In high-power DC/DC switching power supplies, to obtain greater power, especially large current, N units are often connected in parallel. Parallel connection of multiple units offers advantages such as high reliability and standardization of circuit modules. However, the main problem encountered in parallel connection is current unevenness, which can cause serious consequences, especially under heavy loads. The common current sharing method uses independent PWM controller modules, sampling current and feeding it back to the PWM controller's pins FB or COMP (i.e., the input or output pin of the feedback operational amplifier) ​​to adjust the output voltage, thereby achieving current sharing. Clearly, current sampling is a key issue: resistor sampling results in significant losses and large distortion after current amplification; current sensors are costly; current transformer sampling is inconvenient and causes current distortion. This paper proposes a novel, convenient, and lossless current sampling method, and implements current sharing in a parallel system based on this current detection method. 1 A New Current Sampling Method As mentioned earlier, some traditional current sampling methods in current sharing systems have certain drawbacks. The new current sampling method proposed in this paper is simple, convenient, and lossless. The following uses the Buck circuit shown in Figure 1 as an example to illustrate the principle and application of this new current detection method. Figure 1 Simple current detection circuit The current detection circuit consists of a simple RC network. Let the current flowing through L be iL, the current flowing through C be iC, the voltage across L be vL, the output voltage be vo, and the voltage across C be vC. Then we have vL + iLR1 + vo = vc + iCR (1) The average value of equation (1) in one switching cycle is VL + ILR1 + Vo = VC + ICR (2) Where: VL is the average value of the voltage across the inductor in one switching cycle, obviously VL = 0; Vo is the average value of the output voltage; IL is the average value of the inductor current, which is equal to the load current ILoad; IC is the average value of the charging and discharging current of the capacitor in one switching cycle, obviously IC = 0; R1 is the equivalent series resistance (ESR) of the inductor. Therefore, equation (2) can be transformed into ILR1 + Vo = VC (3) IL = ILoad = (VC - Vo) / R1 (4) So, to detect the magnitude of the load current and the inductor current, it is only necessary to detect the magnitude of the voltage on the capacitor of the RC network. This method can conveniently, easily and without loss sample the current. 2 Current sharing principle based on the new current sampling method Take a two-way parallel Buck circuit as an example, as shown in Figure 2. Figure 2 Simple current detection using RC network From equation (3), we know that VC1 = IL1R1 + Vo VC2 = IL2R2 + Vo Where: VC1 and VC2 are the average values ​​of the voltage on C1 and C2 respectively; IL1 and IL2 are the average values ​​of the current flowing through L1 and L2 respectively, that is, the average values ​​of the two output currents; R1 and R2 are the equivalent series resistance of the filter inductor. When the output inductance of each parallel power supply is basically the same in the process design, it can be considered that R1 = R2. Therefore, to control the current sharing of the two currents, i.e., to require IL1 = IL2, it is only necessary to control VC1 = VC2. Thus, the voltages VC1 and VC2 across capacitors C1 and C2 can represent the magnitudes of the two currents IL1 and IL2, and can be used for current sharing control. This yields the control block diagram shown in Figure 3. Figure 3: Block Diagram of a Two-Way Buck Parallel Current Sharing System. 3. Analysis and Comparison of Commonly Used Current Sharing Methods The following are some commonly used current sharing methods in parallel switching power supply systems: Output Impedance Method (Droop Method): Adjusting the droop of the external characteristic of the switching converter (i.e., adjusting the output impedance) to achieve near-current sharing among the parallel modules. This method is a simple approximate current sharing method with relatively low accuracy. Master-Slave Method: Suitable for current-controlled parallel switching power supply systems. This current sharing system has voltage control and current control, forming a dual closed-loop control system. This method requires communication between each module, thus complicating the system, and when the master module fails, the entire power supply system cannot operate. In the average current sharing method, the output of the current amplifier of each parallel module is connected to a common bus with the same resistor, forming the average current sharing bus. When the voltage of a module is higher than the bus voltage, the output voltage drops, and vice versa. The maximum current sharing method is similar to the average current sharing method, except that each current path is connected to a common bus through a diode. This method is essentially a "democratic current sharing" method, where the module with the largest current automatically becomes the master module, and the other modules become slave modules, thus achieving "automatic master-slave control". Disconnecting or opening the current sharing bus in the average current sharing and maximum current sharing methods will not affect the independent operation of each power supply module, and it is an automatic current sharing method with high current sharing accuracy. Figure 4 shows the schematic diagram of common current sharing methods. If the current sharing bus is the average value of the current of the parallel modules, it is the average current sharing method; if it is the maximum value of the current of the parallel modules, it is the maximum current sharing method; if the current sharing bus is the current of the master module in the parallel modules, it is the master-slave current sharing method. However, in these current sharing methods, each module needs an independent PWM control loop. Figure 4 Commonly Used Current Sharing Methods 4 New Current Sharing Scheme The scheme proposed in this paper is based on adding a simple RC network to each channel to detect the magnitude of its allocated current. The average voltage across capacitor C can characterize the current magnitude of this module. Therefore, current sharing control of the system is to equalize the voltage across C of each RC network. Its current sharing principle diagram is shown in Figure 5. Figure 5 Current Sharing System of Two Buck Circuits in Parallel In Figure 5: Vbus is the current sharing bus voltage; Vref is the output voltage reference value; Vs is the sampled value of the output voltage. Its working principle and process are as follows: By detecting the voltage across C in the RC network as a current signal, several current signals (only two in this example) are passed through the same resistor to obtain the average current sharing bus. The average current sharing bus voltage value is related to the load and characterizes the magnitude of the load current. Then, the current signal sampled from each channel is compared with the bus voltage to obtain an error signal, which is used to correct the output voltage reference signal, thereby fine-tuning the duty cycle output of the PWM controller to achieve the purpose of current sharing and voltage regulation. 5. Measured Results The prototype is a 4-channel Buck parallel switching power supply with a DC 5V input and a 2V/40A output, operating at a frequency of 200kHz. The current output of each channel was measured under full load. The current sharing effect was good, with an error of less than 2%, and the power supply output was stable. The higher the output current, i.e., in a high-power parallel power supply system, the better the current sharing effect. 6. Conclusion This paper analyzes and studies commonly used current detection methods and current sharing control, and proposes a method for detecting the power supply output current using an RC network. Based on this current detection method, a simple current sharing scheme for parallel systems is presented. This scheme makes current detection convenient and enables efficient, low-cost, simple, and convenient current sharing in parallel systems.