Failed Repair: MikroTik RouterBOARD 750GL

Joshua R. T. Purba
Electronics

Looks like that portable hot-air station caught me a repair bug 😆.

This time it was MikroTik RouterBOARD 750GL. I salvaged this from the trash bin. When asked why they were throwing this out, they said it stopped functioning 🤷.

As expected of something that was thrown in the trash bin, it was dirty.

Top view of the salvaged MikroTik RouterBOARD 750GL.

The front panel was not dirty.

Front view of the salvaged router.

I brought it home and tried it. I powered it using my DC bench power supply from 8V to 30V. The PWR indicator didn’t light up. No activity whatsoever. It was not functioning all right 🙅.

Visual inspection did not spot any obvious damage.

Top view of circuit board.

I opened the case and visually examined it. No obvious damage. I powered it up and probed it using a multimeter. It appeared that the power supply section did not give out any output. I found a service manual that had nice flowcharts, board layout, and components list.

The service manual had a flowchart to troubleshot no-power condition, and I was down at the last step: checking the buck converter. According to the service manual the power supply section used LSP5502SAC buck converter chip. Reading the datasheet I concluded that the power supply section should output 3.3V. Though there was a test point on the PCB labeled 3V.

Now I adjusted my DC bench power supply to output 3.1V (current limit to 2A), and powered the board directly from the output point of the buck converter. I did this by soldering a wire at the buck converter inductor and connecting the DC bench power supply to it. The LEDs flickered for a split second, immediately turned off again. So it seemed the buck converter chip was the problem.

I probed the LSP5502SAC chip and it seemed that pin 7 (EN / enable pin) was shorted to ground. I fired up the hot-air station and removed the buck converter chip from the board. I powered the board directly from the output point of the buck converter again as above, and the LEDs turned on and stayed on. I examined the buck converter chip again, closer this time, and saw it had a crack. I was happy, thinking that I found the problem.

Under this lighting, the damage to the buck converter chip was not immediately obvious.

Desoldered LSP5502SAC chip laid beside its original position on the circuit board.

With this lighting the crack can be seen.

Close-up of the damaged chip.

Lighting from a different angle to show to crack on the damaged chip.

Same chip, different lighting.

But something weird was going on. The board kept drawing near 1.5A. Usually when devices like this started they draw high current but then it fell down. This board though kept drawing high current. So I probed using my finger when the board was still powered. I touched the the only heatsink on the board. The heatsink was attached on the AR8327-AL1A chip (gigabit ethernet switch chip). It was HOT 😰. I used my temperature probe on the heatsink, and it was 70°C, and still climbing, finally stabilizing around 80°C. The ambient temperature at the time was around 30°C as measured by one of my previous project.

Powered using DC power supply directly to the board, the buck converter chip was already removed. At 3.1V it drew around 1.5A. Heatsink was 78°C and still going up. Ambient was 30°C.

The heatsink on AR8327-AL1A reaching 78°C.

Was that the cause of the original fault? Was that why the buck converter chip cracked? AR8327-AL1A chip drew near 1.5A continuously so the LSP5502SAC gave up the magic smoke? Weird, considering LSP5502SAC should be able to handle 2A according to the datasheet.

I proceed to check the online marketplace and AR8327-AL1A was available. However, the price made buying a replacement not economical 😢.

So I stopped the repair there. This repair project was a failure 😖.

Oh, well.