• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 2 answers
  • Subscribers 32 subscribers
  • Views 295 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.

Original question:

TUSB8042: Compatibility with TUSB8044

TUSB8041: Using available 8041 instead of unavailable 8042 part. Are they swappable?

WST
Intellectual 935 points
Part Number: TUSB8041
Other Parts Discussed in Thread: TUSB8040, TUSB8042

TUSB8041 compatibility...
Since the original question was resolved, I moved this here to ask...

A follow up. We STILL are having issues with getting the TUSB8042RGCT part in the future. I have turned on our board with the 8042 and have no issues, all 4 ports are in use. For the future, our board load shop has the TUSB8041RGCR part available ([41]USB3.0 vs [42]USB3.1). I looked at the differences in the above documents and it looks like I can just load the 8041 instead of the 8042 in my situation (not using any of the charging options). Is the only difference I will see the 5Gbps vs 10Gbps [not sure we would have that speed with our layout anyway, it looks like the TI comparison says they all operate at 5Gbps]? I think that the 3.1 vs 3.0 is the only difference, but I wanted to double check. Should I load one on a prototype to verify operation (it will be difficult because of the thermal pad and removal issues) or am I pretty safe just loading the 8041 instead of the 8042?
I think the 8041 is NRND (not recommended for new designs) but it is the only one that we have sufficient quantities of for future builds at this point.

The 8041 I can get is actually MPN = TUSB8041IRGCR, the industrial version, NON-'A' part. I'm not sure that matters, but I wasn't sure what the 'A' provides, but I think it is USB3.1 to match the 8042.

  • 0
    TI__Guru 60865 points

    Hi,

    You shouldn't have any issues using TUSB8041 instead of TUSB8042/A.  They are both 5 Gbps, just compliant to different versions of the USB 3.x specification.  The TUSB8041 is not marked NRND.  The much older TUSB8040 is marked as NRND, however, but that has a different package and couldn't be used with your design.

    Please note that for best interop performance the TEST pin should be pulled low in the design.  It isn't required for all versions of the hub, but if you are changing them it is best practice to have it pulled down. Also, since the TUSB8041 is compliant to an older version of the specification, it will allow downstream ports to enter USB 3.x compliance mode if they detect the receiver terminations but polling doesn't complete - this is different from later versions of the hub that require a command for a USB host to enable entry to compliance mode. 

    Regards,

    JMMN

  • 0
    Intellectual 935 points

    Thanks for the answer. We currently have TEST pulled low through a 4.64K resistor, so that is already in place.
    I was trying to test the downstream ports for compliance, but since our OS is LINUX I couldn't use the Windows HSETT tool to put the ports into test mode with the 8042. I couldn't find a way to make them go into compliance mode to run the compliance tests, even using the VID[1A0A]/PID[0104] method. I just couldn't get it to work and put out test signals. Maybe the 8041 will get me around this problem.

    Thanks for letting me know the part is still active, I think the NRND was our internal classification since the 8042 was released and materials preferred we use the newer part. However, since it isn't available right now I'll switch to the 8041.

    Warren

  • 0 in reply to WST
    TI__Guru 60865 points

    Hi Warren,

    Just a note, the VID[1A0A]/PID[0104] method is for USB 2.0 testing of embedded hosts so I wouldn't expect it to work for hubs at USB 2.0, or USB 3.0 at all.  You should be able to get USB 3.0 TX compliance running on the TUSB8041 downstream ports on the Linux platform though.

    Regards,

    JMMN

  • 0 in reply to JMMN
    Intellectual 935 points

    Is there a documented procedure for the 8042 or 8041 to do this?

  • 0 in reply to WST
    TI__Guru 60865 points

    Hi Warren,

    The TUSB80xx hubs follow the standard USB 3.x TX/RX test procedures, so we don't have a TI specific document.  The procedure also will vary depending on which test boards / test equipment you are using.  If you can share what your test setup consists of I can help debug test issues, but unlike USB 2.0 there isn't an easy way to get an eye diagram for USB 3.0 without using a scope / test equipment with a dedicated USB 3.x test program.

    Regards,

    JMMN

  • 0 in reply to JMMN
    Intellectual 935 points

    I am using a D9020USBC USB3.2 Test App on a Agilent DSO-X93204A oscilloscope. This is combined with a U7242A test fixture. I tried it in HOST mode but couldn't get the DUT to put out the compliance patterns. I have the TX+/- cabling connected to the test app along with the RX+ cable coming back. It is plugged into one of the 3 HUB downstream ports that I am using on the 8042. I am trying the HUB-downstream instead. Maybe I can get that to work.

    It runs the LFPS test but won't put out the CP0 test pattern.

  • 0 in reply to WST
    TI__Guru 60865 points

    Hi Warren,

    Ok, it looks like you have all the necessary equipment for USB 3.x testing.  The downstream ports of the TUSB8042 won't output the compliance pattern unless the hub receives a command from the USB host to enter compliance mode (usually from the USB-IF XHCI HSETT program running on a Windows based host).  The TUSB8041, however, will be able to enter compliance mode under these conditions without the command.

    Regards,

    JMMN

  • +1 in reply to JMMN
    Intellectual 935 points

    Thanks for the reply! Yes, I had figured that was the problem, I needed to send commands to the HUB. Since we are using the HUB in a system run by a LINUX CPU we can't use the HSETT tool. I have used that before. I suppose I could disconnect the upstream port on the board and kludge it into my Windows desktop through a cable and run the HSETT tool there, but I'm not sure it is worth it. I will likely wait until I get a board with the 8041 loaded and try again. I'm assuming that the 8042/41 ports are electrically similar so that testing on the 8041 would reflect 8042 driving and receiving if we used that in the future. The compliance testing is really testing the board layout and routing and that won't change with a replacement part.

  • +1 in reply to WST
    TI__Guru 60865 points

    Yes, the electrical performance between the TUSB8041 and TUSB8042 is the same, and it is definitely easier to test the TUSB8041 on Linux platforms.

    Regards,

    JMMN