Heating analysis of ECP high frequency transformer

In the design and manufacture of ECP high-frequency transformers, we often encounter situations where all the steps have been done, but the heat generated by the transformer cannot be reduced. At this time, it shows that there are still some things that are easy to be overlooked, which can lead to Other factors related to fever exist.

When the transformer design has "no problem" but the operating temperature cannot drop, we have to shift our attention from the transformer to the "device characteristics" of the peripheral circuit. In fact, it comes from the "interaction with the transformer" caused by the peripheral device. The excessively high operating temperature caused by the "function" also occupies a considerable proportion of the heat generated by the transformer. After all, for the entire power supply, whether it is the primary side switch tube, the secondary side rectifier tube, or absorption compensation, The resonant circuit (inductance or capacitance), even PFC, filter capacitor, PCB wiring, etc., belong to the same whole with the transformer, and their working conditions must be interrelated and affect each other, only the strength of the effect.

Among them, the reverse recovery characteristic of the secondary side rectifier (freewheeling) diode has the greatest influence on the working temperature rise of the transformer. Take the common high-power power supply as an example (it is not difficult to analyze the working condition of the low-power flyback secondary rectifier diode). The two full-wave rectifier diodes on the secondary side of the bridge topology, or the rectifier and freewheeling diodes of the forward topology, will produce instantaneous common-state conduction during the reverse recovery period, thereby causing a sine of decreasing amplitude in the leakage inductance (Sometimes not completely sinusoidal) spike oscillation. This oscillation wave with a much higher switching frequency and a higher voltage peak will be coupled between the primary and secondary sides, which will additionally increase the various losses of the wire package and the magnetic core, especially It is the loss that is exponentially proportional to the frequency, and the increase is more obvious.

Because in the first peak period of the diode "common state conduction" moment, the excitation inductance of the primary side drops to close to: "the leakage inductance value of the primary side measured on the secondary side of the short circuit". The instantaneous peak current of the side will exceed several times to several ten times that during normal operation! At this time, the magnetic swing △B of the magnetic core will increase, and the high-frequency current density of the winding wire will also increase sharply. In the subsequent attenuation oscillation process Although the loss is decreasing, the entire peak attenuation oscillation is generated repeatedly with the operating frequency, so it is not difficult to imagine that the wire temperature and iron temperature will increase a lot. Of course, such spikes will also adversely affect the reliability of the power supply.