An engineer's guide to selecting and using a resonator

4 mins read

Increasing demand for complex and compact consumer electronics applications is driving the need for low power, small and reliable components. The Internet of Things is driving the need for connectivity of the simplest devices, while wearable applications are becoming popular.

An accurate and stable clock is at the heart of today's electronics designs. From mobile phones to in car infotainment systems, you will find at least one resonator in an oscillator circuit providing the clock that keeps the design operating. A few types of resonator are available, with the different control elements shown in fig 1. The cheapest is the humble quartz crystal. Manufactured by shaping and cutting a slice of factory grown quartz crystal and encapsulating it in a metal case, crystal resonators have an exceptionally high Q, which is ideal for a tuned circuit, and are very stable over temperature and time. Oscillation frequency tolerance can typically be ±10 ppm at 25°C, while temperature stability is typically ±10 ppm across the operating range from -30 to 85°C. Examples include Murata's XRCGD/XRCMD series crystals. Size can be a key factor when selecting a quartz crystal. Since the resonant frequency is dependent on the size of the crystal slice, there are constraints on how small the package can be made. Variants of the simple quartz crystal have evolved to suit specific applications. These include the vcxo, tcxo and ocxo, where the whole oscillator circuit is maintained at a constant temperature that is typically higher than the ambient. Another approach uses the mechanical resonance of ceramics. Ceramic resonators also have high frequency stability; typically ±0.1% of nominal. However, they tend to be much smaller compared to their quartz based counterparts; often half the size. For example, the Murata CSTCE surface mount device series measures 3.2 x 1.3 x 0.9mm, has a typical frequency stability of ±0.1% and a temperature stability of ±0.08% in temperatures ranging from 0 to 70°C. Ceramic resonators have different oscillation mode characteristics and these are related to the nominal frequency. Low frequency oscillations – in the range from 100kHz to 1000kHz – typically use area expansion mode, while at frequencies of more than a few MHz, oscillation tends to be by thickness shear mode. The very nature of a ceramic resonator's construction means that it lends itself to mass production techniques. This, in turn, makes them much lower cost compared to the quartz crystal device. When selecting which type of clock device to use, engineers need to consider several technical criteria. Clearly, the operating frequency is key. The type of device ic being driven, and in some cases, the application, will normally determine the choice of resonator. Engineers should avoid using non standard frequencies, since that can impact lead times significantly and, usually, cost. In many cases, the function or application where the clock device is used will indicate the degree of clock accuracy required. For example, the IEEE802.1 network data communications standard sets the required accuracy. Also, for some specific product designs such as Wi-Fi routers, the wireless transceiver and/or microcontroller supplier may stipulate a number of specific criteria for resonator selection based on precertified designs. Another critical factor in today's space constrained designs is board footprint. Frequency stability and accuracy are greatly influenced by the size and shape of the resonator. This will also have an influence on the cost of the device. So there needs to be a balance between the operating specifications and the BOM cost. Other considerations when selecting a resonator can include the device grade and power consumption. The device grade generally relates to the resonator's operating temperature range. Certain resonators may be available in two types: one for general consumer applications; the other for automotive applications. Clearly, the marketing specification and the use case profiles will define the likely operating temperatures and environments of the end design. Automotive qualified parts have a wider operating temperature and are also usually more suitable for more harsh operating environments in terms of humidity, moisture ingress and vibration. Typical operating temperature range will be from –40 to 125°C. Power consumption considerations are getting more and more important, especially for wearable applications such as personal fitness monitors. Selection of the clock device may have an influence on the power consumption. The frequency of the clock device will be one of the parameters associated with power consumption. Generally speaking, using a low clock frequency could reduce the power consumption. Once suppliers and datasheets have been reviewed against the application requirements, the engineer can focus on oscillator circuit design. A reference oscillation circuit consists of an amplifying circuit that uses a cmos inverter or transistor, in addition to a feedback resistor, a damping resistor and two external load capacitors. Figure 2 shows the reference oscillation circuit design. The feedback resistor is connected to the cmos inverter in parallel to the oscillation circuit, but sometimes it might be integrated into the associated microcontroller. The feedback resistor balances the dc voltage between the I/O of the inverter and the inverter can function as an amplifier. When the feedback resistor is not integrated into an mcu, it is best to use a 1M? external feedback resistor.










The damping resistor is applied at output side of the oscillation circuit. It serves to attenuate the oscillation amplitude for reducing the drive level. On the other hand, it is necessary to pay attention to the oscillation margin because excessive damping resistance may cause oscillation to stop. The damping resistance is generally in the range from 0 to 2kO, but this can depend on the characteristics of the mcu. The external load capacitors are applied to the input and output of the oscillation circuit and the capacitance value has to be selected carefully. They are important parts that influence negative resistance and the oscillation frequency directly. Two similar capacitors are usually used as external load capacitors and generally have values in the order of 5 to 10pF, but the exact value depends on the characteristics of the mcu and the parasitic capacitance of the mounting substrate. When considering the layout of the pcb (see fig 3), attention should be paid to avoiding a low oscillation margin, as well as avoiding emi problems. The length of signal patterns in the oscillation circuit should be kept as short as possible in order to minimise stray capacitance and/or inductance. Through holes should not be used in an oscillation circuit because this approach can cause a lot of emi. Meanwhile, the ground or signal path should not be located in the middle layer of multilayer circuit board as the stray capacitance between ground and oscillation circuit becomes large.












Large stray capacitances may cause oscillation to stop due to insufficient oscillation margin. If the signal path is located close to the input of the cmos inverter, this may also generate emi as noise in the waveform is amplified. For crystal unit mounting, it is recommended to employ machines with optical positioning capability to prevent applying excessive mechanical stress. Kazutaka Hori is a product engineer with Tokyo Denpa, a Murata company.