Hi
We have an AGX Xavier carrier board with two type C connectors.
However, the type C connector detected a USB3 device as plug one side type c device and detected a USB2 only as plug the other side type c device.
The typeC connectors are following J512 and J513 of AGX EVB schematic to design. The carrier board uses the same redriver and PD of AGX EVB too.
Thank you for any advice,
attached the log of ccg4 with USB3 and USB2
ccg4_USB2_I2C.sal (123.2 KB)
ccg4_USB3_I2C.sal (24.0 KB)
From internal team:
Decoding the I2C trace you provided:
USB3 I2C log:
write to 0x20 ack data: 0x03
read to 0x20 ack data: 0xBF
write to 0x20 ack data: 0x01
read to 0x20 ack data: 0x40
write to 0x20 ack data: 0x03 0xFF
write to 0x0C nak
write to 0x0C nak
write to 0x74 nak
write to 0x20 ack data: 0x03
read to 0x20 ack data: 0xFF
write to 0x20 ack data: 0x01
read to 0x20 ack data: 0x40
write to 0x20 ack data: 0x03 0x7F
write to 0x20 ack data: 0x01 0xC0
write to 0x74 nak
write to 0x0C nak
write to 0x0C nak
USB2 I2C log:
write to 0x0C nak
write to 0x74 nak
write to 0x20 ack data: 0x03
read to 0x20 ack data: 0xFF
write to 0x20 ack data: 0x01
read to 0x20 ack data: 0x00
write to 0x20 ack data: 0x03 0xBF
write to 0x20 ack data: 0x01 0x40
write to 0x74 nak
write to 0x0C nak
write to 0x0C nak
It looks like the attempts to write to TUSB1046 are failed, so only the default selection works.
I saw two TUSB1046s in their design, maybe you can recheck if there is any I2C slave address confliction?
Hi Kayccc,
Thank you for your prompt support.
The design does have TUSB1046s, however, the two TUSB1046s have each PD with.
Also the two PDs and two TUSB1046s will not work at the same time. Only one PD and one TUS1046s work every time.( Used HW switch to separate it.) ( see attached.)
From the I2C trace:
The mainly differences of usb3 and usb2 looks like as below:
USB3:
write to 0x20 ack data: 0x03
read to 0x20 ack data: 0xFF
write to 0x20 ack data: 0x01
read to 0x20 ack data: 0x40
write to 0x20 ack data: 0x03 0x7F
write to 0x20 ack data: 0x01 0xC0
USB2:
write to 0x20 ack data: 0x03
read to 0x20 ack data: 0xFF
write to 0x20 ack data: 0x01
read to 0x20 ack data: 0x00
write to 0x20 ack data: 0x03 0xBF
write to 0x20 ack data: 0x01 0x40
Do the differences mean any?
Or it’s normal for different size of type C.
Or these could be results not cuased?
We don’t get much information about this from TUSB1046s datasheet.
And don’t have the I2c TRACE of AGX Xavier EVB to compare with.
Thank you for any advice,
From internal team:
They need to figure out why I2C communication doesn’t work on TUSB1046:
read to 0x20 ack data: 0x40
write to 0x20 ack data: 0x03 0xFF
write to 0x0C nak
write to 0x0C nak
Base on above, CCG4 did attempt to write to 0x0c but failed. In the meantime, CCG4 can talk to other slave (0x20).
I will suggest they just attached an external I2C master and verify the I2C communication.
BTW, we also suggest customer to do USB3 compliance test and then try again. If the issue is still present, please request them to reproduce it on developer kit so that we can check further. If it is something wrong in our firmware it should be reproducible on developer kit also.
Guidance of compliance test:
Jetson AGX Xavier USB 3-1 Compliance Test Guide Application Note
Hi Kayccc,
Thank you for your prompt support.
Will transfer the suggest to the team.
However, the I2c trace of USB 3 showed CCG4 did attempt to write to 0x0c but failed too.
Thank you,
Issue is tracking with internal team, topic closed now.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.