9
\$\begingroup\$

Since USB Type-C cables have only a single CC wire for detecting orientation (Microchip USB Type-C diagram:

diagram

can an upstream (device) connect its CC1 and CC2 pins and share a single pull-down resistor?

This would prevent the device from detecting orientation (and thus using the SuperSpeed USB 3.0 lines) and so limit it to USB 2.0. But could shorting the CC pins in a receptacle actually cause damage to other devices or prevent it from operating as a 2.0 device?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Are you trying to save on 0.1c component? Connecting CC1 and CC2 together means "Audio Accessory mode". \$\endgroup\$
    – Ale..chenski
    Commented Jun 28, 2019 at 1:07
  • 1
    \$\begingroup\$ I'm trying to save the cost of not understanding why a component appears in a schematic.The Microchip overview I linked explains in section 3.3 that active cables are signaled by resistance Ra on the CC lines, which is distinct from the resistance Rd used for orientation. \$\endgroup\$
    – lynaghk
    Commented Jun 28, 2019 at 7:44
  • \$\begingroup\$ For better understanding you better look into the original USB Type-C specifications instead of reading some limited interpretations. You will find out that there are active cables that take alternative power from the other CC pin that becomes Vconn, and multitude of "alternate modes" that also use certain combinations of both CC1 and CC2 pins. \$\endgroup\$
    – Ale..chenski
    Commented Jun 29, 2019 at 1:27
  • 1
    \$\begingroup\$ If it's clear in the spec, can you write a full answer so I can accept it? I'm designing a receptacle, not a plug, and my reading of USB Type-C Specification Release 1.4 suggests that shorting the CC lines in my receptacle is fine. Even if a powered cable is used, table 4-42 says that electronically marked cables (of which active cables are a subset) have isolation elements to prevent Vconn from reaching the sink. \$\endgroup\$
    – lynaghk
    Commented Jun 29, 2019 at 23:51

2 Answers 2

Reset to default
6
\$\begingroup\$

No, the CC pins should not be tied together on Type-C receptacle. USB Type-C specifications, V1.4, has several places that deal with use of CC1 and CC2 pins.

If you are making a USb 3.x implementation, the active CC pin will tell you the right direction of Rx/Tx data MUX. Individual CC inputs on device receptacle form the entire idea of detection of plug orientation. The specs describe the process on p.165, Section 4.5.2.2.5.1 as:

If the port needs to determine the orientation of the connector, it shall do so only upon entry to this state by detecting which of the CC1 or CC2 pins is connected through the cable (i.e., the CC pin that is in the SNK.Rp state).

More, on pages 162 and 163 the specs explicitly say, i.e. Section 4.5.2.2.3.1:

Both CC1 and CC2 pins shall be independently terminated to ground through Rd.

Is this enough to convince you that CC pins should not be tied together?

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Thanks for taking the time to explain this; I edited my question to indicate that the device only needed USB 2.0 support and thus didn't need to know orientation. Your finding in section 4.5.2.2.3.1 --- "Both CC1 and CC2 pins shall be independently terminated to ground through Rd." --- makes sounds unambiguous to me. I'm curious, though, do you know what specific failure (besides disobeying the spec) would occur if the CC pins were shorted? \$\endgroup\$
    – lynaghk
    Commented Jul 1, 2019 at 5:32
  • 5
    \$\begingroup\$ The Raspberry Pi 4 shorted CC pins together, which keeps it from being charged when used with an electronically marked cable. See scorpia.co.uk/2019/06/28/… for a full explanation. \$\endgroup\$
    – lynaghk
    Commented Jul 9, 2019 at 20:40
5
\$\begingroup\$

Sharing the pull-down will cause problems with cables that have an IC and need power. These cables have a pull-down internally, Ra = ~1kΩ, on the CC pin that is not being used for the CC wire. You can see it in this diagram, probably from 2: Rp, Rd, Ra diagram

Imagine what happens when Ra is present if both CC pins are tied together on the UFP and connected through a single Rd to ground. The DFP will see the combined pull-down of about 830Ω through both Rd (5.1k) and Ra (1k) in parallel on both CC pins.

It will appear to the UFP that it should supply VCONN on both CC pins and that there is no DFP connected. A charger that doesn't support USB-PD, i.e. 5V only, might not bother to care, but a USB-PD charger will likely refuse to supply power.

Also consider the UFP's problems. Of course you can't sense cable orientation, but for only power and USB2 that doesn't matter.

The UFP will sense 0.07V on CC for a 56kΩ Rp (500mA supply) or 0.38V for a 10kΩ Rp (3A supply). If the UFP is trying to do USB-C power supply sensing, it should detect the first as below the 0.25V threshold and consider no DFP connected, and sense the 2nd as a 500mA supply.

All that said, if the UFP does not care about orientation, does not try to draw over 500 mA and so does not try to sense DFP power capability, and one does not use a cable with Ra (e.g. a 5A charge cable), then there isn't a problem.

\$\endgroup\$
0

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