> Jinx, I've never had the standard "Pye" type circuit NOT work,
> as long as normal cut Xtals were used.

> I would think that it is more a matter of avoiding weird Xtals.

> What problems do you notice?

> ..................... Zim

The immediate need is for a 3.6864MHz clock source that
can drive upwards of 6 PICs for an RS232 system. I'm not
aware of a resonator < 4MHz that could be used for 0%
comms unfortunately. There's a 7.37MHz though (2 x 3.685,
= 3.6864 x 0.9996)

(One option, just to keep things simple, if defeatist, would be
to run one PIC at 14.7456MHz and use Clkout = osc/4 for
the others)

The crystal is a Fox0368S. Maybe the S is for series ? I've
got Fox 10MHz crystals here that I use with the PICs and
they aren't marked with a P (unlike the back of our couch,
damn cat) so perhaps the S is meaningless in this context

In front of me now I have a Pierce circuit. A 3.6864 crystal in
parallel with an LS04 gate, 2 x 22p to ground, 470R across
the crystal (originally a 330k pot), like this one

http://vlsi.wpi.edu/webcourse/ch05/ch05.html

oscillating at 0.4V with a very nice 48MHz sine wave.

[update while off-line]

However, after re-reading the text with the text the link below
I've changed that to an HC and the o/p is now 11.092MHz (3rd
harmonic) at 4.5V, so that's an improvement. By changing the
feedback resistor to 4k7 in series with a 100k pot I can tune
the circuit to o/p the required 3.6864MHz. Grounding the crystal
case helps a lot. But it doesn't take much for the waveform
and frequency to go nutso, and it doesn't start reliably

Now, this one here (bottom of page in black), uses a series
resistor as a filter

http://www.gaby.de/z80/uexosc.htm

but I don't know what calculation he's using to work out the
value of the resistor. Some 1/ sqr rt pi rfc I guess

The others on that page look like the examples in the PIC
manual. The parallel one is what I tried first. But as you read
in my original request, you don't have to twiddle with pots on
a PIC. The best so far is the series circuit. With 470R feedback
resistors, no caps and an HC00, it will start reliably with either
a 3.6864MHz or a 9.8304MHz crystal, at good voltage and
wave quality

What I don't understand is that these are parallel cut crystals,
the type I use all the time with micros (which require parallel
cut), yet the series circuit seems to work so much better

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