Nulling IQ Modulator LO Leakage with Differential Offset Controls

A tone at an IQ modulator’s local-oscillator frequency can remain even when the intended signal is offset from the carrier. Reducing that leakage requires the correct adjustment. Changing both sides of a differential input together is not the same as correcting differential offset, and image-rejection controls address a different imperfection.

Identify the observed spectral component

The Analog Devices application note AN-1039 explains that I and Q differential offsets and direct coupling contribute to LO leakage. It distinguishes that carrier component from the unwanted sideband caused by gain and quadrature imbalance.

Apply a documented test signal and identify the wanted sideband, carrier-frequency component, and image separately. Record LO frequency, drive level, baseband amplitude, common-mode voltage, and output loading.

Use the intended offset controls

Determine which DAC registers or analog adjustments change differential I and Q offset. Confirm their range, polarity, and resolution. Keep baseband common-mode levels within the modulator’s specified operating conditions.

Read back or otherwise verify the applied setting. A software field labeled offset may instead alter an unrelated digital stage, so confirm its effect on the actual differential input.

Search the null reproducibly

Sweep one offset control while holding the other fixed, then refine the second control and repeat as needed. Record carrier power across the explored settings rather than retaining only the best point.

Check that the analyzer has enough dynamic range to resolve the residual. A flat minimum can reflect the measurement floor rather than a broad physical null.

Verify the complete operating range

Repeat at the required LO frequencies and temperature conditions. A setting that minimizes leakage at one point may not remain optimal elsewhere.

For example, a product using several frequency bands may need separate stored settings or a documented residual allowance between calibration points. Interpolation should be validated rather than assumed to follow a simple trend.

After correction, remeasure wanted output, image level, and relevant modulation quality. Confirm that the procedure has not pushed the baseband path toward clipping or changed an unintended operating parameter.

Test startup and configuration restoration so the calibrated offsets are applied before the product relies on the corrected output. Include coefficient identity and units in the stored record.

Save the spectral component map, offset sweep, measurement floor, and operating-range verification. The result should establish a repeatable carrier-leakage reduction with the correct controls. A lower center-frequency marker at one temperature is useful evidence, but it does not establish image rejection or modulation performance throughout the product’s range.