Design of a New Monolithic Switching Power Supply

drly

Design of a New Monolithic Switching Power Supply [Copy link]

The flyback conversion principle derived from Buck-Boost is shown in Figure 1. Since the circuit is simple and can provide DC output efficiently, it is particularly effective for multi-channel outputs and is therefore widely used as the internal power supply of power electronic devices.

Figure 1 Principle of flyback converter

In a flyback converter, there are generally two operating modes: complete energy conversion (discontinuous inductor current) and incomplete energy conversion (continuous inductor current). The small signal transfer functions of these two operating modes are very different, and different treatments are required during dynamic analysis. When the converter input voltage changes within a large range, or the load changes within a large range, it is inevitable to cross the two operating modes. Therefore, the flyback converter is often required to work stably in both complete and incomplete energy conversion modes. The TDA1683x introduced in this article is a current-type IC, so when used in a flyback converter, it can simplify many design issues of the complete energy conversion type. In the incomplete energy conversion type, due to the "right half plane zero" of the transfer function, a 180° phase change is introduced in the high frequency band, and the current mode control cannot eliminate the inherent instability problem, which requires the control loop gain to deviate from the low frequency band and to reduce the transient response speed, all of which must be achieved by adjusting the PID constant. At the same time, the transfer function of the incomplete energy conversion type has a two-pole system with low output impedance. When the output load increases, the pulse width only needs to be slightly increased, which can increase the output load capacity. During the design process, proper control of magnetic parameters can enable the power supply to operate within a larger dynamic range.

When using the incomplete energy transfer method, due to the presence of DC components, an air gap needs to be added to increase its dynamic range and output power. At the same time, a small △B is selected during design to reduce losses and improve the stability of the power supply. Generally, the power transferred by the transformer can be expressed by the following formula: Where: Ve represents the volume of the magnet, Lp represents the length of the air gap, and Ip1 and Ip2 are the instantaneous current values at the beginning and end of the conduction cycle. Obviously, the existence of the air gap increases the power transferred (the shaded part of Figure 2).

(a) Magnetization curve and energy transfer of the transformer in a flyback converter when the core air gap is very small

(b) When the air gap of the magnetic core is large, the magnetization curve and the transferred energy of the transformer in the flyback converter

Figure 2

It is worth noting that increasing the air gap has a significant inhibitory effect on the magnetic saturation caused by the load, but it is useless for the magnetic saturation caused by too high input voltage or too low switching frequency.

In short, the applied volt-second value, number of turns and core area determine the △Bac on the B axis, and the average value of the DC current, number of turns and magnetic path length determine the position of the Hac value on the H axis. In order to prevent magnetic saturation, there must be enough coils and core area to balance the applied volt-second value, and there must be enough air gap to balance the DC component. This is very important in transformer design.

In actual design, the energy transfer method is selected by changing the primary inductance and adjusting the air gap size. Experience has shown that a suitable air gap can make the instantaneous value of current, efficiency, and noise more appropriate. In addition to calculation, the determination of the air gap also requires actual adjustment to achieve the best indicators above.

2 Working Principle

The TDA1683x series of current-mode monolithic switching power supplies newly launched by Siemens are characterized by small size, simple peripheral circuits, and wide temperature range. They are very suitable for working power supplies of power electronic equipment such as inverters and frequency converters. The working principle of this current-mode control device is shown in Figure 3. The magnitude of the current flowing through the MOS tube is compared with a given voltage value, and its output determines the conduction time of the MOS tube; the output voltage is related to the given voltage. In order to simplify the external circuit, the current detection resistor of the MOS tube is directly integrated into the IC. At the same time, the oscillation resistor and capacitor that determine the frequency of the monolithic power supply are also integrated into the IC. In order to minimize the influence of temperature on the operation of the chip, special temperature compensation measures are taken and the error of the resistor is minimized.

Below, we take TDA16833 as an example to analyze the functions of its internal structure.

Figure 3 Working principle of TDA16831-9 series current-mode monolithic switching power supply

(1) Start-up circuit (UVLO): This part is connected to the external Vcc terminal. When the Vcc voltage is higher than the fixed turn-on threshold voltage, the UVLO works, and the bias circuit and soft start circuit are closed. Once the Vcc voltage is lower than the turn-off threshold voltage, the circuit is turned off. At this time, the current consumption of the circuit is limited to a minimum.

(2) Current source (bias): This circuit provides a constant current to the internal part for current limiting, timing, soft starting, etc.

(3) Softstart: This function is limited to TDA16836/7. After the chip starts to oscillate, the on-time of the MOS tube gradually increases during the rise of the soft-start capacitor voltage, avoiding the adverse effects caused by the instantaneous increase of the on-time of the MOS tube to the maximum value, such as preventing the current saturation of the power transformer, limiting the surge current formed by the output filter capacitor, and the instantaneous overshoot voltage of the output. In particular, when the power supply of the chip is provided by the auxiliary winding of the power transformer, the impact caused by the output overshoot voltage will be very large. At this time, when the output power is large, the overshoot voltage generated by the transformer often exceeds the maximum allowable value of Vcc.

(4) Oscillator (OSC): This part has a frequency-dividing trigger that can reduce the oscillation frequency to half to form the switching frequency. At the same time, it can make the maximum duty cycle Dmax = 0.5. If necessary, an auxiliary circuit can be used to make Dmax = 0.75.

(5) Error amplifier (pwmop): amplifies the incoming MOS current signal and converts it into voltage. Generally, when the MOS tube is chopped, a spike pulse will be formed in front of the current signal, which will invalidate the comparison result, especially under light load. This amplifier provides a leading edge blanking function, and the actual current signal is externally biased to limit this phenomenon.

(6) Comparator: This comparator compares the amplified MOS tube current signal pwmrmp with the reference signal pwmin from Vout. When the current signal is greater than the reference signal, the MOS tube is turned off. There is also an auxiliary comparator in the figure, which sets a very low threshold voltage to achieve a 0% duty cycle.

(7) Logic control (logpwm): This circuit consists of an RS trigger and a NAND gate. This part controls the on and off of the MOS tube. The conditions for turning on the MOS tube are that all the following conditions are met: soft start starts, the pwmin signal exceeds the minimum threshold, the pwmin signal is higher than the pwmrmp, the overcurrent shutdown is invalid, the overheat shutdown is invalid, and the tff circuit outputs a start pulse. The conditions for turning off the MOS tube are that any of the following conditions are met: pwmrmp exceeds pwmin, or the duty cycle exceeds 0.5, or pwmcs exceeds Imax, or the temperature of the silicon chip exceeds Tmax, or UVLO is lower than the minimum threshold. The function of the RS trigger is to ensure that only one turn-on operation is performed in each cycle.

3. Design of dedicated driving power supply

Siemens' TDA1683x series single-chip switching power supplies can be applied to power electronic drive systems with high reliability requirements due to their unique performance. The following is a brief introduction to the design of the IPM module (CTM-series) power supply.

3.1 Circuit Principle

The circuit principle is shown in Figure 4. It is suitable for the power supply requirements of three-phase inverters below 3kW. Due to the small number of peripheral components, the circuit is simple and the reliability is very high. TDA16832 itself does not require a heat sink.

Figure 4 Schematic diagram of the switching power supply circuit of the variable frequency air conditioner

3.2 Hardware design and debugging

(1) Inverter transformer: The IVT-28P transformer specially used for inverter power supply by TOKIN is used, and the arrangement order is shown in Figure 4. Due to the special structure of this transformer, it is guaranteed that each winding can withstand the AC1500V (1min) withstand voltage test.

　　(2) Output waveform (see Figure 5): Figure (a) shows the waveform when the input voltage is AC220V and the load is 3RL[Note]; Figure (b) shows the waveform when the input voltage is AC240V and the load is RL. In both cases, the power supply does not operate in continuous current mode.

(a)

(b)

Figure 5 Output waveform

4 Conclusion

The switching power supply adopts the TDA16832 single-chip switching power supply. While meeting the requirements of the inverter power supply of the variable frequency air conditioner, the load has 4 relays, works reliably and stably, generates little heat, and does not require a heat sink. At the same time, the peripheral circuit is simple, debugging is very convenient, and the development cycle is simplified.