5G Base Station(gNB) Conformance Testing: In-Depth Analysis Based on Spec. TS 38.104 & 38.141-1

Conformance testing ensures 5G New Radio (NR) base stations (gNBs) meet the stringent requirements defined by the 3rd Generation Partnership Project (3GPP) for reliable network performance. The 3GPP specifications TS 38.104 and TS 38.141-1 outline the minimum requirements and test procedures for gNBs, focusing on conducted conformance testing for Frequency Range 1 (FR1, sub-6 GHz). This article provides an in-depth exploration of gNB conformance testing, detailing each test parameter, its intuitive significance, practical examples, and the associated threshold values as per Release 16 of the specifications.

Overview of 3GPP TS 38.104 and TS 38.141-1

TS 38.104: Defines the minimum Radio Frequency (RF) and performance requirements for gNBs, covering transmitter, receiver, and performance characteristics. It sets the baseline for acceptable operation, ensuring interoperability and compliance with 5G NR standards.
TS 38.141-1: Specifies the conducted conformance test methods for gNBs, detailing how to measure each parameter defined in TS 38.104. It includes test setups, tolerances, and procedures to verify compliance under controlled conditions.

Conformance testing verifies that a gNB operates within specified limits for output power, signal quality, unwanted emissions, and receiver performance, ensuring robust network operation in real-world deployments.

Key Test Parameters, Intuitive Understanding, Examples, and Thresholds

Below, we explore the primary test parameters for conducted conformance testing, categorized by transmitter characteristics (Chapter 6 of TS 38.141-1), receiver characteristics (Chapter 7), and performance requirements (Chapter 8). Each parameter is explained with its purpose, intuitive significance, a practical example, and the threshold values as defined in TS 38.104.

1. Transmitter Characteristics (Chapter 6)

These tests ensure the gNB transmits signals with the correct power, quality, and minimal interference.

1.1 Base Station Output Power

Purpose: Measures the gNB’s total radiated power to ensure it meets the declared maximum power level, critical for coverage and network planning.

Think of output power as the water in a glass (gNB’s signal). Too low, and devices struggle to connect; too high, and it may interfere with other cells or exceed regulatory limits.

Example: A gNB is configured for a 100 MHz bandwidth TDD signal with a declared maximum power of 43 dBm (20 W)(see the image below). The test measures the actual power at the antenna connector to confirm it’s within tolerance.

Threshold (TS 38.104, Clause 6.2):

For Wide Area BS: Declared power ±2.7 dB (test tolerance included).
For Medium Range BS: Declared power ±3.0 dB.
For Local Area BS: Declared power ±3.0 dB.
Example: For a 43 dBm Wide Area BS, the measured power must be between 40.3 dBm and 45.7 dBm.

1.2 Output Power Dynamics

Purpose: Verifies the gNB’s ability to adjust power levels dynamically, ensuring efficient resource allocation and minimal interference. This is like a LED bulb switch for the gNB’s signal. It must smoothly adjust power for different users or time slots without causing disruptions.
Example: In a TDD system, the gNB switches between high power for downlink slots and low power for uplink slots. The test checks the power difference (dynamic range) and transient behavior.
Threshold (TS 38.104, Clause 6.3):
- Minimum dynamic range: ≥13.4 dB for 100 MHz channel bandwidth (QPSK modulation).
- Transient period (on/off): ≤10 µs for power transitions.
- Test tolerance: ±2.7 dB for power measurements.

1.3 Transmit ON/OFF Power

Purpose: Ensures the gNB’s transmitter is effectively "off" during idle periods to minimize interference in TDD systems.

Imagine turning off a loudspeaker during silent periods to avoid background noise. The gNB must reduce its signal to negligible levels when not transmitting.

Example: During TDD uplink slots, the gNB’s downlink transmitter should emit minimal power. The test measures leakage power at the antenna connector.
Threshold (TS 38.104, Clause 6.4):
- OFF power: ≤-85 dBm/MHz (conducted measurement).
- Test tolerance: +2.7 dB.
- Example: For a 100 MHz channel, the OFF power must not exceed -75 dBm total.

1.4 Transmitted Signal Quality

This category includes three sub-tests: Frequency Error, Error Vector Magnitude (EVM), and Time Alignment Error.

1.4.1 Frequency Error

Purpose: Verifies the gNB’s carrier frequency accuracy to prevent interference with adjacent channels. Think of tuning a radio to a specific station. If the gNB’s frequency drifts, it could overlap with another station, causing distortion.
Threshold (TS 38.104, Clause 6.5.1):
- Frequency error: ≤±0.05 ppm (±0.1 ppm with test tolerance). It can be converted to Hz, using formula Hz= (f*ppm)/10^6.
- Example: For a 3.5 GHz carrier, the error must be ≤±175 Hz (±350 Hz with tolerance).

1.4.2 Error Vector Magnitude (EVM)

Purpose: Measures the accuracy of the gNB’s modulation, ensuring the transmitted signal closely matches the ideal constellation points.
Constellation points are visual representation of data output by QAM(quadrature amplitude modulation) modulator.
QAM is a form of digital modulation that combines changes in both amplitude and phase. QAM derives its name from the fact that QAM symbols are created using something called an IQ modulator. In an IQ modulator, an in-phase ("I") component is combined with a quadrature (“Q”) component to produce a modulated signal.

Each symbol in a constellation has an ideal or reference point that corresponds to a defined magnitude and phase, but received or measured points rarely fall exactly on the ideal point.
Some of the difference may be due to magnitude error – that is, the received vector is too long or two short. And some of the difference may be due to phase error, in which the angle of the received vector is incorrect. This is illustrated in below diagram.

EVM is measured at each symbol time, and larger values of EVM indicate greater distance between the measured and ideal points.
EVM is calculated by comparing the length of the red error vector either to the length of the green vector (in the case of a peak or max power normalization) or to the length of blue vector (for RMS power normalization). Refer the image below.

Higher EVM values mean a greater probability of bit errors.

Threshold (TS 38.104, Clause 6.5.2):
- EVM limits (RMS, averaged over subframes):
  - QPSK: ≤17.5%
  - 16QAM: ≤12.5%
  - 64QAM: ≤8.0%
  - 256QAM: ≤3.5%
- Test tolerance: +1% (e.g., QPSK limit becomes ≤18.5%).

1.4.3 Time Alignment Error (TAE)

Purpose: Ensures synchronization between multiple antenna ports in MIMO or carrier aggregation setups to maintain signal coherence. Imagine a choir where singers must start at the same time. If MIMO streams are misaligned, they interfere destructively, degrading performance.
Example: In a 4x4 MIMO gNB, the test checks that signals from all ports are transmitted within a tight time window.
Threshold (TS 38.104, Clause 6.5.3):
- TAE: ≤65 ns for MIMO or intra-band contiguous carrier aggregation.
- Test tolerance: ±25 ns.
- Example: Total TAE must be ≤90 ns including tolerance.

1.5 Unwanted Emissions

These tests limit the gNB’s interference to other systems or adjacent channels.

1.5.1 Adjacent Channel Leakage Ratio (ACLR)

Purpose: Measures power leakage into adjacent channels to ensure minimal interference with neighboring frequencies.

Think of ACLR as soundproofing between rooms. The gNB’s signal should stay in its "room" (channel) without spilling into others.

Threshold (TS 38.104, Clause 6.6.3):
- ACLR: ≥45 dB (absolute power ≤-13 dBm/MHz in adjacent channel).
- Test tolerance: ±0.8 dB.
- Example: If the main channel power is 43 dBm, the adjacent channel power must be ≤-2 dBm.

1.5.2 Operating Band Unwanted Emissions (OBUE)

Purpose: OBUE measures the emissions close to the assigned channel bandwidth of the wanted signal while the transmitter is in operation. It Limits emissions within the operating band but outside the channel bandwidth to protect other users.

OBUE is like keeping your music system’s noise within your house, avoiding disturbance to neighbors.

Example: For a 100 MHz channel, the test measures emissions in the 10 MHz bands just outside the channel edges.
Threshold (TS 38.104, Clause 6.6.4):
- Category A: ≤-13 dBm/MHz (1 MHz to 5 MHz offset from channel edge).
- Category B: ≤-15 dBm/MHz (same offset).
- Test tolerance: ±1.5 dB.
- Example: Emissions at 1 MHz offset must be ≤-11.5 dBm/MHz (Category A).

1.5.3 Spurious Emissions

Spurious emission measures the emissions across a much wider frequency range to ensure the emission level is under the limit.
The measurement region includes two parts, the first part is from 9 kHz to ΔfOBUE below the operating band, and the second part is from ΔfOBUE above the operating band to 12.75 GHz.
It ensures emissions outside the operating band are minimal to avoid interfering with other radio services (e.g., GPS, LTE).
Spurious emissions are like unwanted echoes from a speaker that disrupt distant listeners. They must be kept low across a wide frequency range.
Example: A gNB at 3.5 GHz is tested from 9 kHz to 12.75 GHz (excluding the operating band) to ensure no significant emissions.
Threshold (TS 38.104, Clause 6.6.5):
- General limit: ≤-30 dBm/MHz (30 MHz to 12.75 GHz).
- Category B: ≤-36 dBm/MHz (stricter for some regions).
- Test tolerance: ±2.0 dB.
- Example: Emissions at 2 GHz must be ≤-28 dBm/MHz (general limit).

1.6 Transmitter Intermodulation

Transmit Intermodulation is a measure of the capability of the transmitter to inhibit the generation of signals in its non-linear elements caused by presence of the wanted signal and an interfering signal reaching the transmitter via the antenna.
It verifies the gNB’s transmitter does not generate excessive distortion when exposed to interfering signals.
Imagine a singer staying on pitch despite background noise. The gNB must maintain signal quality when other signals are present at its output.

Example: An interfering signal at 10 dB below the gNB’s output is injected, and the test checks for distortion products meeting ACLR and spurious emission limits.
Threshold (TS 38.104, Clause 6.7):
- Interferer power: ≤-52 dBm.
- Meets ACLR (≥45 dB) and spurious emission limits (e.g., ≤-30 dBm/MHz).
- Test tolerance: Same as ACLR/spurious emissions.

2. Receiver Characteristics (Chapter 7)

These tests ensure the gNB’s receiver can detect signals reliably without introducing interference.

2.1 Reference Sensitivity

REFSENS is the minimum mean power received at the gNB antenna port at which the specified minimum requirement shall be met. It measures the gNB’s ability to decode weak signals at the minimum required signal level.
This is like hearing a whisper in a quiet room. The gNB must detect faint signals from distant devices without errors.
Example: A QPSK signal at -104.7 dBm is sent to the gNB, and the test checks for ≥95% throughput.
Threshold (TS 38.104, Clause 7.2):
- Reference sensitivity level: e.g., -104.7 dBm for 100 MHz, SCS 30 kHz, QPSK.
- Throughput: ≥95% of maximum.
- Test tolerance: ±1.8 dB.
- Example: Signal level must be detectable between -106.5 dBm and -102.9 dBm.

2.2 Dynamic Range

Purpose: Verifies the gNB can handle a wide range of input signal levels without performance degradation.
Think of adjusting your eyes from bright sunlight to a dim room. The gNB must process strong and weak signals equally well.
Example: A signal alternates between high power (-76.2 dBm) and low power with added noise, ensuring ≥95% throughput.
Threshold (TS 38.104, Clause 7.3):
- Dynamic range level: e.g., -76.2 dBm for 100 MHz, SCS 30 kHz.
- Throughput: ≥95% of maximum.
- Test tolerance: ±2.4 dB.

2.3 In-Band Selectivity and Blocking

Purpose: Ensures the gNB can receive the desired signal in the presence of strong interferers within the same band.
Imagine listening to a friend in a noisy crowd. The gNB must filter out unwanted signals to focus on the intended one.
Example: A wanted signal at -95.7 dBm is tested with an interferer at -52 dBm in an adjacent channel, checking for ≥95% throughput.
Threshold (TS 38.104, Clause 7.4):
- Adjacent channel interferer: -52 dBm.
- Throughput: ≥95% of maximum.
- Test tolerance: ±1.4 dB.

2.4 Out-of-Band Blocking

Purpose: Verifies the gNB’s receiver performance when exposed to strong signals outside the operating band.
This is like ignoring a loud radio from another room. The gNB must focus on its frequency band despite external noise.
Example: An interferer at 1 GHz (-15 dBm) is injected while receiving a signal at 3.5 GHz, ensuring ≥95% throughput.
Threshold (TS 38.104, Clause 7.5):
- Blocking level: -15 dBm (1 MHz to 12.75 GHz, excluding in-band).
- Throughput: ≥95% of maximum.
- Test tolerance: ±1.0 dB.

2.5 Receiver Spurious Emissions

Purpose: Limits emissions from the gNB’s receiver to avoid interfering with other systems.
The receiver shouldn’t "leak" signals, like a microphone picking up its own noise. Emissions must be minimal.
Example: The receiver port is measured from 30 MHz to 12.75 GHz to ensure emissions are below the limit.
Threshold (TS 38.104, Clause 7.6):
- Spurious emissions: ≤-57 dBm (30 MHz to 12.75 GHz, 100 kHz RBW).
- Test tolerance: ±2.0 dB.
- Example: Emissions must be ≤-55 dBm.

2.6 Receiver Intermodulation

Purpose: Ensures the gNB’s receiver can handle multiple interferers without generating distortion that degrades performance.
Like staying focused in a room with two loud conversations, the gNB must ignore combined interference effects.
Example: Two interferers (-46 dBm each) are injected, and the test checks for ≥95% throughput on the wanted signal.
Threshold (TS 38.104, Clause 7.7):
- Interferer levels: -46 dBm (CW and modulated).
- Throughput: ≥95% of maximum.
- Test tolerance: ±1.7 dB.

3. Performance Requirements (Chapter 8)

These tests evaluate the gNB’s ability to decode signals under challenging conditions, such as fading and low signal-to-noise ratios (SNR).

3.1 Performance Requirements for PUSCH

Purpose: Verifies the gNB can decode the Physical Uplink Shared Channel (PUSCH) under fading and noisy conditions.
This is like understanding a phone call during a storm. The gNB must extract data despite signal degradation.
Example: A faded QPSK signal with AWGN is sent to the gNB, simulating a mobile device in a moving car. The test checks for ≥95% throughput.
Threshold (TS 38.104, Clause 8.2):
- Throughput: ≥95% of maximum at specified SNR (e.g., -1.4 dB for QPSK, FRC A3-1).
- Test tolerance: ±0.7 dB for SNR.
- HARQ feedback: Verified via non-standard interface.

3.2 Performance Requirements for PUCCH

Purpose: Ensures the gNB can decode the Physical Uplink Control Channel (PUCCH) for control signaling under adverse conditions.
Think of receiving critical instructions during a noisy event. The gNB must reliably detect control signals.
Example: A PUCCH signal with fading is tested to ensure the gNB detects acknowledgments (ACK/NACK) correctly.
Threshold (TS 38.104, Clause 8.3):
- Detection probability: ≥99% for ACK/NACK at specified SNR (e.g., 1.0 dB for Format 1).
- Test tolerance: ±0.7 dB.

3.3 Performance Requirements for PRACH

Purpose: Verifies the gNB can detect the Physical Random Access Channel (PRACH) for initial access under low SNR.
This is like hearing a knock at the door in a noisy house. The gNB must detect devices attempting to connect.
Example: A PRACH preamble is sent with fading and noise, and the gNB must detect it with high reliability.
Threshold (TS 38.104, Clause 8.4):
- Detection probability: ≥99% at specified SNR (e.g., -8.2 dB for Format A1).
- False alarm probability: ≤0.1%.
- Test tolerance: ±0.7 dB.

4. Test Models in TS 38.141-1

Test Models (TMs) defined in TS 38.141-1 (Clause 4.9) provide standardized signal configurations for testing gNB transmitter characteristics. They ensure consistent evaluation of parameters like output power, EVM, and unwanted emissions by simulating realistic NR signals. Below are the key Test Models with brief details:

NR-FR1-TM1.1 (General Purpose, Single Channel):
- Simulates a fully allocated PDSCH with QPSK modulation across all resource blocks.
- Used for tests like output power, EVM, and ACLR, representing a typical high-load scenario.
- Example: Tests a 100 MHz channel at maximum power with uniform resource allocation.
NR-FR1-TM1.2 (General Purpose, Partial Allocation):
- Configures PDSCH with QPSK but only partially allocated resource blocks (e.g., 50% RB usage).
- Evaluates performance under non-contiguous allocation, stressing dynamic range and linearity.
- Example: Tests OBUE with scattered resource blocks to mimic sparse traffic.
NR-FR1-TM2 (High Spectral Efficiency, 64QAM):
- Uses 64QAM PDSCH modulation with full resource block allocation.
- Designed for EVM and power tests under high-throughput conditions, challenging modulation accuracy.
- Example: Verifies EVM ≤8.0% for 64QAM in a 50 MHz channel.
NR-FR1-TM2a (High Spectral Efficiency, 256QAM):
- Employs 256QAM PDSCH with full allocation, targeting advanced deployments.
- Tests stringent EVM requirements (≤3.5%) and linearity for high-capacity scenarios.
- Example: Ensures gNB supports premium services with minimal signal distortion.
NR-FR1-TM3.1 (Mixed Modulation, Realistic Traffic):
- Combines QPSK, 16QAM, and 64QAM in PDSCH with dynamic allocation.
- Simulates real-world traffic with mixed modulation, testing adaptability and signal quality.
- Example: Measures ACLR under varied modulation to mimic multi-user scenarios.
NR-FR1-TM3.2 (Low Modulation, Robustness):
- Uses QPSK with partial allocation to test robustness under low modulation schemes.
- Focuses on emissions and power control in coverage-limited scenarios.
- Example: Tests transmit ON/OFF power for TDD systems with sparse allocation.

These Test Models cover a range of operating conditions, from basic functionality to high-capacity, real-world traffic, ensuring comprehensive gNB evaluation.

5. Practical Testing Considerations

Test Setup: Conducted tests use equipment like the Keysight Vector Signal Analyzer (VSA) for TX tests and Signal Generator (MXG) for RX tests, R&S®FSW signal analyzer for transmitter measurements and the R&S®SMW200A vector signal generator for receiver and performance tests. These tools support automated setups with presets for Fixed Reference Channels (FRCs) and Test Models (TMs).

Below is the test lab diagram for reference.

Challenges: Performance tests, especially for PUSCH, require complex setups to simulate fading and AWGN. Non-standard HARQ feedback interfaces can complicate testing.
Recent Trends: Advances in test automation (e.g., Keysight P7000A, Rohde & Schwarz QuickStep 5.0) simplify compliance by integrating 3GPP requirements into user-friendly software. xApps/rApps based conformance testing is also performed.

6. Conclusion

Conformance testing per TS 38.104 and TS 38.141-1 ensures gNBs deliver reliable 5G NR performance while minimizing interference. By rigorously testing parameters like output power, EVM, ACLR, and receiver sensitivity, manufacturers verify compliance with global standards. Understanding the intuitive role of each parameter—whether it’s controlling signal "volume," ensuring modulation "clarity," or filtering out "noise"—helps engineers design robust base stations. The specified thresholds provide clear pass/fail criteria, enabling consistent performance across diverse deployments.

References: