Band-limited white noise is generated covering the device bandwidth. A narrow notch (typically 1 to 5 percent of the total bandwidth) is created using a sharp bandstop filter. This notched noise signal is applied to the device under test. At the output, the power spectral density is measured both in the noise band and in the notch. NPR = P_noise / P_notch (in dB). The notch depth at the input should be at least 35 to 40 dB to ensure the measurement is limited by device distortion, not the test setup. NPR is plotted versus input drive level: at low drive, NPR increases with power (noise floor limited). At optimum drive, NPR peaks. Above optimum, NPR decreases as distortion increases. The peak NPR and the input power at which it occurs are the key specifications.

How does NPR relate to other distortion metrics?

NPR is equivalent to C/IM (carrier to intermodulation ratio) for a multi-carrier scenario with a very large number of carriers. For N equal carriers: C/IM approximately equals NPR minus 10 log(N minus 1). NPR captures all orders of intermodulation simultaneously, making it more representative of real-world multi-carrier operation than two-tone IMD testing. For third-order dominated systems: NPR is approximately equal to 2 times (OIP3 minus P_out) minus 10 log(N) plus constant, where N is the number of effective carriers. NPR is also related to EVM in digital systems: higher NPR generally correlates with lower EVM.

Satellite transponders: the TWTA or SSPA must amplify multiple carriers simultaneously. NPR characterizes the transponder's ability to handle multi-carrier loading without excessive intermodulation. Typical TWTA NPR: 15 to 20 dB near saturation, 20 to 25 dB with 3 to 5 dB output back-off. SSPA: 25 to 35 dB. Wideband communication amplifiers: CATV amplifiers carrying hundreds of channels are specified by composite second order (CSO) and composite triple beat (CTB), which are related to NPR. DAC testing: NPR measures the spurious-free dynamic range of wideband digital-to-analog converters. Multi-carrier base stations: NPR-like measurements verify transmitter linearity under realistic multi-carrier loading.

Distortion Metric

NPR

/en-pee-ahr/ — Noise Power Ratio

Measures wideband intermodulation: apply band-limited noise with narrow notch, measure distortion filling the notch. NPR = P_noise/P_notch (dB). Higher = better linearity. Captures all IMD orders simultaneously (more realistic than two-tone). Satellite TWTA: 15-20 dB near saturation, 20-25 dB with back-off. SSPA: 25-35 dB. C/IM ≈ NPR − 10log(N−1). ITU-R SM.1446.

TWTA: 15-25 dB

SSPA: 25-35 dB

Method: Notched noise

Understanding NPR

NPR is the most realistic measure of wideband amplifier linearity because it simulates actual multi-carrier or multi-user loading. Unlike two-tone IMD testing (which captures only third and fifth-order products), NPR captures all orders of intermodulation distortion simultaneously, providing a single metric that directly predicts system performance under real operating conditions.

The measurement concept is elegant: fill the amplifier bandwidth with noise (representing many signals), carve out a narrow notch, and measure how much distortion power leaks into the notch. The more linear the amplifier, the deeper the notch remains at the output. This directly relates to the carrier-to-interference ratio that users will experience in a multi-carrier satellite or cellular system.

NPR Equations

NPR definition:
NPR = P_noise(out)/P_notch(out) (dB)
Notch depth at input: ≥35 dB

Relation to C/IM:
C/IM ≈ NPR − 10log(N−1)
N = number of equal carriers
NPR=20 dB, N=10: C/IM=10.5 dB

NPR from OIP3:
NPR ≈ 2(OIP3−P_out)−10log(N)+K
3rd-order dominated regime

Optimal input drive:
Peak NPR occurs at optimum
back-off from saturation

NPR by Device Type

Device	NPR (peak)	Back-off	Bandwidth	Application
TWTA	15-20 dB	0-3 dB	36-72 MHz	Satellite bent-pipe
Linearized TWTA	20-25 dB	1-3 dB	36-72 MHz	HTS satellite
GaN SSPA	25-35 dB	3-6 dB	100-500 MHz	Satellite, BTS
CATV amplifier	40-50 dB	15-20 dB	50-1000 MHz	Cable distribution
DAC	50-70 dB	N/A	1-4 GHz	Wideband TX

Common Questions

Frequently Asked Questions

How measured?

Band-limited white noise with narrow notch (1-5% of BW, ≥35 dB depth). Apply to DUT. Measure output: NPR = noise PSD / notch PSD. Plot vs. input drive: low power = noise-limited, peak = optimum, high power = distortion-limited. Peak NPR and its drive level are key specs.

Relation to other metrics?

C/IM ≈ NPR − 10log(N−1), N = carriers. Captures all IMD orders (not just IM3/IM5). More realistic than two-tone. For 3rd-order: NPR ≈ 2(OIP3−Pout)−10logN+K. Higher NPR correlates with lower EVM in digital systems.

Where used?

Satellite transponders: TWTA/SSPA multi-carrier loading. TWTA: 15-20 dB near sat, 20-25 dB backed off. SSPA: 25-35 dB. CATV: CSO/CTB related to NPR. DAC: SFDR measurement. Multi-carrier base stations: realistic TX linearity under multi-carrier loading.

Related Terms

← Noise Factor N-Type Connector →

RF Testing

Request a Quote

Need NPR measurement, transponder characterization, or PA linearity analysis? Contact our team.

Get in Touch