Quartz and Programmable All-Silicon MEMS Oscillators Comparison (Part 1)

If many people compare MCU or SoC main chip to the brain of modern electronic system, then the clock component is undoubtedly its heart. Whether it is electronic engineers or component purchasers, they will conduct a comprehensive and rigorous evaluation when selecting clock components. Because a healthy, stable and durable "heart" will directly affect the function and reliability of the electronic system.

　　Clock components can be divided into three categories: passive crystal oscillators, active crystal oscillators, and multi-output clock generators. In the past 60 years, quartz has been the mainstream technology in the clock market and has always occupied a dominant position. However, due to the limitations of traditional manufacturing processes and the monopoly of the downstream raw material (starting circuit and base) market, the cost performance cannot be further improved. In order to meet the electronic market's demand for smaller, more reliable, and more flexible components, clock components must take the path of full siliconization. This article will mainly introduce the differences between all-silicon MEMS oscillators and traditional quartz, as well as the problems solved by all-silicon IC technology.

　　Introduction to Quartz and All-Silicon MEMS Clock Oscillators

　　Traditional quartz oscillators are composed of piezoelectric quartz, simple oscillator chips and metal packages. The production process includes: quartz cutting and silver plating, purchasing bases, oscillator chips, and combining quartz and chips with special adhesives and placing them on the base, then filling with nitrogen and sealing with metal packages. The generation of oscillators with different frequencies and different operating voltages is determined by the different shapes of quartz, the thickness of silver plating and the oscillator chips. Therefore, from the perspective of production technology, the quartz industry is a labor-intensive, semi-automated traditional industry, and its products are also limited by traditional raw materials and processes:

　　1. The complex production process leads to the extension of the delivery period and the difficulty of emergency response to shortages;

　　2. Different oscillator specifications require different raw materials and processes, which makes the finished product lack flexibility and cannot be configured in real time to meet different applications;

　　3. The high temperature sensitivity of piezoelectric quartz causes the temperature drift of the quartz oscillator;

　　4. Quartz’s weakness of being fragile, afraid of falling and aging needs to be addressed through production processes and quality management, and lacks consistency in quality and long-term reliability.

　　In order to fundamentally solve the inherent weaknesses of quartz, electronic system manufacturers have begun to turn to all-silicon MEMS oscillators in the selection of clock components. Different from traditional quartz oscillators, MEMS oscillators use an all-silicon product structure, with an all-silicon MEMS resonator and a programmable Analog CMOS driver chip stack, and are completed in a standard QFN IC package. Figure 1 is a perspective view of the SITIME MEMS oscillator.

　　Figure 1: All-silicon MEMS oscillator display

　　Compared with traditional quartz, all-silicon MEMS oscillators are more in line with the standards of modern electronic products in terms of production process and component design structure, and are also an upgrade to traditional quartz products.

　　* High-performance analog temperature compensation technology enables the all-silicon MEMS oscillator to have excellent full-temperature frequency stability, completely eliminating the problem of temperature drift;

　　* The programmable platform provides the necessary flexibility for system design and shortens the new product development cycle;

　　* A complete semiconductor production chain can shorten the delivery time of all-silicon MEMS and improve the ability to respond to emergencies;

　　* Fully automated production of IC structures with unquestionable consistency in quality and reliability.

　　All-silicon MEMS oscillator performance advantages over all-temperature

　　频率稳定性，特别是在不同温度下的稳定性，是电子工程师在选择振荡器时考虑的主要参数之一。因为每一个设计，都需要保证系统在整个工作温度范围内正常运作。而温飘(频率随温度而显著变化的现象)则是传统石英产品的弱点，难以单纯从制造上克服。图2是石英和全硅MEMS振荡器在频率稳定性方面的比较。

　　Figure 2: All-silicon MEMS oscillator 25PPM frequency stability exceeds quartz

　　The dark black curve in Figure 2 shows the technical difficulty of achieving a full-temperature frequency stability of 25PPM for an industrial-grade (-40°C-85°C) quartz oscillator. As can be seen from the figure, under high and low temperature conditions, the design margin of quartz as a reference clock is relatively insufficient, which also increases the possibility of unstable operation of the overall system under industrial-grade full temperature.

　　Figure 2 also shows the balance lines of various colors, representing the actual total frequency difference of 110 SiTime all-silicon MEMS oscillators in the range of -40°C to 85°C. Compared with quartz oscillators, the frequency stability of these industrial-grade MEMS oscillators can not only be maintained below 15PPM, but the curve is more linear, providing greater design margin for the system.

　　Because all-silicon MEMS oscillators use temperature compensation technology to solve the problem of quartz temperature drift in oscillator design, electronic engineers have more room to choose materials. They can choose 50PPM MEMS oscillators to replace many 25PPM quartz, which can meet the system specifications and reduce costs. Or, they can use 25PPM MEMS oscillators to improve the overall stability of the system.