After spending some more time, using also linux and 32-bit Win-Systems for the test, I came to the conclusion, that it must be hardware relevant.
The facts:
- The SAM3X8E has the SAM-BA bootloader in its ROM, so it can't be an issue with other Arduinos or the digispark, where the bootloader ist just damaged
- The SAM3X8E has native hardware USB support, which is used in that case, so it can't be one of the problems, with Zeners, pull-ups and/or software emulation of the USB protocol
- There is no difference connecting the digiX with an charger-cable or an true USB-cable. There are no signs of connecting a broken USB device on any operation system.
Comparing the digiX with the Due, there is one difference:
The DUE has two USB-ports. The second port is connected to an Atmega16U2, which works similar to an FT232 as Serial over USB and connects to the UART of the SAM3X8E.
The UART on the digiX is however shared with the Wifi module, but there is the Wifi-Disable jumper.
OK, that means: When I disable Wifi and connect an FT232H cable (
http://digistump.com/board/index.php/topic,1127.msg4762.html#msg4762) to Pin 0 and Pin1 (labeled also RX and TX) I have the second USB port also for the digiX.
Reading the documentation of the SAM3X (
http://www.atmel.com/Images/doc11057.pdf) you can see in chapter 22.2, that the chip tries to connect the UART, after USB enumeration was not successful.
If it can read an character from UART at 9600 8N1 it starts the SAM-BA monitor on UART instead of USB.
So I connected the digiX with my FT232H cable and opened first a terminal program with 9600 8N1 on the corresponding port. I just pressed Enter in the terminal, while I pressed reset on the digiX, which was at this point externally powered and had no USB cable in.
After that, I closed the external terminal and went into the Arduino IDE with the blink-sketch. I selected "Arduino Due (Programming port)" and the corresponding port of the FT232H and tried to upload the sketch!
And: Voila, it worked, the digiX is blinking again!
After that I researched the USB plug much more carefully (with lots of light and magnifying glasses) and detected, that the soldering didn't look so good there. After reflowing the connections with the finest tip of my soldering iron, USB is now working as well!