Appendix A: CGM Practical Aspects
68HC(9)12DG128 Rev 1.0
380
Appendix A: CGM Practical Aspects
MOTOROLA
Theoretically speaking, nothing precludes the use of non-identical
values for C1 and C2. As this complicate a bit the management of the
final board device list, this is not recommended. However, if
asymmetrical capacitors are chosen, the value of C1 should be higher
than C2 (because the reflected loading is proportional to the square of
the impedance of C2).
What Is the Final
Oscillation
Frequency
The exact calculation is not straightforward as it takes into account the
resonator characteristics and the loading capacitors values as well as
internal design parameters which can vary with Process Voltage
Temperature (PVT) conditions. Nevertheless, if L is the series
inductance, R is the series resistance, C is the series capacitance and
Cc the parallel capacitance of the crystal, we can then use the following
simplified equation:
How Do I Control
The Peak to Peak
Oscillation
Amplitude
The CGM oscillator is equipped with an Amplitude Limitation Control
loop which integrates the peak to peak ‘extal’ amplitude and in return
reduces the steady current of the transconductor device until a stable
quiescent point is reached. Controlling this final peak to peak amplitude
can be performed by three means:
1.
Reducing the values of C1 and C2. This decreases the loading so
that the necessary gm value required to sustain oscillation can be
smaller. The consequently smaller current will be reached with a
larger ‘extal’ swing.
Using VDDPLL=VSS (i.e. shutting off the PLL). Doing so
increases the starting current by approximately 50%. All other
parameters staying the same, a larger ‘extal’ swing will be
required to reduce this starting current to its quiescent value.
Also, placing a high value resistor (>1M
) across the EXTAL and
2.
3.
Fosc
2
π
C1=C2=Cl yields to
--1
LC
-1
L
Cc
C1
C2
||
+
)
--------------------1
+
=
Fosc
2
π
--1
LC
-1
+
L
Cc
Cl 2
(
)
-----------------1
+
=
2-cgmpa