Re: [avr-chat] XMEGA clocking questions

[Anthony Andriano]

I tend to agree with David. While it should work and I've personally never had a problem, the spec does allow for it to drift far enough to start causing issues. You could use a resonator instead of a crystal to improve the accuracy of your clock while saving cost and area. It's somewhat in the middle of the other two options.

On Thu, Mar 14, 2013 at 2:29 PM, David Kelly <address@hidden> wrote:

On Mar 14, 2013, at 12:04 PM, Rick Mann <address@hidden> wrote:

> I see that the 32.768 kHz crystal on the dev boards could be used to automatically calibrate the 32MHz internal clock, although I'm not sure if they're actually doing that.
>
> I'm laying out a board now, need to know if I should include the crystal or not.

Will it kill you to have to make room for the crystal? You don't have to load it if you don't need it. Plus I find Murphy's Law has a heavy hand in such things, you can make space for a crystal which you will not need, or don't and it will almost be certain the onboard RC oscillator is not accurate enough.

I brought my first Xmega up on an XPlain board. Wanted a 2.500 MHz SPI clock so I programmed the oscillator chains for 30.0 MHz as 15 times 2.000 MHz supposedly corrected using 32 kHz crystal reference. The result was closer to 31 MHz but close enough to 30 that other things were more important to make work before doting i's and crossing t's. Since, Atmel has updated their libraries. I need to drill through the changes to see if the clock error was theirs or mine.

Before my PCB is finished I may simply drop a manufactured oscillator on the board. Brute force. Hit it over the head with a sledgehammer. Its not worth the extra NRE.

--
David Kelly N4HHE, address@hidden
============================================================
Whom computers would destroy, they must first drive mad.

_______________________________________________
AVR-chat mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/avr-chat