Use MEMS-based TCXOs for Precise and Stable Timing in Harsh Environments

Technical Overview: This article examines the use of SiTime SiT5346 temperature-compensated oscillators (TCXOs) in applications exposed to severe shock, vibration, and temperature changes. It explains how microelectromechanical systems (MEMS) resonators avoid several mechanical limitations associated with quartz devices while providing stable timing performance. The discussion covers vibration and shock environments, acceleration sensitivity, temperature compensation, frequency stability, phase noise, and Allan deviation. It then introduces the SiT5346 family and its programmable operating modes, output options, supply voltages, and frequency range. A selected device illustrates how these oscillators can provide reliable timing for industrial, aerospace, defense, and other applications operating under demanding environmental conditions. (Overview courtesy of ChatGPT)

Oscillators are critical components in electronic systems, providing frequency-setting signals for wired and wireless links and timing for circuit and system clocks. While performance metrics such as precision, accuracy, stability, and acceleration sensitivity are important, maintaining them under extreme conditions is crucial for aerospace, defense, and other high-reliability applications.

Traditional quartz-based piezoelectric crystals have served the electronics industry well since the 1940s; however, meeting the performance and functional requirements of modern, advanced systems in harsh environments requires a different approach to timing that minimizes tradeoffs.

This article examines the challenges designers face as systems demand higher timing performance in extreme environments. It then introduces microelectromechanical systems (MEMS) oscillators from SiTime Corp and shows how they can achieve and maintain this performance.

Why use MEMS oscillators?

Since their initial development in the 1930s, when each unit was made and trimmed manually, quartz-based oscillators have advanced by orders of magnitude in performance, size, power requirements, mechanical specifications, and volume production techniques. In the basic design, a pure piece of quartz crystal (often denoted as XTAL) is “stimulated” to oscillate at its natural resonant mechanical frequency using the piezoelectric effect and a circuit in a self-reinforcing feedback scheme.

The oscillator (crystal plus support circuit) can provide frequencies from tens of kilohertz (kHz) up to hundreds of megahertz (MHz). However, performance and functional demands now exceed what these units can meet without added cost and complexity.

Fortunately, MEMS technology and devices provide a viable and attractive alternative to quartz-based designs. This technology first appeared commercially in the 1990s, used as accelerometers to trigger airbag inflation. Innovators have extended this silicon-based solid-state technology to create many new sensor types.

SiTime Corp has applied MEMS technology to oscillators and timing devices, delivering performance that exceeds quartz units across multiple dimensions, along with much smaller size, greater ruggedness, and other desirable features. This disruptive transition is analogous to vacuum tubes versus transistors; steam engines versus diesel locomotives; and propeller aircraft versus jet engines, though less visible and less well known.

Begin with device basics

The oscillator industry uses several standard terms for different oscillator architectures, regardless of fabrication technology. Among the most important are oscillator (XO), temperature-compensated XO (TCXO), voltage-controlled XO (VCXO), and oven-controlled XO (OCXO):

• A basic XO is an active device that combines the resonator and oscillator circuit into a single package, without temperature compensation. The XO designation has been used since the earliest days of quartz-based designs.
• A TCXO incorporates temperature compensation to correct for the frequency-versus-temperature characteristic of the resonator.
• A VCXO includes a voltage-control pin that allows the user or circuit to control (“pull”) the output frequency by a modest amount around the nominal frequency.
• An OCXO provides temperature compensation by using an oven to maintain an almost constant temperature for the oscillator, even as the ambient temperature varies.

As with all components, oscillators have key specifications and figures of merit, regardless of technology and architecture. Key oscillator parameters include nominal frequency, initial accuracy, output jitter, phase noise, startup time, operating voltage and tolerance, operating current, and size. VCXOs also have additional basic parameters, such as pull range and pull linearity.

Because of their critical role, oscillators must remain stable over long-term aging and unavoidable temperature variations. These variations are typically specified in parts per million (ppm) and parts per billion (ppb).

Further, oscillators must be characterized for performance and ruggedness under single-event shocks ranging from modest to extreme, as well as under ongoing vibration at various intensity levels.

The SiTime MEMS-based oscillator solution

SiTime’s oscillators use a radically different approach than classical crystal-based timing. These oscillators use MEMS technology and processes, a mature, field-proven, silicon-based technology with decades of design, production, and field experience.

The SiTime design uses a MEMS resonator coupled with an analog CMOS circuit (Figure 1, top) that integrates multiple sophisticated functions for configuration, compensation, and optimized performance (Figure 1, bottom). According to SiTime, their respective characteristics include:

MEMS resonator

• Extremely resistant to shock and vibration due to its small mass
• Hermetically sealed, with no contamination
• No fatigue mechanism over time
• No possibility of breakage under stress

Analog CMOS die

• High-precision temperature sensing
• High-performance temperature compensation
• Advanced mixed-signal circuitry

Figure 1: The SiTime oscillators use a design that combines a MEMS resonator with an analog CMOS circuit (top); the circuit integrates multiple sophisticated functions for configuration, compensation, and optimized performance (bottom). (Image source: SiTime)

Start with frequency

Frequency is the primary specification of an oscillator. SiTime oscillators are available in frequencies as low as several hundred kHz for low-power devices and as high as 725 MHz (exact numbers depend on the chosen model family). In many models, the frequency is programmable within a device’s range to six decimal places. An important attribute is that the frequency can be factory-programmed in one of three ways: by SiTime, by key partners and distributors, or in the customer’s lab using an oscillator programmer when lower volumes or a single unit are needed for prototype and pilot runs.

This straightforward programming capability is a key benefit of SiTime oscillators. It is accomplished by setting the fractional-N phase-locked loop (PLL) division factor in the basic oscillator. This contrasts sharply with how frequency is established in crystal-based units, where the individual resonator is usually tuned to the desired frequency by ablating the metal electrode with an ion beam, a process ill-suited to fast-turn, low-volume requirements.

Temperature and frequency-related performance

Operating temperature variations affect electronic components, especially oscillators, and challenge even the best designers. They can reduce drift in key specifications of quartz-based units through circuit-compensation schemes or even by using a miniature stabilizing oven to create a steady temperature environment, but these add size, weight, and power consumption to the unit and have limited effectiveness.

Several parameters determine frequency-related performance. The first is frequency stability, which is the deviation of the output frequency due to external conditions. It is typically expressed in ppm or ppb relative to the nominal output frequency.

Second is the frequency-versus-temperature slope (Figure 2). Shown as ΔF/ΔT, it is the rate of frequency change per 1°C change in temperature. It quantifies the oscillator's sensitivity to small temperature variations near the operating temperature. When combined with a specified temperature change, ΔF/ΔT can be used to estimate the resulting frequency error in ppm or ppb.

Figure 2: The change in frequency per degree is used to calculate the frequency-drift error, typically expressed in ppb for high-performance devices. (Image source: SiTime)

Because of the embedded electronic circuitry around the core resonator, along with an integrated temperature sensor, SiTime oscillators can use a variety of circuit arrangements to self-cancel much of the drift error and actively compensate for it, improving performance over crystal units.

For example, the DualMEMS structure used in SiTime’s Endura SiT5346 Ruggedized Super-TCXO family offers an electrically noiseless temperature-compensation arrangement. It consists of two MEMS resonators fabricated on the same die substrate. One resonator is designed to have a flat frequency response over temperature, while the other is designed to be sensitive to temperature changes. The frequency ratio between these two resonators provides an accurate reading of the resonator temperature with 20 microkelvin (μK) resolution.

By placing the two MEMS resonators on the same die, this temperature-sensing scheme eliminates thermal lag and gradients between the resonator and the temperature sensor, overcoming an inherent weakness of legacy quartz TCXOs.

Do not ignore shock and vibration

Although SiTime devices may look fragile, the opposite is true. In fact, MEMS-based devices excel in shock and vibration performance. This characteristic is useful beyond extreme applications; almost all electronic products experience shock and vibration during their lifetime. The key question is how much vibration and shock tolerance the oscillator offers while still operating within specifications.

These vibration (g) forces can range from the motion experienced by mobile consumer products in pockets or backpacks to very high g (acceleration) levels typical of rocket launches, military and aerospace systems, mining and drilling, and other harsh industrial settings (Table 1). Even stationary products in buildings can experience ongoing vibration from nearby fans, other equipment, or passing vehicles.

Acceleration profiles can be simple sinusoidal or random, so the levels will vary. Regardless, even low-level ongoing vibration can compromise performance and reliability.

Table 1: The range of real-world vibrations spans several orders of magnitude; even low-level, continuous vibration can compromise performance and reliability. (Image source: SiTime)

Shock presents another set of challenges. Shock levels range from low to extremely intense (Table 2). As with vibration, standard tests use different shock profiles, with severity determined by factors including peak acceleration, pulse shape, and duration.

Table 2: The various levels of shock to which a device may be subjected span many orders of magnitude. (Image source: data compiled by Bill Schweber)

The SiTime resonator structure behaves like a stiff spring and is therefore highly resistant to external acceleration. According to SiTime, modeling indicates that acceleration greater than one million g would be required before the resonator contacts a surrounding surface. This structural estimate should not be confused with the qualified shock resistance of a complete oscillator.

A SiTime family and a device illustrate achieved performance

The Endura SiT5346 Ruggedized Super-TCXO family is designed for aerospace, defense, and other high-reliability applications that require precision timing under sustained environmental stress. DualMEMS temperature sensing and TurboCompensation maintain stability under airflow, thermal gradients, vibration, shock, and electromagnetic interference (EMI).

The family delivers ±100 ppb over-temperature stability from -40°C to +105°C, a ±1 ppb/°C frequency slope (ΔF/ΔT) (Figure 3, left), and a 1.5 × 10^-11 Allan deviation (Figure 3, right) at a 10 second (s) averaging interval, all in a 5.0 mm × 3.2 mm, 10-pin package.

Figure 3: Devices in the Endura SiT5346 Ruggedized Super-TCXO family demonstrate frequency deviation within tens of ppb across a wide temperature range (left) along with high stability over time as measured by the Allan deviation (right). (Image source: SiTime)

Their low acceleration sensitivity, reaching 0.004 ppb/g (typical), reduces vibration-induced phase noise by up to 20× compared to quartz, helping minimize link and call drops on high-vibration platforms. A power supply noise rejection of 0.2 picoseconds/millivolt (ps/mV) eliminates the need for a dedicated low-dropout regulator (a significant benefit to circuit layout, bill of materials, and cost), while digital-frequency tuning via an I²C port corrects board-level frequency shifts. The device undergoes 100% Endura process screening and meets MIL-PRF-55310 and MIL-STD-883 requirements, with a mean time between failures (MTBF) of 1 × 10⁹ hours and a lifetime warranty.

Factory programmability supports frequencies from 1 MHz to 60 MHz, supply voltages of 2.5 V, 2.8 V, 3.0 V, and 3.3 V, and pull ranges from ±6.25 ppm to ±3,200 ppm, eliminating quartz-related lead times and per-configuration non-recurring engineering (NRE) costs. Stability grades of ±0.1 ppm, ±0.2 ppm, and ±0.25 ppm are available across temperature ranges of -20°C to +70°C, -40°C to +85°C, and -40°C to +105°C. The LVCMOS or clipped-sinewave output optimizes EMI and jitter performance for each system design.

The SiT5346 devices can be ordered in one of three basic configurations:

• Basic TCXO with non-pullable output frequency (Figure 4, left)
• VCTCXO allowing voltage control of output frequency (Figure 4, center)
• DCTCXO enabling digital control of output frequency using an I²C interface, pullable to 5 parts per trillion (ppt) resolution (Figure 4, right)

Figure 4: Devices in the SiT5346 family can be ordered configured as a basic TCXO (left), as a VCTCXO (center), or as an I²C DCTCXO (right). (Image source: SiTime)

Different applications have specific needs or preferences that call for a unique configuration. A close look at an individual oscillator model in the family, such as the SIT5346AE-FQ033JHB40.000000F, along with its ordering form (Figure 5), shows the flexibility and options available.

Figure 5: The ordering options for the SiT5346 oscillator show the many variations that are available in addition to basic output frequency. (Image source: SiTime)

Based on the ordering-number details, this is a Revision A part with an extended industrial temperature range in a 5.0 mm × 3.2 mm package. The model has ±0.1 ppm stability, I²C address mode “0”, a 3.3 V supply, software output-enable control, a ±200 ppm pull range, ultra-low g sensitivity of 0.009 ppb/g, a 40 MHz output, and 12 mm cut-tape carrier packaging.

Conclusion

Maintaining timing performance under extreme conditions is crucial for aerospace, defense, and other high-reliability applications. For these systems, designers can use MEMS oscillators from SiTime’s Endura SiT5346 ruggedized Super-TCXO family for precise and reliable timing with relatively easy design-in, small footprint, minimal EMI sensitivity, a wide variety of options, and short factory lead times.

Related Content

1. Glossary of Oscillator Terminology
2. The top 8 reasons to use an oscillator instead of a crystal resonator
3. Shock and Vibration Performance Comparison of MEMS and Quartz-based Oscillators
4. Beyond Quartz: The MEMS Edge
5. Endura Low Phase Noise Super-TCXOs: Time Synchronization and RF Systems
6. MEMS-Based Resonators and Oscillators are Now Replacing Quartz
7. AN10039 TCXO Frequency Stability and Frequency Accuracy Budget
8. Do you know when to use a crystal or an oscillator? The wrong answer can cost you.
9. Crystal Clear: The Struggle for Reliable Communications Technology in World War II, Richard J. Thompson, Jr.