【正文】
y. Closed loop gain.Only for the crystal there are more then ten parameters. For a practical evaluation it is almost impossible to include all of them. A second reason for this is that many parameter values are not known to those engineersapplying the ponent. Does this mean that oscillation will be a matter of luck? Without going in much theoretical detail the impedance approach can be used to give an indication on the expected oscillation. The idea behind it is that the effective transconductance (1/ohm) should at least make up for the load (ohm)Figure 5. shows a characteristic that indicates a relation between the Rx,C0 and oscillation. When the coordinates of Rx and C0 are in the upper area, oscillation conditions are not met. When the coordinates cross at the characteristic or are in the lower area there will be proper oscillation.Now the difficulty arises in putting actual values on the Rx and C0 axis, as well as determining the right shapeand position of the characteristic. Practical experience however indicates that a circuit like in figure 3: with C1=C2 = 30 pF will be in the safe area, when C0 is about 10 pF and Rx is about 100 ohm or smaller.A lower value of Rx allows a higher value of C0 and visa versa.5. The drive level issue.Once we have oscillation is there something else to worry about?Most of the information so far is based on experience with the standard 5 Volt microcontrollers. Their oscillator stage was designed to drive a crystal with about 1 mW. This means that the crystal itself should be specified to perform proper resonation when it is driven at this power level. This crystal parameter is called the drive level and it is specified with the crystal parameters.6. Oscillator circuits in practice.All the 5V microcontrollers have a Pierce oscillator circuit (Xtal1,Xtal2) that will oscillate with an external Xtal and two capacitors. This standard circuit is presented in most data sheets. In some cases there may be good reasons to modify the standard circuit . to reduce interference and/or to pensate external influences. Reducing amplitudeReducing the oscillator amplitude is a frequently used approach for . reducing interference. The circuit (figure5) is based on the “Remendations to reduce the interference of 558 oscillator” (see References). This circuit generates an (almost) sine wave shape signal on the Xtal2 is achieved by using asymmetrical capacitors, and using the largest capacitor on the INPUT, (XTAL1). In this situation the voltage on XTAL1 is input voltage for the amplifier stage is lower so the stage will not go into saturation and therefore produce a signal with much less harmonics.A customer modified the circuit of figure 5. to the circuit of figure 6.Apparently the loop gain was changed by swapping the two capacitors and then pensated the by shunting aresistor to the Xtal. This circuit also works, but generates an output signal with more distortion. Increasing the output capacitor will result in more current through the output XTAL2. DCoffsetsThe main reason for modifying the circuit with a capacitor in series with the input (see circuit of figure 7) had to do with the influence of moisture on the oscillator circuit ponents. This could effect the basic circuit of figure due to the conductance of moisture a DCoffset is introduced to the XTAL1 input resulting in an asymmetric bias of the oscillator stage which may affect startup and/or event prevent the stage from oscillatin