Some applications require tighter regulation for lower output voltages as it generally requires tighter tolerances. It is assumed that V OUT1 (3.3 V) will be regulated and the higher output voltage V OUT2 (12 V) will remain unregulated. Therefore, V OUT1 is regulated to 3.3 V, and the turns ratio of the transformer winding determines V OUT2. This configuration works well for systems with low parasitics, including low leakage inductance, even at light loads.
However, if the leakage inductance is large, the coupling of the windings is poor, and the overshoot is large, then the cross regulation is no longer good, because the transformer winding voltage ratio is no longer proportional to the winding turns ratio. As a result, V OUT2 can rise very rapidly, easily becoming twice the expected level or more. A resistor or zener diode will limit the voltage but will also significantly increase the standby power. So you need to consider other possibilities.
Therefore, rather than regulating a lower output voltage, it may be helpful to regulate a higher output voltage, VOUT2. If the unregulated output V OUT1 usually does not exceed the value of V OUT2, in principle the low-voltage output can at most reach the level of the high-voltage output. This means that in some cases it can be advantageous to regulate a higher voltage, as doing so will maintain a lower absolute maximum voltage in the system.
As always, there are trade-offs, as unregulated outputs are regulated worse. A compromise is to regulate both outputs simultaneously, as shown in Figure 5. This approach works well as long as you don't need isolation between the outputs, but has a downside as it is impossible to regulate any output with very high precision.
The Figure 5 schematic shows an external error amplifier connected to V OUT1 and V OUT2.
Another option is to use an inner loop connected to the anode of the optocoupler from one output and an outer voltage loop from the other output, as shown in Figure 6, to achieve precise regulation of V OUT2 and improve regulation to some extent V OUT1.
Figure 6 This is how the external error amplifier is connected to the inner loop and VOUT2.
Since the final regulation is highly dependent on the parasitic capacitance and inductance of the power stage components and layout, it is recommended to evaluate alternatives in the laboratory.
Modern flyback controller
Modern flyback controllers can achieve very low standby power consumption because the pulse width modulation algorithm changes the switching frequency and primary current while maintaining the discontinuous conduction mode. The algorithm reduces switching frequency and peak current at light loads. With modern flyback controllers, some applications can even achieve standby power below 20 mW. However, when designing a power supply, causes that increase power consumption must be avoided.
To achieve low standby power consumption, the energy drawn from the input per cycle must be reduced by using an active start-up circuit to reduce switching frequency and primary peak current and reduce secondary side preload resistors. A good layout can also reduce noise in the system, and snubber networks suitable for the primary and secondary switching nodes can further reduce noise and overshoot. Finally, don't overlook the transformer; it's the most important part of the power supply, aside from the controller.
