Measuring crystal frequency requires either a frequency counter or an oscilloscope, with the counter being the more accurate and straightforward option. The method you choose depends on how precise you need to be: a basic frequency counter resolves down to about 1 to 5 parts per million (ppm), while careful technique and better equipment can push accuracy into the parts-per-billion range.
Getting an accurate reading isn’t just about owning the right instrument, though. Crystals are sensitive to how you probe them, what circuit they sit in, and even the room temperature. Here’s how to do it right.
Using a Frequency Counter
A frequency counter is the most direct way to measure a crystal’s output frequency. You connect the crystal’s oscillator output to the counter’s input, and it displays the frequency. Base-model benchtop counters use an internal reference oscillator with 1 to 5 ppm stability, which is sufficient for most general work. For a 10 MHz crystal, that means your reading is accurate to within roughly 10 to 50 Hz.
The gate time setting on your counter matters more than most people realize. Gate time is how long the counter samples the signal before giving you a reading. A longer gate time means finer resolution. At a 100 ms gate time measuring a 20 MHz signal, you get about 5 parts per billion (ppb) resolution, or 0.1 Hz. Drop the gate time to 10 ms and your resolution gets ten times worse. For precision work, use the longest gate time you can tolerate. Modern counters also achieve timing resolution of 20 picoseconds or better, which helps when you’re comparing two close frequencies.
One thing to keep in mind: your counter is only as accurate as its internal reference. If you need better than 1 ppm accuracy, you’ll want a counter with an oven-controlled reference oscillator, or an external reference locked to a GPS-disciplined standard.
Using an Oscilloscope
An oscilloscope lets you see the crystal’s waveform directly and read the frequency from the period of the signal. This is useful for checking whether the crystal is oscillating cleanly, spotting startup problems, or looking at waveform shape, but it’s generally less precise than a counter for pure frequency measurement.
The bigger concern with oscilloscopes is probe loading. Every probe has input capacitance and resistance, and connecting it to a crystal oscillator circuit adds those parasitics to the circuit. This can shift the frequency you’re trying to measure.
Why Probe Loading Shifts Your Reading
When you touch a probe to a crystal circuit, the probe draws current. Its input capacitance and resistance act like extra components added in parallel with the circuit. At lower frequencies, a probe’s capacitance has very high reactance and doesn’t disturb much. But as frequency increases, that capacitive reactance drops and the loading effect grows significantly.
Most passive 10X probes that ship with oscilloscopes have 10 megaohms of input resistance, which is usually fine for resistive loading. The overlooked specification is input capacitance, typically 8 to 15 picofarads for a standard passive probe. For a crystal oscillating at tens of megahertz, that extra capacitance can pull the frequency noticeably off its true value.
Active probes solve this problem. They use transistors or amplifiers at the probe tip to present much lower input capacitance (often under 1 picofarad) and higher impedance to the circuit. If you’re measuring crystal frequencies above a few megahertz and need accurate readings, an active probe is worth the investment. At minimum, use a 10X passive probe rather than a 1X, since the 10X setting reduces the effective capacitance seen by the circuit.
Series vs. Parallel Resonance
Every quartz crystal has two resonant frequencies, and knowing which one you’re measuring is essential. These are called series resonance and parallel resonance (sometimes called antiresonance).
At series resonance, the crystal’s impedance drops to its minimum and current flow through it is greatest. At parallel resonance, impedance reaches its maximum and current flow is at its lowest. Both frequencies produce a zero-phase condition where the crystal behaves as a pure resistance, but at very different impedance levels.
The two frequencies are close together but not identical. Which one matters depends on your circuit. Crystals specified for series resonance are designed to oscillate at the frequency where impedance is minimal. Crystals specified for parallel resonance (also called “load capacitance” operation) oscillate at a slightly higher frequency determined by the external capacitors in the circuit. The crystal’s datasheet will tell you which mode it’s designed for and, if it’s a parallel-resonance crystal, what load capacitance it expects.
How Load Capacitance Affects Frequency
If your crystal is designed for parallel resonance, the actual oscillating frequency depends on the total capacitance the circuit presents to the crystal. This is called the load capacitance. Changing it shifts the frequency.
The relationship is predictable. The frequency at a given load capacitance relates to the series resonant frequency through the crystal’s internal capacitance values. In practical terms, increasing the load capacitance pulls the frequency lower (closer to series resonance), and decreasing it lets the frequency rise. This is why adding probe capacitance to the circuit can shift your measurement: you’re effectively changing the load capacitance.
When you change the load capacitance from one value to another, the resulting frequency shift depends on both capacitance values and the crystal’s own electrical characteristics. Crystal datasheets specify the intended load capacitance, commonly 12 pF, 16 pF, or 20 pF. If your circuit’s actual load capacitance doesn’t match the spec, the crystal won’t oscillate at its rated frequency. A mismatch of just a few picofarads can shift the frequency by tens of ppm.
Temperature and Environmental Effects
Temperature is the single largest environmental factor affecting crystal frequency. The relationship between frequency and temperature is nonlinear, meaning it doesn’t shift at a constant rate as the temperature changes. The shape of the curve depends on the crystal’s cut.
AT-cut crystals, the most common type, have an S-shaped frequency-temperature curve with a relatively flat region near room temperature. Within that flat region, frequency stays reasonably stable, but outside it the drift accelerates. SC-cut crystals are more advanced: their frequency stays within plus or minus 1 ppm over a 25°C range around their turnover point, and their dynamic temperature sensitivity is roughly 100 times smaller than AT-cut crystals.
A complication called “activity dips” can occur at specific narrow temperature ranges where an unwanted vibration mode interferes with the desired one. When this happens, the temperature coefficient can jump by an order of magnitude and even reverse direction. These dips can appear over temperature ranges as small as a few millidegrees. In practice, this means that if your readings seem erratic at a particular temperature, moving the crystal a degree or two warmer or cooler may help.
Humidity also plays a role. The temperature coefficient of a crystal can change depending on humidity levels, making the two effects interdependent rather than separate. For the most stable measurements, control both temperature and humidity, or at minimum let your crystal reach thermal equilibrium in a stable environment before taking readings.
Practical Measurement Tips
For a quick check of whether a crystal is working and roughly on frequency, a frequency counter connected to the output of a simple oscillator circuit is all you need. Power up the circuit, let it stabilize for a few minutes, and read the counter. Compare the result to the crystal’s rated frequency. Most consumer-grade crystals have a tolerance of plus or minus 20 to 50 ppm at 25°C. A 10 MHz crystal rated at 50 ppm tolerance could legitimately read anywhere from 9.9995 MHz to 10.0005 MHz and still be within spec.
For more precise work, pay attention to these factors:
- Warm-up time: Let both your measurement equipment and the crystal oscillator circuit run for at least 10 to 15 minutes before recording data. Crystals and counter references both drift as they warm up.
- Probe choice: Use an active probe or a high-impedance buffer if you’re probing directly at the crystal. If you’re measuring at a buffered output, a standard 10X passive probe is usually acceptable.
- Load capacitance matching: Make sure your oscillator circuit provides the load capacitance the crystal expects. Mismatched capacitance shifts the frequency and makes your measurement meaningless for judging whether the crystal is in spec.
- Counter gate time: Use the longest gate time practical for your measurement. A one-second gate time gives you 1 Hz resolution directly.
- Stable environment: Avoid drafts, direct sunlight on the circuit, or touching the crystal with your fingers during measurement. Body heat transfers quickly to a small quartz package.
If you’re evaluating aging, which describes how a crystal’s frequency drifts over months and years, a high-quality oscillator ages at roughly 1 ppm per year. Measuring aging requires repeated readings over weeks or months against a stable reference, not a single snapshot.