Re: CKOPT (was Re: [avr-chat] avrdude-ATMega32 fuse bits issues)

[Richard Urwin]

On Thursday 22 Dec 2005 00:33, Graham Davies wrote:
> Richard Urwin wrote:
> > I found the datasheets ambiguous when I built for the
> > ATMega16. It says CKOPT should be programmed
> > for fast resonators, but it does not mention crystals.
>
> Underneath the paragraph about resonators there is a table, Table 4.
> in my copy of the data sheet.  This table tells you how to set
> CKSEL3..1 for various frequency ranges.  It also gives a binary value
> for CKOPT.  From this and knowing that 1 is unprogrammed and 0 1s
> programmed you can figure out that CKOPT should be programmed for
> frequencies above 8MHz, just as Brian says.  I would have presented
> it differently, but the information is there.

Yes, if you read it as applying to crystals, the information is indeed 
there. And I certainly decided that it did apply to crystals. However 
the operative sentence is "For resonators, the maximum frequency is 8 
MHz with CKOPT unprogrammed and 16 MHz with CKOPT programmed." The only 
reference to crystals is the title of Table 4. and even then the last 
row - which applies for fast clocks - uses an obscure way to say "clock 
>= 1 MHz". It would have been a lot clearer if they had said 
"resonators and crystals" in the text and "1 to 16" in the table.

I'm only making the point because this has come up several times, and 
the usual response is "RTF datasheet". I think it is worth keeping in 
mind that this particular datum is rather inaccessible and easily 
missed.

-- 
Richard Urwin