Power, Linearity, and Distortion Intermodulation and Spurious Informational

What is the effect of amplifier nonlinearity on adjacent channel leakage ratio?

Q: How do I measure ACLR?

Use a spectrum analyzer or VSA: set the channel bandwidth to match the signal (e.g., 20 MHz for LTE). Measure the integrated power in the desired channel. Measure the integrated power in the adjacent channel (offset by the channel spacing). ACLR = P_desired - P_adjacent (in dB). Ensure the spectrum analyzer dynamic range exceeds the ACLR requirement (use low noise floor settings and appropriate attenuation).

Q: Can I improve ACLR without DPD?

Yes, through: better PA topology (Doherty, envelope tracking: these improve efficiency at back-off without degrading linearity), higher OIP3 PA (allows operating closer to P1dB while meeting ACLR), output filtering (a bandpass filter after the PA removes out-of-band energy, improving ACLR at the expense of filter insertion loss), and crest factor reduction (CFR: reduces the signal PAPR, allowing higher average power without compression).

Q: What about EVM and ACLR simultaneously?

Both EVM and ACLR degrade with compression, but at different rates. ACLR typically becomes the binding constraint before EVM for lower-order modulations (QPSK, 16-QAM). For higher-order modulations (64-QAM, 256-QAM): EVM may be the binding constraint. The PA designer must meet both simultaneously: operating point selection balances EVM and ACLR.

Amplifier nonlinearity causes spectral regrowth, which increases the Adjacent Channel Leakage Ratio (ACLR) by spreading signal energy into the adjacent frequency channels: (1) Mechanism: a modulated signal has a defined bandwidth (e.g., 20 MHz for LTE). When this signal passes through a nonlinear amplifier: the third-order nonlinearity generates IM3 products that extend the signal spectrum beyond its original bandwidth. The IM3 products fall in the adjacent channels (first adjacent) and alternate channels (second adjacent). This broadened spectrum is called spectral regrowth. (2) ACLR definition: ACLR = P_channel / P_adjacent (in dB). Where P_channel = power in the desired channel, and P_adjacent = power leaked into the adjacent channel. Requirements: LTE/5G: ACLR < -30 dBc (first adjacent), < -33 dBc (second adjacent) (3GPP specifications). Wi-Fi 6: ACLR < -25 dBc (less demanding). FM broadcast: ACLR < -25 to -40 dBc. (3) ACLR vs back-off: backing off the PA (reducing the output power below P1dB) improves ACLR: 1 dB additional back-off improves ACLR by approximately 2 dB (because the IM3 drops 3 dB while the fundamental drops 1 dB). For LTE (ACLR < -30 dBc): typically requires 6-10 dB back-off from P1dB without DPD. With DPD: 2-4 dB back-off is sufficient (the DPD corrects the remaining nonlinearity). (4) ACLR and IP3: ACLR is directly related to the PA OIP3: ACLR (dBc) ≈ 2 × (OIP3 - P_out) - 10 × log10(BW_adj / BW_meas). Where OIP3 = output IP3, P_out = average output power, and BW terms adjust for the measurement bandwidth.

Category: Power, Linearity, and Distortion

Updated: April 2026

Product Tie-In: Amplifiers, Filters, Connectors

ACLR and PA Nonlinearity

ACLR is the primary spectral emission metric for modern wireless transmitters, directly impacting the coexistence of adjacent-channel users.

DPD Impact on ACLR

Without DPD: the PA must operate well below P1dB to meet ACLR. PAE at the back-off point: 5-15% (very inefficient). With DPD: the DPD applies an inverse nonlinearity: the PA input is pre-distorted so that the PA output is linear. The effective ACLR improves by 15-25 dB (DPD correction depth). The PA can operate 5-8 dB closer to P1dB, improving PAE to 25-40%. DPD bandwidth: the DPD must process 3-5× the signal bandwidth (to capture and correct the IM3 and IM5 spectral regrowth). For a 100 MHz 5G signal: the DPD bandwidth must be 300-500 MHz.

ACLR Parameters

ACLR = P_channel/P_adjacent (dBc)
1 dB back-off → ~2 dB ACLR improvement
LTE: ACLR < -30 dBc required
Without DPD: 6-10 dB back-off needed
With DPD: 2-4 dB back-off sufficient

Common Questions

Frequently Asked Questions

How do I measure ACLR?

Use a spectrum analyzer or VSA: set the channel bandwidth to match the signal (e.g., 20 MHz for LTE). Measure the integrated power in the desired channel. Measure the integrated power in the adjacent channel (offset by the channel spacing). ACLR = P_desired - P_adjacent (in dB). Ensure the spectrum analyzer dynamic range exceeds the ACLR requirement (use low noise floor settings and appropriate attenuation).

Can I improve ACLR without DPD?

Yes, through: better PA topology (Doherty, envelope tracking: these improve efficiency at back-off without degrading linearity), higher OIP3 PA (allows operating closer to P1dB while meeting ACLR), output filtering (a bandpass filter after the PA removes out-of-band energy, improving ACLR at the expense of filter insertion loss), and crest factor reduction (CFR: reduces the signal PAPR, allowing higher average power without compression).

What about EVM and ACLR simultaneously?

Both EVM and ACLR degrade with compression, but at different rates. ACLR typically becomes the binding constraint before EVM for lower-order modulations (QPSK, 16-QAM). For higher-order modulations (64-QAM, 256-QAM): EVM may be the binding constraint. The PA designer must meet both simultaneously: operating point selection balances EVM and ACLR.

Keep Reading