I had a sketch running on an UNO with a couple nested for loops. the loops would change outputs on i2c DAC's and then read values from an i2c ADC. the values read from the ADC were then serial printed to the serial monitor.
I had USB and 12v going into the power jack. I tapped power out from the 5v pin and from the Vin pin for a 12v supply. besides that i connected gnd and the SDA SCL lines to my breadboard. no other connections to the UNO were made.
these loops were pretty big, so after about 20,000 lines were printed to the serial monitor, it suddenly stopped. the sketch was still running because the output voltages on the DAC's were changing as expected. after cycling the power, unplugging and replugging USB, rebooting the computer, the UNO wont communicate with windows anymore. I plug it in and the red power led comes on. if I hit reset, the 'L' red led blinks 3 times and then no change. the sketch continues to run, but I cant communicate with the board. Windows does not recognize any device being plugged in at all.
So, I suspected bad bootloader or 8u2. When trying to reburn the bootloader using the MKII connected by usb, and the UNO powered with external 12V supply, and the arduino IDE, the MKII lights up green as it should when the UNO is powered, but the UNO board never lights green. when I hit burn bootloader in IDE i get the error:
avrdude: usbdev_open(): did not find any USB device "usb"
Any suggestions on how to get around this? it seems the main chip works since the sketch seems to be still running, and I thought the MKII would bypass the 8u2. So I don't know what gives. And would it have died just because I serial.print too many lines too fast?
Check the MKii driver. It's provided with the IDE, look for the folder with this file: avrisp2.inf
You don't want the Jungo driver that is supplied with Atmel Studio.
Ok, there's a seemingly pretty good explanation for getting the drivers to work on windows x64 here, but I couldn't get that to work for me.
I shifted to a windows 32 bit machine and got the programmer working, bootloader burned, and uploaded 'Blink' via the programmer. All seems to be working fine, but there's still no USB connection on either machine.
Does this sound like the 8u2 is fried? can this happen just from sending too many serial.println??
To get it work on a 64 bits windows machine you may need the libusb filter, at least that was necessary to me to install USBTinyISP.
And about serialprint, I doubt that's the reason to burn the atmega8U2, maybe it's an opportunity to solder a newer atmega16U2.
archy587:
..and 12v going into the power jack.
I cannot enumerate the number of times I've read people having Arduino issues after powering it with 12V (I burned myself one too, but it was worse because the atmega328p was burnt too). That's why I started to use 9V at max, if I need 12V I use external LM7805.
I cannot enumerate the number of times I've read people having Arduino issues after powering it with 12V
Wow, really?!? that's totally disappointing considering its the recommended input!! wtf. the high end of recommended voltage sure...but still not even close to the 20V max they specify... : /
Lots of DC adaptors claim to be 12V, but under load. Without big load most of them have 14-17 V.
That's why you must use adaptors with switching regulation.
