Troubleshooting of CVF-G1 type switching power supply for CONVO frequency converter

Received 3 small power machines from Kangwo CVF-G1, all of which had faults due to no output from the switching power supply and no screen display. The IC of the power supply for this machine is 3844B, and I do not have this model of IC at hand. It is impossible that all three machines are damaged by 3844B, right? So start by checking its peripheral circuits.

All switch mode power supplies have the following branches: 1. The power on start branch is often composed of several resistors with larger resistance values connected in series. When powered on, 500V DC is led to the 3844B power supply pin to provide the starting voltage of the switch tube; 2. The positive feedback and working power supply branches are composed of feedback windings and rectifier filtering circuits (some machines are composed of two winding power supply branches, while others are used in combination); 3. The stabilizing branch is usually powered by a secondary 5V power supply branch, which compares the change in 5V voltage with a reference voltage. Its variable is fed back to the 2 pins of the primary 3844B through the optocoupler, but the voltage feedback of this model is taken from the primary.
The conditions for the circuit to vibrate are: 1. The 500V power supply circuit is normal, the 500V DC is added to the drain of the switch through the main winding, and the source of the switch is formed by a small resistance current sampling resistor to form the power supply circuit; 2. The power on startup branch is normal, providing sufficient amplitude of starting voltage (current); 3. Positive feedback and working power supply branches are normal, providing positive feedback voltage (current) and working power supply that meet amplitude requirements; 4. There is no short circuit on the load side, and the short circuit on the load side cannot establish sufficient amplitude of the feedback voltage, so the circuit cannot vibrate. The above circuit can be called an oscillation circuit.
To minimize the fault, the voltage stabilizing branch should be opened to see if the circuit can vibrate. Voltage reduction and regulation power supply should be implemented, and circuits that are susceptible to voltage impact damage should be cut off to ensure safety. If it can vibrate, it indicates that the four branches that meet the vibration conditions are generally normal, and the faulty components of the voltage stabilizing branch can be further investigated. If the vibration still cannot start, it indicates that the fault is in the oscillation circuit. You can search for the four branches mentioned above.
According to the above inspection sequence, the faults in the switching power supply of machines A, B, and C are all in the oscillation circuit. Check that there are no abnormalities in the four branches of machine A and the peripheral components of 3844B. Try replacing a 3845B and the power output is normal. Repair it; Machine B, after switching to 3845B, still cannot vibrate, and all four branch components are normal. After connecting the 300k resistor of the power on starting branch in parallel with the 200k resistor, the power on is restored to normal; Machine C was also damaged for 3844B, and the fault was resolved after replacing it with a new block.
Only the malfunction of machine B is slightly interesting. The analysis is as follows:
On the surface, it appears that the second machine could not detect any faulty parts, leading to difficulties in maintenance. But after reducing the resistance value of the starting branch, it can work normally. Where exactly is the abnormality of machine B? It may be a slight change in the performance of components that leads to changes in electrical parameters, such as a slight decrease in the amplification ability of switching tubes, a change in Q value of switching transformers due to mild moisture, an increase in internal resistance of 3844B output, or a slight variation in resistance capacitance components. Finding and confirming the above reasons is indeed difficult, or there may be one or even multiple reasons involved. But the various reasons mentioned above only lead to one consequence: the switch tube cannot be effectively started, and the circuit cannot vibrate! The solution is to transform the existing state and exert efforts to promote the oscillation of the switching tube. Parallel connection of resistors in the starting branch is the most labor-saving and effective method.
By the way, the starting resistance of the machine is 300k, and with the addition of resistance from other links, the actual starting current applied to the gate of the switching tube is only slightly over 1mA. Although field-effect transistors are voltage controlled devices that theoretically do not absorb current, the charging current of the junction capacitor that can make them conductive is precisely the hard indicator that makes them conductive. From this perspective, field-effect transistors are still current drivers. When the circuit parameters change, the supply current of the original starting branch is not enough to make the switching transistor conductive or even slightly conductive, so the circuit cannot vibrate. By slightly increasing the starting current value, the circuit may start to vibrate. There is a suspicion that the resistance value of the 300k starting resistor is too high. I believe that slightly reducing its resistance value is beneficial and not harmful.

Therefore, an efficient repair method can be taken as follows: check that the switch tube is not damaged, and that the four branches are generally normal. First, conduct a parallel resistance test on the starting branch, and if it fails, switch to 3844B. If it fails again, then focus on carefully checking the circuit. Often, the fault is already resolved in the first or second step.