Rectifier spike absorption circuit

Flyback's secondary side rectifier diode has an RC spike absorption problem. I feel that everyone is still too traditional in dealing with this kind of spike problem. In fact, using RCD to absorb here will be better than using RC. Using RCD to absorb the rectifier peak voltage The voltage can be lower (with reasonable parameter matching, it can be completely absorbed and the peak voltage is almost invisible), and the absorption loss is also smaller.

Rectifier diode voltage waveform (RC absorption)

Rectifier diode voltage waveform (RCD absorption)

Judging from these two simulation diagrams, the absorption effect is equivalent. If the high voltage drop when the diode is turned on is not considered, the absorption can be considered to be complete.

After the test, you should be pleasantly surprised. SMD diodes can be used (fast switching diodes, if the parameters are suitable, 1N4148 is good), and SMD resistors and capacitors can be used.

The RCD absorption design here can be thought of as follows: In order to absorb the oscillation spike, C should have sufficient capacitance, so that after absorbing the spike energy, the voltage on the capacitor will not be too high. In order to balance the energy on the capacitor, the resistor R The leakage inductance energy stored in the capacitor C needs to be consumed, so the ideal parameter combination is that the energy consumed by the resistor is exactly equal to the energy in the leakage inductance peak (at this time, the voltage at the terminal of the capacitor C is exactly equal to Uin/N+Uo), because There are many uncertain factors in the leakage inductance peak energy, and the calculation method is difficult to be effective. Therefore, an experimental method is introduced below to design

1. Choose a larger capacitor (such as 100nF) as capacitor C, and choose an ultra-fast recovery diode (such as 1N4148) with enough voltage resistance >1.5*(Uin/N+Uo) for D;

2. You can choose a smaller resistor of 10K or 1W as the absorbing R;

3. Gradually increase the load and observe the voltage at the C terminal of the capacitor and the peak voltage of the rectifier:

If the voltage ripple on C is greater than 20% of the average value, the C value needs to be increased;

If at full load, the voltage at terminal C is too much higher than Uin/N+Uo (more than 20%, depending on the withstand voltage of the rectifier), it means that the absorption is too weak, and the resistance R needs to be reduced;

If the voltage on C is lower than or equal to Uin/N+Uo at full load, it means the absorption is too strong and the resistor R needs to be increased;

If the voltage on C is slightly higher than Uin/N+Uo (5%~10%, depending on the withstand voltage of the rectifier) ​​at full load, the design parameters can be considered reasonable;

Under different input voltages, verify whether the parameters are reasonable and finally select the appropriate parameters.

Let’s take a look at the absorption losses corresponding to the two absorption circuits (taking Flyback as an example):

Using RC absorption: the voltage on C reaches the steady state after the primary MOS is turned on is Vo+Ui/N, (Vo is the output voltage, Ui input voltage, N is the primary-to-secondary turns ratio of the transformer), because the RC we designed The time parameter is much smaller than the switching period. It can be considered that within an absorption cycle, RC charge and discharge can reach a steady state, so the energy lost in each switching cycle is: secondary leakage inductance peak energy + RC steady state charge and discharge energy , which is approximately RC charging and discharging energy = C*(Vo+Ui/N)^2 (energy is consumed on R, charging and discharging once in each cycle), so the energy consumed by RC absorption is fsw*C*(Vo+Ui/ N)^2, taking DC300V input, 20V output, transformer turns ratio 5, switching frequency 100K, absorption capacitance 2.2nF as an example, the energy loss is 2.2N*(20+300/5)^2*100K =1.4w;

RCD absorption is used. Because RCD absorption is used, the absorbed energy includes two parts. One part is the DC energy on the capacitor C, and the other part is the peak energy converted from the leakage inductance energy to C. Because the leakage inductance is very small, the peak current is impossible. is too large, so the energy is also very limited. Relatively speaking, just consider the DC energy consumed by R. With the same parameters above, the DC voltage on C is Vo+Ui/N=80V, and the resistor R is 47K. The energy consumption is 0.14W, compared with the above 1.4W, the "low carbon" effect is extraordinary.

Let’s talk about the characteristics of these two absorption circuits and other absorption circuits:

RC absorption: While absorbing peaks, it also absorbs the square wave energy output by the transformer. The absorption efficiency is low and the loss is large, but the circuit is simple and the absorption period is consistent with the switching frequency. It can be used in low standby power circuits;

RCD absorption: suitable for all applications where RC absorbs leakage inductance spikes (including forward, flyback, full-bridge, half-bridge and other topologies). The absorption efficiency is higher than that of RC, but there are situations where the energy stored in the capacitor (generally larger) is always consumed. , not suitable for application in low standby power consumption circuits (including leakage inductance absorption of primary MOS tubes);

Let’s discuss ZENER absorption again: it can be applied to primary MOS leakage inductance spike absorption, secondary rectifier voltage spike absorption, and can also be applied to low standby power consumption circuits. It has the highest absorption efficiency and high cost, but ZENER voltage stabilization parameters vary greatly. Requires careful design.

The reverse recovery of the rectifier will only occur in the continuous working mode. In the power supply topology of the intermittent working mode, there will be no reverse recovery problem of the rectifier;

The capacitance effect of the rectifier, secondary stray capacitance and secondary leakage inductance will cause oscillation. This oscillation occurs when the rectifier has a large dv/dt (voltage change rate between the transformer and the rectifier end) and the diode reverse recovery current (continuous mode ), it manifests as the resonance of the transformer output + output voltage through the secondary leakage inductance and stray capacitances such as the rectifier, thus causing a reverse voltage spike on the rectifier.

Generally speaking, the reverse recovery of a diode refers to a dynamic process in which a conducting diode transitions from a conductive state to a reverse cut-off state. There are two prerequisites here: the diode must have a certain forward current before reverse cut-off ( The current size affects the maximum peak current and recovery time of reverse recovery. The original cut-off state is not included in this list, so only the continuous mode has reverse recovery problems); in order to allow the diode to quickly enter the cut-off state, there will be a reverse voltage Added to both ends of the diode (the size of this reverse voltage also affects the reverse recovery current and recovery time of the known diode). So to see if there is a reverse recovery problem, you can compare whether it meets these two conditions.

The advantage of the quasi-resonant circuit is that the maximum terminal change voltage of the discontinuous mode rectifier diode N*Uo+Uo becomes N*Uo-Uo, which reduces the voltage change rate of the rectifier diode when the primary MOS tube is turned on, thereby reducing The excitation source of leakage inductance oscillation can reduce the oscillation peak generated by it. If the amplitude does not affect the voltage resistance safety of the rectifier, the absorption circuit such as RC can be completely omitted.