Skip to main content

MEMS Oscillators Make Inroads

Introduction

Cell phones, computers, radios, watches, and many other devices rely for their success on an electronic oscillator that produces an output with a precise frequency to generate timing pulses and synchronize events.

The rise in portable and wearable electronics is driving the need to reduce both the energy consumption and footprint of all types of electronic components, including oscillators. Oscillators based on Micro-Electro-Mechanical Systems (MEMS) technology combine accurate frequency generation with low power consumption and are becoming increasingly popular in clock circuits.

This article will look at MEMS technology, MEMS oscillators and why they're replacing more traditional solutions in portable and non-portable applications alike.

Overview of Traditional Oscillator Types

Before the advent of MEMS devices, designers used several approaches to generate a clock signal, depending on the application requirements.

The lowest-cost option is an RC oscillator, which uses a network of passive components and an amplifier to produce an oscillating signal using positive feedback. The phase-shift oscillator, for example, uses three cascaded RC sections with a cumulative phase shift of 180°. When added to the feedback loop around an opamp this generates positive feedback, giving an oscillating output.

An integrated silicon oscillator uses similar circuitry, but all of the components are on a single die, giving more precise operation and improved performance over temperature. Silicon oscillators are available with a choice of factory-programmed operating frequencies.Maxim Integrated's MAX7375, for example, can be factory-trimmed from 600kHz to 9.99MHz and features 2% initial accuracy and ±50ppm/°C drift over temperature.

For precision applications, the traditional solution has been a circuit based on a vibrating quartz crystal. It is a piezoelectric device; when a voltage is applied across it, it behaves like an RLC circuit with a precise resonant frequency. A ceramic resonator uses a similar operating principle, but its vibrating element is made from a ceramic material such as lead zirconium titanate (PZT).

To make an oscillator, the crystal or resonator is combined with an analog circuit that drives it at the resonant frequency. Many embedded processors include internal circuitry to make it easy to accommodate either type of device. Alternatively, a crystal oscillator module combines the crystal and the support circuitry into a single package.

Table 1 compares these various options.

Clock Source

Accuracy

Cost

Comments

RC Oscillator

Very low

--

May be sensitive to EMI and humidity. Poor performance over temperature.

Integrated Silicon Oscillator

Low to Medium

-

Insensitive to EMI, humidity, and vibration. Small size; no additional components required. Temperature sensitivity better than RC, worse than crystal or ceramic resonators.

Ceramic Resonator

Medium

+

Sensitive to EMI, humidity, and vibration. External drive circ uitry required.

Crystal

Medium to High

++

Sensitive to EMI, humidity, and vibration. External drive circuitry required.

Crystal Oscillator Module

High

+++

Insensitive to EMI and humidity. No additional components needed. High power consumption; sensitive to vibration; large package.

Table 1: Comparison of traditional oscillator options (Source: Maxim )

MEMS Technology & MEMS Oscillators

MEMS technology uses standard semiconductor manufacturing processes such as lithography, deposition, and etching to produce miniature electro-mechanical elements varying in size from less than one micron up to several millimeters.

Harvey Nathanson of Westinghouse invented the first MEMS device in 1965 to serve as a tuner for microelectronic radios. In the 1990s, MEMS pressure sensors and accelerometers started to become widely used in applications such as automotive airbags and medical respirators, spurring widespread development and helping to drive down the cost of MEMS technology.

A MEMS resonator is a small structure (0.1mm or less) that is designed to vibrate at high frequencies under electrostatic excitation. During manufacturing, the resonator structure is first etched in a silicon on insulator (SOI) layer. The wafer surfaces are then planarized by filling the trenches with oxide. Next, contact holes are formed to allow electrical connections to be made. Finally, the oxide is removed with hydrofluoric acid to create freestanding resonator beams with the ability to vibrate.

Fig. 1
Figure 1: Depending on the desired frequency, MEMs resonators have different sizes and shapes (source: SiTime)

The resonant frequency of a MEMS resonator is inversely proportional to its size, and both kHz and MHz frequency resonators are available. Resonators in the kHz range are optimized for low power consumption. They are typically used for time-keeping applications such as real-time clocks, or provide sleep and wake-up functions for power management systems. Resonators with MHz frequencies provide precise references for serial and parallel communication where data transfer speed is critical.

As shown in Figure 2, a MEMS oscillator combines a MEMS resonator die and a programmable oscillator IC; the resonator is driven by the circuit blocks on the analog oscillator IC. The resonator sustaining circuit drives the resonator into mechanical oscillation. Both die are mounted together in a stacked-die or flip-chip configuration and packaged in either a standard or chip-scale package.

Fig. 2
Figure 2: A MEMS oscillator includes a resonator and a separate oscillator die in a single package. Precision applications often require integrated temperature compensation. (source: SiTime)

The output frequency is set by a Fractional-N phase-locked loop (PLL) block, which generates an output signal that is a multiple (N) of the MEMS resonator frequency. An on-chip one time programmable (OTP) memory stores the configuration parameters. Many devices also include output drivers with configurable drive strength for impedance matching or emissions reduction.

For precision timing applications, MEMS oscillators often include temperature compensation by means of an on-chip temperature sensor.

MEMS Packaging

Like other semiconductor devices, MEMS oscillators come in a variety of packages. For designers looking to replace quartz oscillators in existing designs, MEMS oscillators are available in compatible 2.0 x 1.2mm (2012) SMD packages. Since a MEMS oscillator requires two extra pins for power and ground, these are placed between the existing SMD end caps, as shown in Figure 3.

Fig. 3
Figure 3: In addition to traditional semiconductor packages, MEMS oscillators are available in both SMD and CSP sizes (source: SiTime)

In addition, a MEMS oscillator can be combined with another device such as an ASIC or microcontroller into a single package using chip-scale packaging (CSP) techniques.

MEMS Oscillator Performance

Early MEMS resonators were not stable enough to be used as timing references, but current-generation devices can achieve stability as low as ±5ppm. For portable use, low-power devices can attain a frequency tolerance of ±20ppm and a stability of ±100ppm.

The use of semiconductor packaging allows the MEMS oscillator to withstand high levels of shock and vibration, which is particularly valuable in portable and wearable devices such as digital cameras, cell phones and watches, which are always subject to being dropped.

Examples of MEMS Oscillator Products

Several manufacturers offer low-power MEMS oscillators and support products. SiTime's SiT1533, for example, is an ultra-small and ultra-low power 32.768 kHz oscillator optimized for mobile and other battery-powered applications. The SiT1533 features maximum operating current of only 1.4μA and is pin-compatible and footprint-compatible to existing 2012 XTALs when using the recommended layout. The part's factory programmable output reduces the voltage swing to minimize power. An operating voltage of 1.2V - 3.63VDC makes it suitable for mobile applications that incorporate a low-voltage coin-cell or super-cap as a battery back-up.

The DSC1001 from Microchip is a MEMS-based oscillator offering excellent jitter and stability performance as low as 10ppm over a wide range of supply voltages and temperatures. The device operates from 1MHz to 150MHz with supply voltages between 1.8 to 3.3 VDC and temperature ranges up to -40ºC to 105ºC.

MEMS oscillators are available over an extremely wide frequency range. The Abracon ASTMK-0.001kHz, for example, can operate down to 1Hz with 20ppm tolerance while consuming 1.4μA. At the other end of the spectrum, IDT's 4H family of ultra-low jitter MEMS oscillators can operate up to 625MHz.

Designing With MEMS Oscillators

In keeping with a high-frequency clock device, designers should follow best practice layout techniques such as limiting trace lengths, paying attention to routing, limiting the use of vias, and using ground planes.

In addition, proper use of capacitors can help in several areas:

Decoupling: Fast switching devices such as clock oscillators can place a significant demand on the power supply, with voltage sag a possible consequence. A decoupling capacitor close to the power supply can act as a local reservoir to ensure sufficient charge is always available.

Bypassing: To limit the amount of noise propagating through the system, bypass capacitors are needed to provide low-impedance paths that shunt this transient energy to ground.

Power Supply Noise Reduction: In most applications, a single 0.1 μF capacitor between the power supply voltage and power return will shunt much of the power supply noise to ground. For greater noise reduction designers may implement RC or LC power supply filtering strategies.

Options To Reduce EMI

As processor speeds increase and more devices are crammed into smaller spaces, ensuring electromagnetic compatibility (EMC) between devices is assuming greater importance.

A signal generated in one device can couple into other devices, causing errors or malfunctions. The oscillator clock is often a major source of electromagnetic interference (EMI) because it consists of a repetitive square wave with high-frequency harmonics, and is often widely distributed across the board.

Filtering, shielding, and good layout practices can limit EMI, but add cost and consume board space. An alternate approach is to reduce clock-generated noise by modulating the clock frequency slowly over time. This variation reduces the peak spectral energy in both the fundamental and harmonic frequencies. This reduction also helps in FCC certification, which uses the peak power within a specific bandwidth to determine the EMI.

A programmable spread-spectrum MEMS oscillator such as the SiT9003 reduces EMI by modulating its PLL with a 32kHz triangular wave to vary the center frequency of the output. The amount of frequency spreading is user-selectable; varying the output frequency between 98MHz and 100MHz, for example, can give an average EMI reduction of 13dB.

Conclusion

MEMS oscillators offer an attractive combination of low power consumption, small size, high performance and physical robustness that make them ideal for numerous applications, especially in portable and wearable electronics.

Their ability to leverage standard semiconductor manufacturing and packaging methods means that their cost and performance will continue to improve, ensuring that they will continue to make inroads into applications traditionally reserved for quartz crystals and ceramic resonators.

About the Author

As a freelance technical writer, Paul Pickering has written on a wide range of topics including: semiconductor components & technology, passives, packaging, power electronic systems, automotive electronics, IoT, embedded software, EMC, and alternative energy. Paul has over 35 years of engineering and marketing experience in the electronics industry, including time spent in automotive electronics, precision analog, power semiconductors, embedded systems, logic devices, flight simulation and robotics. He has hands-on experience in both digital and analog circuit design, embedded software, and Web technologies. Originally from the North-East of England, he has lived and worked in Europe, the US, and Japan. He has a B.Sc. (Hons) in Physics & Electronics from Royal Holloway College, University of London, and has done graduate work at Tulsa University.