One socket dead: the diode behind it

3:21

The symptom that names itself

One socket does everything: powers the board, lights the LED, appears as a port, uploads. The other does nothing at all, with the same cable that just worked.

Nothing else on this board produces that. A cable fault would follow the cable to the other socket; a driver problem leaves the board powered; a dead chip takes both sockets down. One live socket and one completely dead one points at the one part that sits between each socket and the rest of the board, and there is one per socket.

What the part is doing there

Look beside each USB-C socket on the front of the board. The small black component with a bar printed on one end is a Schottky diode, rated 40 V and 2 A (the parts list gives a K24; the marking on the case varies between batches). It is wired in series: the socket's 5 V goes in one end, and the board's 5 V rail comes out of the other.

A diode passes current one way and nothing the other, and that is the whole job. It is there for two reasons:

• The sockets and the 5V pin must not feed each other. Both sockets and the 5V pin meet on the same rail. Without a diode in each path, a charger in one socket would push current into the other socket and into whatever is plugged into it, and a supply wired to the 5V pin would push back into your computer.
• A supply connected backwards stops at the diode. USB-C cables cannot be reversed, but a bench supply, a battery lead or a homemade adapter can. The diode blocks it, the board sits dark, and the part that took the abuse costs cents.

It is not a fuse, and the board does not rely on it to limit current. But it is the one part every amp of that socket's supply has to pass through, and it is rated for 2 A. When something asks for more, it is a likely weakest link, and that is why the failure shows up here first.

Why it dies on a bench

A computer's USB port hands over half an amp to a bit under one. A USB-C charger or a power bank is another matter: the two 5.1 kΩ resistors on the socket tell it a device is attached, and it may offer 1.5 A to 3 A at 5 V. A short circuit takes whatever it is given.

On a bench, that happens in the ordinary course of making things:

• a jumper wire that slips from 5V to GND while the board is powered
• a motor, a servo or a strip of LEDs run from the 5V pin instead of its own supply
• a module plugged in the wrong way round, or a loose wire resting on a neighbour

Push 3 A through a 2 A diode and it heats until something inside melts. Usually that leaves it open — permanently non-conducting — and the socket behind it finished while everything else still works perfectly. This is common in a kitchen-table lab and nothing to be embarrassed about: the cheap part died so the board did not.

Check it: the tweezers test

You do not need a multimeter. Metal tweezers will tell you.

1. Plug the cable into the dead socket, and the other end into the computer or charger. The board stays dark.
2. Touch the tips of a pair of metal tweezers across the diode — one tip on each end of the small flat part, only those two. Hold them there steadily.
3. Watch the red power LED.

If the LED lights and the board comes to life, the diode is open and you have found the fault. Everything else on that socket is fine.

If nothing changes, it is not this diode. Go back to the cable and the connector below.

With a multimeter on the diode setting, the same fault reads OL in both directions across the part. A healthy one reads about 0.2–0.4 V one way and OL the other — see which way round a diode goes.

Fix it

The tweezers already showed you the repair: put a conductor across the diode.

Bridge it — the quick one. Lay a short piece of wire across the two ends and solder both, or run a blob of solder from end to end. It takes two minutes and the socket is live again. A bare wire or tweezers held there is the same thing, only temporary.

Replace it — the proper one. Any 40 V, 2 A Schottky in the same flat SOD-123FL package will do; the DSK24 has the same ratings as the K24. Lift the dead one with a hot iron and flux, and fit the new one with the bar toward the board, away from the socket: current runs from the socket through the diode to the board, and the bar marks the end it leaves by.

Choose by what the board is for. A board on your own desk, run from one known-good charger: bridge it and carry on. A board that goes into a classroom, a kit, or a project other people will plug things into: replace it.

The part that matters more than the repair

The diode did not die of old age. Something drew more than 2 A, and if you bridge or replace the diode and leave that something where it is, the next part in line takes the hit: the 3.3 V regulator, the connector, or the board's own copper.

So before you plug it in again, unplug whatever was connected to the 5V pin and the header, and look for a wire that has fallen across two pins. Run a motor or a strip from its own supply, with a shared ground. And switch the board off before you rewire it — most of these happen with the power on and a jumper in one hand.

Before you reach for the tweezers

Two minutes of checks, because everything above assumes the diode is the failure:

1. The same cable, both sockets. One works, one does nothing at all — that is the pattern. If both are dead, this article is not your answer.
2. A different cable. Two dead cables in a row is rare and not impossible.
3. Look at the connector. A USB-C socket with a bent or pushed-back contact, or with a pin lifted from the pad, fails the same way and is visible under a loupe. Fix the visible thing before the invisible one.

And if the board is new, stop: talk to us before you solder or hold anything against it. A diode that failed on its first day is a warranty question, not a repair job.