Voltrcraft LSP-1403 / Maplin N27GG / Manson SSP-7080 / Extech 382260

I bought a Voltcraft LSP-1403 power supply on ebay. I intentionally choose the broken one so I could challenge myself and try fixing it. This power supply is a product of Manson but it's sold under at least three other brands which I named in the title of this note.

Voltcraft LSP-1403, Maplin N27GG, Manson SSP-7080 and Extech 382260

The power supply was turning on, display lit and showed reasonable numbers which could be altered by turning knobs, but it didn't provide any output voltage. 

Block diagram of the LSP-1403 power supply

One look at the PCB shown that Q2, U4 R26, R27 and D9 were burned component. I replaced those damaged components and added a light bulb in series, to prevent a small explosion in case I overlook any other damaged/shorted component. Additionally, all the time the power supply was connected to mains over an isolating transformer, this way I was able to connect an oscilloscope to whatever point in the circuit I wanted.

Part of the circuit diagram of the main power supply

After putting this all together power supply provided correct output voltage but after loading it a light bulb lit up which was a sign that something is going wrong with the circuit.

I measured voltage on a shunt resistor R28 ant it turned out to look like this:

Voltage on R28 before fixing the power supply (low output current)

Voltage on R28 before fixing the power supply (higher output current)

Which is wrong because of too high current value and too steep rise. The UC3843 controller was detecting this high current and was turning the MOSFET off to prevent damage, this was the reason the power supply wasn't able to provide demanded output current. Such a steep current rise in the primary winding of a transformer is not possible because of primary winding inductance. My first guess was a damaged transformer (shorted primary winding) but with another transformer, from a similar power supply symptoms were exactly the same.

My second guess was that the snubber circuit is damaged. I measured all snubber components for an obvious damage but they all seemed to be fine. I added two shunts just to see if the current observed on the low side of the primary winding is flowing through the snubber circuit. or through the primary winding. I measured that there is no such a high current in any of the upper shunts. I started scrutinizing the (single sided) PCB for a short but there was none.

The solution, the MOSFET's radiator is connected to GND over a 10nF capacitor (a green capacitor on the circuit diagram above). The isolating layer between MOSFET case and its radiator was shorted and this 10nF was connected to the drain of the MOSFET. The capacitor charged to almost 400V over primary winding and when FET turned on it had to discharge this capacitor first what caused heavy current spikes.

After isolating the MOSFET from the radiator power supply works correctly, voltage on a shunt during normal operation looks like this:

Voltage on R28 after fixing the power supply

Lesson learned: mica is not always isolating, before starting with wild theories check very basics. I did some research on google and it seems that there are often problems at the primary side of the main power supply.

I do have circuit diagram of this power supply but I am not supposed to publish it. However, I will answer all questions to components.