[avrdude-dev] Differences between programmers?

[Rick Mann]

I had written this email to send to the list. The behavior was experienced with 
an AVR Dragon. Then I tried with an AVR ISP MkII, and it worked correctly.
----------------------------------
I built a new board around an ATmega1281. I soldered a brand-new IC onto it, 
connected power and ISP, and I'm getting a device signature of 0 with avrdude 
and Dragon.

I tried it on an OSEPP ATmega2560, and it works correctly.

Looking at the PDO waveform of my '1281 on the oscope, it seems to visually 
match the timing of the OSEPP '2560.

I have a Beagle SPI protocol analyzer, but I'm not sure it can properly capture 
the transaction (it wants proper SPI modes 0-3, and I can't quite tell from the 
data sheet what mode programming is). It does show this for MOSI:

AC 53 00 00 AC 53 00 00 30 00 00 00 30 00 01 00
30 00 02 00 30 00 00 00 30 00 01 00 30 00 02 00
30 00 00 00 30 00 01 00 30 00 02 00 00

Which looks like an attempt to read bytes 0, 1, and 2 of the signature, three 
times.

It shows this for MISO:

FF FF 53 00 00 AC 53 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00

Which doesn't look great at all. It does look like the start of a proper 
response, but all the results are 00?

Is it just a bad IC?


-- 
Rick Mann
address@hidden