Signal generators have long been a cornerstone of telecommunications testing, but the arrival of 5G has raised the bar dramatically. With its reliance on millimeter-wave frequencies, massive MIMO arrays, and complex modulation schemes, 5G demands test equipment that can replicate real-world conditions with unprecedented precision. The latest innovations in signal generator technology are not incremental improvements; they represent a fundamental shift in how engineers approach network validation, from lab benchtop to field deployment. These advances are enabling faster certification cycles, more accurate performance assessments, and a level of flexibility that keeps pace with the rapidly evolving 5G standards.

Evolution of Signal Generators for 5G Networks

The transition from 4G to 5G forced a rethinking of signal generation. Earlier generations operated primarily below 6 GHz, but 5G NR introduces frequency range 1 (FR1, sub‑6 GHz) and frequency range 2 (FR2, mmWave from 24 GHz to over 50 GHz). Conventional signal generators could not handle the wide bandwidths required—up to 400 MHz per channel in FR2. Engineers needed instruments capable of generating highly stable, low‑phase‑noise signals across these bands while supporting the full flexibility of 5G numerology, including variable subcarrier spacing and slot formats.

Modern signal generators have risen to the challenge. They now integrate multiple radio frequency (RF) chains, digital pre‑distortion, and real‑time signal processing to produce the complex waveforms used in 5G. Beyond simple continuous waves, these instruments can generate 5G NR conforming test models (NR‑TM) and fading profiles that mimic realistic propagation environments. The evolution has been driven by the need to test not only the physical layer but also the higher‑layer protocols and network slicing capabilities central to 5G. As network operators push toward standalone 5G core architectures, the signal generator's role has expanded from a simple stimulus source to a sophisticated platform that simulates entire network scenarios.

Key Innovations Driving Performance

Wideband and High‑Fidelity Signal Generation

One of the most significant breakthroughs is the ability to generate wideband signals with exceptional spectral purity. For 5G NR, a test signal must maintain low error vector magnitude (EVM) across a bandwidth that can be 100 MHz or more in FR1 and up to 400 MHz in FR2. New digital‑to‑analog converter (DAC) architectures, operating at sampling rates exceeding 100 GS/s, allow signal generators to create these wideband waveforms directly without complex frequency conversion schemes. High‑fidelity generation ensures that impairments measured in a device under test are due to the device itself, not the test equipment.

Additionally, phase noise performance has improved dramatically. At mmWave frequencies, phase noise can degrade the error rate of high‑order QAM modulations such as 256‑QAM and 1024‑QAM. Modern signal generators employ low‑noise frequency synthesis and digital cancellation techniques to achieve phase noise levels below −140 dBc/Hz at 10 kHz offset. This level of purity is essential for accurately evaluating receiver sensitivity, adjacent channel selectivity, and blocker rejection—all critical for 5G base stations and end‑user devices.

AI‑Driven Optimization and Adaptation

Artificial intelligence is beginning to transform how signal generators operate. Instead of relying on manual configuration of all parameters, new instruments can learn from the environment and automatically adjust the test signal. For example, during over‑the‑air (OTA) testing of a 5G handset, the signal generator can analyze the received power at the device antenna and adapt the beamforming angle to maintain a stable link. This dynamic optimization reduces test time by eliminating the iterative manual tuning that previously consumed hours.

Machine learning algorithms also assist in generating worst‑case interference scenarios. By examining logs from field trials, an AI‑enhanced signal generator can reproduce specific fading conditions and co‑channel interference patterns that were problematic in live deployments. Engineers then use these realistic test cases to validate both hardware and software improvements. The result is a testing process that is not only faster but also more representative of real‑world network behavior.

Software‑Defined Signal Architectures

The shift toward software‑defined signal generators has been a game changer for 5G testing. Traditional hardware‑based generators were tied to specific air interfaces; upgrading to support a new release of 5G NR often required a new instrument. Today, software‑defined architectures decouple the signal creation from the physical hardware. A field‑programmable gate array (FPGA) or a general‑purpose processor runs a flexible software stack that can be updated as 3GPP specifications evolve. This approach allows a single instrument to support not only 5G NR but also LTE, NB‑IoT, and even legacy 2G/3G standards—all without changing hardware.

Software‑defined generators also make it easier to implement custom waveforms for research and development. Engineers can write their own 5G NR slot configurations, pilot patterns, and control channel placements, which is invaluable for testing new algorithms for beam management or link adaptation. The same flexibility enables rapid prototyping of 5G‑Advanced features, such as enhanced carrier aggregation with 32 component carriers or full‑duplex operation. As Keysight’s signal generators demonstrate, software updates now deliver new capabilities long after the hardware has been deployed, extending the instrument’s useful life.

Advanced Waveform Generation and Pre‑Distortion

Creating a true 5G NR waveform requires more than just setting a power level and a frequency. The signal generator must accurately generate the OFDM symbols with correct subcarrier allocation, cyclic prefix, and windowing. Newer instruments incorporate dedicated waveform generation engines that can build the complete baseband representation of a 5G NR transmission in real time. They can also simulate the complete downlink and uplink structures, including synchronization signal blocks (SSB), physical broadcast channel (PBCH), and demodulation reference signals (DM‑RS).

Digital pre‑distortion (DPD) has also been integrated into signal generators. DPD compensates for nonlinearities in the generator’s own power amplifier, ensuring that the output signal remains clean even at high output power. This is particularly important when testing power amplifiers for base stations, where the test signal must be a clean representation of the intended modulated waveform. Without DPD, the amplifier’s own distortion would mask the performance of the device under test.

Phase Noise and Jitter Improvements

In 5G networks, timing accuracy is paramount. The signal generator’s internal clocks and sampling systems must produce extremely low jitter to support the tight timing budget of 5G NR. Jitter in the test signal directly translates to increased EVM, which can cause a device to fail specification limits even if the device itself is performing correctly. Modern generators use integrated clock multipliers with phase‑locked loops (PLLs) that achieve sub‑picosecond jitter. Some instruments also offer external reference inputs that can be locked to a network‑grade cesium or rubidium standard, enabling lab measurements that replicate the synchronization accuracy of a live network.

Enhancing Testing Flexibility and User Workflow

Beyond raw performance, modern signal generators are designed to streamline the testing workflow. Engineers no longer need to laboriously program each parameter via a front‑panel interface. The latest instruments feature intuitive touch‑screen displays, drag‑and‑drop waveform editors, and API‑driven remote control that integrates seamlessly into automated test sequences. Many generators support SCPI commands, Python scripting, and industry‑standard test automation frameworks like LabVIEW and MATLAB.

Modularity is another defining trend. Rather than purchasing a single box that covers all frequencies, engineers can now use a mainframe with interchangeable modules for different bands and applications. This approach is especially beneficial for R&D labs that must test devices across multiple 5G bands concurrently. A module that handles 28 GHz can be swapped for a 39 GHz module within minutes, reducing the capital expenditure that would otherwise be required to buy multiple dedicated instruments.

Remote collaboration has also become a priority. With distributed engineering teams and the rise of “lab as a service,” signal generators now include built‑in web servers and secure remote access capabilities. A team in San Diego can control a signal generator in a lab in Munich, sharing the same test environment and data in real time. This capability was vital during the pandemic and continues to accelerate global product development cycles.

Automation is further enhanced by cloud‑based test orchestration. Some signal generator platforms now allow test engineers to define sequences on a cloud dashboard, which then sends the configuration to a fleet of generators located in different test chambers. This “lights‑out” approach reduces manual intervention and increases the throughput of regression testing. As Rohde & Schwarz signal generators illustrate, the convergence of hardware and software is enabling a new level of operational efficiency in 5G validation.

Impact on Key 5G Testing Scenarios

Over‑the‑Air (OTA) Testing

One of the most demanding applications for signal generators in 5G is over‑the‑air testing. Because 5G makes extensive use of beamforming, it is no longer sufficient to test devices through a conductive cable connection. OTA testing requires the signal generator to produce a known wavefront that impinges on the device’s antenna array from a specific angle. Innovations in multi‑channel signal generators allow the creation of multiple independent phase‑coherent signals, each feeding a different antenna in a test chamber. By precisely controlling the phase and amplitude of these signals, engineers can emulate a base station transmitting from any direction, validating the device’s beam steering, beam management, and spatial multiplexing capabilities.

Wideband and high‑fidelity generators are essential for OTA because any amplitude or phase error directly affects the angle of arrival measurement. The tightest specifications call for phase coherence within ±1° across the test bandwidth. New generators that incorporate phase‑synchronized outputs across multiple modules make it possible to scale OTA test systems from a few antenna elements up to the dozens used in massive MIMO. This scalability is a direct result of the modular and software‑defined architectures discussed earlier.

Massive MIMO and Beamforming Validation

Massive MIMO (multiple‑input multiple‑output) base stations can have 64, 128, or even 256 antenna elements. Testing such a system requires a signal generator that can simultaneously stimulate all receive chains with known test signals. Modern generators offer parallel RF outputs that are phase‑coherent and frequency‑locked, enabling full‑array characterization. The ability to generate different beamforming weights for each output path allows engineers to inject a known spatial profile and then use a network analyzer to measure the resulting beam pattern. Without this multi‑channel capability, testing a massive MIMO array would require hundreds of separate generators, which is impractical both in cost and complexity.

Network Slicing and Protocol Testing

5G’s network slicing feature allows operators to partition the same physical infrastructure to support different services (e.g., ultra‑reliable low‑latency communications, enhanced mobile broadband, massive IoT). Signal generators now incorporate higher‑layer protocol emulation that can simulate multiple slices simultaneously. For testing a core network slice manager, the generator can produce traffic flows with different QoS profiles and then verify that the base station and core properly route each slice. This moves signal generators beyond simple physical‑layer test tools into the realm of end‑to‑end validation.

Carrier Aggregation and Dual Connectivity

5G NR operates in both FR1 and FR2, and devices are expected to aggregate carriers across these bands. Testing carrier aggregation requires a signal generator that can produce multiple independent signals, each with its own frequency offset, power, and channel configuration. Advanced generators can generate up to 16 or 32 component carriers in a single instrument, with independent fading profiles for each. This enables realistic throughput and handover testing that mirrors actual network deployments.

Similarly, LTE‑NR dual connectivity (EN‑DC) scenarios demand that the generator produce both an LTE anchor carrier and a 5G NR carrier, synchronized in time and frequency. Modern generators meet this need by using a common reference clock and digital baseband engines that align the timing of the two signals to within the tolerances set by 3GPP. As a result, device manufacturers can verify EN‑DC performance without having to physically connect two separate test stations.

Future Directions: Preparing for 5G‑Advanced and 6G

The innovations in signal generation are not a finished story. 3GPP Release 18 and beyond, collectively termed 5G‑Advanced, will introduce even more complex features: integrated sensing and communication, artificial intelligence in the air interface, and enhanced support for extended reality (XR). Signal generators must be ready to test these capabilities. The trend toward software‑defined instruments ensures that many of these features can be added through firmware updates, but hardware headroom is also critical. Next‑generation generators are being designed with extra bandwidth (up to 2 GHz internal bandwidth), higher dynamic range, and faster switching speeds to handle the demands of carrier aggregation with up to 64 component carriers.

Looking further ahead to 6G, which is expected to operate above 100 GHz, signal generators will need to produce signals with bandwidths exceeding 10 GHz. Research prototypes already exist that use photonic‑assisted generation to create mmWave and sub‑THz signals. These systems leverage optical frequency combs and photodiodes to generate frequencies beyond the reach of conventional electronics. While still in the lab, the principles demonstrated by these photonic signal generators will eventually be commercialized, ensuring that test equipment remains aligned with the trajectory of radio technology.

Furthermore, the integration of digital twin technology may allow signal generators to be programmed based on a virtual model of the network environment. Instead of manually configuring fading profiles, an engineer could feed the generator a digital twin of a city, and the generator would produce signals that match the exact propagation conditions for any given location. This is a natural extension of AI‑driven optimization and would dramatically reduce the time needed for field simulation.

Conclusion: A Critical Enabler for 5G Success

The rapid deployment of 5G networks around the world would not have been possible without parallel advances in test equipment. Signal generators have evolved from simple sine‑wave sources to complex, software‑defined platforms that can simulate the full richness of a 5G air interface. Innovations in wideband generation, AI‑driven optimization, software‑defined architecture, and modular design have made these instruments faster, more accurate, and more versatile than ever before. They enable engineers to thoroughly test every aspect of a 5G system—from individual chips to integrated base stations and complete network slices—before it reaches the field.

As 3GPP continues to define 5G‑Advanced and begins work on 6G, signal generator technology will need to keep pace. The foundations laid by today’s innovations ensure that the test and measurement industry will rise to the challenge. For network operators and equipment manufacturers, investing in state‑of‑the‑art signal generation is not merely a cost; it is a strategic necessity to ensure that 5G networks deliver the performance, reliability, and flexibility that users and applications demand. The journey from the lab to live deployment is paved with precise, repeatable signals—and modern signal generators are the tools that pave that road.