Crystal Oscillator

Boulder, CO (October 2024) - This experiment explored how crystal oscillators ring at a fixed frequency when excited by a voltage. To do this a ring oscillator was created from an SN74AHC14 hex inverter which creates a 50MHz signal. The crystal was then attached to the ring oscillator’s feedback loop to drive the oscillation to the crystal’s resonant frequency as shown in Figure 1. A circuit is shown in Figure 2. The experiment shows that crystal oscillators are effective at accurately maintaining their advertised frequency when used in conjunction with a large-value resistor and small capacitors.

The crystal is a slice of quartz that exhibits piezoelectric properties. This means that when the crystal deforms slightly under stress, it generates a voltage. The inverse effect is also true: when an electric field is applied, the crystal will deform and vibrate at a specific rate. This ability to reliably vibrate at a particular frequency is a useful property that can be used to keep time in a circuit. These crystals are manufactured to have a specific fundamental frequency (the first harmonic frequency) corresponding with the advertised frequency of the crystal. The crystal oscillates at this frequency, or possibly one of the higher-order harmonic frequencies, because the fundamental mode requires the least energy to sustain oscillations.

First, the output of the 5-stage ring oscillator was measured to confirm a frequency of approximately 50MHz (Figure 3). Next, a 1M resistor was attached in series with the feedback loop to determine the impact of additional impedance on the ring oscillator’s output frequency. The expectation is that the system will move to a lower energy state, which is confirmed by the reduced output frequency shown in Figure 4. The crystal will apply a high impedance at all non-resonant frequencies; the expectation with the crystal applied in series is that the output will oscillate at the lowest impedance state, which is the fundamental frequency of the crystal. 

The frequency response of the 12 and 16MHz crystals were analyzed before inserting the crystal into the feedback loop. Detailed response characteristics are available in XXXX which shows a large increase in the amplitude of the oscillations near the resonant frequency of both crystals. XXXX depicts the response for the 16MHz crystal, showing the increase in efficiency at the resonant frequency indicated by the spike in the amplitude.

Finally, the crystal was attached to the feedback loop of the ring oscillator as depicted in Figure 1 and Figure 2 using a 1M resistor and two 22pF capacitors. Because the impedance of the crystal is high at all but the resonant frequencies, when the circuit is first powered on, it can act like an open on the feedback line which can keep the crystal from oscillating. Adding a large value resistor in parallel with the crystal allows some current to flow and jump-start the process. During this experiment, the crystal was able to oscillate without the resistor, but that cannot be guaranteed every time the circuit powers up.

With only the crystal and resistor in the feedback loop, a higher-order mode was observed during one instance when applying power to the circuit. Although this only happened once, adding the capacitors to the circuit suppresses these higher-order modes. Adding both the large value resistor and the capacitors creates a more reliable and robust circuit that is guaranteed to oscillate at startup at the fundamental frequency. Table 1 shows the output of the ring oscillator using the 16MHz crystal.