Confirmed that is the setup. I’ve double checked the cabling tx to rx. Even tried swapping.
If I have the screen/minicom active before I power up the R3 I see a blank screen with an unresponsive cursor. Then plug the power in. There is no activity on the screen at all. As if no data turned up at the serial port from the power on.
As I’ve said, it just worked once. I’ve no idea what changed such that it now does not.
Many thanks once again.
Does it matter how the Dupont connectors are inserted onto the pins? I’d have thought not.
Each connector is 4 sided and one side has a metal square (I used this for the various electrical tests with the meter). I have them all oriented the same way with the sides with metal exposed all pointing to the outside of the board. (same direction as the DIP switches point). Surely this doesn’t matter?
Orientation does not matter on the Dupond. But they are very cheap connectors and the metal curved part that keeps the tension between the connector and the pin gets bend sometimes, resulting in a very poor connection. Surely could have happended after trying so many times. Re-bending or replacing the connector can fix this issue, it happened to me more then once… You can feel the connector fit with some tension and is not too easily removed.
If you want to fix by bending, then you need to take the metal part from the housing first…
thanks. I’ve just tried lightly touching them. There is some looseness between the connector and pin while being completely inserted and it can be lifted a couple of millimetres and there is some wriggle room. but after that its quite firm and needs effort to remove. This is the same with all three pins.
So I tried lots of jiggling and moving around of the 3 connectors. I expected to maybe see some random data shown in the terminal session just as the voltages come and go as signals. But nothing at all. As I say, it’s quite unresponsive.
Disconnecting the USB connector causes /dev/ttyUSB0 to be removed and it gets created afresh every time its connected.
Getting very hard to think of options.
You probably checked the wires are 0 Ohm?
Are the settings saved correctly in minicom?
on this latest USB FTDI adapter? no I can’t. It’s hardwired so I can’t get to the other end of the three conductors (connected to the USB enclosure).
In the other two adapters yes I did check. just did again with one of the other adapters. Conductivity is fine for all of them.
yes. the only one that needed to change from your suggestion was Hardware Flow Control set to no. Did that and restarted minicom to inspect the setting and it showed ‘no’. Much as it originally showed yes prior to changing.
Thanks again. As per your note the Dupond connectors are fairly flimsy and that it does not matter which way they are slid onto the pins I re-seated each of them in the orientation that feels the tightest, with the least movement when they are attached. I used the FTDI adapter as its the newest and has been used the least. No change at all.
I suppose I now need to buy a new R3, unless there is some further creative testing anyone can suggest.
Is there maybe a cheap uART based device I could connect to test that connection? Something that just flashes lights for correct operation or some such. That would show the connection is ok and the device functioning. I previously posted what voltages I could see but don’t know how to interpret.
Sure, you already have it, another usb-uart device.
If you just run an image from sd catd, then you don’t really need the uart.
oh. I don’t quite understand.
Do you mean I could connect the Dupond connectors from one adapter to the pins on another adapter? So I have USB to USB each end, talking to screen -L /dev/ttyUSB0 at one end? what would I see?
If boot the R3 from sd-card (all DIP switches up) rather than NAND or eMMC, how does that help me get a console session?
Thanks.
So do screen command on both ends. Preferably use 2 different hosts. If you plug both adapters in the same host, you are not checking the most important gnd.
If you run a functional image (on any medium, but sd is easiest) you can use ssh. Only when things break down you need uart.
I’m revisiting this issue as despite my best efforts (on an on and off basis) it is still unsolved. I have acquired a brand new Raspberry Pi to use to connect to the Banana Pi R3 UART. I also now have 3 USB - UART adapters. Two (including the original which worked on day one) are CP2102, while the third is an FTDI based chip. The new RPi has a fresh sdcard install + screen installation. NO matter which USB3/2 port nor which USB adapter I use, I cannot get any response from a terminal with ‘screen -L /dev/ttyUSB0 115200’. Just like before.
Unless there is some other explanation the community can offer I think the UART circuit on the Banana Pi R3 has failed or been damaged. Therefore the solution is to buy a new one.
Except. I tried the test suggested here in May of connecting the USB adapters together. So one in my original RPi connecting the TX/RX/GND to another adapter in the other RPi via a second adapter. I made sure and checked TX/RX were crossed. With ‘screen -L /dev/ttyUSB0 115200’ at both ends there was no communication at all. no data ever seen. I tried installing the usbmon kernel module and monitored the usb bus via /sys/kernel/debug/usb/usbmon/2u - where 2u matches the bus in use for the FTDI adapter (from /sys/kernel/debug/usb/devices). When starting the screen session the output was a continuous stream of data, seemingly a small number of repeated values. Needless to say it offered no insights.
So now I wonder is a new R3 needed? Any other tests possible with it? Or is the lack of any connectivity between two RPis via serial UART USB adapter and the screen connection a reliable test and indicative of an issue elsewhere?
Many thanks.
Easiest check is using 1 UART, connecting rx to tx. Without the connection, you see nothing when typing (checking echo not configured), with connection you should see what you are typing.
You can do this with usb-uart or with the UART on bpi board .
This way you can check which UART is not functional.
Thank you. I was a little confused at first. But I think you mean that I take the tx and rx pins on the USB adapter and connect them to one another? I’m confused how any signal could be encoded between the two pins (without ground) but anyway I tried this in turn with each of the three USB UART adapters I have. Nothing worked. I ensured that local echo was off (using explicit settings in minicom). I got no response at all. It was a little more difficult with the FTDI adapter as the cables are permanently attached - there are no pins to which I could connect tx to rx (bit of scrap wire worked). The others were easy to do so.
So I’m dubious that all three UARTs attached to a USB interface are faulty. What can the lack of any output from this test mean?
I could not test with is the Banana Pi as I have no access to it. That is why I am trying to get the UART console to work. There is no working network connection I can find.
In terms of the network connection if I boot the Bpi R3 from a new sdcard install (all dip switches up) It seems to have booted at least to some extent as I can ping the address and get replies but ssh does not respond. A port scan of the address reveals no tcp ports are listening as far as I can tell. So how I can do a headless network install remains unknown.
(If I try and boot from the suspected install on eMMC then I get no response to ping, nor does any light except red remain steady on. Green flashes then goes out. SDCard boot red and green are steady).
Appreciate the assistance.
Hello again.
I have made some progress in terms of troubleshooting, but no fix yet. I discovered that a Raspberry Pi 4B that I have is equipped with a UART and is accessed via 3 pins on the GPIO. So I connected each of the three USB-TTL adapters I have to these pins and I can successfully obtain a connection via the serial command ‘screen -L /dev/ttyUSB0 115200’ in each case. I get login and all proceeds perfectly. That surely means the adapters are all working perfectly?
Then for each adapter I then moved the 3 pin connection to that of the Banaana PI R3 UART and in all cases I obtained no connection or response at all.
Is this sufficient to conclude the R3 has a now faulty UART and needs to be replaced? Is there any other explanation? Is there a hardware fix for the R3 if it is faulty in this way?
Thanks and regards.
It is strange this does not work. Did you try this with the screen command?
yes. I only used the 'screen -L /dev/ttyUSB0 115200’command in every case, Exactly as you described No response. I tried it with the three different UBS-TTY adapters I have.
With sudo before it?
Well no. It was from a root shell. So real and effective UID=0
