This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS65994AD: [Public/TPS65994BF] USB-C with DP device cannot be recognized.

Part Number: TPS65994AD
Other Parts Discussed in Thread: TPS65994

Hello PD support team,

This is Dave, FAE supporting Taiwan Industrial PC customers.

My customer has experienced the type-C port problem that related to the figure below:

Here are some notes for this bug:

1. Only happening at DP to USB-C dongles (There is no issue when asserting USB-C storage and pure DP to type-C cable)

2. I've logged the I2C2 and I2C3, it seems normal, please see the ellisys log if needed:

0123_I2C2.ctrt0123_I2C3.ctrt

We did tell chipset with updating 5F register and also send the message through I2C3 to retimer(Hayden bridge).

However, my customer has asked Intel for this one, they told us that it should be dealt with PD firmware.

Can you please help to check this issue? I am also sending the link to RD Tony who is taking care of this project, he will help to share some info if I missed.

Thanks for the support.

Best Regards,

Dave

  • Hi Dave,

    Can you let us know who is the customer you are supporting?

    To confirm my understanding of the issue:

    • The TPS65994AD PD is in the industrial PC and is the DFP host. 
    • The far-end UFP is a DP monitor connected to the PC host using a USB-C to DP dongle.
    • The DP monitor is not enumerating on the PC after connection to type-C port.
    • Is there DP display?

    Please let me know if any of my understanding above is incorrect.

    I took a look at I2C logs, and it looks like the 0x5F data status register of the PD is indicating a DP Pin assignment C/D connection. This is not correct since a Type-C to DP cable or dongle should be DP pin assignments E/F. Unless there is a dock in between the PC and the monitor, which in that case, DP pin assignment C/D would be ok. Is there a dock in between the PC and the monitor? Or is the monitor connected directly to the PC using the type-C to DP dongle?

    Since the customer has access to Ellisys, could you have them collect PD/CC logs of the connection using Ellisys? Can they also provide the project file so I can check the configuration?

    Best,

    Alex

  • Hi 

    I am Avalue RD Tony, Correction, the PD I use is TPS65994BF

    tks

  • Hello Alex,

    My customer is Avalue, and RD Tony is in charge of this case.

    For Ellisys part, I don't have the pass-through cable, however, it can still be logged by the normal cable?

    The device is a USB to DP dongle, and the cable should be the DP-to-DP cable, and use the dongle to transfer to USB-C. No dock is presented.

    The DP is displaying and can be enumerated, but the problem is that why is the unknown warning at the device tree?

    Thanks.

    Best Regards,
    Dave

  • Hi Dave,

    The ellisys acts as a pass through between the UFP and the DFP. There should be two type-C ports on one side of the ellisys tracker. You can connect the USB-C end of the USB-C to DP dongle to the ellisys tracker, and connect the ellisys tracker to the PC under test using a normal type C to type C cable. This should allow for pass through of PD communications to be captured by ellisys.

    Does the warning at the device tree happen on every connection? Or just once. Having the PD logs will allow me to check the DP alt mode negotiations. I predict this is ok since there is display, but it will be good to see if anything is wrong that might be causing the warning message.

    Best,

    Alex

  • Hello Alex,

    Thanks for your reply here.

    I will capture PD/CC log for you next week.

    However, as Intel replied to this case today:

    What in byte 1 (Bit 07-00) should we send to SOC at 5F register?

    Currently, the U3 is present as result, please see the red square part in the figure.

    Thanks.

    Best Regards,

    Dave

  • Hi Dave,

    I would expect the below:

    Can you send me the project file so I can check configuration?

    Best,

    Alex

  • Hello Alex,

    The log shown means that the USB3 connection at bit 5 should not be asserted at this case, right?

    Please see the attached pjt file, thanks.

    Avalue_EEV-EX16-TPS65994BF_20231206.pjt

    Dave

  • Hi Dave,

    Yes, DP dongle monitor connection to PD should not assert the USB3 bit in data status register. I will check the project file and get you an update. It may have something to do with the configuration.

    Best,

    Alex

  • Hello Alex,

    Thanks, please let me know if there is the update after you check.

    Best Regards,

    Dave

  • Hi Dave,

    Nothing looks obviously wrong with the project file for DP. I would need to check PD logs to see what pin assignment is used for DP configuration. I believe the DP connection was configured as 2 Lane + USB3, pin assignment C. If 2 Lane DP is used, then USB3 bit will be asserted. If you can access debug mode using the PD GUI tool, you can take a register snapshot and I can look at the DP status from there as well.

    One thing I noticed was that in the DP configuration register, since you checked "Multifunction Preferred", we will prefer 2 Lane DP + USB3 connections. However, this should not be an issue here, depending on what the monitor supports.

    Does the PC under test support USB4 or USB3 at all? If not, you can disable whichever USB speed that is not supported in the 0x47 TX Identity register, DFP VDO. 

    Please provide the PD logs and/or PD register snapshot in debug mode when you can. We will need this to proceed with debug, as the next step is to check DP alt mode configuration upon entry.

    Best,

    Alex

  • Hello Alex,

    I will close the thread here.

    After checking, if "Multi-function" preferred is asserted, then the PD controller will identify the dongle as the 2laneDP+2laneU3 device, which may be incorrect.

    And this will be the reason for the system error, thanks!

    Best Regards,

    Dave

  • Hi Dave,

    Great! Glad this was resolved.

    Best,

    Alex