Oscillator

From OpenCircuits
Revision as of 08:32, 8 January 2016 by DavidCary (talk | contribs) (mention the "negative resistance test")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

The vast majority of electronic systems operate at some fixed frequency. Usually the component that fixes that frequency is a quartz crystal, sealed in a metal can. (Some very-low-cost devices might use a "ceramic resonator" (made of high-stability piezoelectric ceramics, generally lead zirconium titanate) or a resistor and a capacitor to fix the frequency. ).

The entire circuit that generates the frequency is called an "oscillator", includes the resonant part (crystal, resonator, or RC), some capacitors, and a silicon chip and is therefore called a hybrid device. (An oscillator that uses a resistor and a capacitor to fix the frequency is called a "RC oscillator". An oscillator that uses a crystal to fix the frequency is called a "crystal oscillator".)

The oscillator includes, in addition to the frequency-fixing component just mentioned, an amplifier and capacitors.

Oscillators usually come in a metal can, but Epson also encapsulates them in plastic. Typically an oscillator can has 4 pins. Inside the can is all the components of the oscillator. One applies DC power on 2 of the pins, and the oscillating signal (the "CLK OUT") appears on another pin. (The remaining pin is unused).

An oscillator can also be made from scratch using crystal in a (2-pin) metal can, a couple of capacitors, a resistor and an inverter. In either case, the frequency is printed on the top of the crystal or oscillator.

Many microcontrollers have 2 pins (typically labeled "XTAL1" and "XTAL2", or "OSC1" and "OSC2", or something similar) that are designed to be directly connected to the 2 pins of a crystal. (Capacitors from those pins to VCC and GND are also part of the recommended circuit). An inverter inside the microcontroller acts as the amplifier, and the crystal and capacitors make up the rest of the oscillator. Microcontrollers when connected directly to a crystal need to have capacitors attached that need to be percisely matched with other circuit components on the board. Resonators tend to have the capacitor buildin reducing the amount of trial and error of capacitor selection.

In systems with multiple CPUs, it is often simpler, cheaper, and more reliable (avoiding metastability problems) to use a single crystal (rather than a dedicated crystal for each CPU). CPUs (and many other components) often have a single "CLK" pin designed to be connected to the wire used to send that "clock signal" (a fixed frequency) everywhere.

Unfortunately, many people confuse the "clock signal" generated by an oscillator (a simple metronome beat, tone, at constant frequency) with far more complicated "clock system"s that keep track of seconds, minutes, hours, and sometimes days, weeks, months, and years.

The "negative resistance test" measures the oscillation allowance of a crystal oscillator circuit. Most manufacturers recommend that the oscillation allowance of the circuit be at least five times the maximum resistance value of the quartz crystal unit used in the circuit.<ref> "MSP430 32-kHz Crystal Oscillators". p. 5. </ref><ref> "Oscillator Theory of Operation". p. 109. </ref><ref> "Oscillation Circuit Design Overview". </ref><ref> "AVR4100: Selecting and testing 32kHz crystal oscillators for Atmel AVR microcontrollers". p. 12. </ref><ref> "Oscillator Circuit Design Considerations". p. 8 </ref>


Further reading[edit]