The latest technology in quasi-resonant control ICs! Part 1

Narrow down by specifying conditions

現在2149件がヒットしています。check

Basic analog discrete power on semi

Introduction

Among the many ACDC switching power supplies, flyback power supplies are widely used as general-purpose power supplies that can cover an output range from several W to about 100W.

The flyback power supply also has various drive methods. The control IC on the primary side has a fixed frequency method in which the switching element is turned on at a constant cycle (here, it also includes a method in which the frequency is variable according to the load), and the valley of the ringing waveform of the drain voltage. There is a quasi-resonant control method (Quasi-resonant, QR) that detects and turns on the switching element at that timing. The quasi-resonant control method is known as a drive method that improves efficiency and effectively suppresses EMI noise. ONSEMI's quasi-resonant control ICs have many achievements in the market, and their functions are updated year by year.

This article introduces the state-of-the-art technology in ONSEMI's quasi-resonant control IC, NCP1342.

<Contents>
1. How is the switching frequency determined in quasi-resonant control?
2. Advantages and disadvantages of quasi-resonant control
3. Introducing the latest functions of NCP1342
1) HV start circuit
2) X2 capacitor discharge
3) Brownout protection
4) Valley detection
5) Valley Timeout

1. How is the switching frequency determined in quasi-resonant control?

A fixed frequency control IC has an oscillation circuit that generates a fixed frequency signal inside the IC. The signal of this oscillator circuit determines the turn-on of the switching element. The turn-off timing is determined by the FB signal (usually a coupler current signal) generated according to the secondary output voltage.

Then, how is turn-on and turn-off determined in quasi-resonant control? In quasi-resonant control, turn-on is determined according to the drain voltage waveform of the switching element, not by the oscillator circuit built into the control IC.

Figure 1: QR operating waveform of Flyback power supply

Figure 1 shows the waveforms of each part of the quasi-resonant control.
From the top, blue indicates the primary side AUX voltage, green indicates the primary side switching element (MOSFET) drain current, red indicates the secondary side diode current, and light blue indicates the primary side gate drive signal. While the primary MOSFET is on, the drain current increases as shown, and when the FB signal reaches the set level, the MOSFET is turned off. Turn-off control is the same as for fixed frequency control, with the FB signal determining its timing.

After the primary MOSFET turns off, the secondary diode current begins to flow. At the timing when the secondary-side diode current stops flowing (called Demagnetizatoin), the drain voltage VDS of the MOSFET begins to drop, creating an LC resonance waveform as shown in the figure. The figure shows the auxiliary winding voltage VAUX, but the drain voltage waveform is VAUX multiplied by the turns ratio. The LC resonance waveform of this drain voltage becomes a ringing waveform that decays as shown in Figure 1.

Quasi-resonant control causes the next turn-on to occur at the valley of this ringing waveform. Therefore, the turn-on timing is determined by the operating waveform of the flyback power supply, not by the oscillation circuit built into the control IC. Specifically, the L value, turns ratio, input voltage, output voltage, and output current of the transformer are mainly involved.

2. Advantages and disadvantages of quasi-resonance control

Quasi-resonant control is a very effective way to improve efficiency and suppress EMI noise in flyback power supplies.

The advantage of quasi-resonant control is that the turn-on loss is small because the drain voltage is low when the switching element turns on. Unlike CCM operation, the drain current starts at zero amperes, which also helps reduce turn-on losses. This feature reduces switching loss and at the same time contributes to EMI noise suppression.

The disadvantage of quasi-resonant control is that it basically operates in DCM, so the peak current flowing through the switching element is larger than in the fixed frequency method that operates in CCM under the same load conditions. In other words, a switching element with a higher current rating is required. At the same time, it also has the disadvantage of increasing the output voltage ripple and output current ripple.

3. Introducing the latest functions of NCP1342

1) HV starting circuit

The NCP1342 has a 700V tolerant HV terminal. By connecting the HV terminal to the high voltage line of the input, the circuit current at start-up is supplied. In the case of a product that does not have such a high withstand voltage HV pin, for example, the VCC pin with a withstand voltage of 30V is connected to the input high voltage line. By inserting it, overvoltage is prevented from occurring at the VCC pin. Therefore, it is necessary to keep the current flowing through the starting resistor at all times.

The high withstand voltage switch element (e.g. JFET) provided inside the HV terminal of the NCP1342 is turned off except at startup, so there is no wasteful loss, making it an essential function for applications such as adapters that have strict regulations on standby power. Become. However, in general, a controller with a HV pin is more suitable for products because a special process is added to incorporate a 700V withstand voltage device, and because the 700V device itself occupies a larger size on the chip. The unit price will be higher.

2) X2 capacitor discharge

Figure 3: Operating waveform when AC input is removed

Electrical Appliance and Material Safety Law J60335-1 stipulates that "the voltage between the pins of the plug must not exceed 34V one second after disconnecting the device from the power supply." This is because there is a risk of electric shock if a person accidentally touches between the input terminals of the device. Usually, this is countered by providing a discharge resistor in parallel with the input electrolytic capacitor. However, since this discharge resistor constantly consumes power, there has been a demand for a countermeasure to replace the discharge resistor, especially in applications where standby power is emphasized. That's why we developed the X2 capacitor discharge function.

For this function, connect the HV terminal to the AC input through a diode and resistor as shown in Figure 2. When the slope (change over time) of the AC input voltage is negative or falls below a certain value, the built-in timer starts to operate, and if this state continues for 100ms or more, it judges that the AC power supply has been disconnected from the equipment, and the HV terminal A discharge of the input electrolytic capacitor to GND is performed. In addition, since there is a risk that the chip temperature of the NCP1342 will rise if the discharge continues for a long time, it will be an intermittent discharge operation in which discharge is restarted after a certain period of discharge and a rest period. In addition, in a power supply with a small output, the loss caused by the discharge resistor may be tolerable. In that case, the X2 capacitor discharge function is unnecessary, and the two diodes and resistors required for this function are nothing more than a cost increase factor.

The NCP1342 offers many options with different capabilities. Among them, if the X2 Discharge Disabled option is selected, these two diodes are unnecessary and the HV terminal can be used by connecting it to Bulk after rectification and smoothing.

3) Brownout protection

Figure 4: Waveform during brownout operation

Brown-out protection is said to have originally been introduced in CRT TVs to prevent screens and indicators from flickering even after the power is turned off. The NCP1342 monitors the input voltage on the HV pin. Connect the HV terminal to the AC input through a diode as shown in Figure 2. The HV terminal voltage has a full-wave rectified waveform as shown in Fig.4. Every time the HV pin voltage decreases and falls below VBO(stop), the internal brownout timer circuit operates.

Then, when the HV pin voltage increases again and reaches VBO(start), this brownout timer circuit is reset. If the HV pin voltage does not reach VBO(start) and is held at a low level, the brownout timer circuit continues to operate. is stopped.

Next, the HV pin voltage rises above VBO(start), and then restarts when VCC reaches VCC(on).

Figure 5: Each terminal waveform at start-up

Figure 5 shows the waveform of each pin at startup.
When the HV pin voltage rises during start-up, the start-up current is supplied internally from the HV pin. This start-up current has two stages, a smaller current of Istart1 (0.5mA) while VCC is low and Istart2 (2mA) when VCC rises to VCC(inhibit). This function reduces the risk of heat generation due to continuous flow of start-up current when, for example, the VCC pin is shorted to GND. At startup, switching operation starts after the HV pin voltage exceeds VBO(start) and VCC reaches VCC(on).

The NCP1342 does not require an AC input voltage detection resistor, but the Brownout detection level is fixed inside the IC. Onsemi offers two types of detection levels to meet the various needs of our customers.

4) Valley detection

Valley detection connected to auxiliary winding ZCDs This is done with a pin. ZCDs voltage on the pin VZCD(trig) (55mV), Valley is detected as

old quasi-resonant control I C If, MOSFET can be turned off first Valley (1 ^st Valley), but at light load, MOSFET Since the ON time of the IC becomes shorter, the frequency rises as the load becomes lighter, resulting in an increase in switching loss and a drop in efficiency. Recent quasi-resonant control I C In order to reduce switching loss at light load, 1st Valley, 2nd Valley, 3rd Valley A common technique is to suppress the increase in frequency by shifting the turn-on timing.

The NCP1342 can detect up to ^6th Valley and selects the Valley according to the FB voltage.
Since the valley is selected according to the load level, the relationship between output power and frequency has a sawtooth wave characteristic as shown in Fig.6. In FIG. 6, in the conventional quasi-resonant control IC, when the load is about 21 to 22 W, for example, it is near the boundary between the 2nd valley and the 3rd valley, so there is a phenomenon that both valleys go back and forth every switching cycle. could have occurred. Such behavior causes audible noise.

The NCP1342 incorporates a Valley lock out (VLO) circuit that eliminates such behavior. Once a valley is selected, it remains locked to this selected valley until the output power changes significantly, thus maintaining stable operation and suppressing audible noise.

5) Valley Timeout

Valley is detected by the built-in comparator of the ZCD pin, but if the drain voltage ringing suddenly decays, the ZCD pin comparator may not be able to detect it. In such cases, the controller will wait for the Valley to be detected, thus stopping switching.

NCP1342 incorporates a maximum timeout period from the Demagnetization timing to handle such cases. Even if Valley is not detected, the next turn-on is forcibly executed after the maximum timeout period has elapsed.
The maximum steady-state timeout period is set to 6us.

On the other hand, at startup, the output voltage has not yet risen, and the drain voltage is small when the MOSFET is off. This makes valley detection more difficult, and the maximum timeout period drives switching. Setting the same maximum timeout period as steady state at start-up results in CCM operation, so a longer maximum timeout period (100us) is applied. After repeating switching with this maximum timeout time several times, when the output voltage rises and the valley can be detected, it becomes QR operation by turn-on in the valley.