Working principle and design introduction of airborne high-frequency switching power supply

Airborne high-frequency switching power supply products are specially used for input AC 400Hz occasions. They are specially designed to meet the special purposes of military radar, aerospace, ships, locomotives and missile launches. In response to user requirements, the development of airborne high-frequency switching power supply products is of great significance to the localization of electronic weapon equipment systems, breaking international blockades, and improving the mobility and high performance of our military equipment.

There are two input modes for the power supply available on the aircraft: 115V/400Hz medium frequency AC power supply and 28V DC power supply. Both input modes have their own advantages and disadvantages. The 115V/400Hz power supply has small fluctuations and requires relatively high withstand voltage of the device; while the 28V DC power supply is the opposite. It cannot be directly provided to the equipment components. The power supply must be isolated and stabilized to become the required DC power supply before it can be used. The use environment of the airborne power supply is relatively harsh. It must adapt to a wide range of temperatures to work normally and withstand stress screening tests such as shock, vibration, and humidity. Therefore, the reliability of the design of the airborne power supply puts higher requirements on us. The following mainly introduces the switching power supply developed by the 115V/400Hz medium frequency AC input mode. Its output voltage of 270~380Vdc can be adjusted, the output power is not less than 3000W, and the ambient temperature can be as wide as -40℃～+55℃, which fully meets the needs of military-grade power supply.

System composition and main circuit design

Figure 1 shows the block diagram of the whole circuit. Its design is mainly completed by two parts: boost power factor correction circuit and DC/DC conversion circuit. The 115Vac/400Hz medium frequency AC power supply is input filtered, and the boost power factor correction (PFC) circuit completes power factor correction, boost pre-stabilization, and energy storage. Then, the DC/DC half-bridge conversion, high-frequency rectifier filter, output filter circuit, and feedback control loop are used to achieve the performance requirements of 270~380Vdc adjustable output voltage regulation.

Figure 1: Schematic diagram of the whole circuit

The boost power factor correction circuit mainly makes the input power factor meet the index requirements and realizes the boost pre-stabilization function. This part of the design takes into account the requirement that the power factor circuit meets 0.92, and makes the DC/DC input voltage appropriate so as not to overload the power factor correction circuit, so it is set at 330~350Vdc.

The isolated DC/DC converter circuit topology structures mainly include the following: forward, flyback, full-bridge, half-bridge and push-pull. Flyback and forward topologies are mainly used in small and medium power supplies and are not suitable for the 3000W output power requirement of this power supply. Although the full-bridge topology can output a large power, its structure is relatively complex. The voltage stress of the switch tube in the push-pull circuit structure is very high, and unidirectional magnetic saturation may occur in both the push-pull and full-bridge topologies, causing damage to the switch tube. The half-bridge circuit is a reasonable choice because it has automatic anti-unbalance capability and is relatively simple, with a small number of switch tubes and moderate voltage and current stress.

The DC/DC conversion circuit is mainly designed for power transformers, using IGBT/MOSFET parallel combination switch technology and half-bridge circuit balance control technology. After analysis and calculation, a double E65 core is used, with 12 turns of primary coil and 15 turns of secondary winding.

Key technology design

1 Power factor correction technology and passive lossless buffer circuit

Single-phase input power factor correction circuits with sinusoidal input current are increasingly used in switching power supplies. Figure 2 shows a boost power factor correction and passive lossless buffer circuit.

Figure 2 Power factor correction and new passive lossless snubber circuit

The passive and lossless snubber circuit is adopted, and all components are passive devices such as L, C, and D. It has both zero current conduction characteristics and zero voltage shutdown characteristics, which is 30% less than the traditional lossy snubber circuit components. The snubber circuit components include L1, C1, C2, D1, D2, and D3.

UC2854A can be used to control the main switch SWB. Its buffer circuit does not need to be controlled and has the characteristics of simple circuit. Its principle is to transfer the energy of diode DB reverse recovery and the energy stored in C2 when SWB is turned off to C1 when SWB is turned on. When SWB is turned off, the energy stored in L1 charges C2 and is transferred to CB through D1, D2, and D3. At the same time, it also discharges to CB. This circuit realizes zero voltage shutdown and zero current conduction, effectively reducing losses and improving circuit efficiency and reliability.

The main features of this circuit are:

The maximum voltage on the switch SWB is the output voltage VL.

The maximum reverse voltage on the Boost diode DB is VL+VE, and the VE value is determined by the relevant values ​​of IR, L1, C1 and C2.

The maximum current rise rate on switch SWB is determined by L1 and V1, and the conduction loss and stress are very small.

The maximum voltage rate on switch SWB is determined by C2, and the turn-off power consumption and stress are very small.

During the switching cycle, the energy stored in L1 and C2 to obtain control of the current and voltage rise rate is eventually returned to the output power supply, thus ensuring true lossless operation of the circuit.

2 IGBT/MOSFET parallel combination switch technology

Figure 3 shows the IGBT/MOSFET parallel combination switch circuit and operating waveform. Compared with MOSFET, the IGBT on-state voltage is very low, and the current quickly drops to 5% of the initial value when it is turned off, but it takes a long time to reduce to zero, about 1 to 1.5μs, which will cause great switching losses in hard switching mode. In the combination switch, the parallel MOSFET is turned off only when the tail current has been reduced to nearly zero 1.5μs after the IGBT is turned off.

Figure 3 IGBT/MOSFET parallel combination switch circuit and operating waveform

这种技术因通态损耗很低而使得DC/DC变换器的效率很高。但需工作频率相对较低，一般选取20～40kHz。由于半桥组合开关只需两个开关，总的开关器件的数目少，使可靠性显著提高。

3 Half-bridge circuit balance control technology

By controlling and adjusting the delay time of the IGBT/MOSFET gate drive, the half-bridge can be balanced to avoid transformer saturation overcurrent and burnout of the switch tube. This is easy to achieve when the pulse is wide. However, when there is a light load or no load, the pulse width is very narrow (for example, less than 0.3μs), and the IGBT/MOSFET delay is canceled at this time. Therefore, in order to maintain its balance when the pulse width is narrow, we use a low-frequency oscillator. When the pulse width is less than 0.3μs, the oscillator starts to make the PWM generator work intermittently, keeping the pulse width not less than 0.3μs to maintain the balance of the half-bridge and enable it to work normally when there is no load.

Due to the low operating frequency, the switching loss of the combination switch is very small and the conduction loss is also very small.

Figure 4 Half-bridge circuit balance control circuit

4 Multi-loop control circuit

The average current mode control system uses a PI regulator, and needs to determine two parameters: the proportional coefficient and the zero point. The calculation principle of the regulator proportional coefficient KP is to ensure that the slope of the rising phase of the current regulator output signal is smaller than the sawtooth wave slope, so that the current loop will be stable. The zero point is selected in the lower frequency range, at 1/10 to 1/20 of the angular frequency corresponding to the switching frequency, to obtain a sufficient phase margin at the open-loop cutoff frequency.

In addition, a pole located near the switching frequency is added to the PI regulator to eliminate the interference of noise generated during the switching process on the control circuit. The structure of such a PI regulator is shown in FIG5 .

Figure 5 PI regulator with filtering function

The core of the control circuit is the design of voltage and current feedback control signals. In order to ensure the response speed under the premise of system stability, a multi-loop control technology with voltage loop as the main factor is designed. The current loop responds to the change of load current and has a current limiting function. The design circuit adds differential amplification after sampling the output inductor current, and adds it to the feedback loop to participate in the control after isolation. The regulator gain can be adjusted through the amplifier link with a potentiometer in the later stage. In this way, the power supply works under high-precision constant voltage state, and the output dynamic response makes the power supply have no large output voltage overshoot in the case of sudden load change.

5. Improve heat dissipation effect and reduce thermal resistance

In order to reduce the size of the whole machine and achieve a reasonable power density, forced air cooling is adopted. For air-cooled radiators, the wind speed is directly related to the quality of the heat dissipation effect. Since front and rear ventilation is required, the following should be considered during design:

Ensure that the wind speed meets certain requirements (V = 6m/s) and consider the influence of wind pressure. When the wind pressure is lower than the radiator head loss, the cooling wind cannot blow through or the wind speed is very low, and the purpose of improving the heat dissipation rate cannot be achieved.

Since the gap between the radiator and the fins is very different from the gap between the air duct and the radiator, when the wind pressure is too low, a baffle plate or a trumpet-shaped inlet can be added between the gap between the radiator and the air duct at the air inlet to force the wind to flow between the fins of the radiator.

The boost inductor, main transformer, and output filter inductor are fixed in a row on the upper half of the radiator, and the mainboard is fixed on the lower half. The power devices on the mainboard, such as the power switch tube and the output rectifier tube, are fixed to the radiator through steel strips. Low-quality components are placed on the upper half of the mainboard, and high-quality components are placed on the lower half. The fan is placed in the upper middle position in front of the radiator and fixed on the front panel, adopting a front-in and rear-out air flow method.

Military high-frequency switching power supply products should not only consider the power supply parameters themselves, but also electrical design, electromagnetic compatibility design, thermal design, structural design, safety design and three-proof design. Because even the slightest negligence in any aspect may lead to the collapse of the entire power supply, we should fully realize the importance of reliability design of military high-frequency switching power supply products.

Test results

The design parameters were tested and the test results are shown in Figures 6 to 8.

Figure 6 DC/DC primary voltage waveform (full load)

Figure 7 DC/DC secondary voltage waveform (full load)

Figure 8 High-frequency inductor current simulator waveform

It can be seen from Table 1 that the test results are in compliance with the protocol requirements, among which the power factor, efficiency, line regulation, load regulation, output noise and other parameters are better than the protocol requirements.