22
\$\begingroup\$

I'm working on a project which uses a PCBA which basically acts as a USB-C extension. It has 2 USB-C receptacles (female) connectors. The mapping between the two receptacles is one to one. The pair of D+ and D- are shorted (meaning both orientation should work for USB2.0)

An SSD USB-C stick is connected at one end of this PCBA. On the other end I'm using a USB-C to USB-C cable connected to a PC. From looking on the web and reading other answers on the subject, I expected the connection to happen in half the orientation configurations, fine.

My question is: Why? Why or how does this setup violate the USB-C standard? I though it would be a cable orientation detection issue but I'm not sure.

I have disassembled a USB-C cable I've been using (PC to my extender PCBA) and it seems it has both USB CC lines existing inside as well as both SS pairs, so all path should be possible.

Below are some diagrams I made to try understand the CC lines path in several orientation (not all of them included.)

Study of some USB-C connection orientation for CC lines

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Consider the mayhem that would ensue if you plugged two of these into such a thing. \$\endgroup\$
    – Joseph Sible-Reinstate Monica
    Commented Jun 25, 2022 at 20:13

3 Answers 3

Reset to default
22
\$\begingroup\$

A single USB-C cable will not pass all wires through it, so you can't make an USB-C extension cable.

As there are two CC pins. The cable, or rather, both connectors terminate separately one of the CC pins at both ends, and passes the other through cable.

So even if you have only the host connected to a cable, it knows there is cable connected and cable orientation based on one CC pin used by connector.

Then as cable passes only one of the CC pins through the connector, it is used to detect if a device is connected to the other end and what type device it is.

So a passive extender that works in all orientations can't be built.

And extender would need to be smart and detect host cable orientation, and the device cable orientation, and multiplex the CC pins (and others?) correctly to pass through.

But since the cable terminates one CC pin, the host only knows about the first cable connected to it. It can't know about different cables or adapters connected after the extender plug.

So, for example, if you have passive USB-C headphones, they can be properly detected by a mobile phone from both CC pins when it is directly connected to the phone. It can't be detected if one of the CC pins is taken by cable connector and only one CC pin passes to headphones.

Therefore USB-C can't support extender plugs.

\$\endgroup\$
4
  • \$\begingroup\$ Why would a pure USB C cable have only one CC line going through ? It is intended to link a host an a device. The device would pull down one of the CC and the host pull up both on its end. Also, why would a host want to know a cable is connected of nothing is at the other end ? \$\endgroup\$
    – Frenetique
    Commented Jun 24, 2022 at 19:44
  • 4
    \$\begingroup\$ How else would you know which way the connector is, flipped or unflipped? Or if you instead of standard cable, you have a debug adapter cable? Or USB-C to MicroUSB adapter? Or USB-C headset? All these require knowing which way the plug is inserted and what is inserted. Just look up USB specs or USB-C cable wiring diagrams. \$\endgroup\$
    – Justme
    Commented Jun 24, 2022 at 20:43
  • 1
    \$\begingroup\$ Ah my bad. I thought the UFP (device) had only one CC pin pulled down when it's actually both of them Here is a nice schematic from microchip: microchipdeveloper.com/local--files/usb-i:cable-connection/… \$\endgroup\$
    – Frenetique
    Commented Jun 24, 2022 at 21:23
  • 8
    \$\begingroup\$ This is a good explanation for why the specification does not support extension cables, but it's worth noting that consumer products violate technical specifications all the time. I've added another answer that elaborates. In general I don't encourage deviating from a spec, but in the case of USB-C, I'd say the cat's already out of the bag. \$\endgroup\$
    – MooseBoys
    Commented Jun 25, 2022 at 9:35
20
\$\begingroup\$

USB-C extension cables may violate the standard, as it breaks some required functionality as mentioned by @Justme. But that does not seem to have stopped a variety of manufacturers from building and selling them anyway. They may not be permitted to put "USB Type-C®" and accompanying logos on the box, but people still buy them, which means devices need to support them.

From what I've seen, these cables are implemented as basic pass-through cables, where all wires are connected directly from one end to the other. For cables that are likely to be connected via adapter to USB-A or other connections, the cable is shipped with a label attached, with text along the lines of "if it doesn't work, try flipping the cable over".

USB-C is great from a technology perspective, but it's also a great case study in how not to manage an ecosystem. Between Nintendo Switch's not-quite-USB-C-sized interconnect, to "USB-C Alt-mode" and my favorite, sounding like something straight out of HG2G's Vogon Bureaucracy, "USB™ 3.2 Gen 2x2 SuperSpeed 20Gbps", there is already a huge amount of fragmentation in the ecosystem.

To answer your original question, yes - you can build a device that functions as a USB-C coupler. You just won't likely be able to pass USB-IF certification (and so can't use the USB logo on the box/product), if that's important to you. It's also important to make sure that if the connection supports power delivery, it has ample current capacity (>>5A) because you won't have any way to negotiate down the PD protocol.

\$\endgroup\$
4
  • 4
    \$\begingroup\$ +1 for the H2G2's reference :) \$\endgroup\$
    – Frenetique
    Commented Jun 25, 2022 at 20:40
  • \$\begingroup\$ Remember that USB-C, USB4 and USB Type-C are also protected (word) trademarks. So if you read this on a bogus product that doesn't charge properly from a standard USB charger or has other non-compliant issues return it under warranty, because the seller is lying to you when they sell you a USB product that turns out to be fake incompatible USB Type-C. Don't accept that bs as a consumer! (not legal advice) \$\endgroup\$
    – user643011
    Commented Sep 19, 2022 at 17:21
  • \$\begingroup\$ @user643011 According to USB-IF only the graphical logos require or imply certification. The Word Marks "USB-C" and "USB Type-C" only imply compatibility with the "connector specification" which is the baseline profile, and almost entirely defined using "should" instead of "must". If you're a major top-tier device manufacturer like Apple or Microsoft, you're going to be certifying anyway; if you're a small manufacturer, nobody's going to care. \$\endgroup\$
    – MooseBoys
    Commented Sep 21, 2022 at 4:59
  • \$\begingroup\$ @MooseBoys Good point! However the "Universal Serial Bus Type-C Cable and Connector Specification" is still pretty comprehensive and includes a lot of "shall" and normative sections. So common errors like not using the right resistors to signal a device / sink, not supporting reversible connectors or not using proper USB voltages or voltage/current signaling are not allowed. If a device is claimed to use USB-C and then that's a false representation I would absolutely return that for a fixed replacement or my money back when I discover that, certified or not. \$\endgroup\$
    – user643011
    Commented Sep 22, 2022 at 12:35
2
\$\begingroup\$

Also, USB-A/USB-C cables are legal. If you connect two of those to your device, you end up with an A-A cable, which is definitely illegal.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ That was an extemely good point which I did not think when writing my answer. Unfortunately, USB-A to USB-A cables also exist, even if they should not exist. \$\endgroup\$
    – Justme
    Commented Jun 28, 2022 at 16:21
  • \$\begingroup\$ I know this. I have one in my room, and I know it's not a fast transfer cable (looks A-A, has a circuit in the middle to fix it up). How do I know? Because the idiots who designed my laptop cooling pad decided to give it two USB-A sockets and provide an A-A cable to power it. If I weren't worried about the cable breaking, I'd epoxy one end into the pad so it couldn't be misused. \$\endgroup\$
    – NoLongerBreathedIn
    Commented Jun 30, 2022 at 19:42
  • \$\begingroup\$ Of course, C-socket/C-plug cables (true USB-C extension cables) would cause no issues beyond signal loss and voltage drop. \$\endgroup\$
    – NoLongerBreathedIn
    Commented Jun 30, 2022 at 19:43
  • \$\begingroup\$ @NoLongerBreathedIn Y-cables aren't even that bad. The automatic window shades I have use micro-USB to connect the motor to the 12V battery pack. The pack has a red warning label next to the plug saying "DO NOT PLUG IN USB DEVICES". And my old electric train set used a telephone cord to attach the power supply to the rails. And "HDMI-over-ethernet" adapters just use ethernet cables for cheap conductors - they don't actually send ethernet frames. \$\endgroup\$
    – MooseBoys
    Commented Jul 1, 2022 at 1:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.