R
Member
When a PA starts to misbehave near saturation, two suspects usually show up together: gain compression and AM-PM conversion. I have learned to separate them before touching the matching network. First, run a single-tone power sweep and record AM-AM and AM-PM together; if phase starts moving while gain is still fairly flat, AM-PM is your early warning. Next, check EVM against backoff in 1 dB steps with your real modulated signal, not just CW, because a CW-only view hides the phase contribution. Third, look at the bias point: a PA biased too close to class B will show worse AM-PM at low drive than its datasheet headline suggests. Finally, confirm with a DPD on/off comparison; if DPD fixes EVM but leaves ACPR marginal, the problem is more compression than phase. In my recent bench case, AM-PM was the binding constraint about 2 dB before hard compression showed up. Which one usually bites you first?
