USB-C on a PCB: the two 5.1 kΩ resistors, and why boards get no power from a C-to-C cable

A USB-C cable has the same plug at both ends, so the two ends have to work out which one supplies power. They do it on the CC (configuration channel) pins. The charger, the source, pulls CC up through a resistor called Rp. The device, the sink, pulls CC down to ground through 5.1 kΩ, called Rd. The charger watches the voltage on CC, and only when it sees that divider does it turn VBUS on. No Rd, no 5 V.

Old USB-A chargers hide the mistake. A USB-A port has 5 V on all the time, and an A-to-C cable carries the 56 kΩ pull-up in its USB-C plug, so a board with no CC resistors runs fine from it. Then someone plugs in a phone charger with a C-to-C cable, and the board is dead.

The receptacle has two CC pins, CC1 on A5 and CC2 on B5, because the plug can go in either way up. An ordinary cable connects one of them. An electronically marked (e-marked) cable, the kind sold for 5 A or for fast data, also puts its own 800 Ω to 1.2 kΩ resistor, Ra, on the other one. That is how the charger knows to power the chip inside the cable.

Tie CC1 and CC2 together through one 5.1 kΩ and, with an e-marked cable, the cable’s Ra lands in parallel with it: 5.1 kΩ and 1 kΩ in parallel is about 836 Ω. The charger reads that as Ra, not as a device, and keeps its 5 V off. That is what happened to the Raspberry Pi 4 in 2019: its two CC pins shared one resistor, and chargers with e-marked cables identified it as an audio adapter accessory and refused to power it.

A USB-C receptacle has D+ twice (A6 and B6) and D− twice (A7 and B7). A USB 2.0 plug has D+ and D− in one row only, so the plug orientation decides which pair is live: one way up it meets A6 and A7, turned over it meets B6 and B7. Wire only the A row to your microcontroller and USB data works one way up, while power works both ways. It looks like a flaky cable.

Join A6 to B6 and A7 to B7 right at the connector, then route one D+ and D− pair to the chip. ST’s application note on USB-C protection keeps the joining traces under 3.5 mm so the unused branch stays a short stub. The SBU pins are not used by USB 2.0 and can stay unconnected.

The charger’s pull-up says how much current it offers, and the voltage on your CC pin shows which one it is:

A board that draws less than 500 mA can ignore this. One that needs more should read the CC voltage with an ADC or comparator, or use a sink controller that does it, before drawing 1.5 A or 3 A. Either way the voltage stays 5 V: a charger offers 9, 15 or 20 V only after a USB Power Delivery negotiation, which takes a PD controller, not resistors.

• CC1 (A5) to ground through 5.1 kΩ, and CC2 (B5) to ground through its own 5.1 kΩ.
• Every VBUS pin (A4, A9, B4, B9) on the supply, and every GND pin (A1, A12, B1, B12) on ground.
• D+ on A6 and B6 joined, D− on A7 and B7 joined, at the connector, for USB 2.0 data.
• SBU1 and SBU2 unconnected unless you know you need them.
• A power-only receptacle still gets both CC resistors.
• Test with a USB-C charger and a C-to-C cable, both ways up, and with an e-marked cable if you have one. An A-to-C cable proves nothing about CC.

Boards PCBEditor generates are checked for this. A USB-C receptacle that powers the board needs its own 5.1 kΩ to ground on each CC contact, and a CC pin left open, tied straight to ground or to a supply, given the wrong value, or sharing one resistor with the other CC pin is reported as an error. Generated boards have come back with CC “pull-downs” drawn as 0 Ω, CC pins wired to microcontroller GPIOs and a CC pin used as the supply, which is why the check exists. If you draw the schematic yourself, the circuit diagram maker shows the connector’s pin names, so you can check each line of the list above against it.