Software-Defined Radio and Practical System Evaluation

Introduction and Context

Software-defined radio (SDR) is a radio architecture in which functions traditionally implemented by dedicated analog circuits or application-specific hardware are performed, wholly or partly, by programmable digital hardware and software. Modulation, demodulation, filtering, synchronization, channel coding, spectrum monitoring, and protocol processing can therefore be modified without redesigning the complete RF signal chain.

SDR is particularly important in postgraduate wireless engineering because it connects communication theory with experimentally verifiable systems. A link budget, waveform design, or channel model is incomplete unless its assumptions can be tested using a transmitter, receiver, channel emulator, measurement instrument, or field trial. SDR provides this bridge. It also exposes practical nonidealities—frequency offset, phase noise, nonlinear amplification, quantization, finite sampling rate, interference, latency, and imperfect synchronization—that are often hidden in ideal simulations.

A practical SDR experiment should be regarded as a controlled measurement study rather than merely a software demonstration. The experiment must define the research question, hardware configuration, waveform, channel conditions, metrics, calibration procedure, data-recording format, and acceptance criteria. Reproducibility is as important as achieving a favorable result.

Learning Objectives

After completing this chapter, a student should be able to:

1. Explain the architecture and signal flow of an SDR transceiver.

2. Select suitable RF front-end, converter, clock, and processing resources for an application.

3. Implement baseband DSP operations and identify their practical impairments.

4. Distinguish link-level simulation, system-level simulation, channel emulation, and over-the-air testing.

5. Calculate and interpret important KPIs such as received power, SNR, EVM, BER, throughput, latency, and outage probability.

6. Design a measurement campaign using laboratory instruments or drive testing.

7. Diagnose common RF, synchronization, DSP, and software problems.

8. Report results in a reproducible, technically defensible, and regulatory-compliant manner.

1. SDR Architecture and Signal Flow

A generalized SDR transceiver contains the following functional blocks:

• An antenna and RF filtering network.
• A low-noise amplifier (LNA) in the receive path.
• A frequency converter or direct-conversion receiver.
• Analog filtering and automatic gain control (AGC).
• An analog-to-digital converter (ADC).
• Digital downconversion, filtering, synchronization, and demodulation.
• Channel decoding and higher-layer processing.
• In the transmit direction, encoding, modulation, digital filtering, digital-to-analog conversion (DAC), frequency conversion, power amplification, and antenna filtering.

In a zero-intermediate-frequency architecture, the received RF signal is mixed directly to in-phase (I) and quadrature (Q) baseband components. In a low-IF architecture, the signal is first converted to a small intermediate frequency and digitally translated to baseband. Direct conversion offers integration and flexibility but is sensitive to DC offset, I/Q imbalance, and LO leakage.

The complex baseband representation is commonly written as

x(t) = I(t) + jQ(t),

where I(t) and Q(t) are real-valued orthogonal components and j is the imaginary unit. The analytic representation is useful because amplitude and phase can be processed independently, although the actual RF signal remains real-valued. For a carrier frequency f_c, a passband signal can be represented conceptually by

s_RF(t) = Re{x(t) exp(j2πf_c t)}.

This expression assumes that x(t) is a complex low-pass equivalent signal and that the bandwidth is sufficiently narrow relative to the carrier for the representation to be useful.

1.1 Sampling and bandwidth

The ADC sampling rate must support the occupied bandwidth and the chosen frequency-conversion architecture. For a real low-pass signal bandlimited to B Hz, ideal Nyquist sampling requires f_s > 2B. In practice, the rate is selected with margin to accommodate transition bands, anti-alias filtering, clock tolerances, and implementation constraints. A bandpass signal can sometimes be sampled below twice its highest RF frequency using bandpass sampling, but only when spectral replicas do not overlap and the analog front end adequately limits the input spectrum.

The ADC’s effective number of bits, full-scale voltage, input bandwidth, clock jitter, and dynamic range determine the quality of digitization. Quantization noise is not the only limitation: a strong interferer can saturate the converter even when the desired signal is weak, and clock jitter becomes more damaging as input frequency increases.

1.2 RF front-end considerations

The front end determines much of the practical performance of an SDR. Important properties include:

• Frequency range and instantaneous bandwidth.
• Maximum safe input power and receiver noise figure.
• Transmitter output power and spectral purity.
• Gain control range and gain-step resolution.
• Phase-noise performance and reference-clock stability.
• I/Q amplitude and phase imbalance.
• Isolation between transmitter and receiver.
• Spurious emissions, image rejection, and filter selectivity.

A front end is not an ideal frequency translator. Its gain may vary with frequency, its filters may distort the waveform, and its amplifier may become nonlinear near compression. The designer must therefore characterize the complete chain rather than relying only on nominal specifications.

2. Baseband DSP Processing Chain

A representative digital receiver performs the following operations:

1. Sample acquisition and metadata validation.

2. Digital downconversion and decimation.

3. Matched filtering or pulse-shaping filtering.

4. Timing recovery.

5. Carrier-frequency and phase recovery.

6. Frame detection and synchronization.

7. Channel estimation and equalization.

8. Demapping, deinterleaving, and channel decoding.

9. Packet checking and KPI extraction.

2.1 Pulse shaping and matched filtering

Raised-cosine and root-raised-cosine filters are widely used to control intersymbol interference. The roll-off factor α determines excess bandwidth. For symbol rate R_s, the ideal raised-cosine occupied baseband bandwidth is

B = (1 + α)R_s/2

for a one-sided low-pass representation. This relation assumes ideal Nyquist pulse shaping and defines the bandwidth between DC and the highest positive-frequency component. A root-raised-cosine filter is usually placed at both transmitter and receiver; their cascaded response is raised cosine.

A matched filter maximizes the signal-to-noise ratio at the sampling instant in additive white Gaussian noise when the receiver knows the pulse shape and timing is correctly aligned. In practice, filter truncation, timing error, frequency offset, and multipath reduce the ideal benefit.

2.2 Synchronization

Synchronization is often the difference between a waveform that works in simulation and one that works over the air. Major synchronization tasks include:

• Frame or preamble detection.
• Symbol-timing recovery.
• Carrier-frequency-offset estimation.
• Carrier-phase tracking.
• Sampling-clock-offset compensation.

A frequency offset Δf causes a complex-baseband phase rotation that evolves as

φ(t) = 2πΔf t + φ_0,

where φ_0 is an initial phase. Thus, even a small frequency error produces a large phase change over a long packet. Frequency offset arises from oscillator mismatch, Doppler shift, and frequency-converter error. Pilot symbols, repeated preambles, phase-locked loops, and decision-directed tracking are common remedies.

2.3 Channel estimation and equalization

A known training sequence or pilot pattern allows the receiver to estimate the channel response. For a frequency-domain subcarrier k, a simple estimate is

Ĥ[k] = Y[k]/X[k],

where X[k] is the transmitted pilot, Y[k] is the received pilot, and Ĥ[k] is the estimated complex channel response. This expression assumes that noise is not explicitly removed and that X[k] is nonzero. Interpolation is needed when pilots occupy only selected subcarriers.

A zero-forcing equalizer forms

Ŝ[k] = Y[k]/Ĥ[k],

but can strongly amplify noise where |Ĥ[k]| is small. A minimum mean-square error equalizer introduces a noise and signal-power term and usually provides a better compromise, especially near deep fades. The selected equalizer must be consistent with the channel model, pilot density, computational budget, and mobility.

3. From Simulation to Hardware

SDR development is most effective when organized into progressively realistic stages.

3.1 Algorithm or link-level simulation

Link-level simulation models a single transmitter-receiver link and estimates metrics such as BER, packet-error rate, throughput, or EVM. It is useful for comparing modulation, coding, equalization, and synchronization algorithms under controlled channels.

Typical assumptions include AWGN, Rayleigh fading, ideal oscillators, perfect timing, and perfect channel knowledge. These assumptions are valuable for establishing reference curves but can produce optimistic results. A valid simulation report must state which impairments are included and which are absent.

3.2 System-level simulation

System-level simulation models multiple users, cells, scheduling, interference, mobility, propagation loss, and resource allocation. It generally uses abstracted link-performance curves rather than individual samples. It answers questions such as coverage, capacity, fairness, handover performance, and interference coordination.

Link-level and system-level results should not be confused. A BER curve for one link does not directly establish network throughput, and a system-level throughput result may conceal the waveform or receiver assumptions used to generate its link abstraction.

3.3 Hardware-in-the-loop and channel emulation

A channel emulator applies controlled delay, Doppler, path loss, fading, and sometimes spatial correlation to real transmitted signals. It permits repeatable tests without the uncontrolled variability of an outdoor environment. Important emulator parameters include sampling rate, delay resolution, Doppler spectrum, number of paths, attenuation range, and synchronization method.

Hardware-in-the-loop combines real SDR hardware with simulated or emulated channels. This approach reveals converter, clock, buffer, and RF impairments while retaining controlled propagation conditions. It is especially appropriate for testing adaptive modulation, coding, synchronization, and mobility algorithms.

3.4 Over-the-air testing

Over-the-air (OTA) testing uses antennas and a physical propagation environment. It captures real multipath, blockage, interference, polarization mismatch, and installation effects. Its disadvantages are reduced repeatability, regulatory constraints, and difficulty separating causes of performance degradation. Shielded enclosures, conducted connections, calibrated attenuators, and channel emulators should be used where isolation and repeatability are essential.

4. Performance Metrics and Their Interpretation

A useful evaluation plan combines physical-layer, link-layer, system-level, and operational metrics.

4.1 Received power and link budget

A simplified received-power relation in decibels is

P_r(dBm) = P_t(dBm) + G_t(dBi) + G_r(dBi) - L_p(dB) - L_c(dB),

where P_t is transmit power at the reference point, G_t and G_r are antenna gains, L_p is propagation loss, and L_c represents cable, connector, filter, and other losses. All terms must refer to compatible reference planes. Additional gains or losses, such as feeder gain or polarization loss, should be included explicitly rather than hidden in an unexplained margin.

For thermal noise, the available noise power over bandwidth B is

N(dBm) = -174 dBm/Hz + 10 log10(B) + NF,

where -174 dBm/Hz is the approximate room-temperature thermal-noise density, B is in hertz, and NF is receiver noise figure in decibels. This approximation assumes a physical temperature near 290 K and an idealized noise reference. The signal-to-noise ratio is then SNR = P_r - N when both quantities are expressed in dB or dBm at the same reference point.

Worked example. Suppose a receiver observes P_r = - eighty? No; use numerical notation: P_r = -80 dBm, bandwidth B = 1 MHz, and NF = 5 dB. The noise power is

N = -174 + 10 log10(1,000,000) + 5 = -174 + 60 + 5 = -109 dBm.

Therefore, SNR = -80 - (-109) = 29 dB. This is a thermal-noise estimate; interference, phase noise, quantization noise, and implementation loss may reduce the effective SINR.

4.2 EVM, BER, and packet error rate

Error vector magnitude (EVM) measures the distance between received symbol estimates and ideal reference symbols. A common RMS definition is

EVM_rms = sqrt(Σ|r_k - s_k|² / Σ|s_k|²),

where s_k is the ideal symbol and r_k is the corresponding received symbol after the specified synchronization and equalization operations. If expressed as a percentage, multiply the ratio by 100. EVM may also be reported in dB as 20 log10(EVM_rms), which is negative for EVM below 1. The measurement bandwidth, reference constellation, filtering, and whether common phase error is corrected must be documented.

BER is the number of erroneous decoded bits divided by the total number of compared bits. Packet-error rate (PER) is the number of failed packets divided by the number of transmitted packets, according to a stated packet-failure criterion. BER and PER are not interchangeable; a system with strong coding may have a low BER within successfully decoded packets but a significant PER due to frame-level failures.

4.3 Throughput, latency, and reliability

Goodput is successfully delivered payload bits per unit time. It excludes retransmitted, corrupted, or protocol-overhead bits according to the stated definition. A nominal physical-layer rate may therefore be much higher than application goodput.

Latency should be decomposed into processing, queuing, transmission, propagation, and retransmission components. Reporting only an average can hide unacceptable tail behavior. The 95th or 99th percentile is often more informative for control and interactive applications.

Outage probability can be estimated as

P_out = N_bad/N_total,

where N_bad is the number of samples, locations, or time intervals below a defined service threshold and N_total is the total number evaluated. The threshold must specify the metric, such as received power, SINR, throughput, or packet success.

5. Test Equipment and Measurement Practice

A professional SDR laboratory commonly uses:

• Spectrum or vector signal analyzers for occupied bandwidth, spurious emissions, ACLR, carrier leakage, and EVM.
• Signal generators for calibrated input power, frequency-offset tests, and sensitivity measurements.
• Power meters for independent verification of RF power.
• Oscilloscopes for clock, trigger, envelope, and transient observation.
• Programmable attenuators, splitters, combiners, filters, and isolators.
• Reference clocks or disciplined oscillators for coherent measurements.
• Shielded boxes or anechoic chambers for controlled OTA tests.
• Temperature and supply monitors when hardware stability is under investigation.

The measurement plane must be defined. For example, “transmit power” could mean DAC output, connector output, power after a cable, or effective radiated power. These are different quantities. Calibration should include cable loss, attenuator value, instrument correction factors, and antenna gain where relevant.

A typical conducted sensitivity test connects a calibrated signal generator to the receiver through a known attenuator. The signal level is reduced until a specified PER or BER threshold is reached. The result is meaningful only if the receiver is not overloaded, the waveform and coding are specified, and the noise bandwidth and measurement duration are controlled.

6. Drive Testing and Field Evaluation

Drive testing evaluates a radio system along a route while recording location, time, radio configuration, received signal measurements, serving cell or transmitter identity, throughput, latency, packet loss, and environmental observations. The route should include representative urban, suburban, indoor-outdoor transition, rural, and obstruction conditions when applicable.

A robust campaign should define:

• Route, direction, speed, stopping points, and test duration.
• Antenna type, mounting height, orientation, cable arrangement, and vehicle configuration.
• SDR gain, center frequency, sample rate, bandwidth, firmware, and software version.
• GPS source, timing accuracy, and sampling intervals.
• Traffic generation pattern and packet sizes.
• Weather, temperature, interference, and unusual events.
• Data-quality checks and rules for removing invalid samples.

Drive-test data are spatially and temporally correlated. Hundreds of adjacent samples do not necessarily represent hundreds of independent observations. Statistical summaries should therefore include route segmentation, percentile plots, confidence intervals where justified, and clear treatment of missing or invalid data. A map colored by signal strength is useful diagnostically but does not by itself prove causal relationships.

7. Troubleshooting Framework

Troubleshooting is most efficient when symptoms are separated by domain.

RF and analog symptoms

Unexpected low power, broad spectral regrowth, or receiver desensitization may arise from incorrect gain settings, damaged cables, poor connectors, amplifier compression, inadequate filtering, or external interference. Check the signal at successive reference planes using known test signals. Never infer RF power solely from digital sample magnitude unless the ADC scale and gain calibration are known.

Clock and synchronization symptoms

A drifting constellation, periodic packet loss, or increasing phase rotation may indicate frequency offset, sampling-clock offset, unstable references, or incorrect timestamp handling. Test with a common reference clock, then deliberately introduce a known offset to verify the correction algorithm.

Digital and algorithmic symptoms

A receiver that works with stored ideal samples but fails in real time may suffer from buffer underruns, overruns, thread starvation, incorrect sample-format conversion, endian errors, or latency beyond the protocol limit. Compare recorded samples at each DSP stage. Use deterministic test vectors and unit tests for filters, FFT scaling, synchronization, encoding, and decoding.

Configuration and experimental symptoms

Many apparent algorithm failures are caused by mismatched center frequencies, sample rates, gains, channel bandwidths, antenna connections, or software versions. Every experiment should save a machine-readable configuration file and a human-readable test log.

A useful diagnostic sequence is: verify safety and connections; validate clocks and frequencies; confirm sample rates and data types; measure the analog spectrum; inspect raw I/Q data; test synchronization; validate demodulation with known vectors; and only then investigate higher-layer behavior.

8. Reproducible Experimentation

Reproducibility requires more than publishing a graph. Record hardware model and serial number, firmware, driver, operating-system version, commit identifier, configuration parameters, calibration files, random seeds, antenna and cable information, environmental conditions, raw or sufficient intermediate data, and analysis scripts.

Experiments should distinguish:

• Repeatability: the same team obtains similar results using the same equipment and procedure.
• Reproducibility: an independent team obtains consistent results using the described method, possibly with equivalent equipment.
• Replicability: an independent study reaches consistent conclusions under a suitably related experiment.

Use version control, automated experiment scripts, immutable raw-data archives, checksums, metadata schemas, and documented preprocessing. Predefine KPI calculations and exclusion rules where possible. If a result changes after post-processing, retain both the original data and the transformation code.

9. Applied Case: Illustrative SDR Evaluation of an Adaptive Link

Background

Consider an illustrative university research team evaluating an adaptive QPSK/16-QAM SDR link for a 2.4 GHz short-range data application. The objective is to determine whether adaptation improves goodput without causing unacceptable packet loss under controlled fading.

Problem

Initial simulation predicts high throughput in a Rayleigh channel, but the hardware link performs poorly. The team must determine whether the cause is propagation, oscillator mismatch, RF nonlinearity, receiver implementation, or an incorrect simulation assumption.

Analysis

The team first constructs a conducted setup using a transmitter, programmable attenuator, channel emulator, and receiver. A shared reference clock is used to remove uncontrolled oscillator drift. The transmitter output is measured with a spectrum analyzer, and the receiver input is verified with a power meter and calibrated attenuation. Results show that 16-QAM has excessive EVM at high transmit gain, while QPSK remains within the intended error range. Spectral regrowth increases near amplifier compression.

The channel emulator is then configured with static AWGN, flat fading, and frequency-selective fading separately. For each condition, the team records received power, estimated SNR, EVM, PER, goodput, and latency. The receiver logs the selected modulation and the reason for each adaptation decision.

Approach and solution

The team reduces transmitter gain to create linear operation, adds an RF back-off margin, and improves pilot-based phase tracking. The adaptation algorithm uses a hysteresis margin so that it does not switch modulation repeatedly near a threshold. The team also compares results with and without ideal channel knowledge in simulation.

Outcome

The corrected experiment shows that the original hardware failure was caused mainly by transmitter compression and insufficient phase tracking, not by the adaptation concept. QPSK achieves lower peak goodput but much lower PER in deep fades. 16-QAM provides higher goodput only when EVM and SINR remain within its operating region. The final report includes conducted and OTA results separately, with complete configuration files and raw-data identifiers.

Learning points

1. A favorable link simulation does not validate an RF implementation.

2. Gain must be optimized for both sensitivity and linearity.

3. EVM, PER, and goodput explain different aspects of performance.

4. Channel emulation can isolate impairments before field testing.

5. Adaptation thresholds require hysteresis and validation across mobility and fading conditions.

10. Limitations, Governance, and Contemporary Developments

SDR offers exceptional flexibility, but flexibility introduces risks. An improperly configured transmitter can cause harmful interference, violate emission masks, or operate outside authorization. Experiments must comply with national spectrum regulations, licensing requirements, power limits, approved bands, and institutional safety procedures. Conducted testing with attenuation or shielded environments is preferred when the signal is not intended for public reception.

Security is also relevant. Programmable radios can support spectrum sensing and resilient communications, but they can equally enable unauthorized transmission, spoofing, jamming, or interception. Access control, secure firmware, logging, responsible disclosure, and ethical review are appropriate for advanced projects.

AI and machine learning are increasingly used for modulation recognition, channel prediction, anomaly detection, resource allocation, and adaptive waveform selection. However, learned models may fail under distribution shift, rare interference, changed hardware, or adversarial signals. A responsible evaluation should compare AI methods with interpretable baselines, report training and test conditions, and quantify uncertainty. AI should not replace RF calibration or physical reasoning.

Contemporary SDR platforms increasingly combine FPGA processing, heterogeneous CPU/GPU resources, high-speed data converters, open-source radio frameworks, containerized software, and cloud-assisted experiment management. The resulting systems can support massive bandwidth and real-time learning, but they also increase timing, data-movement, thermal, and reproducibility challenges.

Best-Practice Checklist

Before an experiment, define the hypothesis, KPIs, thresholds, reference planes, and regulatory conditions. During setup, verify cables, antennas, clocks, gains, sample formats, and calibration. During acquisition, record metadata and raw data, monitor clipping and buffer errors, and repeat measurements. During analysis, separate conducted, emulated, and OTA results; distinguish noise from interference; report uncertainty and exclusions; and preserve scripts and configurations. During dissemination, state limitations and avoid presenting a single favorable trial as universal performance.

Chapter Summary

Software-defined radio integrates RF hardware, high-speed data conversion, programmable DSP, and protocol software into a flexible experimental platform. Its value lies not only in waveform programmability but also in exposing the practical gap between ideal communication theory and physical implementation.

A complete SDR evaluation considers sampling, RF dynamic range, oscillator stability, filtering, synchronization, channel estimation, equalization, coding, and real-time processing. Link-level simulation provides detailed waveform insight; system-level simulation evaluates networks and resource allocation; channel emulation enables controlled hardware tests; and OTA or drive testing reveals environmental behavior.

Reliable assessment requires correctly defined reference planes and KPIs. Received power and noise estimates support link-budget analysis, while EVM, BER, PER, throughput, latency, and outage probability describe complementary aspects of performance. Calibration, controlled experimentation, systematic troubleshooting, regulatory compliance, and reproducible data practices are essential. Contemporary AI-enabled SDRs offer new capabilities but must be evaluated against distribution shift, interpretability, and security risks.

Key Terms

Software-defined radio: A radio whose functions are implemented partly through programmable digital hardware and software.

RF front end: Analog circuitry connecting the antenna to the data converters.

I/Q signal: Complex representation using in-phase and quadrature components.

ADC: Device that converts an analog signal into digital samples.

DAC: Device that converts digital samples into an analog signal.

Digital downconversion: Translation and rate reduction of a sampled signal in digital processing.

EVM: Normalized error between received and ideal constellation symbols.

BER: Ratio of erroneous bits to compared bits.

Packet-error rate: Fraction of packets failing a defined reception criterion.

Goodput: Successfully delivered payload rate.

Noise figure: Degradation of signal-to-noise ratio introduced by a receiver.

Link budget: Accounting of transmit power, gains, losses, and received power.

Channel emulator: Equipment that applies controlled propagation impairments.

Link-level simulation: Detailed simulation of a single communication link.

System-level simulation: Network-scale simulation involving users, interference, and resource allocation.

EVM: Error vector magnitude used to assess modulation accuracy.

Frequency offset: Difference between transmitter and receiver carrier frequencies.

Drive testing: Field measurement of radio performance along a defined route.

Repeatability: Consistency of results under the same conditions.

Reproducibility: Ability of an independent party to obtain consistent results using the reported method.

Review Questions

Short Answer Questions

1. What are the principal functional blocks of an SDR transceiver?

2. Why are synchronization and clock quality important in practical SDR systems?

3. Distinguish BER from packet-error rate.

4. What is the purpose of a channel emulator?

5. Define repeatability and reproducibility.

Long Answer Questions

1. Explain the complete SDR signal chain from antenna to decoded data, including major analog and digital impairments.

2. Compare link-level simulation, system-level simulation, channel emulation, and OTA testing.

3. Explain how EVM, SNR, PER, goodput, and latency provide complementary performance information.

4. Develop a laboratory procedure for measuring receiver sensitivity using a calibrated signal generator.

5. Discuss the regulatory, security, and ethical responsibilities associated with programmable radios.

Analytical / Application Questions

1. A receiver has P_r = -92 dBm, bandwidth 200 kHz, and noise figure 6 dB. Estimate the thermal-noise power and SNR, stating assumptions. How might interference alter the conclusion?

2. An SDR link has good simulation results but high OTA PER. Design a troubleshooting plan that separates RF, synchronization, channel, and software causes.

3. Design a reproducible drive-test campaign for an adaptive modulation system. Specify hardware metadata, route controls, KPIs, data-quality rules, and reporting methods.

Multiple Choice Questions with Answers

Question 1

Which SDR block converts a continuous-time analog waveform into digital samples?

A. Power amplifier

B. ADC

C. Antenna filter

D. DAC

Correct Answer: B. ADC

Question 2

For a one-sided ideal raised-cosine baseband bandwidth, which expression is applicable for symbol rate R_s and roll-off factor α?

A. B = R_s/(1+α)

B. B = αR_s/2

C. B = (1+α)R_s/2

D. B = 2(1+α)R_s

Correct Answer: C. B = (1+α)R_s/2

Question 3

A frequency offset Δf produces which baseband phase evolution?

A. φ(t) = Δf/t

B. φ(t) = 2πΔf t + φ_0

C. φ(t) = 2πt/Δf

D. φ(t) = Δf + φ_0t²

Correct Answer: B. φ(t) = 2πΔf t + φ_0

Question 4

What is a principal disadvantage of a zero-forcing equalizer near a deep channel fade?

A. It always removes thermal noise

B. It can amplify noise

C. It requires no channel estimate

D. It prevents frequency correction

Correct Answer: B. It can amplify noise

Question 5

For B = 1 MHz and receiver noise figure NF = 5 dB, using -174 dBm/Hz thermal density, what is the estimated noise power?

A. -179 dBm

B. -114 dBm

C. -109 dBm

D. -60 dBm

Correct Answer: C. -109 dBm

Question 6

Which metric measures the normalized distance between received and ideal constellation symbols?

A. EVM

B. Goodput

C. Outage probability

D. Noise figure

Correct Answer: A. EVM

Question 7

What is the main purpose of a channel emulator in SDR testing?

A. To replace all RF filters permanently

B. To apply controlled propagation impairments

C. To increase antenna gain automatically

D. To decode every modulation without configuration

Correct Answer: B. To apply controlled propagation impairments

Question 8

Which practice most directly improves reproducibility?

A. Reporting only the best trial

B. Changing software versions between trials

C. Saving configurations, calibration data, and analysis scripts

D. Using undocumented manual adjustments

Correct Answer: C. Saving configurations, calibration data, and analysis scripts

Question 9

Why can a high transmitter gain reduce practical link performance?

A. It may drive the power amplifier into compression

B. It always reduces received power

C. It eliminates all spectral components

D. It prevents ADC sampling

Correct Answer: A. It may drive the power amplifier into compression

Question 10

Which distinction is correct?

A. System-level simulation models only one ideal symbol

B. Link-level simulation primarily evaluates detailed single-link behavior

C. OTA testing is always more repeatable than conducted testing

D. Drive testing does not require metadata

Correct Answer: B. Link-level simulation primarily evaluates detailed single-link behavior