I'm not familiar to the oscillator you wrote about but if you use a ordin=
ary
crystal you can go on reading.
The advantages with a crystal is that its less costly and that you get a
sinus output. A caned oscillator generates square wave and that can gener=
ate
frequencies not desirable in the design. (A sharp edge contains over stor=
ed
amount of sinus frequencies)
You cant toss away the caps, they are needed. You can experiment with the=
ir
values to get the oscilator design to fit your microcontroller. Chose
capacitors from what the crystal manufacturer suggest and then test it in
your system with the PICmicro. Test the design in absolute maximum and
minimum conditions. Max and min Vdd. Test the Vdd min by toggle an I/O fa=
r
away from de oscillator and measure with a oscilloscope (crystal oscillat=
or
is very sensitive and can be affected of your probe or even by dirt on th=
e
PCB). It should generate a even square signal. If it is unreliable you ha=
ve
to play with the C1 and C2. Do the same test at Vdd max but then you need=
 to
measure on the oscillator. The signal should be a nice and smooth sinus
signal. If its cut in the top it can generate unwanted frequencies and
during long time use even damage the crystal. If the amplitude is to big,
try to lower it with a serial resistor or change  c1 and c2. After this i=
s
done go back to test the Vdd min as it might have generated to low amplit=
ude
in that range of the voltage. If you can combine this two extreme in your
design you are quite sure you have done a good oscillator design.
If you search Microchips application notes you find a good oscillator des=
ign
guide that if I remember right is called AN588.

Niklas Wennerstrand

-----Ursprungligt meddelande-----
Fr=E5n: pic microcontroller discussion list [mailto:PICLIST@MITVMA.MIT.ED=
U]
F=F6r Nick Veys
Skickat:        den 11 juni 2001 03:04
Till:   PICLIST@MITVMA.MIT.EDU
=C4mne: [PIC]: Downsides to these oscillators?

Hey, I've been looking for a nice, simple oscillator to use in my project=
s.
My last one, a 16F84 project used a 4MHz full can style oscillator, didn'=
t
use any caps on it, and it was a fairly time-critical project and it work=
ed
just fine.

Now I'm on a 16F870, using a 20MHz CTS Reeves MXO45 1/2-can oscillator.
Again w/o caps, testing my busy waits shows it to be very accurate (not a
perfect test, but still).  I'm wondering what the downside to this is.
Anyone care to shed some light on the adv/disadv of doing this?

I'm assuming the caps are for voltage stability?  So in my final build I'=
ll
probably toss them in.  But how necessary are they really?

nick@veys.com | www.veys.com/nick

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads