Xtal oscillators, why the capacitors??

[Steve]

Given a microprocesser that has a osc in and out, connected to
this:


----------------------------osc in
| |
--- |
22pf--- ----
| xtal
GND ----
| |
--- |
22pf--- |
| |
----------------------------osc out


Why the capacitors?? would it work to simply connect the xtal across the IC pins??

 

[Michael Black]

Steve (steven.fox@nuigalway.ie) writes:

Given a microprocesser that has a osc in and out, connected to
this:


----------------------------osc in
| |
--- |
22pf--- ----
| xtal
GND ----
| |
--- |
22pf--- |
| |
----------------------------osc out


Why the capacitors?? would it work to simply connect the xtal across the IC pins??

Because the oscillator is of such a type that it needs capacitors. A
different scheme would get away without capacitors.

But once the oscillator inside the IC is chosen to be a type that needs
capacitors, then you don't have much option. Capacitors, at least
of the value needed, are difficult or impossible to put on the IC.
So they have to be external.

Michael

 

[Terry King]

But once the oscillator inside the IC is chosen to be a type that needs
capacitors, then you don't have much option.

Additionally, crystals are designed for a specific capacitance and will
only be accurately on frequency with that load. You'll see some circuits
where one of the two capacitors in the circuit you showed is a variable
capacitor used to precisely set the frequency where this is needed
(mainly in communications equipment, not microprocessor clocks).

--
Regards, Terry King ...In The Woods In Vermont
terry@fredking.us
The one who Dies With The Most Parts LOSES!! What do you need?

 

[Joe McElvenney]

Hi,

Remembering that the crystal and the two capacitors need
to provide a phase shift of 180 degrees, another way to
look at things is this.

The capacitor at OSC_OUT is in series with the gate
output resistance, and in some oscillators an additional
crystal current-limiting resistor, giving a small part of
the phase shift. At the oscillating frequency the crystal
has an inductive reactance which together with the second
capacitor provides the rest (in fact most) of the phase
shift. This is because, as the crystal has a very high 'Q',
small changes in frequency will produce a large changes in
its series inductive reactance. However, the crystal and
second capacitor are not enough on their own and so the
need for the first capacitor.


Cheers - Joe

 

[ddwyer]

In article <VA.00000085.0049ac5c@btinternet.com>, Joe McElvenney
<ximac@btinternet.com> writes

Hi,

Remembering that the crystal and the two capacitors need
to provide a phase shift of 180 degrees, another way to
look at things is this.

The capacitor at OSC_OUT is in series with the gate
output resistance, and in some oscillators an additional
crystal current-limiting resistor, giving a small part of
the phase shift. At the oscillating frequency the crystal
has an inductive reactance which together with the second
capacitor provides the rest (in fact most) of the phase
shift. This is because, as the crystal has a very high 'Q',
small changes in frequency will produce a large changes in
its series inductive reactance. However, the crystal and
second capacitor are not enough on their own and so the
need for the first capacitor.


Cheers - Joe


Agree with the above.

Remember the total phase shift through the crystal and caps/IC amplifier
must be zero.
The crystal will adjust its frequency to satisfy the zero phase
condition.
The amplifier will have slightly in excess of 180 degrees and each
capacitor in conjunction of the source resistance will add approx < 60
degrees .
Less well known is the fact that there is a frequency of maximum
activity/negative resistance for any source resistance/capacitor
combination and there will be a locus of optimum cap/resistance for
32kHz.
The smaller the crystal the smaller the maximum allowable value of
current through the crystal before frequency shift and ultimate
disaster. The series r at the output of the gate is the means of
limiting current.
Many IC oscillators are non optimum for stability and reliability and do
not allow enough negative resistance to allow for increase in crystal
resistance with time i.e. they fail to start a year or so after
manufacture.

..

--
ddwyer