• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 101 subscribers
  • Views 864 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

SK-AM68: the board arrived dead

yuli kaplunovsky
Prodigy 140 points
Part Number: SK-AM68

I was ordered by the TI support to post here in order for them to replace a damaged eval board.

I have more than 15 years experience doing this kind of development, dealing with booting linux and similar, so I'm 99% sure that my conclusion that the board was dead on arrival is CORRECT. but here it is:

it never booted from the microSD. It never showed any output on the HDMI. It never showed an output into any of the virtual serial outputs.

I reformat the microSD and put on it the correct partitions. It is not trivial, but as I said before, I did that for so many years, so it is all very familiar to me.

I tried to change boot switch to boot from different media. I tried with 2 different power suppliers. Nothing helped.

The microSD that arrived with the board - looked like somebody already did some experiments with it, and it DID not contain anything.

The package when arrived, looked like somebody already opened it And my guess is that this person already returned it to TI and they mailed this failed board to me.

  • 0
    TI__Mastermind 19680 points

    Hi Yuli Kaplunovsky,

    The SD card that is packaged with the board is always blank. Did you flash the SD card with the latest image? If not, this is where you download it: https://www.ti.com/tool/download/PROCESSOR-SDK-LINUX-AM68A

    When you power on the board and/or connect through UART, do you see the LEDs light up? What power supply are you using?

    Please let me know if flashing the SD card resolves your issue. In the meantime, I will consult with our hardware team to see if the reason why the package was opened was due to some hardware validation practices or some other reason.

    Thank you,

    Fabiana Jaimes

  • 0 in reply to Fabiana Jaimes
    Prodigy 140 points

    Thanks for the reply. I did program the SD with what i downloaded exactly as you specified in your response.
    Again - for more than 10 years I was doing these things, it is complicated, I know, but i already executed all your suggestions originally:  The Windows recognize all 4 com ports, i tried to connect putty to them all at both 11520000 and 9600 speed, and never have seen anything. I tried two different power supplies, and the pwr button of the board works as described (3 seconds to power off, and the other button turns it on). I tried to move the boot switches to force different type of boot, ethernet, and few more and powerup the system, just to see if I see anything coming out of the serial port, and nothing came out. 
    I am 99% I know to recognize bad hardware when I see one. I really would like to ship it back to you guys for your inspection, and I would really want to receive another eval board. 

    thanks

  • +1 in reply to yuli kaplunovsky
    TI__Intellectual 1160 points

    Hi Yuli Kaplunovsky,

    Sorry to hear that the board you received is not functional. As a SW engineer, I can understand the pain to see the new HW DOA.

    We will help you in getting the board replaced if it is defective. Before you start with return process, which is time consuming, I would like to request you to do couple of checks to assess the exact state of the board.

    Can you please keep the boot switch to setting to UART boot (SW1.1 - ON, SW1.2 - OFF, SW1.3 - ON, SW1.4 - OFF) and check if you are seeing any logs on UART consoles when you power ON the board.

    Connect all the four ports of the UART to serial console application while checking above. I am sure you are familiar with the default UART settings but listing the configurations that we typically use - 115200 baud rate, 8-bit data length, no parity, 1 stop bit, no HW flow control. 

    - Pratap.

  • 0 in reply to M V Pratap Reddy
    Prodigy 140 points

    after a lot of tinkering i found the problem, and TI guys please correct this on the release of the SDK 08_06_00!

    The SK-AM86 suppose to be the GP processor but it does NOT run with tiboot3-j721s2-gp-evm.bin and runs ONLY with the tiboot3-j721s2-hs-fs-evm.bin  - very strange ....

    The file that it reads while it boots from the SD is actually tiboot3.bin, which is a copy of one of the files above.

    The full blown .wic file tisdk-edgeai-image-j721s2-evm.wic.xz which you can download directly from TI\s web site above has it set correct, and this image works! (use BalenaEtcher)
    BUT
    The files inside the SDK 08_06_00  have the tiboot3.bin as a copy of tiboot3-j721s2-gp-evm.bin instead of tiboot3-j721s2-hs-fs-evm.bin!!! which causes the system NOT to boot and NOT to display anything.

    TI guys - Please fix it for the convenience of everybody....

    Another observation - the short test above is an excellent way to see if the board is alive or not - to set the boot switches (SW1.1 - ON, SW1.2 - OFF, SW1.3 - ON, SW1.4 - OFF) - then the SECOND terminal shows some characters on booting....

    When I connect the usb to my windows it appears as COM10, COM11, COM12, COM13. I was expecting the first COM (COM10 in my case) to show the console terminal of booting, BUT - pay attention - it comes on the 3rd terminal (COM12 in my case)....   115200 rate.

    And thanks the TI support guys helping me in spite of my complains...

    And here are my additional 2 cents ...

    I also was able to compile the SDK with yocto and create the SD from the generated files, and it did boot well!

    Here how i did that:

    prepare SD with fdisk /dev/sdb to look like that:

    Device     Boot  Start      End  Sectors  Size Id Type
    /dev/sdb1  *      2048   264191   262144  128M  c W95 FAT32 (LBA)
    /dev/sdb2       264192 30984191 30720000 14.7G 83 Linux

    sudo mkfs.vfat -F 32 -n "BOOT" /dev/sdb1
    sudo mkfs.ext4 -L "rootfs" /dev/sdb2

    build the yocto (takes half a day to compile... and 250GB of disk space...)

    cd /opt/ti-processor-sdk-linux-j721s2-evm-08_06_00_10/yocto-build/build/arago-tmp-external-arm-glibc/deploy/images/j721s2-evm
    cp u-boot.img /media/yuli/BOOT
    cp uEnv.txt /media/yuli/BOOT
    cp tiboot3-j721s2-hs-fs-evm.bin /media/yuli/BOOT/tiboot3.bin
    cp tispl.bin /media/yuli/BOOT

    sudo tar -xf tisdk-default-image-j721s2-evm.tar.xz -C /media/yuli/rootfs

  • 0 in reply to yuli kaplunovsky
    Prodigy 91 points

    I had the exact same issue and the above helped, thanks

  • 0 in reply to Richard Haar
    Prodigy 140 points

    Glad it helped!

    I'm designing my own HW with this chip...