Re: [avr-chat] Crystals (was ATmega32 @ 16MHz : fuse bits ??)

[Vincent Trouilliez]

On Tue, 2005-11-15 at 14:25 +1000, Neil Davey wrote:
> It is possible for the strip tracks to have cracks in them, either from 
> manufacture, fatigue or stress, so it might not hurt to run
> a line of solder over the unused tracks between the micro and the crystal...
> 
> Also, I know it's a simple thing.. but we have all done it... are you 
> *sure* your caps are attached to ground.... :)

Yeah... I resoldered everything, putting solder all over the tracks.
They are very short, as I put the AVR chip on the edge of the
prototyping board, leaving just enough space to squeeze teh crystal and
capacitors. I replaced the crystal by another one (same part number
though). I tried with CKOPT enabled and disabled, still won't start...
ohhhhhhhhhhhhhh... this is getting crazy. 

About the fuse bits that have a reversed logic, I am not sure why the
data sheet keeps insisting that 0 means programmed and 1 unprogrammed.
It's starting to confuse me more than anything. No need to tell the user
about this reversed logic is it ?  I am really starting to find it
confusing: "when they write "1" in a truth table, does it really mean
"1", or does it mean "enabled"/programmed, hence I must actually write a
0 not a 1 ?
Brrr...
Why don't they just give tables and let the user program this into the
fuse bytes, completely omitting any and all comment about this reversed
logic.

They say to program CKSEL3:0 to 1111 for a high speed crystal, so I set
all these bits to one in the lower fuse byte, not to 0.

Looks kinda hopeless now. I am getting bored with it :-/
Since the on-board RC oscillator at 8MHz is enough to keep me going, I
guess I should better move on and resume the development of the
software. Once finished, I will get a proper PCB manufactured, and try
one last time to get the crystal going...

--
Vince, completely clueless.