Because of the holidays, TI E2E™ design support forum responses will be delayed from Dec. 25 through Jan. 2. Thank you for your patience.

  • State TI Thinks Resolved
  • Locked Locked
  • Replies 17 replies
  • Answers 2 answers
  • Subscribers 34 subscribers
  • Views 2038 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TUSB8041: Cannot detect USB disk on certain port

yf liao
Expert 1080 points
Part Number: TUSB8041
Other Parts Discussed in Thread: AM5728

Hi, we are using TUSB8041 with processor AM5728 for a 4 port downstream USB3.0 application. The design is basically following AM5728 GP EVM. We found a same issue on both our board and EVM board: the Kingston USB3.0 disk, PN DTSE9G2, both 64GB and 128GB cannot be detected. The USB disk information is as below. 

On our board, the issue persist happening on certain port, however it is board to board deviation. For instance, issue persist happened on port 1 and port 3 and 4 on one of our board(port 2 is OK), but happened on port 2, and port 3(port 1, 4 is OK) on another our board.

On AM5728 GP EVM board, issue persist happend on USB port P14's lower port(this is a stacked USB connector) and on port P13. 

We have connected some other devices on our board: Transcend JF780 USB 3.0 disk,  SanDisk SDCZ48 USB 3.0 disk, they all work normally on all the 4 ports and the speed is USB3.0.

HP USB 1.1 mouser, also able to work normally.

Our schematic is mostly same as EVM board, but layout is totally different. However the issue happened on both EVM board and our board.

Below is debug message from AM5728 serial port upon the issue happened and some part of our schematic. 

root@am57xx-evm:~# [  848.440849] usb 2-1.4: new SuperSpeed USB device number 20 using xhci-hcd

[  848.471047] usb 2-1.4: device descriptor read/8, error -71

[  848.590795] usb 2-1.4: new SuperSpeed USB device number 20 using xhci-hcd

[  848.622183] usb 2-1.4: unable to get BOS descriptor set

[  848.627834] usb 2-1.4: unable to read config index 0 descriptor/start: -71

[  848.634911] usb 2-1.4: can't read configurations, error -71

[  849.240874] usb 2-1.4: Device not responding to setup address.

[  849.460852] usb 2-1.4: new SuperSpeed USB device number 21 using xhci-hcd

[  854.550865] usb 2-1.4: device descriptor read/8, error -110

[  854.670800] usb 2-1.4: new SuperSpeed USB device number 21 using xhci-hcd

[  854.701157] usb 2-1.4: device descriptor read/8, error -71

[  854.820892] usb 2-1-port4: attempt power cycle

[  855.760850] usb 2-1.4: new SuperSpeed USB device number 22 using xhci-hcd

[  860.870851] usb 2-1.4: device descriptor read/8, error -110

[  860.990802] usb 2-1.4: new SuperSpeed USB device number 22 using xhci-hcd

[  861.021143] usb 2-1.4: device descriptor read/8, error -71

[  861.520865] usb 2-1.4: new SuperSpeed USB device number 23 using xhci-hcd

[  866.630870] usb 2-1.4: device descriptor read/8, error -110

[  866.750802] usb 2-1.4: new SuperSpeed USB device number 23 using xhci-hcd

[  866.781057] usb 2-1.4: device descriptor read/8, error -71

[  866.900898] usb 2-1-port4: unable to enumerate USB device

Please advise possible root cause?

  • 0
    TI__Guru 60870 points

    Hello YF,

    Can you confirm if you only see this behavior with a single instance of the USB 3.0 Kingston drive?  We use this drive in our lab very frequently for testing and we don't see issues.  This issue you are reporting sounds like a signal quality issue where the SuperSpeed connection is dropping.  Is it possible this drive was damaged?

    Regards,

    JMMN

  • 0 in reply to JMMN
    Expert 1080 points
    Hi JMMN,
    We saw this issue on one 64GB and one 128GB Kingston drive, both of them can work on laptop USB port. We don't have a chance to test more drive of this MPN.
    We will measure the signal quality on the superspeed differential trace. However I still doubt it is signal quality issue, as we see on the EVM board, the trace between the TUSB8041 and USB port is short, and the layout design looks good as well. However the issue happened on one of the stacked USB port where the two port USB 3.0 trace routing are somewhat identical.

    The controller of your Kingston drive is same as ours, Phison PS2307 08.04.5D? Is it possible there is compatiblity issue between TUSB8041 and Phison on particular firmware version?
    And is it possible you are able to test this Kingston drive with a AM5728 GP EVM board as well?
  • 0 in reply to yf liao
    TI__Guru 60870 points
    Hi YF,

    What drivers are loaded on the USB host? The Kingston drive implements U1/U2 power states and some USB driver do not properly handle these power states when a USB hub is in the path. Is there an external EEPROM for the hub?

    Regards,
    JMMN
  • 0 in reply to JMMN
    Expert 1080 points

    Hi JMMN,

    The drivers loaded on host is AM57xx Linux SDK 5.2, there is no EEPROM on I2C bus, the Hub pin 39 is floated. Some description from the driver code:

    CONFIG_USB_DWC3_OMAP:                                                    

      │ Some platforms from Texas Instruments like OMAP5, DRA7xxx and           │ 

      │ AM437x use this IP for USB2/3 functionality.                            │ 

      │                                                                        

      │ Prompt: Texas Instruments OMAP5 and similar Platforms 

  • 0 in reply to yf liao
    TI__Guru 60870 points

    Hi YF,

    I will check if anyone in my lab has this loaded so I can check the driver.  This sounds like a USB intermediate power state issue in the driver.

    Regards,

    JMMN

  • 0 in reply to JMMN
    TI__Guru 60870 points
    Hi YF,

    Can you try disabling power management in the driver?

    Thanks,
    JMMN
  • 0 in reply to JMMN
    Expert 1080 points

    Hi JMMN,

    Could you elaborate how to disable power management?

  • 0 in reply to yf liao
    TI__Guru 60870 points
    I am checking with the team.
  • 0 in reply to JMMN
    TI__Mastermind 27890 points

    Hi,

    This is taking a bit longer than expected, but it is actively being worked. We will post updates as they become available.

  • 0 in reply to yf liao
    TI__Guru*** 147816 points

    yf liao said:

    Hi JMMN,

    Could you elaborate how to disable power management?

    The usb driver in the SDK v5.2 kernel should already have usb3 lpm feature disabled. Can you please check the usb bus to see U1/U2 happens or not?

  • 0 in reply to Bin Liu
    Expert 1080 points

    We probed the USB bus signal on good Sandisk USB drive and problematic Kingston drive. I am not sure how to determine whether U1/U2 happened or not. Below are what we saw during the measurment.

    Sandisk drive:

    • After insertion, there is data transmission in both SSRX pair and SSTX pair.
    • The VBus voltage may drop to 3.3V and last around 6ms then return back to 5V.
    • The hub is able to detect the Sandisk drive.

    Kingston drive:

    • After insertion,  there is only data transmission on SSRX pair, but the differential signal amplitude is lower than that on Sandisk SSRX pair.
    • There is no data on SSTX pair on both side of the AC couple capcitor(hub side and connector side).
    • The VBus voltage drops to 3.3V with a very narrow pulse duration(~10 us)then return to 5V.
    • The hub is NOT able to detect the Kingston drive.

    Attached some measurment waveform, due to the oscillscope bandwith limitaion, the waveform in SSRX and SSTX pair may be distorted, we just want to see if there is data transmission occurred.

    Can you help to identify where the problem may be?

    New Microsoft Word Document (2).docx

  • 0 in reply to yf liao
    TI__Guru 60870 points
    Hi YF,

    Ok, this does not appear to be a U1/U2 issue. Can you share the layout of the USB 3.0 signals on your board? Is the EVM you duplicated the issue on the AM5728 EVM or a TUSB8041 EVM?

    Thanks,
    JMMN
  • 0 in reply to JMMN
    Expert 1080 points

    Hi JMMN, attached is the screenshot on one stacked type A connector J1. We have another stacked connector J2 beside, please leave the email address if you need whole layout review, I can send you the ODB++ file.

    layout.docx

  • 0 in reply to JMMN
    Expert 1080 points
    The issue is duplicated on AM572x GP EVM, USB port P14's lower port(this is a stacked USB connector) and on port P13.
  • 0 in reply to yf liao
    TI__Guru 60870 points
    Hi YF,

    I'm asking the AM572x GP team to comment on if they have seen any similar issues. We have not seen this on the TUSB8041 EVM and I have retested the Kingston drive today. Is it possible that your Kingston drive is damaged?

    Regards,
    JMMN
  • 0 in reply to JMMN
    Expert 1080 points
    We are now investigating on the Kingston drive connector plug compatibility with the receptacle we use. We just realized the there is some mechnical dimension in the both Kingston plug and receptacle are not fully compliance with USB standard, this will lead to connection issue during the insertion.
  • 0 in reply to yf liao
    TI__Guru 60870 points

    Hi YF,

    Thanks for the update.  I am closing this thread for now, if you have any additional questions you can start a new thread or post on this one to reopen it.

    Regards,

    JMMN