I guess I've been lucky, because I don't remember this sort of thing happening to me before, but I somehow managed to clobber the fuse bytes while programming an ATmega1284P in the TQFP package.
I was able to set it straight using Jeff Keyzer's MightyOhm HV Rescue Shield. The wiring was pretty sketchy but it worked! I did remove the 10K pullup and 1N4148 diode on the reset pin, and I had to solder a wire to the board to access the 1284P XTAL1 pin.
So you couldn't just reburn the fuses using a regular AVR SPI interface programmer?
Was it one of these that got you?
JTAGEN, 6, Enable JTAG, default = 0 (programmed, JTAG enabled)
SPIEN, 5, Enable Serial Program and Data Downloading, default = 0 (programmed, SPI prog. enabled)
CrossRoads:
So you couldn't just reburn the fuses using a regular AVR SPI interface programmer?Was it one of these that got you?
JTAGEN, 6, Enable JTAG, default = 0 (programmed, JTAG enabled)
SPIEN, 5, Enable Serial Program and Data Downloading, default = 0 (programmed, SPI prog. enabled)
Tried that first and couldn't get a response with AVRISP MkII and Atmel Studio even at very low clock speeds. JTAGEN and SPIEN were OK. While all the fuse bytes were changed (!), the high and extended fuses were only slightly different. What got me was the oscillator config in the low fuse byte, it was set to low freq osc (i.e. 32kHz crystal)! The chip didn't have a valid clock, so no SPI programming.
