At 02:38 PM 10/2/99 +0200, you wrote:
>I think somebody here complained before about build in caps that does
>not work.
>Never used resonators with that build in myself. Try use resonators with
>external 15pF caps.
>Or swap the resonator with one from the working boards.
>
>How far away is your resonator from the OSC. pins? Best to have it as
>close as possible with direct tracks to the pins.
>
>Normally when I get intermitted startups it, mostly it's noisy supply
>lines. put some cap filters in. And since 90% of the time my PICS is
>used in the industrial field, I always use external Power On Reset.
>Sometimes as long as a one second delay.
>
>Also, are your selection of oscillator type correct in the Config
>register?

Hi
Been following the thread, and one thing comes to mind.
Most of my problems have been with watch crystals in LP mode.
Power was very clean, caps were all the right values straight out of the
PIC data books, etc., and it still wouldn't kick into oscillation.
Also worked with OLD revisions of the '54s, but not the new die shrinks.
Solution cost less than two cents.  Add a CARBON composition resistor in
series with the oscillator as shown in the pic data book.
Experiment with the values, but I found 100K to be about right.
The problem is caused (in my case) by having too clean of a system--the
oscillator needs some noise to get it going, and the carbon composition is
inherently noisy enough to get things going.

Pick a value that keeps the P-P oscillations just within the power supply
rail voltages.

Let me know if this helps with the resonators too.

Kelly

William K. Borsum, P.E. -- OEM Dataloggers and Instrumentation Systems
<borsum@dascor.com> & <www.dascor.com>San Diego, California, USA