SW reboot issue on TK1

• Plug in power supply DC jack to power on M/B, check all function work properly.
• Execute the command: “reboot”, then TK1 system will be hanging.

We have some finding as following

  1. The +1.05V_RUN power will lose after the command “reboot” was being excuting.
  2. Monitor I2C interface between TK1 with PMIC, TK1 does not complete the I2C command procedure to PMIC.
    Then M/B hang.
  3. Upgrade SW to R21.7 can fix the “sw reboot fail issue” but there are side effects when we continuously reboot M/B by SW command more than 200 times: the USB3.0 interface will be no working.
  4. Above failure symptom follows CPU, we duplicated it by AB swap approach.
    Please let me know detail about what’s modified on R21.7, why it impacts USB3.0 and how to fix this bug.

Hi, please clarify:

  1. R21.7 SW can solve “sw reboot fail issue” on failure board, but USB3.0 fail?
  2. Swap Tegra chips of failure board and good board, the “sw reboot fail issue” will go with chip to good board, and the failure board will have no issue?
  1. yes, but R21.7 will impact USB3.0 function.
  2. yes. After change NG TK1 to good board, this board will have “sw reboot issue” and the original NG board become as good board.

Does the good board have USB3.0 issue with R21.7?

yes, USB3.0 issue always follow R21.7.

Did yo swap failure chip to devkit board? That should be able to confirm if any chip hw issue.

I will check if devkit is available and function ok first, then swap failure chip. It will take some time becasue we must have BGA rework process internally.

We have found two chips with reboot issue. Can we ship one to you for RMA process first? We will try to rework the other one on TK1 devkit and this will take some time.

Sure, see here for info on the RMA process - https://developer.nvidia.com/embedded/faq#rma-process

This RMA process will be handled by our factory, I will rework failed chip on your TK1 devkit.

please ignore this.

We have started RMA process with your team but they asked that we must provide serial number. Actually, we don’t keep it because the serial number is outside the package, the package is discarded after we use it. It’s difficult to keep all package on our production line for TK1 chip. Could you advise me other way without serial number?

Hi ken6_lai,

Please contact with NVIDIA region sales team if they could help.

I get feedback from your RAM team as following. FA and RMA are different teams, right?

RMA team:
we do not have the option available to share FA for the failed chips

Yes, they’re different team, and you need to contact with our region sales team if they could help on the FA things.

It seems no RMA option for TK1 chip only. Do you advise me other way?

---------------RMA team feedback----------------
We do not have option to just replace the microprocessor for Jetson TK1 development from our end.

Hi ken6_lai,

As my suggestion at comments#15, please contact with our region sales team. That’s out of forum support scope.


Hi Sir,
Could you let me know detail difference between R21.5 and R21.7? Do you have any release note for that?

Hi ken6_lai,

Please find the release notes from each release packages at https://developer.nvidia.com/embedded/linux-tegra-archive