Snap, crackle and pop! Another board to repair....

Put the kettle on, make a brew, and get comfy....I have another PCB repair and design analysis for you.

The other day my friend over in Wales sent me these two horrible little boards to repair. He has sent me various boards to repair over the years, such as central heating clocks, an i-pad, an electric gate controller, a cooker controller, and an electric vehicle ECU, and various other boards. So far I have managed to repair everything apart from the ECU. My friend took great delight in this when I told him that I couldn't fix it. The bl**dy thing was potted inside a metal enclosure, and I simply couldn't get the board out, or decent access to it. Arrghh! Anyway, these two horrible little things turned up, and I had an opportunity to redeem myself. Can you spot what is wrong with the bottom board? No prizes for guessing correctly.

I have no idea what these board are from, my friend didn't tell me, but you can tell that they are power supply boards from the ferrite transformer, Y-mode cap, and pi filter, etc.

Before we look at the repair, let's review the design together....

If we start by looking at the mains supply input area, (the brown and blue wires on the bottom left), you can see the black fuse. This protects against short-circuits on the board. After the fuse, there is a basic common mode filter block using the yellow X2 mode capacitor, and the common mode choke, this helps to prevent conducted EMI from being injected onto the mains supply, which could interfere with other equipment. After the fuse there is a through-hole bridge rectifier which produces a DC link voltage. This is then fed through a relay contact to the motor, (I think), output, M+ and M-, the red and black wires.

A 10ohm metal film power resistor, R2, feeds the switched mode power supply. The topology, or type, of switching power supply is most likely to be a flyback converter. This is because they are cheap and simple to make. Notice however that there isn't a voltage feedback opto-isolator on this board? This is because the output voltage regulation is done on the primary side by the control IC looking at the transformer reflected voltage when the primary isn't being driven, ie, the switch is in the OFF state. This saves the cost of the opto, but also helps to improve reliablility as the opto-isolator is often a reliability weak-point in a converter. The other weak point tends to be the electrolytic capacitors as they eventually dry out over time. The warmer they are, the quicker they dry out. The electrolytics are looking particularly delicious in this design - made by ChongX, mmmm.

The input stage to the flyback converter typically consists of a pi filter comprised of two electrolytic capacitors and an inductor. You can see these on the photo above. This filter generates a smooth DC link voltage for the converter and also helps to minimize EMI on the mains supply.

The blue Y-mode capacitor next to the transformer is usually called a bridging capacitor as it forms a bridge across the primary and secondary side of the transformer. Its purpose is to bypass, or short-out, any common mode noise currents generated, and prevent them from leaving the board. Whenever a capacitor is required to be place from one of the live lines, either live or neutral, to earth, or to the secondary low voltage side, a Y-mode capacitor must be used.

Let's now have a look at the bottom side of the board.... You can see the flyback control IC, U1, surrounded by its peripheral components. The converter is fed by its own surface mount bridge rectifier, DB1. The resistors, R12 and R14, are bleed resistors, which are placed across the yellow X2-mode capacitor. This is to discharge the capacitor voltage to safe levels, (usually around 50V), when the power and lid are removed in order to prevent an electric shock hazard. The standards state that these capacitors have to be discharged to a safe voltage within 1 second. So, worst-case, the capacitor could be charged to 325VDC, and we would need to discharge down to 50V with a second. This is about two time constants. The capacitor value is 470nF and the resistors are both 390k ohms. Therefore, the time to discharge is 0.73s. Just in!

The secondary side of the transformer consists of a schottky diode rectifier, D1, which has an RC snubber network across it, C1 and R1, to damp any ringing, and a minimum load, R5 and decoupling capacitor, C3. There is also an electrolytic bulk capacitance to provide output voltage ripple reduction. The relay coil hangs from the 12V output rail and is switched to ground by something off the board via the little 3 way harness.

The slots in the board are there to increase the creepage distance between various points in the circuitry. The minimum creepage and clearance distances required will depending on the voltage between the nets, the board material, the pollution degree, and even the altitude, (air pressure). These minimum distances are defined in British Standard, BS EN 60664-1. A bedtime must-read!

The controller IC, for those playing along at home, is the On-Bright, OB24136, a high precision PWM controller with primary side sensing. Datasheet below:

So, let's get to the repair.... The first board was an easy one, which I spotted almost straight away. When I fault find a board, the first thing I do is carry out a visual inspection. A lot of failures can be found this way. If not, then the real fun begins! In this case, the root cause was a fractured solder join on the relay output pad. This is a common occurrence with lead-free solder.

All I had to do was remove the old solder and apply some fresh stuff. Job done!

As for the second board, this was fairly easy too. You can clearly see the that the 10ohm power resistor had gone boom-bang-a-bang, but also, the input fuse had failed open circuit.

Bang, and the resistor is gone!

I temporarily fitted some alternative components to see if the converter would work, but before powering up the board I checked in the input impedance of the converter to ensure that there were no other short-circuits looming, such as the internal switch in the controller IC. All was good and I powered the board up and all is well, yay! So, what was the root cause in this case? I believe that the power resistor failed initially, which then caused the fuse to blow open circuit. The power resistor is used to limit the inrush current to the pi filter. During worst case conditions, the power dissipated in the resistor is several thousand watts for up to about 80us. I suspect that the resistor is not really rated for this pulse power, and it eventually died. I have ordered some lovely 3W fusible pulse withstanding resistors which are specifically designed for high pulse power applications. Check out the datasheet excerpt regarding this: X marks the spot....

So there you go, both boards repaired, and I'm still winning! Have a lovely evening.