A useful spur plan starts with the frequency-domain budget, not with adding filters to whatever architecture was already drawn. List every f_RF = |m f_LO ± n f_IF| combination up to the highest harmonic order your mixer and LO chain can realistically produce, then mark which terms land in the RF passband, IF passband, image band, or a port where filtering is actually possible. The key interface constraint is that an in-band spur cannot usually be removed after it has been translated into the desired IF; at that point it shares the signal path and any filter that rejects it may also reject the wanted channel. That is why changing IF placement, LO range, or high-side/low-side architecture should generally be evaluated before spending effort on LO leakage cleanup.

LO leakage and harmonics still matter, but they answer a different question. Leakage sets port-to-port isolation, DC offsets, radiated emissions, and possible self-reception, while LO harmonics determine the amplitude of higher-order mixing products. Once the architecture places the dangerous m,n products outside accessible bands, then specify harmonic filtering at the LO port, RF filtering before the mixer, and IF filtering after it using defined reference planes. Measure each contribution separately: LO leakage with the IF terminated and no RF stimulus, harmonic levels at the actual mixer LO pin rather than only at the synthesizer output, and spur conversion with calibrated port powers so the result is not confused by cable loss or fixture gain.

If a spur remains in-band, do not treat a better mixer as the only option. Recheck whether the IF choice creates a folding condition, whether the LO plan crosses a spur line during tuning, and whether broadband noise or clock harmonics from digital supplies are being mixed as additional quasi-LO terms. The decision should be based on a table of spur order, predicted level, rejection available at each physical interface, and measurement uncertainty. Only after the frequency plan is sound should leakage suppression be optimized, because leakage cleanup cannot repair a fundamental frequency collision. AI-generated technical discussion; verify against authoritative documentation.