Why a USB-C board works on USB-A but not C-to-C
The prototype arrives. You plug in the USB-A to USB-C cable from the bench drawer and the board boots. Then you plug in a proper USB-C charger with a C-to-C cable and nothing happens. No LED, no current, no enumeration. When this happens, check two small pins on the connector and the resistors that belong on them first.
Why the A-to-C cable hides the bug
A USB-A port is a source that always puts power on VBUS. It does not wait for the device to announce itself. A legacy USB-A to USB-C cable is built around that: per onsemi’s application note on USB Type-C, the plug’s CC pin is pulled up to VBUS through a resistor inside the cable, which lets a sink connect safely to an A-port that supplies VBUS regardless. Your board sees 5 V and runs, whether or not it has done anything correct on its CC pins.
A USB-C source is different. It starts with VBUS off. Each of its CC pins has a pull-up (Rp), and it watches for the voltage to drop below the unterminated level. In TI’s description of the Type-C termination model, a source monitors CC1 and CC2 for a voltage lower than its unterminated voltage to detect a connection, and a sink advertises a pull-down (Rd, 5.1 kΩ) on both of its CC pins. No Rd, no drop, no attach detected, no VBUS. The board is not broken. The source is correctly refusing to power something that has not asked.
So the A-to-C cable works because it skips the conversation. The C-to-C cable fails because it requires one and the board never joins it.
The usual causes, most common first
No Rd at all. The designer treated the connector as a plain power jack and routed only VBUS, GND and the data pair. The CC pins (A5 and B5 on the receptacle) are left floating. This is the textbook case.
One resistor shared by both CC pins. Because the plug can be inserted either way, the cable only connects one CC wire end to end. The termination model has each CC pin carrying its own Rd, so give each one its own resistor and do not tie CC1 and CC2 together through a single resistor. Two resistors, one from each CC pin to ground.
The wrong value. The sink termination in TI’s document is 5.1 kΩ. Use that value on both pins, and read it in the schematic, the BOM and on the assembled board. They are three different documents and they drift apart, so a part from the wrong reel or the wrong line of the BOM, or a leftover from a previous revision, is easy to miss.
Resistors fitted, connector not soldered. The CC pins on a USB-C receptacle sit among fine-pitch signal pins. A missing joint on A5 or B5 produces exactly the symptom above, with a schematic that looks perfect. This is an assembly defect, not a design one, and it is why the next section starts with an ohmmeter.
An e-marked or powered cable in the path. A cable with active electronics advertises its own termination, called Ra, on one CC pin. In TI’s words, a powered cable advertises Ra on only one of the CC pins of the plug, and it tells the source to apply VCONN. Your board still only needs Rd on its own two pins, but this is why one C-to-C cable can behave differently from another while the board stays the same. If a charger works with one cable and not another, swap cables before touching the board.
The bench checks that find it in minutes
Work from unpowered to powered, and write each reading down.
- Unplug everything and measure. With the board unpowered, measure from each CC pin to ground at the connector or the nearest test pad. Both should read close to the value on the schematic. Floating means a missing resistor or open joint. Both pins reading the same low value in parallel with each other points to shorted or shared termination.
- Inspect the receptacle under magnification. Look at A5 and B5 and their neighbors. Check for bridges between CC and an adjacent pin, which makes the resistor see VBUS or a data line.
- Try a known-good C-to-C charger and cable. Plug in one way, then flip the plug. A board that works in only one orientation points to one CC resistor or one side of the connector’s wiring. A board that works in neither points to both CC lines or the supply path behind them.
- Check that VBUS reaches the board. If the source turns VBUS on and the board still does not boot, the CC network is fine and the fault is downstream: a reverse-polarity FET, a load switch, or a regulator. Probe VBUS at the connector first, then after each stage.
- Use a CC termination adapter to isolate the problem. An inline adapter that adds the 5.1 kΩ resistors tells you quickly whether the board’s CC network is the whole story. It is a diagnosis tool, not a design fix.
Do these in order. Jumping to the firmware or the regulator wastes a day on a fault that two resistance readings would have named.
What to design in, so it never shows up
For a device that only needs default USB power, the fix is two resistors and a correct footprint. Put one Rd from each CC pin to ground, as close to the receptacle as is practical, and give both a reference designator that makes them hard to omit from a cost-reduced BOM. Add the CC pins to the connector’s test points so the ohmmeter check above is possible on a finished board. Our note on what test points a board needs covers where those pads earn their space.
Two decisions sit beyond the resistors.
Do you need more than default power? Resistors alone only make the device attach. If it needs more current than a default USB source offers, the board has to read the source’s advertised current from the Rp level on CC, or negotiate with USB Power Delivery. That means a CC controller or a microcontroller with a Type-C port peripheral. We walk through the choices in USB-C Power Delivery input for a product that isn’t a phone. Do not draw more than the source advertises.
Does it also need to work on legacy A-to-C cables? It will, as long as the board draws only what a default source offers. The A-to-C cable is a common case for field users, so the board should be tested on both cable types before release, not only on whichever one was on the bench.
Protection on the connector matters too. CC lines face the outside world and are exposed to hot-plug events. Where to put the protective parts is covered in where to put ESD protection on a USB or connector port.
Checking the specification, not the forum
Forum threads are good at telling you the symptom. They are not a reliable source for resistor values, tolerances or current levels, and plenty of them repeat a wrong number with confidence. The authoritative document is the USB Type-C Cable and Connector Specification, published through the USB-IF’s USB Charger and Power Delivery page. Check which release is current before you design to it. For a specific design, use the datasheet of the CC controller or connector you chose, and use the Infineon knowledge base article on Type-C termination resistors or a similar vendor note as a cross-check. If the part you picked asks for something different from what a blog says, follow the part.
How we approach it
We check the CC network, connector footprint and protection against the specification and the chosen controller’s datasheet, then test the first boards on more than one charger and cable type. If you have a board that powers from some cables and not others, send the schematic, the BOM and a clear photo of the connector area, and we will tell you which of the causes above fits.
If your USB-C board only powers from some cables or chargers, send us the schematic and a photo of the board and we will find which CC or connector detail is wrong.
Tools for this topic
- Regulator feedback resistors Feedback pairs for any Vref, for feedback-pin parts and three-terminal adjustables, with the bias-current term, worst-case range and divider-current guidance.
- PCB trace impedance Trace width for a target impedance, or the impedance of a width: microstrip and stripline, single or differential pair, with 4-layer stack-up presets.
- Voltage divider Vout from Vin and two resistors, or the best standard pairs for a target Vout — with tolerance range and load effect.
Related notes
- Logging to an SD card without corrupting it on power loss How to log to an SD card without corrupting it when power drops: when to sync FAT, where littlefs fits, and the hold-up hardware that matters.
- Why a microcontroller ADC reading is noisy or wrong Find out whether a noisy or offset ADC reading comes from source impedance, sample time, reference, layout or firmware, with a check order for each.
- Why an IMU reads wrong once it's in the enclosure Mounting stress, aliasing, clipping and offset calibration: the four reasons an IMU that worked on the bench misbehaves in a product, and the checks.