What Is Amplifier Compression?
An ideal RF amplifier maintains constant gain regardless of input power. In practice, as input power increases, transistor nonlinearities cause the output to grow more slowly than expected — the amplifier compresses. The 1 dB compression point (P1dB) is the input (or output) power at which the small-signal gain drops by exactly 1 dB from its linear value.
P1dB Definition
Two conventions are used:
| Notation | Reference Plane | Typical Range (LNA) | Typical Range (PA) |
|---|---|---|---|
| IP1dB (input) | Amplifier input port | −20 to −5 dBm | +10 to +30 dBm |
| OP1dB (output) | Amplifier output port | −10 to +5 dBm | +20 to +40 dBm |
Relation: OP1dB = IP1dB + G − 1 dB where G is small-signal gain.
Visualizing Compression on an S21 Plot
For small-signal S-parameter measurements (linear regime), S21 magnitude is flat with input power. The compression curve is swept separately using a signal source + power meter. However, RF View can overlay S21 magnitude across frequency to verify the small-signal gain baseline, and the S21 phase slope indicates group delay linearity.
AM-AM and AM-PM Distortion
Compression introduces two distortion mechanisms:
- AM-AM: Output amplitude deviates from linear — gain is amplitude-dependent
- AM-PM: Output phase shifts with input amplitude — causes EVM degradation in modulated signals
AM-PM is especially harmful in QAM systems. A 1° phase shift at high power can degrade EVM by 1–2%.
Back-Off Strategy
To keep a PA linear, designers operate below P1dB by a back-off margin:
| Modulation | PAPR (dB) | Required Back-off from P1dB |
|---|---|---|
| QPSK | 3–4 | 4–6 dB |
| 16-QAM | 6–8 | 8–10 dB |
| 64-QAM | 8–10 | 10–12 dB |
| OFDM (LTE) | 10–12 | 12–15 dB |
Relationship to IP3
The third-order intercept point (IP3) is related to P1dB by approximately:
IIP3 ≈ IP1dB + 9.6 dB (theoretical for memoryless polynomial)
In practice the offset is 8–12 dB depending on device technology and bias point.
Using RF View for PA Analysis
RF View reads S-parameter files from PA characterization. Load a 2-port .s2p file, read S21 for gain, S11/S22 for input/output match. Compare files measured at different drive levels to observe gain compression trends. The batch SNP processor lets you load 10–20 power-swept files simultaneously for compression analysis across frequency.