(I) Introduction
A switching power supply, also known as a switching converter or switching power supply, is a high-frequency power conversion device and a type of power supply. Its function is to convert a voltage at a certain level into the voltage or current required by the user through different architectures. The input of a switching power supply is mostly AC power (such as mains power) or DC power, while the output is mostly for devices that require DC power, such as personal computers. The switching power supply performs the voltage and current conversion between these two sources.
(II) Main Applications of Switching Power Supplies
Switching power supplies are widely used in industrial automation control, military equipment, scientific research equipment, LED lighting, industrial control equipment, communication equipment, power equipment, instruments and meters, medical equipment, semiconductor refrigeration and heating, air purifiers, electronic refrigerators, LCD displays, LED lamps, communication equipment, audio-visual products, security monitoring, LED light strips, computer cases, digital products and instruments, and other fields.
(III) Main Classifications of Switching Power Supplies
1. Miniature low-power switching power supply
Switching power supplies are becoming increasingly mainstream and miniaturized. They will gradually replace transformers in all everyday applications, with low-power miniature switching power supplies initially appearing in digital meters, smart meters, and mobile phone chargers. Currently, the country is vigorously promoting the construction of smart grids, significantly increasing the requirements for electricity meters, and switching power supplies will gradually replace transformers in their application.
2. Reverse-type series switching power supply
The difference between a reverse-type series switching power supply and a regular series switching power supply is that the output voltage of the reverse-type series switching power supply is a negative voltage, which is exactly the opposite polarity of the positive voltage output of the regular series switching power supply. Furthermore, since the energy storage inductor L only outputs current to the load when the switch K is turned off, under the same conditions, the output current of the reverse-type series switching power supply is half that of the series switching power supply.
(a) Parallel connection of rectifier bridges
In low-power designs, parallel connection of rectifier bridges is rarely used. However, in some high-power output applications, where adding new components is not desired and the current of a single rectifier bridge may not meet the input power requirements, parallel connection of rectifier bridges is necessary. Parallel connection of rectifier bridges cannot be achieved by connecting two rectifier bridges individually in parallel (Figure 1), because the rectifier bridges are not paired. Relying solely on their individual VI characteristics, they generally cannot achieve current sharing, leading to uneven heating between the two rectifier bridges. The method shown in Figure 2, however, is suitable. It is generally assumed that two diodes within a single package are perfectly matched and can share the current equally, thus enabling parallel connection of rectifier bridges.
(II) Floating Drive
In driver circuit design, it's often mentioned that MOSFETs need to be driven on a floating ground. So what is floating ground driving? Simply put, it means the source (S) of the MOSFET is not directly connected to the ground of the control IC; in other words, they don't share a common ground. Taking a commonly used BUCK circuit as an example, as shown in the diagram: the control IC's ground is generally shared with the input power supply ground, but there's a diode between the MOSFET's source and the input power supply ground. Therefore, the control IC's drive signal cannot be directly connected to the MOSFET's gate. An additional driver circuit or driver IC is required, such as a transformer-isolated driver or a driver chip with a bootstrap circuit like the IR2110.
Of course, there is another way, which is to use another method to power the control IC, and then connect the ground of the control IC to the source (S) terminal of the MOSFET. In this way, it is no longer a floating ground, and the output of the control IC can directly drive the MOSFET.
(III) Hysteresis Comparator
In protection circuits, a certain amount of hysteresis is usually added to prevent the protection circuit from oscillating back and forth near the protection point.
The 1M resistor acts as a hysteresis loop. Without it, the op-amp outputs a low level when VF reaches 2.5V and a high level when it falls below. With the 1M resistor, the voltage at pin 6 is 0.7 + (2.5 - 0.7) * 1000 / 1010 = 2.48V when the op-amp outputs a low level. When VF falls below pin 6, pin 7 outputs a high level (if the op-amp is powered by 15V, the output at pin 7 can be calculated as 14V). The voltage at pin 6 at this point is 2.5 + (14 - 2.5) * 10 / 1010 = 2.61V. If this is an input undervoltage protection circuit with a 100:1 sampling rate for VF, the circuit will operate normally when the input voltage is above 261V, and undervoltage protection will only be applied when the voltage is below 248V. This enhances the anti-interference capability of the protection circuit.
Hysteresis comparators are commonly used in applications such as over/under voltage protection circuits and indicator light circuits.
(iv) Error amplifier output clamping circuit
In power supply design, whether it is a constant voltage source or a constant current source, as long as it is closed-loop control, an error amplifier is always required. Before entering the closed loop, the output voltage of the error amplifier is at its highest value. Normally, the power supply of the error amplifier is around 15V, so the output of the error amplifier is around 14V in the open loop. As the input signal increases, after reaching the voltage (current) regulation point, the error amplifier starts to decrease from the highest point until the value required for the closed loop. During the process of the error amplifier output decreasing, the longer the time, the larger the output overshoot and the more difficult it is for the circuit to enter stability.
Adding a diode and a Zener diode can improve this problem to some extent. As shown in the figure below, if the Zener diode is 5V, then in the open loop, the output of the error amplifier is clamped at about 6V. In this way, when entering the closed loop, the output of the error amplifier will not start from 14V but from about 6V. The time required to drop to the voltage value required for the closed loop will be shorter, and the circuit will be easier to stabilize.
You can take a look at the output of the error amplifier inside the IC. No matter what voltage the IC is supplied with, the maximum value of the error amplifier's output voltage should not be the IC's supply voltage, but rather around 6V. I wonder if this is the reason.
(V) Switching of the dual-loop control system
In circuit design, constant voltage sources with current limiting function and constant current sources with voltage limiting function are probably familiar to everyone. When designing circuits, many users sometimes use the circuit shown in the figure below, with a voltage regulator loop and a current regulator loop. As the load is gradually increased, the current regulator loop outputs a low level to enter current limiting. When the load decreases and exits current limiting, the voltage regulator loop needs a switching time. This creates a blank area where neither loop works. During this time, the circuit is equivalent to an open loop, which is not a good thing for the circuit.
However, this problem doesn't exist with the second circuit. During current limiting, the current-regulating loop lowers the reference value of the voltage-regulating loop. Both loops are active during this process. Even if the load is suddenly disconnected during current limiting, the circuit will stabilize very quickly because the voltage-regulating loop is constantly working, thus avoiding the dead zone found in the aforementioned circuit.
(vi) Measurement of leakage inductance
In the design of power transformers, everyone is familiar with how to measure the leakage inductance. Many netizens often mention in posts that their transformer has a 1mH inductance and a 600uH leakage inductance. If you also measure this, it's best to double-check, because we know that the energy stored in the leakage inductance cannot be transferred to the secondary side. If your transformer parameters are as described above, think about how efficient your transformer will be. Some netizens may wonder why their transformers, which have low leakage inductance in tests, show such large spikes in application. This is because in actual operation, not only the transformer's leakage inductance plays a role, but also the wiring inductance.
The correct method for testing leakage inductance is to first solder the transformer onto the PCB without soldering any other components. Then, short-circuit the MOSFET and output rectifier diode with a thick, short wire, and short-circuit the output filter capacitor. The leakage inductance measured from the input filter capacitor is the input leakage inductance. Conversely, short-circuiting the input filter capacitor and measuring from the output filter capacitor gives the output leakage inductance. This testing method takes into account the distributed inductance of the PCB and is closer to the actual situation.
(vii) Driving of MOSFETs
The over/under voltage and overcurrent protection circuits are controlled by two optocouplers. Under normal conditions, the optocouplers are on, and the MOSFET is on. In case of an abnormality, the optocouplers are off, and the MOSFET is off. This diagram has at least two obvious errors; please check where they are. (R6 and R7 are 1kΩ, R25 and R26 are 10kΩ)
(viii) Selection criteria for the two resistors in the feedback circuit
Taking the 384X circuit as an example, there are two common ways to connect the optocoupler isolation feedback circuit. One way is to ground pin 2 and connect pin 4 of the optocoupler to pin 1, and then pull the level of pin 1 low to achieve voltage regulation.
Some people feel that this method is unreasonable and will adopt it; the same principle applies to this method.
In the circuit, R7 and R8 are connected in a proportional amplification configuration with a gain of 1, meaning R7 = R8. Capacitor C2 mainly serves as a filter; I usually choose a very small 100pF capacitor. If the current sampling signal is within the 0-1V range, the circuit operates normally, corresponding to a COMP terminal voltage of 1V-4.4V (the internal diode voltage drop is considered to be 0.7V, and 1V is the minimum operating voltage provided by the PDF). Therefore, the voltage across R6 should vary between 0.6V and 4V. If the optocoupler transfer ratio is β, then the following formula can be obtained: 4 ≤ R6 * (V0 - 2.5 - 1.1) * β / R5
In other words, when the primary side of the optocoupler carries the maximum current, the voltage drop across R6 due to the secondary side current should be no less than 4V. Regarding the selection of R5, I mentioned in another post that the primary side current of the optocoupler is generally controlled at 5mA, which allows for the selection of the appropriate value for R6.
(ix) Commissioning of low-power flyback power supplies
For experienced designers, low-power flyback power supplies are typically powered on under no-load or light-load conditions without much trouble. The main challenge lies in parameter optimization. However, for beginners or novices who may not fully understand the circuit principles and want to reinforce their understanding through hands-on practice, the chances of their self-built power supply failing are likely to exceed 50%. Therefore, a gradual approach is preferable.
First, power the control IC separately to see if it works properly. Pay special attention to the frequency and the drive signal of the MOSFET. If the IC doesn't work properly when powered alone, the consequences of powering it on directly are obvious. After the IC works properly when powered alone, I usually find a DC output power supply with current limiting function to power my designed power supply, and then power it on under no-load to see if the output voltage is normal. Since the DC output power supply has a current limiting function, even if there is a problem, the power supply will be protected by current limiting. Once the no-load output voltage is normal, I gradually load it.
If you don't have a DC power supply with current limiting, I suggest you don't rush to add AC directly. You can connect an incandescent lamp in series at the AC input to limit the current, and then check if it works normally under no-load. If it does, then remove the incandescent lamp and add AC. This is safer.
