Transconductance Amplifier
A transconductance amplifier converts an input voltage into an output current rather than an output voltage, with the ratio between them — the transconductance gm, in siemens or milliamps per volt — set not just by fixed device geometry but often by a bias current the designer can vary in real time, which turns gm itself into a control knob and makes this topology the building block behind voltage-controlled filters, variable-gain amplifiers, and current-output DACs rather than a conventional voltage-gain stage.
Current output, not voltage output, is the defining difference from an op-amp. An op-amp's output stage is built to drive a low impedance and hold a commanded voltage regardless of load; a transconductance amplifier's output behaves like a current source with deliberately high output impedance, so it doesn't fight to hold a voltage — it pushes whatever current gm times the input voltage dictates into whatever impedance is connected, which is the exact behavior needed to directly charge or discharge a capacitor at a controlled rate rather than needing a separate current-sense-and-feedback loop.
Making gm itself adjustable is the real reason this topology exists. Because the same bias current that sets the input pair's operating point also sets its transconductance, varying that one bias current changes gm continuously without touching any other component, and that single tunable parameter is what makes a transconductance amplifier driving a fixed capacitor into a voltage-controlled low-pass or band-pass filter — the gm-C filter technique used throughout RF and analog IC design, including in commercial parts descended from the classic LM13700 operational transconductance amplifier.
It's the natural current source for a current-mode DAC output stage. Because the output is already a current proportional to an input signal, transconductance amplifiers map directly onto current-steering DAC architectures and current-output ADC front ends, where the signal needs to exist as a current all the way to a resistive or transimpedance output stage rather than being converted to voltage early and reconverted later — skipping that unnecessary voltage-current-voltage round trip is a meaningful precision and bandwidth win in high-speed data-converter design.
Voltage-controlled amplifiers and audio synthesizers were built on this circuit decades before "gm-C" became a chip-design term. Analog synthesizer VCAs, used by instruments built around chips like the LM13700 and the CA3080, exploit exactly the same bias-current-controls-gm mechanism to let one control voltage set the gain applied to an audio signal, which is the same physics now used inside modern RF transceiver ICs to build tunable channel-select filters — the application changed from analog music synthesis to gigahertz radio, but the underlying transconductance-as-a-knob trick did not.
Read a transconductance amplifier through a *voltage-controlled-knob* lens rather than a *weaker-op-amp* lens: the output isn't trying to be a well-behaved voltage source at all, it's a controllable current source whose gain constant can be dialed up or down by a bias current, which is precisely the property that turns a simple amplifier into a tunable filter, a variable-gain stage, or a clean current-mode link in a data converter.