Hi,
I just installed 0022 on my Ubuntu 10.10 machine. I have an Uno board and can't seem to get it work. Whenever I connect the board dmesg tells me that port /dev/ttyACMx where x is always different:0, 1, 2, 3... It shows up in the Tools menu but serial monitor reports that the port cannot be found.
I've been reading about similar issues on earlier version, I thought it has been resolved?
Any solutions? Can I brick the board by playing like that?
Thanks,
Aleksandar
I had previously installed 0020 from ubuntu package, but that didn't work either. Afterwards, I just downloaded the code from arduino.cc and copy those files over the old ones in /user/share/arduino. Didn't try to run arduino as root. Also didn't upload any code to Uno board. I just tried it on some windows machine, just to check the board, and it's working.
Do you think that the problem is access to port as normal user? Shoul I try to run it as root?
I've noticed that arduino did recognize the port when I run it for the first time, but when I try to open serial monitor, it disappears from the Tools menu.
My experience of the Uno on Linux is rather disheartening - if I'd not had success with the Duemillenove and the Mega I'd probably have given up by now. I'm running Ubuntu 10.04LTS kernel version 2.6.32.28: the latest kernel. Basically it sometimes works and sometimes doesn't. The only consistent behaviour is that you need to ensure long pauses between any serial output from your sketch otherwise you will have great problems communicating with it.
However that is far from a solution to the problem, and I've wasted countless hours trawling the forums, trying fixes, thinking I've found a solution only for it to fail again.
Symptoms as reported widely in the forums:
User interface runs very slowly indeed.
Tools-Serial Port greyed out or absent
Where you can select /dev/ttyACM0 an upload produces error messages (which vary).
This is constant across Arduino software versions 0021 and 0022 and across the current and last but one kernel.
In my opinion there is a problem with the Linux kernel support for the ttyACM0 device (ATmega8U2). I've managed to locate a source for some Duemillenoves and ordered a couple. The Uno is going in a drawer until it's properly supported.
Sorry to be negative, but find an Arduino using the FTDI device because they work perfectly every time.
You'll probably transform the serial by upgrading the 8U2 firmware. Upgrading it on mine made it indistinguishable from the FTDI interface on my Duemilanoves, it was next door to unusable before hand.
Thanks for that, although flashing the firmware does sound a little hairy. Like everyone else I had problems with serial output but the major problems I am experiencing are independent of the sketches I'm trying to run. The interface locks up or misbehaves and prevents me from uploading even the simplest demo sketch. This behaviour seems to appear and disappear entirely unpredictably.
Did you experience this behaviour and, if so, did the firmware flash fix it? If so I might be tempted to take the risk!
I thought the same having killed pieces of kit in the past through a firmware update that wasn't quite right.
If you program the arduino with a sketch that uses serial its difficult verging on impossible to get the board to accept new programming. Its like the serial takes over and the programmer can't get a word in edgeways. Its a while since I flashed mine, so I can't remember the errors etc it came up with, but I do remember it was a real pain in the neck. The reflashing procedure went without a hitch for me, and it completely transformed the board. I've since updated the bootloader on the Uno as well, and it cured different problems with a large sketch I run. But one thing at once. If you read through the thread I posted, the author of the 8U2 firmware assures readers that the flash procedure is safe and you can't do any damage if it fails. (post 38). You do need to be careful you download the actual code and not just a Javascript link which is quite easy to do but harmless if you do try to flash the chip with it.
Yes, the firmware update procedure is much safer than I originally thought. It isn't like updating a mobile phone or a BIOS, where if it goes wrong you are stuck. Instead, the 8u2 works in a similar way to the 328. If it doesn't upload right the first time, you can just try again until it does - the code for doing the firmware update is in a protected bootloader area of memory and the firmware update process can't touch it.
I must have reflashed the firmware on mine hundreds of times by now, trying out alternative USB functions. It still works
Thanks guys - I plucked up the courage to flash the firmware and it does seem to have fixed the problems. Too early to say for sure as some of the problems were highly intermittent, but it's looking good so far!