Features and Applications of MEMS Clock Generators

History of Clock Generators

Clock Generators are timing components which integrate the clock functionality of many different discrete devices into one semiconductor component. They offer the traditional semiconductor benefits of integration, lower cost, smaller size, and full customization with minimal additional expenses. Historically, the use of clock generators started with the expansion of the PC motherboard market in the early 1990s. At that time, the functionality of a PC motherboard was increasing dramatically – audio, networking, high-end-graphics, video and various interfaces were being added to a PC. Each of these additional components (in addition to the core processor and memory) required a clock, which necessitated the use of as many as 7 clocking devices on a single PC motherboard. Semiconductor companies started developing clock generators which integrated the clocking function of these 7 devices into one or two clock generators, which resulted in the benefit of lower cost and smaller footprint.

MEMS-based Clock Generators are completely Integrated

Every clock generator requires multiple PLLs (Phase Locked Loops) which are used to generate any specified frequency from a standard fixed frequency reference, which is usually an external clock source such as a quartz crystal or oscillator. With the advent of MEMS resonators (which are available in the form of semiconductor die and can be completely integrated inside a plastic package), the need for having an external crystal or clock source goes away. Thus, a MEMS clock generator provides a completely integrated solution with no external reference clocks. A MEMS clock generator also eliminates the matching of the crystal with the clock generator circuit, which is a time-consuming problem to solve and may sometimes affect the performance of the system.

Features and Benefits of MEMS Clock Generators

* Completely integrated solutions, no external components required. MEMS resonator die (reference) is integrated inside package with analog circuit.

* Functionality of 3 – 6 clock generators in one 7.0x5.0mm package, results in up to 66% board space savings.

* Independent operating voltage on each of the PLLs, eliminates the need for external level translators, thus reducing component count and cost

* Mixed differential and LVCMOS outputs in the same device, addressing the need for different clocking devices in complex systems.

* Available spread spectrum capability to reduce system EMI and pass compliance testing.

About the Author

Learn more about :- Voltage Controlled Oscillators, Silicon Oscillator

High Frequency Crystal Oscillators

The heartbeat of an electronic system is the timing component. For the past 50 years, the timing reference in all electronics has been based on a quartz crystal. Components such as crystal resonators and crystal oscillators (also known as quartz oscillators) have been extensively used. Tens of billions of these devices are shipped every year.

In spite of their popularity, there are some significant limitations to crystal resonators and crystal oscillators. The most basic limitation is that typically, a crystal resonator cannot resonate at higher than 75 MHz in fundamental mode. To overcome this limitation for high frequencies, there are three alternatives:

* Use overtone mode – such as 3rd overtone, 5th overtone, etc. In this case, the oscillator operates at a frequency that is the third or fifth harmonic of the fundamental frequency. The disadvantage of this mode is that the costs are higher, and startup is not guaranteed over the entire temperature range.

* Use Phase Locked Loops to multiply the frequency – this is a very popular mode for MEMS-timing companies, but is not very popular with the crystal-based companies as they do not have analog circuits expertise. The benefits of this mode are that the intelligence is now in the electronics, leading to more features and higher performance at lower cost.

* Use Surface Acoustic Wave (SAW) technology – in which the crystal device operates differently than a standard AT-cut crystal.

Below is a comparison of the key parameters between SAW oscillators and MEMS Oscillators

SAW oscillators

* Frequency Range- Each device has to be cut to operate at a unique frequency

* Frequency Stability - Generally, 50 PPM, occasionally 25 PPM

* Package & Size – Large, 5032 and 7050 packages only

* Power Consumption- Typically more than 40mA

*1.8V Operation - Limited availability

* Aging - ± 3 PPM per year

* Shock & Vibration sensitivity – Sensitive, can break easily. Typical shock resistance is 2,000 to 5,000G

* Available Add-on features – Varies with part and frequency. Very few standard options available.

MEMS Oscillators

* Frequency Range- Programmable up to 800 MHz

* Frequency Stability - Can support as good as 0.5 PPM over industrial temperature range

* Package & Size - Small 2520 and 3225 in addition to 5032 and 7050 packages

* Power Consumption- Typically more than 20mA

* 1.8V Operation – Easily available

* Aging -Typically less than ± 1 PPM per year

* Shock & Vibration sensitivity – Extremely robust – 50,000G shock

* Available Add-on features - VCXO, SSXO (spread-spectrum) available in all frequencies and voltages, programmable drive

As can be seen from the above comparison, MEMS oscillators can offer superior characteristics compared to SAW crystal oscillators for high frequency applications.

About the Author

Learn more about :- Voltage Controlled Oscillators