How does an active FET mixer differ from a passive diode mixer?

A passive diode mixer uses Schottky diodes acting as voltage-controlled switches. Because it is passive, it suffers from conversion loss, typically 6 to 8 dB. That means if a minus 50 dBm RF signal enters, a minus 58 dBm IF signal exits, degrading the receiver's noise figure. An active FET mixer uses transistors that consume DC power. The transconductance of the FETs amplifies the signal during the mixing process, resulting in conversion gain (typically 5 to 15 dB). A minus 50 dBm RF signal might exit as a minus 40 dBm IF signal. Active mixers also require significantly less LO drive power (e.g., 0 dBm instead of the plus 7 to plus 15 dBm required by diode rings), simplifying the LO synthesizer design.

What is a Gilbert cell mixer?

The Gilbert cell is the standard double-balanced active mixer topology used in almost all modern RFICs and CMOS transceivers. It consists of a lower transconductance stage (two FETs) that converts the input RF voltage into an RF current, and an upper switching quad (four FETs) driven by the LO signal. The LO switches steer the RF current back and forth between the differential output loads at the LO frequency, effectively multiplying the RF current by a square wave. The double-balanced nature provides excellent isolation between the RF, LO, and IF ports, meaning the strong LO signal does not leak heavily into the IF output. Because it can be implemented entirely with transistors (no bulky baluns or transformers required), it is perfect for monolithic silicon integration.

Why are passive mixers still used if active mixers provide gain?

Active mixers have three significant drawbacks. First, they add their own active noise to the signal. While a passive mixer's noise figure is roughly equal to its conversion loss (e.g., 7 dB), an active mixer with 10 dB of gain might have a noise figure of 12 dB or worse due to flicker noise (1/f noise) and thermal noise from the biasing circuitry. Second, active mixers have much lower linearity (lower IP3) because the transistors compress under strong input signals. A high-level passive diode mixer can handle plus 20 dBm inputs without severe distortion, while a CMOS active mixer might compress at minus 10 dBm. Third, they consume DC power. For base station receivers where dynamic range and linearity are critical, passive mixers are still preferred. For battery-powered mobile devices where integration and low LO power matter most, active Gilbert cells dominate.

Active Components

Active FET Mixer

A smartphone transceiver chip must downconvert a 2.4 GHz WiFi signal to baseband. A traditional passive diode ring mixer would cause 7 dB of signal loss, require a bulky external transformer, and demand +10 dBm of power from the local oscillator—draining the battery rapidly. Instead, the RFIC designer uses an active FET mixer based on the Gilbert cell topology. It provides 10 dB of conversion gain (boosting the signal instead of losing it), requires only 0 dBm of LO drive, and integrates perfectly onto a microscopic square of CMOS silicon without magnetic components. While it sacrifices some linearity compared to passive diodes, the active mixer's combination of gain, low LO drive, and monolithic integration makes it the absolute standard for all modern consumer wireless silicon.

Category: Active Components

Advantage: Conversion gain, low LO drive

Standard Topology: Gilbert Cell

Active FET vs. Passive Diode Mixer

Parameter	Active FET (Gilbert Cell)	Passive Diode Ring
Conversion Gain/Loss	+5 to +15 dB (Gain)	−5 to −9 dB (Loss)
LO Drive Required	−5 to +5 dBm	+7 to +23 dBm
Linearity (IIP3)	Low to Moderate (0 to +10 dBm)	High (+15 to +30 dBm)
Noise Figure	High (10 to 15 dB, high 1/f noise)	Low (Equal to conversion loss, ~7 dB)
DC Power Consumption	10 to 50 mW	0 mW (Passive)
Silicon Integration	Excellent (no transformers needed)	Poor (requires bulky baluns)

Gilbert Cell Conversion Gain:
G_c ≈ (2/π) · g_m · R_L
Where g_m is the transconductance of the RF input FETs, and R_L is the load resistance. The 2/π factor comes from the Fourier fundamental of the square-wave switching action.

System Noise Figure Impact:
NF_sys = NF_mixer + (NF_next − 1) / G_c
Because G_c is a gain (>1) rather than a loss (<1), the noise contribution of subsequent IF stages is suppressed.

Common Questions

Frequently Asked Questions

How does it differ from a passive mixer?

Passive diode mixers have conversion loss (~7 dB) and require very high LO drive power (+15 dBm) to switch the diodes. Active FET mixers use biased transistors to provide conversion gain (+10 dB), amplifying the signal during frequency translation, and they require much less LO power (0 dBm) because the LO only drives FET gates.

What is a Gilbert cell?

The Gilbert cell is the dominant active mixer topology. A lower pair of FETs (transconductor) converts the RF voltage into current. An upper quad of FETs, driven by the LO, acts as commutating switches, steering that current back and forth into the load. This multiplies the RF current by the LO square wave, generating the IF.

Why use passive mixers if active ones have gain?

Linearity and noise. Active FETs compress under strong signals, giving them lower IP3. They also generate high thermal and flicker (1/f) noise, making their noise figure worse than their passive counterparts. Base stations prioritize linearity and use passive mixers. Smartphones prioritize battery life, integration, and gain, so they use active Gilbert cells.

Related Terms

← ACLR (ET) Active Load Modulation →

Receiver Design

Mixer Topology Selector

Enter your system linearity requirements, available LO power, and noise figure constraints. Compare cascaded performance between active Gilbert cells and passive diode rings for your specific transceiver architecture.

Select Mixer Topology