@guido4321 The more I think about the problem I seem to have with booting up the Xavier board, the more I think it is a result of not having a dual screen. I have a Samsung large display hooked to a Mac Pro with a VGA cable. I also have an HDMI cable that connects this display with the Xavier board. How do I turn the latter in a serial console or how do I get a dual display on my display? When I have a dual display do I need a terminal window or dos do I save a log from the serial window? Final question: when downloading jetson-nx-developer-kit-sd-card-image.zip from NVIDIA website…when unzipped and uncompressed I end up with sb-bog-img which is the file I flashed to the SD…is that the right file?
FYI, VGA is not supported (though it won’t hurt if you are only using this on your Mac). HDMI is needed for video auto configuration to succeed (and it sounds like this is what you have so it shouldn’t be a problem). Dual screen is not needed for the actual NX.
Serial console uses a serial UART cable (usually USB at the host PC side). For serial console, see:
(note that the NX and Nano use the same carrier boards, and thus the same instructions for serial console)
A serial console is a text terminal which runs on a different computer than the computer which is actually being controlled. On Linux I prefer to run
gtkterm on my host PC. Some people using Windows use
PuTTY. I’m not sure what serial console program works on a Mac, but I’m thinking
PuTTY might be available there. Note that
PuTTY also supports some ssh protocols, so you’d have to be careful to use the serial protocol setup.
The serial console program allows you to type in commands and to view responses just like you were on a text-only console on the NX. The nice thing about that is that if something fails on the Jetson, then the output is still running on your host PC and is not lost. Most serial console programs also allow logging of entire sessions.
I’m not sure about all of your questions, but if you use a pre-built image instead of flashing via SDK Manager/JetPack, then the image has to be installed to your SD card with a program intended to do this, e.g., “
Etcher” in Windows or “
dd” in Linux. An explanation as to why this is the case is that normally you would create a partition and format this, and then copy content to the formatted partition; however, the underlying filesystem is itself already included in an image, and so there is no format step…the implication is that to also copy the filesystem you need different software.
If you were to format a filesystem, then the host PC doing the format needs to understand the filesystem type. Windows definitely does not understand “
ext4” (the Windows Subsystem for Linux might, but directly in Windows). Having the
ext4 filesystem already in the image implies Windows only needs to treat the image as binary data and no longer requires knowledge of
Thanks for your reply. I clearly wrote that my display is connected to a Power Mac using VGA cable, hence all references to linux and windows are just not useful.
You wrote HDMI is needed for video auto configuration to succeed (and it sounds like this is what you have so it shouldn’t be a problem)…what do you means by video auto configuration? When I go to System Preferences and open Display, there is no Arrangement options that would allow dual screen mode.
You wrote “I’m not sure what serial console program works on a Mac, but I’m thinking
PuTTY might be available there’ what does this mean? how do I turn my display into a single console if that is necessary?
You rote “if you use a pre-built image instead of flashing via SDK Manager/JetPack, then the image has to be installed to your SD card with a program intended to do this, e.g., “
Etcher” in Windows or “
dd” in Linux.” What about Mac….I used balanaEtcher to flash the image I got from downloading the Jetpack from NVIDIA website.
The formatting of the SD card was done by the SD formatter for Mac which is on the web (I did not use the Mac for formatting)
Please spare me from linux and windows advise.
I think we should make some explanations here.
First, we don’t need to care about what monitor is connecting to the Mac. The monitor should be connecting to the Jetson NX. During your first boot up, there should be NVIDIA logo showing up on the screen. And then the GUI on monitor will ask you to configure user account/timezone and password. That is what “video auto configuration” is doing.
Please note that the GUI is output from tegra… I don’t know why you are trying to do System Preferences and open Display. Your description sounds like you are trying to configure monitor on Mac? That sounds not a normal way to configure monitor… Why do you want to configure it this way? How many monitors do you have there?
Second, serial console is not opened on the monitor. Tegra will need to connect to your host through tty-usb cable. You can google search “tty-usb” and it will show you some pictures like:
Then use some console tool to open it on your host. “Putty” is such tool. But we don’t know which kind would work on Mac so cannot provide precise suggestion.
And I would still post this page again. This page has some links to show you the setup of tty-usb cable and how to dump log.
I have a Samsung large display hooked to a Mac Pro with a VGA cable. I also have an HDMI cable that connects this display with the Xavier board. How do I turn the latter in a serial console or how do I get a dual display on my display?
This description sounds not related to tegra…
“Dual display” is just like the word itself… you need two displays to achieve dual-display…
The multiple ports on monitor are there to let you switch between each source but not display two sources simultaneously… unless your monitor has such capability to do PBP.
I am glad you provided some explanations.
My monitor is connected with HDMI cable to the Xavier board. However during boot up I NEVER got to NVIDIA logo; in consequence the GUI on monitor was never reached for me to be able to configure user account….
As I never got to the NVIDIA I started to think that it was because of my Monitor that is connected at the same time with VGA to a Mac as well as with HDMI to Xavier. Hence I was opening System Preferences to see if I could get a dual mode display…If this is wrong what really should be addressed is WHY I never get to the NVIDIA logo?
What you wrote about serial console and Tegra is still unclear to me. Do I need to get a cable like the one you showed? Do I need Putty?
I have read through the GeneralDebug page you posted.
What remains unanswered is whether the file I flashed to my SD is the right file “ sb-blog-img “. Has anyone at NVIDIA downloaded the Jetpack development kit on a Mac, unzipped it and uncompressed it and then verified what file ought to be flashed to a SD? I think you might get surprised that the jetpack-…download zip produces two files of which only the sb-blog-img can be flashed.
Thanks for your continued support.
Actually, according to your description, I guess your board is not powering up at all…
Do you see any led light up when you connect the power cable?
Also, what is your power cable? Is it through micro usb or power jack barrel?
Yes there is a green LED when I put the power in.
The power cable I used is the one that came with the Xavier board.
My suspicion is that the file I flashed to the SD is not the right one. I will today go to a windows PC, go to the download center, download the Xavier Developer kit, unzip it and then flash the file I obtained to the SD using my Mac. I wish someone on your side would do the same using a Mac instead of a Windows based PC.
Ok. If you can see the led, then using Windows to install the sdcard again is worth trying.
I went to my wife’s window based PC; downloaded the Jetson Xavier Developer kit and unzipped it on her PC! I moved both files the zipped and unzipped version to my Mac. Used balaunetcher to flash the unzipped file “sd-blog-img" to the SD. Plugged the card into the Xavier, powered up and waited…(.I understood that my monitor would auto configure)….waited some more time…No NVIDIA logo came up. After a while the fan got warmer and warmer. I unplugged the power supply.
When I used my wife’s window based PC I noted that her display had ion the front panel a menu and one button was to switch the input source. Returning to my display — which does not have front panel bottoms — I searched the manual of the Samsung SyncMaster T260HD monitor and found that with the remote control I could switch to HDMI source. I plugged in the power supply and EUREKA saw the NVIDIA Logo in no time. I went through the initial configurations and now have a NVIDIA screen whenever I put the input source to HDMI.
Why did anyone else tell me this? Even Paul McWhorter failed to mention this in his tutorial.
Wayne I thank you for sticking with me and eventually leading me to the solution. I am now ready to get back to Paul’s tutorial and learn about the other initial configuration changes he suggested.
Many Thanks Wayne!