What Is a Chopper-Stabilized (Zero-Drift) Op-Amp, and When Do You Need One?

Detailed Explanation

Every real op-amp has an input offset voltage, covered in the foundational what is an op-amp page, a small internal mismatch that behaves as if a tiny DC voltage were added between the two inputs even when they're at the same potential. A standard precision op-amp reduces this at manufacture through laser trimming, reaching offset voltages in the tens of microvolts range on the best conventionally trimmed parts, but that trimmed value isn't fixed forever: it drifts with temperature, and it drifts further, more slowly, over the device's operating life as the silicon itself ages. For a DC or slowly varying precision measurement amplified by a large closed-loop gain, that drift can become a meaningful fraction of the measurement itself, especially in a product that must hold its calibration across a wide temperature range for years without recalibration.

A chopper-stabilized (zero-drift) op-amp solves this differently: instead of relying on a single trimmed value that then drifts, it continuously re-measures its own offset internally, many times per second, and applies a correction in real time. Because the correction is continuously refreshed rather than fixed at manufacture, both the offset itself and its drift with temperature and time are held to a level (commonly under 1 µV of offset and a small fraction of a µV/°C of drift on the best parts) that a conventionally trimmed op-amp, however well trimmed initially, cannot sustain over the same operating conditions.

How Chopping Works

The core mechanism uses an internal switching network to alternate the amplifier's inputs between the actual signal and a zeroed reference state at a fixed internal chopping frequency, measuring the offset during the zeroed state and applying a correction during the signal state. Some parts implement this as pure chopping, others combine it with auto-zeroing (a related but distinct periodic-correction technique), or layer both techniques together specifically to reduce the residual switching artifact chopping alone leaves on the output, covered in the FAQ above. The specific architecture varies by manufacturer and part family, so the datasheet's noise and ripple specifications, not just the headline offset figure, matter when comparing candidate parts.

Practical Examples

A precision current-shunt measurement using a very low-value shunt resistor (to minimise power dissipation) produces a very small voltage across the shunt at typical currents, which then needs a large amplifier gain to reach a usable ADC input range. See measuring current with a shunt resistor for the base measurement principle this amplification stage builds on. A standard op-amp's offset voltage, multiplied by that large gain, can become a significant fraction of the reading at low currents; a zero-drift part keeps that error small enough that the measurement stays accurate across the shunt's full intended current range.

A thermocouple or strain-gauge bridge measurement that must hold calibration across an industrial product's full temperature range for years without field recalibration is a classic zero-drift application. See temperature sensor error budget and calibration for the broader error-budgeting approach a zero-drift amplifier stage should be evaluated within, alongside the sensor's own accuracy specification.

Design Considerations

• Run the actual error-budget calculation before defaulting to a zero-drift part. Confirm what a standard precision op-amp's offset and drift specification, amplified by the actual required closed-loop gain and evaluated across the product's real temperature range and service life, would contribute to total measurement error. Only move to a zero-drift part when that number is unacceptable for the application, since zero-drift parts typically cost more and, per the FAQ above, bring their own noise and ripple trade-off.
• Check the datasheet's chopping frequency and residual ripple specification against the signal path's noise budget, particularly in a design where the amplifier shares a board or supply with sensitive high-frequency circuitry. A part's zero-drift DC performance doesn't guarantee it is quiet at all frequencies.
• Verify the part's bandwidth and slew rate are still adequate for the application, since some (not all) zero-drift architectures trade off some AC performance for their DC accuracy advantage; don't assume a zero-drift part is a drop-in upgrade for every characteristic of an existing precision op-amp design.
• Zeus Design designs precision analog front ends, including zero-drift amplifier stages for current-sensing and sensor-interface applications, as part of complete product electronics development.

Common Mistakes

• Reaching for a zero-drift part by default without confirming the application actually needs microvolt-level long-term accuracy. As covered in the FAQ above, the large majority of analog signal-conditioning applications are well served by a standard precision op-amp at lower cost and simpler design margin.
• Ignoring the chopping artifact's interaction with a sensitive nearby signal path, assuming a zero-drift part is simply "better" in every respect than a standard op-amp rather than checking its specific ripple and noise characteristics against the application.
• Comparing parts on headline offset voltage alone without checking drift over the product's actual temperature range and service life. A part with a slightly higher initial offset but meaningfully better drift performance can outperform a lower-offset part over the product's real operating conditions.
• Assuming AC performance is unaffected when substituting a zero-drift part into an existing design, without confirming bandwidth, slew rate, and any chopping-frequency-related datasheet caveats against the specific signal path's requirements.