A new 200kHz/200W environmentally friendly switching power supply

　　Nowadays, there are more or less some troubles in the environmental performance evaluation of high-frequency practical switching power supplies with rated power above 200W. They either have large EMI noise, or input current harmonics exceed the standard, or the temperature characteristics are not good at a certain power packaging density, the reliability is poor, etc. To solve these problems, one way is to find a new converter topology with more advanced performance, and the other way is to choose new processes and new devices to meet the requirements of environmental performance evaluation as much as possible.

　　In recent years, some well-known foreign semiconductor companies have spent a lot of effort to transform device technology and develop a series of targeted new devices with superior performance. For example, Infineon, formerly Siemens, has launched a package of devices dedicated to solving the above-mentioned problems of high-frequency switching power supplies in recent years. They include CoolMOS tubes with high voltage resistance of 600V and low on-resistance (Rdson) (extremely low temperature rise when used at high frequency, suitable for Boost switch), OptiMOS tubes with high current, low withstand voltage and low Rdson (especially suitable for Buck converters), PFC ?PWM two-in-one ICTDA16888 (can save space and components), high voltage (600V) SiC Schottky diodes (especially suitable for Boost diodes), etc. These devices have special characteristics. If used properly in the design of switching power supplies, they will solve the problem with half the effort and the cost will be controlled.

　　As an example, this article intends to introduce to readers a practical switching power supply with an operating frequency of 200kHz and a power of 200W that meets environmental protection requirements, which is synthesized using the above devices. It uses the second-generation CoolMOSC2 as the power switch for PFC and PWM, SiC Schottky diodes as PFC diodes, OptiMOS as synchronous rectification switches, and the control of PFC and PWM is implemented by the same ICTDA16888. The power supply has a wide input voltage range (90V~275V) and an AC/DC conversion efficiency of more than 80%. There are two groups of output voltages: +5V/20A and +12V/8.3A, with output overload protection and output short-circuit protection. All power devices do not require heat sinks, nor do they require a minimum output load.

　　2 Circuit block diagram

　　Figure 1 shows the working block diagram of the overall power supply. It consists of two parts: PFC and PWM. The first part is an AC/DC converter for power factor correction (PFC), and the second part is a DC/DC converter with forward pulse width modulation (PWM) composed of two power switch tubes. The PFC stage is a Boost converter, which provides a 380Vd.c. at its output while keeping the input current at the input end as a sine wave to obtain a power factor approximately equal to 1. Another feature of the PFC stage is that it allows the power supply to operate in a wide voltage input range (90V~275V) without adding a voltage range switch to reconfigure the rectifier circuit. The power devices used are two CoolMOS type SPB11N60C2 used in parallel and a SiC Schottky diode SDB06S60 (6A/600V).

Figure 1200WSMPS block diagram

　　The dual-tube forward converter is isolated from the grid through the coupling transformer T1. In the primary of the transformer, the power devices are two CoolMOSSPB11N60C2 and two EMCON diodes SDD04E60 (4A/600V). The secondary has two sets of outputs (5Vd.c. and 12Vd.c.), but their rectification principles are different. The 12V output uses a traditional Schottky diode rectification circuit, while the 5V output uses a low-voltage MOSFETSSPB80N03S2L?03 for synchronous rectification.

　　The functional control of both PFC and PWM parts is completed by a single chip integrated circuit TDA16888.

　　3Structure/heat sink design

　　One of the advantages of this power supply is its small size. It consists of two double-sided PCB boards of different sizes. The larger one (18cm×15cm) is the main board, which is equipped with various power devices and passive components, and uses SMD components with small footprint as much as possible. The device does not use any heat sink, and the heat dissipation is achieved by the main copper sheet on the PCB board transferring the heat energy to a metal plate below. The smaller one (6cm×3cm) is the control board, which is equipped with the control circuit and is inserted vertically into the main board.

　　4. Component Function Description

　　4.1 Power supply motherboard

　　The principle of the power supply motherboard is shown in Figure 2, which includes the following parts:

　　(1) AC input/EMI filter

　　The input voltage of SMPS is 90V～275V (50Hz/60Hz). The fuse is used to prevent further damage to the power supply when a circuit failure occurs. The input EMI filter (C86, L1, L4, C24, C25, C26, C2) is used to suppress the high-frequency noise generated by the conversion of the two power switches. The varistor R30 is used to resist high-voltage surges from the power grid. The input power rectifier (D1?D4) uses conventional silicon diodes.

Figure 2 200WSMPS mainboard electrical schematic

　　(2) PFC converter

　　This is a boost converter topology with continuous inductor current flowing through full load. The switching frequency is 200kHz. The output voltage is approximately 380Vd.c.

　　The core of PFC is boost inductor L2, switch tube Q1A/Q1B, boost diode D5 and large electrolytic capacitor C3. In order to reduce parasitic capacitance, L2 is made of a single copper wire wound on a toroidal iron powder core. The parallel tube Q1A/Q1B is made of SPB11N60C2 using the new CoolMOS process. They have high switching speed and extremely low on-state resistance. This advantage is particularly important at 90V low input because the circuit is running at high current and high duty cycle. The purpose of the two tubes in parallel is only to expand the heat dissipation area so that the heat distribution on the PCB board is more balanced. Boost diode D5 is a 600V SiC Schottky diode, which has very good switching characteristics because it has no charge storage (no reverse recovery and no temperature interference on the switching characteristics). D82 is a traditional silicon diode used to charge the electrolytic capacitor from the initial rectified voltage to avoid the SiC diode D5 from being subjected to excessive surge current at the moment of power on. Electrolytic capacitor C3 is used to store energy to reduce the voltage ripple of the second harmonic, and it must also withstand the current of the switching frequency. Capacitor C3A is specifically used to bypass high-frequency harmonic currents.

　　(3) PWM converter (dual-switch forward type)

　　The PWM converter is a two-transistor forward converter topology. Its operating frequency is also 200kHz. The main parts on the primary side are Q2A/Q2B and D22/D27. ​​When the forward transistors Q2A/Q2B are turned on at the same time, energy is transferred to the output through the transformer. Q2A/Q2B selects CoolMOSSPB11N60C2 with high switching speed. D22/D27 selects EMCON sample diodes. During the cut-off period of Q2A/Q2B, D22/D27 is used to clamp the feedback spike voltage generated by the transformer leakage inductance during the transformer flux reset period. Transformer T1 is powered by the DC voltage on the electrolytic capacitor C3 and isolates the output from the input. It uses the RM combined magnetic core RM14/N87 produced by EPCOS (see Figure 3). Its primary winding is made of stranded wire Litz and the secondary is wound with thin copper tape.

Figure 3 Transformer structure

　　In order to reduce leakage inductance, the primary and secondary windings can be wound in an interleaved manner.

　　The secondary is D20/D21, L3A, L6 and C36/C37 of the 12V channel and Q19/Q21, L3B, L5 and C15, C28 of the 5V channel. Among them, D20/D21 is a 45V standard Schottky diode, which works in two timings: D20 acts as a rectifier diode when Q2A/Q2B is turned on, and D21 acts as a freewheeling channel for the load current when the Q2A/Q2B transistor is turned off.

　　(4) Synchronous rectification

　　In the 5V channel, a synchronous rectifier made of three low-voltage 30V/80A OptiMOSSPB80N03S2L?03 is used. Its control signal is generated by the secondary. Two OptiMOSQ19 and Q19A are connected in parallel, and they jointly provide a freewheeling current channel for the "low-state" PWM. The OptiMOSQ21 is used for series rectification. At the moment of transformer primary reset, the PWM pulse output disappears, and the synchronous rectifier Q19/Q19A is turned on through the body diode of Q18. When the primary turns on, the gate of Q18 (previously in negative bias) is driven by the secondary winding voltage through resistor R97, and Q18 turns on, causing Q19/Q19A to turn off. Q21 turns on under the linkage of R96, L3A and L3B, starting a new round of synchronous rectification cycle.

　　4.2 Control circuit

　　The control board circuit of 200WSMPS is shown in Figure 4, which is composed of a hybrid dual ICTDA16888 and its peripheral components.

Figure 4 200WSMPS control board electrical schematic diagram

　　(1) Hybrid dual ICTDA16888

　　TDA16888 is a new product developed by Infineon in recent years. It provides full control of SMPS with PFC. By using the internal synchronous PFC and PWM functions, it can adapt to the voltage input worldwide and is suitable for two-stage offline converters. Its PFC function can meet the provisions of IEC1000-3-2 on the harmonic limit of AC input current. It has fewer peripheral components, so it can reduce the cost of the entire power supply.

　　TDA16888 has the following PFC characteristics:

• 　　Dual-loop control (sensitive to both average current and output voltage);
• 　　As an additional application of auxiliary power supply;
• 　　Fast soft-switching push-pull gate drive (1A);
• 　　Leading-edge pulse width modulation;
• 　　Peak current limitation;
• 　　Overvoltage protection.

　　The determined PWM characteristics are:

• 　　Improved current mode control;
• 　　Fast soft-switching push-pull gate drive (1A);
• 　　Soft start arrangements;
• 　　Trailing edge pulse width modulation;
• 　　To prevent transformer saturation, the maximum duty cycle is limited to 50%.

　　(2) PFC control

　　TDA16888 uses average current control to provide active power factor correction. The "heart" of its PFC part is an analog multiplier. It generates a programmable current reference signal for the current error amplifier OP2. This signal is obtained by multiplying the rectified input power supply voltage with the output of the output voltage error amplifier. Therefore, this current reference signal has the shape of the input voltage (double half-sine wave) and the function of controlling the output voltage amplitude. Through the subsequent OP2, pulse width modulator and driver, the AC input current of the PFC will become a near-sine wave with a power factor close to 1. The output voltage of the PFC is also stable at 380V. In the circuit of Figure 4, the external circuit of the voltage error amplifier (with voltage sensitivity and compensation) is composed of R13, R14, R16, C5 and C6. Resistor R4 (R4A, R4B) is used to monitor the actual rectified input voltage. R5, R7, R8, C7 and C8 are components of the current error amplifier. The inductor current can be monitored by the voltage drop of R6 on the main board. R3 and R26 determine the current limit of the PFC (approximately 6.5A). R11 and R12 determine the overvoltage threshold.

　　(3) PWM control

　　TDA16888 provides an improved current mode control, which brings effective slope compensation and enhanced suppression of voltage spikes. The converter primary switch current can be transmitted to the PWMCS (11) pin through the voltage drop on R15 of the main board, low-pass filtered by R32 and C21, and then amplified internally and input into the internal PWM comparator C8 together with the output voltage control loop feedback signal XS on the PWMin (14) pin for comparison, and the actual duty cycle is determined by them together. C14 provides soft start of the PWM part. The components R20, R19, IC2, etc. of the output voltage control loop are all placed on the secondary side of the main board converter. The transmission of its feedback signal is completed through a low-value optical coupler IC3.

　　(4) Gate drive circuit

　　Considering the high operating frequency, we use a discrete high-speed, high-current driver stage consisting of small-signal bipolar transistors (Q6, Q7, Q10, Q11) and MOSFETs (Q8, Q9, Q12, Q13) to drive the power tubes (Q1A, Q1B) of the PFC part and the low-end power tube (Q2A) of the PWM stage. This is why the Schmitt trigger and subsequent discrete driver amplifier are inserted at the original gate drive signal output of PFCOUT/PWMOUT. For the gate drive of the high-end power tube (Q2B), its signal is also output from PWMOUT, transmitted through the high-speed optocoupler IC8 (SFH6711), and amplified by IC9, Q14~Q17 before inputting Q2B. In order to obtain the floating supply voltage Vcctop for Q2B, we set an additional independent winding on the core of the PFC choke L2.

　　5 Test Results

　　5.1 Efficiency

　　The efficiency tested at nearly full load and different input voltages is shown in Table 1.

Table 1 Efficiency test results

　　As can be seen from Table 1, the highest efficiency is obtained when the input voltage is the highest, and the lowest efficiency is obtained when the input voltage is the lowest. The reason is that when the input voltage decreases, the input current will increase, resulting in increased conduction losses in the input rectifier, EMI filter, PFC choke and PFC current sensitive resistor. When the switch tube must pass a higher peak current, the effective value of the PFC switch tube current will increase under low input voltage conditions. Furthermore, in order to make the PFC stage have a faster settling speed at startup, the transistor switches at a ton time that is twice the effective duty cycle, that is, the longer the conduction time of the transistor causes its current to increase, which also causes the switching loss of the PFC stage to increase. Since the power supply voltage of the PWM is output from the PFC and is pre-regulated, the characteristics of the PWM stage are independent of the input AC voltage.

　　Furthermore, since PWM uses an optocoupler and variable voltage regulator ICTL431 as the output voltage regulation feedback circuit, its load-voltage regulation rate is also very good, and there is no need to make additional requirements on the load size to obtain a stable output voltage.

　　5.2 Distribution of power consumption

　　The maximum power consumption occurs at full load and low AC input voltage. The operating point at this time is: Vin=90V, Pin=224W, Pout=180.5W, power consumption Ploss=43.5W

　　The power consumption distribution can be estimated using the temperature of the measured component, see Table 2.

Table 2 Power consumption distribution

　　5.3 Conducted EMI Testing

　　To test the conducted noise of the whole switching power supply, we used the EMI receiver FMLK1518 and a source impedance stabilization network (LISN) NSLK8128 to test according to the EMI noise test method specified in CISPRPublication16,1977. The results are shown in Figure 5. Test conditions: Vin = 230V, Pout = 181.4W, the whole power supply is placed in a metal box.

(a) AV average detection noise spectrum

(b) QP quasi-peak detection noise spectrum

Figure 5 Conducted noise test

　　As can be seen from Figure 5, the measured EMI noise spectra are all below the normal limit.