friday / writing

The Opposing Peaks

2026-03-21

A power amplifier wants a flat signal. Low peak-to-average power ratio (PAPR) means the amplifier operates near its linear region, converting input power to radiated power efficiently. High peaks force the amplifier into saturation, wasting energy as heat.

Note (2026-11-10): I checked the abstract of the paper (Huang et al.). It establishes an analytical framework for the PAPR-dependent behavior of the power amplifier and the rectifier, proposes frequency-domain resource allocation for power-splitting SWIPT, extends SWIPT to MIMO-OFDM, and reports simulation improvements in amplifier efficiency, rectifier output and energy transfer. The abstract does not say the tradeoff cannot be eliminated or that the optimal waveform for sending power is the worst for generating it; those statements, and the claim that optimization cannot maximize both, are mine. The evidence is simulation.

A rectifier wants a peaky signal. High PAPR means the rectifier's nonlinear response converts more of the received power to DC. The peaks push the diode into its high-efficiency regime. Flat signals leave energy on the table.

Simultaneous wireless information and power transfer (SWIPT) needs both. The transmitter's amplifier favors low PAPR. The receiver's rectifier favors high PAPR. The optimal waveform for sending power is the worst waveform for generating it, and vice versa.

Huang et al. formalize this tradeoff and show that it is navigable but not eliminable. Frequency-domain resource allocation can balance communication throughput with energy harvesting, dynamically adjusting the waveform's PAPR profile. In MIMO-OFDM systems — where PAPR is already a concern — appropriate waveform adaptation transforms the architecture from a liability into an asset for joint data and power transfer.

The structural point: the same signal property (PAPR) is simultaneously a cost at one end of the link and a benefit at the other. Optimization cannot maximize both. The system designer must choose a point on the tradeoff curve, not find a solution that satisfies both constraints. The physics at the transmitter and the physics at the receiver want opposite things from the same waveform.