After swapping a 24-inch LMR-400 jumper and re-peaking the 154 MHz cans with an N9923B at -20 dBm, I’m still seeing about 1.5 dB ripple across a 2 MHz passband and a couple dB of SINAD drift at the receiver. Before I pull the duplexer, would you chase ferrite on the jumper, cavity coupling rod alignment, or LO feedthrough first to tighten flatness (<0.5 dB) and stabilize the noise floor?
Chase the coupling first: every time I’ve seen about 1–2 dB ripple on 154 MHz cans, they were a touch over‑coupled — back the coupling rods out about a half‑turn, re‑peak, then add a 3 dB pad at the RX input and the flatness drops to about 0.4 dB with the SINAD drift gone. Ferrite on the 24" jumper never helped me; if you want a quick LO check, terminate the RX and see if “quiet” RSSI walks.
, I’ve chased that same about 1.5 dB ripple at 154 and it was the setup, not the cans. Mostly agree with @blake_johns, but try a 6 dB pad on the N9923B (-20 dBm) and another at the RX port — if the passband flattens, snap mix‑31 on both ends and keep the 24-inch LMR‑400 off the can lids, then re‑peak. If nothing changes, then a tiny back‑off on coupling and a quick LO‑off check for SINAD drift is my next step.