Differences between TX2 Hardware Revisions


We’re using TX2s on a custom carrier board and running into some display issues. On previous versions of our boards we haven’t had any issues, but on our latest revision only TX2s with the model number 699-83310-1000-D01 display properly on an HDMI monitor. We have several TX2s that are 699-83310-1000-B02 or B01, and when these are used the GUI will not load. The boot text will come up and we will see the mouse cursor and are able to get to a terminal interface, but we can’t start the GUI. What differences are there between the D01 and B01/2 models that could cause this?

Hi, could you please check and tell the difference between your latest carrier boards and old ones first? Basically, if customer followed the latest OEM Design Guide well, there should be no such module compatible problem.

  1. Are you able to reproduce this issue if you put these modules on devkit?

  2. For your board, please compare the syslog (not kernel log) between the working modules and NG case.

  3. Are you able to init the GUI manually? If it is rel-28 based, could you try to use “sudo service lightdm start” ?

One more thing to check, could you compare the node under /proc/device-tree/chosen/plugin-manager/ids ?


  1. The modules all work on a devkit with no issues
  2. The modules that don’t bring up the GUI on the newer boards have this in their dmesg log:
    iso emc max clk=0KHz
    tegradc 15210000.nvdisplay: sanitize_flip_args: WIN 3 invalid size:w=0,h=0,out_w=0,out_h=0 (repeated about a dozen times)
    usb usb1: usb_suspend_both: status -16 (also repeated)

These don’t show up in the modules that are able to bring up the GUI
3. I cannot start the GUI manually

Once I’m able I’ll update with the node under /proc/device-tree/chosen/plugin-manager/ids ?

@WayneWWW A D01 TX2 has the node 3310-1000-D01 which when cat’ed gives /i2c@c250000:module@0x50
A B02 TX2 has the node 3310-1000-B02 which outputs the same thing

Please check your design of strapping pins setting as listed in OEM DG, those pins should be same as reference design. Especial the level shift used on those pins, they can affect the pins status if not same to the ones used on reference design.

Turns out it was something we were doing with UART1_TX on boot, thanks!