FINALLY SOLVED:
apparently on newer versions pins 8/10, it has become ttyTHS1
Also most pinout diagrams are shown 180 degree wrong pin 2 is NOT on the top right corner
trying to use /dev/tty/THS0 for doing serial comm (expansion pins UART1_TX (pin8) _RX(pin10) ), but the port does not show up for python (gives no such..blah blah.. error),nor does the port show up a selection in a terminal app.
Tried systemctrl stop nvgetty and systemctrl disable nvgetty,with no luck (and reboot too)
ls -l /dev/ttyT*
crw–w---- 1 root tty 239, 0 Apr 1 14:40 /dev/ttyTCU0
crw-rw---- 1 root dialout 240, 1 Nov 21 2023 /dev/ttyTHS1
crw-rw---- 1 root dialout 240, 2 Nov 21 2023 /dev/ttyTHS2
is all that shows
what next??? Where are the instructions for configuring/using the port?
My understanding is on some Jetson Orin Nano versions pins8/10 will be accessed as ttyTHS1, not ttyTHS0. The pinout documentation neglects to say this. Also for stupidity, the drawing is created rotated 180 deg from the physical pinout on the right hand side of the Jetson.
The number for serial interface is assigned dynamically and may defer between releases so that it would not be mentioned in the document.
For the UART(PIN8 and PIN10) from 40-pins expansion header, please just check uarta@3100000 in dmesg:
It seems you moved from JP5.x to JP6.x
In JP5, it is enumerated as /dev/ttyTHS0
In JP6, it is enumerated as /dev/ttyTHS1
Please just refer to the number printed on the devkit board to get which one is the first pin.
Please just refer to the number printed on the devkit board to get which one is the first pin. True, but is a poor explanation why it was drawn in a non-sensible way.
This document seem from JetsonHacks, which is not maintained from us.
Please refer to Jetson Orin Nano Developer Kit Carrier Board Specification in Jetson Download Center | NVIDIA Developer
This is still stupidly shown 180 deg reversed, relative to looking at the jetson–WHY? Here is something drawn sensibly–was NVIDIA too incompetent to do so??
I don’t think it is 180 deg reversed.
There’s also the label on the devkit board.
Please refer to the following image that the upper left pin is the first pin on Orin Nano devkit.
I can confirm this, just recently integrated an HM-11 BLE 4.0 module directly to the 8/10 UART1 pins and have been able to access via /dev/ttyTHS1 in code successfully. It would be great if there were half-way markers (PIN 19/20) or so - its really hard to count those tiny pins!
The drawing IS cluelessly drawn 180 deg reversed from looking at the front off the jetson —WHY WAS IT DRAWN IN THIS MANNER?? Your response kind of shows why (people that are clueless). If you want to mark it why not use a font that is readily noticeable from at least a foot away?? Here is the connector drawn properly and smartly, not cluelessly(as seen from front of Jetson)
If an engineer can’t figure this out, maybe they shouldn’t be engineers? I mean the diagram is labeled and the board is labeled with pins 1, 2, 39, 40.
So your problem is the orientation of the two images together? Forget about the drawings and print out the pinout diagram and orient your board correctly next to it, problem solved!
I mean the diagram is labeled and the board is labeled with pins 1, 2, 39, 40. …why cough up a dumb excuse----especially when the labels are almost invisible when the board is installed in the case. You still have not said WHY it would not be drawn in a user-friendly orientation. Print it out and rotate it? Stupid response. Why not draw it so it doesn’t need rotated & the lettering won’t be upside down either. With all of the MANY <MANY complaints about poor NIVIDA implementations, this is just astrong textnother example of their poor engineering. You don’t seem to “get it” and that is why such issues are rampant with Jetson.
There were two ways to orient the drawing vertically. Apparently, smart selections take a back seat at NVIDIA
[quote=“KevinFFF, post:14, topic:328934”]
Sorry that I don’t know where do you get this figure.
It seems not from our official document.
[/quote] I drew it, as it should have been drawn properly, if NVIDIA knew how to do it sensibly–why wasn’t it???