I need to use the serial console as a console, and then also use a 3.3V serial port to talk to a separate board.
The J17 UART port contains level shifted TXD/RXD signals, which is just what I need!
There’s a bunch of talk on the board about it also being “routed to the camera,” and recommendations for removing the camera module when using this UART.
Does the camera use the UART at all?
I hooked up an oscilloscope and checked, and nothing goes out on the UART TXD when the camera is used for capture.
(I can’t check the RXD because the camera port is after the level shifter input, which I can’t probe.)
Or is the recommendation just for signal quality reasons? I’m not worried about two additional inches of unterminated traces when running at 115200 bps, so if that is all, then I’m all set.
I’d like to know, for sure, what all the hubbub is about, though. Maybe I’m missing something?
For the default camera the wiring touches the camera, but I have not seen any software which actually uses this. It would be hard to imagine traffic on just RX or just TX.
It would be harmless to test…wire the TX and RX together as loopback, and then test sending/receiving from “/dev/urandom” and see if any bits are lost or corrupt. Let it run for a few hours while the camera also runs.
Thanks linuxdev, but I really do want an authoritative answer!
A schematic for the camera module would be acceptable, or an actual “yes, the UART is used for X” or “no, the camera module does not actually use the UART” from NVIDIA would be good.
The camera board is whatever board came with the Jetson TX2 devkit. I don’t know how to get a part number from it.
I’m using the UART through J17 to talk to an external board, and so far it works fine, so my GUESS is that the camera board doesn’t actually use the UART, even though the connections route to the camera-board connector on the mainboard.
But ideally, I’d like to get this confirmed or explained.