Is It Too Difficult to Choose a Crystal Oscillator? Master These 5 Core Indicators to Easily Handle Frequency Tolerance/Temperature Drift/Aging/Phase Noise/Jitter

Table of Content [Hide]

    On every circuit board, that rice sized crystal oscillator forever hides the heartbeat of the entire electronic system. From fast charging of mobile phones to 5G base station signal transmission, from Beidou navigation to industrial PLC control, all places that require precise timing rely on this small quartz crystal to set the rhythm. Once the crystal oscillator becomes unstable, it can result in communication lag and packet loss, or even data disorder and system shutdown. Even if a precision measuring device makes a mistake, the entire experimental result may have to be overturned and restarted.


    Many engineers choose crystal oscillators based on their nominal frequency and price, flipping through data manuals filled with incomprehensible English parameters and unable to grasp the core. In fact, to determine whether a crystal oscillator is stable enough, it is enough to grasp the five core indicators without being overwhelmed by unnecessary parameters.


    Is It Too Difficult to Choose a Crystal Oscillator? Master These 5 Core Indicators to Easily Handle Frequency Tolerance/Temperature Drift/Aging/Phase Noise/Jitter


    01. Frequency tolerance depends on the factory's accuracy


    Frequency tolerance, also known as initial frequency accuracy, refers to the deviation between the actual output frequency and the nominal value under standard conditions (25 degrees, rated voltage, matched load) when the crystal oscillator is just manufactured. The unit is expressed in ppm (parts per million). For example, if a crystal oscillator with a nominal frequency of 10MHz is labeled with ± 10ppm, it is allowed to operate at 10MHz. The corresponding maximum deviation is (10 * 10 ^ 6) X (10 * 10 ^ -6)=100Hz.


    The smaller the value, the higher the initial accuracy. Many people think that initial accuracy is not important? In fact, for smart meters and communication modules that require networking upon startup, the initial frequency is slightly different, and the frequency matching is slow at the beginning, and the base station cannot directly search for signals. The first time you turn on the power and don't connect to the internet, often it's because the initial deviation of the crystal oscillator is too large. When choosing, depending on the scenario, ordinary consumer electronics can be relaxed to ± 20ppm, but for industrial communication, it needs to be reduced to within ± 10ppm to be safe.


    02. Temperature stability, whether it is stable or unstable due to hot and cold fluctuations


    The biggest impact on the stability of crystal oscillators is not time, but temperature. The frequency of quartz crystal itself follows a cubic curve with temperature, with the inflection point around 25 ℃ being the most stable and drifting as the temperature deviates. Frequency temperature stability refers to the maximum drift of the frequency relative to 25 ℃ throughout the entire operating temperature range of the crystal oscillator.


    The temperature stability of different types of crystal oscillators varies greatly: ordinary standard crystal oscillators have a temperature drift of ± 10 to ± 30ppm, which is only suitable for household appliances and ordinary industrial controls placed in air-conditioned rooms at home, and the temperature difference is not large enough. If it is placed in outdoor base stations or car navigation systems, with temperatures ranging from minus tens of degrees in winter to over 100 degrees in summer, a temperature compensated crystal oscillator TCXO must be selected. The built-in compensation circuit flattens the temperature drift, which can be compressed to ± 0.1 to ± 2.5ppm, fully meeting the product requirements. If you want to do things like Beidou timing and 5G base stations that require extreme accuracy, you must use a constant temperature crystal oscillator OCXO. Simply place the crystal in a small tank at a constant temperature, and the temperature drift can reach several hundred ppb (1ppm equals 1000ppb, which can be almost ignored)


    03. Aging rate, it depends on whether it floats after prolonged use


    Do you think only people get old? Crystal oscillators can also 'age' over time. After long-term electrical operation, the internal stress of quartz crystals will gradually release, and the material will slowly change. These changes are cumulative and irreversible, causing the frequency to drift slowly, and this drift rate is the aging rate.


    Generally, a better quality crystal oscillator ages the fastest in the first year, with a drift of approximately ± 1 to ± 3ppm per year. After the first year, the aging rate slows down, and a constant temperature crystal oscillator can achieve several ppb per year. This indicator has little impact on the drift of ordinary consumer electronics after three to five years of use, but communication base stations, unmanned industrial gateways in the field, and smart meters hung on power poles have been unattended for more than ten years. If they age too quickly and the frequency drifts out of the range after three to five years, the entire equipment will be useless. For devices that operate for a long period of time, it is necessary to choose low aging crystal oscillators to be reliable.


    04. Phase noise, whether the signal is clean or not


    The first three are all about frequency accuracy. In the fields of high-speed communication and precision measurement, optical accuracy is not enough and needs to be clean. Phase noise refers to the random disturbances around the output frequency in the frequency domain, measured in dBc/Hz. The smaller the value, the purer the signal and the less clutter disturbance there is. For example, when you listen to a radio, low phase noise is like a clear signal without any noise, while high phase noise is like the entire screen is filled with noise, making it impossible to hear clearly.


    In scenarios such as 5G communication, phased array radar, and high-precision ADC acquisition that require extremely high signal-to-noise ratios, if the phase noise is a little higher, the signal-to-noise ratio will decrease, and the sensitivity of the entire system will be poor. Not to mention inaccurate measurements, even the communication distance will be compromised. So in this scenario, it is necessary to choose a crystal oscillator with low phase noise. If it is a few dB worse, the performance will be one level worse.


    05. Jitter, check if the clock is accurate


    Jitter and phase noise are actually the same thing. From two perspectives, phase noise is a description in the frequency domain, while jitter is a description in the time domain. What does that mean? The clock signal is a square wave, and each rising edge should be on time, but in reality there is always a deviation, and the magnitude of this deviation is jitter. The unit is usually measured in ps (picoseconds) to calculate the RMS value, and the smaller the value, the smaller the deviation.


    The current high-speed serial interfaces, such as PCIe 5.0 and SerDes, have a speed of over ten Gbps. If the jitter is a little larger, the eye diagram will close and the bit error rate will directly increase. If one bit is transmitted incorrectly, the entire data packet will be invalidated. So when choosing a crystal oscillator for high-speed interfaces, it is necessary to choose a low jitter differential output crystal oscillator, such as LVDS or LVPECL output. If there is a difference of a few ps, the bit error rate can be reduced by several orders of magnitude, and the transmission will be much more stable.


    Key Tips


    Crystal oscillator selection does not need to pursue extreme parameters, as long as it is sufficient. In fact, choosing a crystal oscillator does not require blindly pursuing the lowest ppm and lowest phase noise. Different scenarios have different requirements. Buying a constant temperature crystal oscillator for 300 yuan in consumer electronics is completely unnecessary, as the cost increases by more than ten times and the performance improvement is not necessary. As long as you have a clear understanding of these five indicators and choose the one that should be loose or tight for your application scenario, you can select a stable and suitable crystal oscillator, and the entire system can run steadily. Engineers don't have to worry about crystal oscillator problems every day.


    References