Mitigating the Effects of Crystal Oscillator Aging

With such a wide range of electronic timing devices to choose from, why is the classic high-performance crystal oscillator still the best choice in many applications? A partial answer is their benefits in high-performance environments where a “service call” is simply out of the question. Space is one such area, from low Earth orbit (LEO) to geosynchronous orbit and deep space satellites. Another involves high-temperature applications like down-hole drilling.

Other applications where quartz-crystal timing devices are the better choice include military and avionics applications, such as missile guidance and navigation systems, where temperature and frequency stability are critical. Crystal oscillators are also a good choice for high-definition video displays, where anything but exceptional performance is plain to see.

Crystal oscillators are also used to drive high-frequency, field-programmable gate arrays (FPGAs). In these applications, the lower jitter of quartz oscillators results in wider “eye” diagrams and lower bit error rates (BER). Yet another application is high-speed analog-to-digital converters (ADCs). In this case, the crystal oscillators’ lower jitter translates to lower phase noise.

In all of these applications, nothing we know of will ever have the “Q,” spectral purity, and stability of quartz. Nonetheless, it’s essential to acknowledge and address the intrinsic aging effects of crystal oscillators on timing performance.

This article discusses this phenomenon and the methods employed by oscillator manufacturers like Q-Tech to mitigate these effects.

Causes of Crystal Aging

Aging in crystal oscillators directly impacts timing applications by causing a gradual and irreversible drift in the oscillator’s output frequency over time. In precision timing applications requiring radiation tolerance, low noise and jitter, and/or high-temperature operation, exceptional mitigation techniques must be employed in the manufacturing process to minimize the impact of aging.

Aging refers to the gradual change in the crystal oscillator’s output frequency over time. This phenomenon is primarily due to physical and chemical changes within the quartz crystal and its environment. This drift, typically measured in either parts per million (ppm) or parts per billion (ppb) per year, can lead to significant issues in applications that require precise and stable timing references.

There are many causes of crystal aging, but two of the most significant — by far — are mass loading and stress changes. Mass loading refers to subtle changes to the quartz resonator’s mass due to absorption or desorption of contaminants on the crystal surface. It alters the mass and, thus, the resonant frequency.

All quartz crystals are processed in clean rooms and sealed in very clean environments, either in a vacuum or in a non-reactive, inert nitrogen-helium atmosphere. Still, some level of undesired material always remains; typically, it’s a very small amount of water vapor that can cause mass-loading changes, resulting in frequency aging.

In terms of stress changes, great lengths are taken to minimize the stress on the crystal. However, some residual stress is always there, especially in the attachment of the crystal to its package and the oscillator circuitry.

Such a mounting structure usually consists of a very-low-stress metallic ribbon connected to the quartz via a very-low-outgassing epoxy or adhesive. Over time, though, some changes in the stress from this mounting structure will affect the crystal and its frequency. Moreover, changes may occur in the internal stresses of the crystal’s metal electrodes and even within the quartz itself.

While mass loading and stress changes are the leading causes of crystal oscillator aging, other instigators include quartz outgassing, diffusion effects, chemical reactions, pressure changes, and oscillator circuit aging (especially load reactance and drive-level changes).

What are the Main Effects of Aging on Timing Applications?

• Frequency drift: As the oscillator ages, its frequency can increase or decrease, making the system clock run slightly faster or slower than intended. This is particularly problematic in systems where even minor deviations can accumulate into substantial timing errors over hours, days, or years.
• Holdover performance: In critical timing systems — such as those for satellites, network infrastructure, and measurement equipment — oscillators often serve as backup clocks when external references (like GPSs) are unavailable. During these "holdover" periods, the local oscillator must maintain accurate timing. Aging-induced frequency drift can degrade holdover accuracy, leading to synchronization errors in networks or data loss in time-sensitive applications.
• Long-term reliability: Over extended operation, the cumulative effect of aging can push the oscillator’s frequency outside acceptable limits for the application, necessitating recalibration, replacement, or compensation mechanisms.
• Environmental sensitivity: Aging effects can be exacerbated by environmental factors such as temperature fluctuations and power cycling, further destabilizing timing performance.

Strategies to Combat Aging

Mitigation strategies include using oscillators with inherently low aging rates, implementing compensation algorithms, and designing systems to allow for recalibration or redundancy in timing sources.

To manufacture very-low-aging crystals, cleanliness and contamination control during their processing are most important. Next, the crystal must be sealed in an extremely hermetic environment, either in a resistance-weld or, even better, a cold-weld package. In addition, lengthy initial burn-in at high temperature can accelerate the initial frequency changes.

Pre-aging is another approach. Before shipment, precision crystals are often factory "pre-aged" to accelerate the initial rapid aging phase. This ensures that the oscillator achieves its specified aging rate more quickly in the field.

High-quality, stress-compensated (SC-cut) or AT-cut crystals are chosen for their superior long-term stability and lower intrinsic aging rates. SC-cut crystals offer better aging performance in crystal oscillators than the typically used AT-cut types (Fig. 1).

The aging process of crystals slows and improves over time, with the first months and years showing more change than subsequent time periods. For example, with standard crystal clock oscillators, where the quartz crystal is enclosed in the same environment as the rest of the oscillator circuitry, a typical aging rate might be ±1 to ±5 ppm for the first year, and then ±0.5 to ±2 ppm for subsequent years.

The standard military specification for quartz crystals, MIL-PRF-55310 (and its predecessor MIL-O-55310), carefully specifies crystal aging and calls for a 30-day aging test at an elevated temperature of 70°C, with strictly defined limits on how much the frequency can change and still be acceptable. In addition, the military specifications give mathematical equations wherein the 30 days of aging data can be fitted to predict worst-case aging over longer time periods. At Q-Tech, crystal oscillators intended for use in timing-critical applications are tested for compliance with MIL-PRF-55310.

In timing applications that require even better performance, the quartz crystal is enclosed in its own, separate hermetic environment, usually a cold-weld crystal package (Fig. 2). Cold-weld is a solid-state welding process whereby joining takes place without fusion or heating at the interface of the two parts to be welded. Unlike fusion welding, no liquid or molten phase is present in the joint, and little heat is involved, thus minimizing outgassing.

Using this type of crystal, temperature-compensated crystal oscillators (TCXOs) and oven-controlled crystal oscillators (OCXOs) can achieve typical aging rates of ±0.2 to ±1 ppm the first year, and then ±0.1 to ±0.5 ppm, or even better, for subsequent years.

Summary

Aging, which is an unavoidable characteristic of crystal oscillators, can cause gradual frequency drift and reduce long-term stability in critical timing applications. Understanding and mitigating these effects through design choices and carefully controlled manufacturing processes is essential for ensuring reliable oscillator performance in precision applications.