At 10:48 AM 2/26/00 -0800, you wrote:
>Don
>Microchip has an extensive ap-note on dealing with crystals.  Whether or
>not the resistor is required in HF has to do with the cut of the crystal.
>In most cases it is not required in HF mode. Also the XT and HF modes use a
>higher power on the oscillator.  I've never needed one on any of the
>designs we've done so far.  However, the pads are always there on the board
>"just in case."
>

Kelly,

I am familiar with the low freq 32 Khz watch xtals needing Rs, etc.
And I looked at Mchp appnote AN588. They say things like "if you are
experiencing overdrive ....", but don't seem to say whether there is
a really problem with "high freq" xtals, per se.

And it is *because* the HS mode oscillators use "higher power on the
oscillator", as you say, that I was worried about xtal overdrive. The
Mchp 16C7x datasheet says to the effect that some manuf xtals may
require use of Rseries. Has anyone found hi-speed xtals (eg, 20 Mhz
range) where this was a problem?
---------

>
>However, I have had experience with watch crystals, and LP mode.  The
>crystal needs a goose to get started, and if your board and environment is
>very clean and quiet, sometimes there is not enough ambient noise to kick
>it into oscillation.  In which case a 100K Carbon Composition (note--carbon
>comp--not metal film) will introduce enough noise to just about guarantee
>start up.
>

Regarding this issue, I came to the conclusion that a large resistor
(1M) *across* the oscillator terminals is the way to go. Since the osc
is usually an inverter, this resistor sends an out-of-phase signal from
the output back to the input.

- Dan Michaels
Oricom Technologies
http://www.sni.net/~oricom
==========================