Method provides overpower protection for quasiresonant supplies

The main characteristic of quasiresonant switch-mode power supplies is that they exhibit a varying frequency when the input voltage changes. For a flyback power supply, the power delivered to the output obeys the following equation: P_OUT=1/2(L_P×I_P²×F_SW×η), where L_P is the primary inductance, I_P is the peak primary current, F_SW is the switching frequency, and η is the efficiency. Because L_P and η are fixed with a moving switching frequency F_SW, I_P has to move in the opposite direction to maintain a constant output power. So, when input voltage V_IN rises, F_SW increases; as a result, I_P needs to decrease in response to the feedback-loop requirements. For a widely varying line-voltage application, the peak current almost doubles between high and low input voltages for a constant output power. But quasiresonant controllers feature only overcurrent protection. This limitation is part of a structural problem. The controller monitors the peak current, and, when they reach the maximum allowed value, the controller circuitry detects an overload. Unfortunately, if the power supply delivers its nominal power at the lowest worst-case input voltage, it delivers more power for a higher input voltage. For a widely varying line-voltage application, this power could be more than three times higher. This fact is the consequence of the flyback equation.

A classic way to compensate this variable-power effect is to create an offset on the current-sense pin that compensates for the peak-current variations as a function of the input voltage, V_IN. You obtain this effect using an overpower-protection scheme—wiring a compensation resistor from the high-voltage rail to the current-sense information (Figure 1a). Unfortunately, you cannot always implement this scheme. Whether you use the CS (current-sense) pin for another function or you need to keep the pin impedance low for noise purposes, it forces you to adopt a low value for resistor R_CS in series with the current-sense information. It then requires a low-value compensation resistor, R_COMP, wasting a lot of power. When you need low standby power, this approach is unacceptable. To overcome the problem, it might be useful to use a fraction of the input voltage to lower the voltage drop on R_COMP. The power the resistor wastes would then become negligible.

You achieve this method by using the forward voltage of an auxiliary winding. On a forward winding, a voltage proportional to V_IN occurs during the on-time, which is the scenario you are looking for. Usually, you use a flyback auxiliary winding to supply the controller and to detect the core-reset event. By modifying the arrangement of the winding, you can generate the flyback information for the demagnetization detection during off-time and combine, on the same winding, the forward information for the overpower compensation during on-time. By adding a diode in series with the auxiliary winding, you can access the forward voltage (Figure 1b). This forward voltage is proportional to N×V_IN, where N is the turns ratio between the primary and the auxiliary windings. You add R_FWD to supply the reverse current during the forward activity.

Knowing the value of the forward voltage and the series resistor, R_CS, you can then easily calculate the value of compensation resistor R_COMP to create the desired offset on the current-sense signal at high input voltage. On a demonstration board built on the NCP1207 from On Semiconductor (www.onsemi.com), D₁ is a 1N4448 diode, R_CS=680Ω, R_COMP=18 kΩ, and R_FWD=4.7 kΩ, the protection toggles at 60W at 100V dc and 70W at 365V dc, instead of 55W at 100V dc and 165W at 365V dc without compensation (Figure 2). Figure 3 shows circuit waveforms of the line compensation at V_IN=365V, and Figure 4 shows the waveforms at V_IN=100V.

Check out our Best of Design Ideas section!