Dynamic Range: Spurious-Free DR, BDR, Linear DR

Why Dynamic Range Matters in Receiver Architecture

A receiver spec sheet rarely lists one dynamic range number. Instead you find three: spurious-free dynamic range (SFDR), blocking dynamic range (BDR), and linear or noise-limited dynamic range. Each describes a different failure mode, and a receiver optimized for one is not automatically good at the others. Understanding which number governs your application is the difference between a design that works on the bench and one that falls apart the first time it sees a crowded band.

Signal detection floor vs. saturation ceiling

Every receiver lives between two limits. At the bottom is the noise floor — the level below which a wanted signal is indistinguishable from thermal and electronic noise. At the top is the saturation ceiling — the input level at which amplifiers and mixers stop behaving linearly and start generating compression, gain reduction, and intermodulation products. Dynamic range is the usable span between these two limits, but "usable" means different things depending on what's degrading the signal: distortion products from other strong signals, compression from a single blocker, or just noise.

Role of dynamic range in link budget

In a link budget, dynamic range sets the margin between the weakest signal you must detect and the strongest signal (wanted or interfering) the front end must tolerate without corrupting that detection. A receiver with 10 dB of margin at the bottom and 10 dB of headroom at the top on paper can still fail if the headroom is consumed by a single strong adjacent-channel blocker, or if two moderate interferers combine to produce a spurious tone sitting exactly on the wanted channel. The link budget needs to account for which of these mechanisms applies.

Impact on system sensitivity and interference immunity

Sensitivity is a noise-floor problem; interference immunity is a linearity problem. A system can have excellent sensitivity in an anechoic test and still desensitize or generate false detections in the field because real RF environments are full of strong out-of-band and in-band emitters. That's why SFDR, BDR, and linear DR are reported separately — a single aggregate "dynamic range" number would hide which failure mode is dominant.

Spurious-Free Dynamic Range (SFDR)

Definition and measurement methodology

SFDR is measured with a two-tone test: two equal-amplitude CW tones, closely spaced in frequency, are injected at the receiver input. As their level increases, third-order intermodulation products appear at 2f1−f2 and 2f2−f1, falling in-band near the original tones. SFDR is the span, in dB, between the noise floor and the input level at which these third-order products just reach the noise floor. Below that input level, the spurs are buried in noise and don't matter; above it, they become visible, potentially masquerading as real signals.

Relationship to third-order intercept point (IIP3)

SFDR is derived algebraically from IIP3 and noise floor, not measured as an independent compression point. Because third-order products grow at three times the rate (in dB) of the fundamental tones, the standard relationship is:

SFDR (dB) = (2/3) × (IIP3 − NF_floor)

where IIP3 and the noise floor are expressed in dBm referred to the same reference plane. A 10 dB improvement in IIP3 buys roughly 6.7 dB of SFDR — not a 1:1 trade. This is why linearity budgeting across a chain, covered in detail in Cascade IIP3: System Linearity from Gain and IIP3 Budget, has an outsized effect on SFDR specifically.

Noise floor, signal tone, and spurious products

The noise floor itself is set by kTB, receiver noise figure, and bandwidth. Narrow the bandwidth and the noise floor drops, which — for a fixed IIP3 — increases SFDR. This is a key reason wideband receivers (software-defined radios covering hundreds of MHz) report worse SFDR than narrowband channelized receivers with the same front-end components: the wideband noise floor is simply higher.

SFDR vs. receiver noise figure trade-off

Improving noise figure (more gain, lower-NF LNA) pushes the noise floor down and improves SFDR, but added gain ahead of a mixer or ADC also compresses the effective input IIP3, since later stages see larger signals. This is the central SFDR trade-off: NF and IIP3 pull against each other through gain distribution, and neither can be optimized in isolation.

Blocking Dynamic Range (BDR)

In-band and adjacent-channel blocking scenarios

BDR describes how large an interfering signal — in-band or adjacent-channel — a receiver can tolerate while still demodulating a much weaker wanted signal. Unlike SFDR, this is often a single-tone problem: one strong blocker, not two tones generating a third-order product.

1 dB compression point and gain compression

The practical ceiling for BDR is the 1 dB compression point (P1dB) of the front-end chain. As a blocker approaches P1dB, gain starts to compress, reducing the effective gain seen by the wanted signal and raising the noise figure referred to the input. BDR is typically quoted as the ratio between the minimum detectable signal and the blocker level that causes a specified amount of sensitivity degradation (commonly 1 dB or 3 dB desense), not necessarily full compression.

Blocking vs. desensitization mechanisms

Blocking and desensitization are related but distinct: desensitization is the gradual loss of effective gain/NF as a strong out-of-band signal pushes the LNA or mixer toward compression, while blocking failure is the point at which the wanted signal is no longer recoverable at all — either because gain has collapsed or because the blocker has driven a stage into hard saturation, generating broadband noise and spurious mixing products of its own.

Filter selectivity impact on BDR

Because BDR is dominated by single large interferers, front-end filtering has a much larger effect here than on SFDR. A preselector or duplexer that attenuates an adjacent-channel blocker by 20 dB directly adds 20 dB of effective BDR, since the blocker never reaches the active devices at full strength. This is why BDR specs are frequency- and channel-plan-dependent: the same radio can have very different BDR numbers in-band versus one channel away, purely due to filter rolloff.

Linear (Noise-Limited) Dynamic Range

Thermal noise floor and minimum detectable signal

Linear DR, sometimes called noise-limited dynamic range, is the simplest of the three: it's the ratio between the maximum input level the receiver can handle while remaining linear (usually P1dB or a defined compression threshold) and the minimum detectable signal set by thermal noise (kTB), NF, and required SNR for detection.

Linearity region before nonlinear distortion

Within this range, output power tracks input power 1:1 in dB — no compression, no significant intermodulation. Linear DR says nothing about spurious products from multiple tones; it only describes single-signal headroom.

Relationship to noise figure and available gain

Linear DR = P1dB(input-referred) − MDS, where MDS (minimum detectable signal) = kTB + NF + required SNR. Both halves of this equation move with gain distribution: more gain lowers MDS (better NF) but also lowers input-referred P1dB, so, as with SFDR, there is no free lunch — you cannot simultaneously maximize headroom and sensitivity from the same gain stage.

Comparison with SFDR and BDR

Linear DR is typically the largest of the three numbers because it imposes the loosest condition — it doesn't care about third-order products (SFDR's concern) or desensitization from out-of-band blockers (BDR's concern), only single-tone compression. A receiver can have excellent linear DR and mediocre SFDR if its IIP3 is poor relative to its P1dB, which happens in some mixer architectures.

System-Level Trade-Offs and Budget Analysis

Cascading gain and linearity through receiver stages

In a cascade, NF is dominated by the first stage (Friis' equation) while IIP3 is dominated by the last, highest-gain stage referred back to the input — the two figures of merit are set by opposite ends of the chain. This asymmetry is precisely why you can't simultaneously maximize all three dynamic range metrics: pushing gain early to minimize NF (helping linear DR and SFDR's noise term) degrades the input-referred IIP3 of later stages (hurting SFDR and BDR).

Front-end LNA vs. mixer linearity contributions

LNAs are usually chosen for low NF at acceptable IIP3; mixers are usually the weak link for IIP3 because of their inherent nonlinear operation (switching or multiplying). In most superheterodyne or direct-conversion receivers, the mixer's IIP3, attenuated by preceding LNA gain, dominates the cascaded IIP3 budget — a relationship quantified using two-port parameters and gain/noise cascading methods described in Two-Port Network Parameters: Z, Y, H, ABCD.

AGC impact on dynamic range measurement

Automatic gain control complicates all three measurements because it changes gain distribution dynamically in response to input level, which shifts both NF and effective IIP3 referred to the antenna. SFDR, BDR, and linear DR specs should state whether AGC was frozen, free-running, or at a defined gain state during the test; comparing numbers measured under different AGC conditions is a common source of spec-sheet confusion. The loop dynamics themselves — how fast gain adjusts and at what thresholds — are covered in Automatic Gain Control (AGC): Loop, Attack, Release Time.

Design choices: noise vs. linearity vs. integration

System designers choosing between discrete front ends and highly integrated RF ICs are implicitly choosing a point on the noise/linearity/integration trade curve. Integrated solutions often sacrifice some IIP3 headroom for lower power and board area; discrete LNA/mixer/filter chains can be tuned stage-by-stage for a specific BDR or SFDR target at the cost of size and cost.

Measurement and Specification in Practice

Test signal configurations for each metric

SFDR uses two closely-spaced equal-level CW tones. BDR uses one CW blocker plus one low-level wanted signal, often modulated, with sensitivity degradation measured directly. Linear DR uses a single swept CW tone against the noise floor, with no second tone involved. Confusing these configurations when reading a datasheet — assuming a quoted "dynamic range" is SFDR when it was measured with a single tone — is a frequent source of design error.

Standards and conformance requirements

Cellular, radar, and EW receiver specifications (3GPP, MIL-STD, or similar domain standards) define specific blocker offsets, levels, and desense thresholds for BDR-type tests, and specific tone spacings for SFDR-type tests. Always confirm which standard's test conditions a datasheet number was measured against before comparing parts.

Frequency-dependent effects and bandwidth

All three metrics are bandwidth-dependent through the noise floor term, and IIP3/P1dB are frequency-dependent across a receiver's tuning range. A number quoted at one frequency and bandwidth should not be assumed to hold at another; narrowband channel filtering downstream, including in the digitization stage, materially changes effective SFDR, which is one reason ADC-referred SFDR is specified separately, as discussed in ADC in Receiver: ENOB, SFDR, Sampling Rate Requirements.

Temperature and supply voltage sensitivity

NF, IIP3, and P1dB all drift with temperature and supply voltage — NF typically worsens and IIP3 typically degrades at temperature extremes or reduced supply headroom. Dynamic range specs taken at room temperature and nominal supply should be treated as best-case; field performance over the full operating range can be several dB worse on all three metrics simultaneously.