|
Integrated circuit mixers provide an alternative to traditional diode based passive mixers. The typical implementation uses a structure called a Gilbert cell. An emitter-coupled transistor pair is used to switch the current path between the inner and outer transistors of a collector-cross-coupled quad; this switching creates a double-balanced mixing action.
Gilbert cell-based IC mixers differ significantly from diode mixers. The Gilbert cell is an amplifier, so the mixer now has conversion gain rather than the approximately 7 dB loss of a passive mixer. The switching action of the mixer takes relatively little energy to initiate, so although the IC mixer must be DC biased, its total power budget is often less than that required for a passive mixer driven by a much higher-powered LO. The Gilbert cell is often less sensitive to loading, and doesn't require any bulky transformers, making it an extremely convenient device to use. In trade for these advantages, Gilbert cell-based mixers have higher noise figures (14 dB vs. 7 dB typ.) and lower intercept points (less linearity) than do diode-based mixers.
A second implementation for active mixers is the cascode type. Also referred to as a transconductance mixer, this topology ultimately can be traced back to dual gate designs. The resulting mixer is single balanced, has conversion gain, and typically has lower a noise figure than a Gilbert cell structure. Although in this structure the mixer itself provides no isolation, the matching structures used on the RF and LO ports are high pass, while the match on the IF port is low pass, so the isolation of the overall structure can be better than that of a balanced mixer (40 dB vs 30 dB). These features make the cascode mixer ideal for down-converting applications. |