Radio-frequency (RF) technology is revolutionized almost daily. This loud and noisy revolution has attracted lots of attention. Most of that attention focuses on the glamorous DSPs and highly integrated RF transceiver chips that are now ubiquitous in cell phones and wireless local-area network cards. But the revolution marches on in other areas of product development, albeit more quietly. Usually ignored in this hubbub are such components as mixers, the humble devices that make it all possible.
A mixer translates a signal up or down in frequency. As such, it enables the "superheterodyne" receiver, invented in 1917 by Edwin Armstrong, from which the modern radio directly descends. The heterodyne principle stems from the discovery that signals can often be more clearly detected if they are shifted to a different frequency.
The mixer is an intrinsically nonlinear device that can be thought of as a multiplier, thus the origin of the term "mixer." Any nonlinear device can be used for frequency translation, but the most common and best mixers until recently were passive mixers based on high-speed switching diodes. Here, the diodes are configured as a bridge and driven by a high-frequency local oscillator (LO). "Passive" means that the input signal and the LO supply the energy for the frequency-translated output signal.
Microwave designers cut their teeth with these mixers. They are easy to use and supplied with impedance-matched, single-ended inputs and outputs. They also offer low power dissipation, low noise, high linearity, and straightforward specifications. On the downside, passive mixers are lossy devices. Typically, the output signal has only about 15% of the energy level of the input signal. The LO needs to be a large amplitude signal to switch the diodes, and it produces spurious electromagnetic radiation that must be filtered and shielded from other circuitry—a leading cause of headaches among microwave folk.
The active mixer answers these problems. Typically, it is a double-balanced differential stage in IC form. The active mixer output power comes from its dc power supply, rather than the LO. Consequently, smaller, more easily managed LO levels can be used, and active mixers offer gain.
Although the principle is grand, until recently, these devices just weren't suitable for high-performance infrastructure applications like cellular basestations. They had too much noise and substandard linearity performance. They also needed external matching elements and an external differential-to-single-ended conversion circuit. These problems ranged from the merely annoying to devastating in scale.
However, improvements in active mixer technology have lately produced ICs with all the advantages of the traditional passive mixer. Fabricated in silicon bipolar, silicon biCMOS, or silicon-germanium technology, these active mixers incorporate low-noise, high-linearity transistors; RF transformers for differential to single-ended conversion; and broadband, on-board matching elements.
These mixers offer ease of use and performance much like passive mixers, while simultaneously providing everything that already makes active mixers popular in lower-performance applications. Features include low LO drive levels, gain, reduced filter requirements, and endless potential for integration. Even today, more functionally complex devices, like high-linearity I/Q modulators and demodulators, are appearing.
As integration follows its usual path, we will see a continuing stream of ICs offering higher performance and greater functionality. The benefits will appear in high-frequency infrastructure systems, continuing the trend to higher performance at steadily lower costs. As such, you will find the benefits of this technology not just in your pocket where your cell phone sits, but also in the other pocket—where you keep your wallet.