Why does Tegra fail to read and program EEPROM

Robotics & Edge Computing Jetson & Embedded Systems Jetson Orin Nano
,
1

I have tried to ensure the flash is correct by triggering recovery mode and then trying to reload the system from the sdk-manager. I am continually having the sdk-manager tell me the system is in recovery mode and cannot be flashed. It then tells me to power cycle the system and re-trigger recovery mode. Which makes no apparent difference. A power cycle alone is supposed to clear the recovery-flag and does not seem to do so. I see Tegra marking errors in its log which do not suggest actual cause.
Full log possibly attached.
Flash Updatae Fail SDKM_logs_JetPack_6.2_Linux_for_Jetson_Orin_Nano_[8GB_developer_kit_version]_2025-02-28_15-31-39.zip (120.9 KB)

13:41:23 INFO: [ 0.1132 ] Assuming zero filled SBK key

13:41:23 INFO: [ 0.1249 ] Warning: pub_key.key is not found

13:41:23 INFO: [ 0.1242 ] Parsing dev params for multi chains

13:41:23 INFO: [ 0.1242 ] Generating br-bct

13:41:23 INFO: [ 0.1246 ] Updating dev and MSS params in BR BCT

13:41:23 INFO: [ 0.1302 ] Signing BCT

13:41:23 INFO: [ 0.1312 ] tegrasign_v3.py --key None --list bct_list.xml --pubkeyhash pub_key.key --sha sha512

13:41:23 INFO: [ 0.3211 ] tegrahost_v2 --chip 0x23 0 --align applet_t234_aligned.bin

13:41:23 INFO: [ 0.3235 ] tegrahost_v2 --chip 0x23 0 --magicid MB2A --appendsigheader applet_t234_aligned.bin zerosbk

13:41:23 INFO: [ 0.3242 ] adding BCH for applet_t234_aligned.bin

13:41:23 INFO: [ 0.3411 ] tegrasign_v3.py --key None --list applet_t234_aligned_sigheader.bin_list.xml --pubkeyhash pub_key.key --sha sha512

13:41:23 INFO: [ 0.3413 ] Assuming zero filled SBK key

13:41:23 INFO: [ 0.3436 ] Warning: pub_key.key is not found

13:41:23 INFO: [ 0.3433 ] tegrahost_v2 --chip 0x23 0 --updatesigheader applet_t234_aligned_sigheader.bin.encrypt applet_t234_aligned_sigheader.bin.hash zerosbk

13:41:23 INFO: [ 0.3450 ] Sending mb2_applet…

13:41:23 INFO: [ 0.3455 ] tegrarcm_v2 --instance 1-4 --chip 0x23 0 --pollbl --download applet applet_t234_sigheader.bin.encrypt

13:41:23 INFO: [ 0.3459 ] BL: version 1.4.0.4-t234-54845784-e89ea9bc last_boot_error: 0

13:41:23 INFO: [ 0.5441 ] Sending applet

13:41:23 ERROR: [ 0.5884 ] ERROR: might be timeout in USB write.

13:41:23 ERROR: Error: Return value 3

13:41:23 INFO: Command tegrarcm_v2 --instance 1-4 --chip 0x23 0 --pollbl --download applet applet_t234_sigheader.bin.encrypt

13:41:23 ERROR: — Error: Reading board information failed.

13:41:23 INFO: 14

13:41:23 INFO: command finished successfully

13:41:23 ERROR: Bootrom status check failed.

2

In case it matters, the SD Card I am using is a 64G Sandisk Extreme
marked as U3 and A2. read speed 170 MB/s Write 80 MB/s

4

Hi,

Some questions to confirm?

  • Is your device a dev kit or a custom board?
  • Could you provide the link you bought?
  • Is your host a native ubuntu or a virtual machine(e.g. VMware…)

Thanks

5

The device I have is a dev kit, purchased by the home office and I am remote. The source may have to wait a bit, time zones and all.
SN on the box is 1424524602041 PN 945-13766-0000-000 SN on the box is 1424524602041 PN 945-13766-0000-000 , The QR Code has an R5 after the part number.

I am using a dual boot machine (HP Zbook G3) with ubuntu as I was unable to get the USB conenction through to the Oracle VM. MS was calling it a disk device and there did not seem to be any way to conenct it as a USB port. I admit my Unix/Linux skills are decardes out of date, but I should be able to repeat actions and suceed.

6

The Dev Kit was purchased on Amazon, from a vendor called Launched Central.

7

If you have dual boot ubuntu, could you install sdk manager in your ubuntu not a Oracle VM and reflash again?

Thanks

8

When I try to connect from the stand-alone ubuntu system. It is telling me that the GPU is in recovery mode and that needs to be corrected prior to reflashing. What I have seen here says that usuallymeans just a power cycle. The sdkmanager, forces a reboot, and then comes back to telling me the system is in recovery mode and needs a restart. I have tried multiple restarts. With and without separate power cycles. I have entered the boot manager page and told it to reset, but see no documentation on what that does. I have reloaded the SD card with Jetson6.2 which boots to operation. For some reason the Tegra base seems to be not clearing the recovery mode flag. Do I need to revert to an earlier version of the jetson SW, and if I do will that work with the 36.4 currently flashed.

9

I have reloaded the Micro SD at least twice and still seem to be stuck on verifying the flash load.
By letting the system come up and log in after the “system in recovery mode” error with a complaint that the “command to set device into recovery mode expected error catched”
And then apparently complains that it did, and did not, find the system in recovery mode a couple of times before proceeding. Totally unclear if this is an issue or not.
Then dealing with running an applet “applet_t234_sigheader.bin.encrypt” it reports a USB timeout.
Shortly followed by a “Bootrom status check failed.”

The dialogs about the failure are saying it “might” be a USB cable issue. I have tried two different new USB cables, purchased for this task. I have tried every port on my ubuntu laptop PC, and one on the Dock that I am using for power to the laptop. All get the same result. The log mixes error messages with “command finished sucessfully” and ends with BootRom status check failed.
ERROR Return Value 3, ought to be usable information if there was a guide to what the return values mean.

If the USB timing is so close that a 2m cable is a problem and 1.1 m cable is required that might be listed somewhere. 3 nanoseconds should not make a difference.

The Logs collection will be attached

10

SDKM_logs_JetPack_6.2_Linux_for_Jetson_Orin_Nano_[8GB_developer_kit_version]_2025-03-05_15-22-55.zip (203.3 KB)

11

Hi,

Could you execute below commands after you put the device in recovery mode?

lsusb

Thanks

12

after several tries, I saw another note that wanted two jumpers, both recovery and reset. After which I was able to program the device

1 Like
13

Hi,

If you succeed to put device in recovery, do you try to flash your device using sdk manager?

Thanks

14

I was able to get the sdk manager to complete the flash operation. Whether it was correct or not is still an open question. Once operational, I was getting error messages about segment violations (SIGSEV) and complaints about the libraries being obsolete.
I have since then migrated to using an NVME 2230 memory as the boot source, and I am adding a 2280 large memory for applicaiton storage. My HW team is making a custom module holder and I need to be aligned with the future hardware. I still got an error message about a camera not being found, I am not sure what might have caused that as I have not indicated a camera is installed anywhere. Both of those errors were reported to cannonical.

You might want to consider having the label sticker on the Dev Kit Carrier board recessed so it is flush with the surface. That edge may have been causing me issues in getting into Forced Recovery Mode, by keeping the double female jumper I was using from seating properly. I have not had a desktop in over a decade and laptops do not use those jumpers. Or you could consider including one or two with the dev kit.

15

And No, I did not use the sdkmanager, I used the command line options instead