Electronics Guide

The Role of Test Signals

Test signals provide the input stimulus that reveals how circuits respond to various conditions. A sine wave applied to an amplifier input produces an output that indicates gain, bandwidth, distortion, and noise characteristics. A step function reveals settling time and transient response. Complex modulated signals exercise communication systems under realistic operating conditions.

The quality of measurements depends directly on the quality of the stimulus signal. Phase noise in a test oscillator obscures the phase noise of the device under test. Harmonic distortion in the generator appears in the output, potentially masking or combining with device distortion. Understanding generator specifications and limitations prevents measurement errors and misinterpretation of results.

Beyond measurement, signal generators serve as replacement sources during system development and debugging. When a particular module is unavailable or malfunctioning, a generator can substitute its output, enabling work to continue on other portions of the system. This application requires careful matching of signal characteristics including frequency, amplitude, timing, and any modulation present in the original signal.

Signal Characteristics and Parameters

Several fundamental parameters characterize any generated signal:

Frequency specifies how rapidly the signal varies, expressed in hertz (cycles per second) for periodic waveforms. Generator frequency range, resolution, and accuracy determine which applications the instrument can address. Crystal-referenced synthesizers achieve frequency accuracies of parts per million or better, essential for precision measurements.

Amplitude describes the signal magnitude, typically specified as peak-to-peak voltage, RMS voltage, or power into a specified impedance. Output amplitude range limits both the minimum detectable signal (constrained by noise floor) and maximum drive level. Flatness specification indicates how amplitude varies across the frequency range.

Waveform shape defines the instantaneous voltage as a function of time or phase. Standard waveforms include sine, square, triangle, sawtooth, and pulse. Arbitrary waveform generators extend this to any shape that can be defined mathematically or sampled from real signals.

Spectral purity characterizes undesired frequency components in the output. Harmonic distortion measures power at multiples of the fundamental frequency. Spurious signals appear at unrelated frequencies, often arising from synthesis mechanisms. Phase noise describes random frequency fluctuations that spread the signal spectrum around its center frequency.

Output Stage Considerations

The generator's output stage must deliver signals with adequate voltage and current drive capability while maintaining signal integrity. Most instruments provide 50-ohm output impedance, enabling direct connection to 50-ohm systems and cables. When driving high-impedance loads, the open-circuit voltage doubles compared to the terminated voltage, requiring appropriate amplitude adjustment.

Output protection circuits guard against damage from external voltages, ESD events, or accidental short circuits. These protection mechanisms may limit output current or voltage slew rate, affecting performance when driving reactive loads or generating fast edges.

Differential outputs, available on some generators, provide balanced signals that reject common-mode interference and enable direct connection to differential inputs. Differential signaling is essential for testing high-speed serial interfaces and precision analog circuits.

DDS Architecture

The fundamental DDS architecture consists of three main blocks: a phase accumulator, a waveform lookup table, and a digital-to-analog converter (DAC). The phase accumulator adds a frequency tuning word to its contents on each clock cycle, generating a phase ramp that represents the instantaneous phase of the output waveform. This phase value addresses the lookup table, which contains samples of the desired waveform shape, typically a sine wave. The DAC converts each digital sample to an analog voltage, and a reconstruction filter removes sampling artifacts.

The frequency tuning word determines the output frequency according to the relationship: output frequency equals the tuning word times the clock frequency divided by the accumulator bit width (as a power of two). For a 32-bit accumulator clocked at 100 MHz, the frequency resolution is 100 MHz divided by 2^32, approximately 0.023 Hz. This extraordinarily fine resolution enables precise frequency control impossible with analog VCO techniques.

Phase-continuous frequency switching occurs because the accumulator simply begins adding a different increment without any transient in the accumulated phase. This capability enables frequency hopping, phase modulation, and complex frequency profiles without the settling time delays inherent in analog synthesizers.

DDS Performance Characteristics

DDS performance depends on the quality of its components and the fundamental limitations of sampled systems:

Spurious-free dynamic range (SFDR) measures the ratio of the fundamental output power to the largest spurious signal. DAC nonlinearity, phase truncation in addressing the lookup table, and clock imperfections all contribute spurious content. Modern DDS chips achieve SFDR exceeding 80 dB, suitable for demanding communications and instrumentation applications.

Phase noise in DDS derives primarily from the reference clock, scaled by the frequency ratio. A 10 MHz reference with -140 dBc/Hz phase noise at 10 kHz offset produces -120 dBc/Hz at 100 MHz output (20 dB degradation for 10x frequency multiplication). Additional phase noise from DAC and digital circuitry may dominate close to the carrier.

Nyquist limitations restrict the maximum output frequency to less than half the clock frequency. Practical implementations typically limit output to 40% or less of the clock to simplify reconstruction filtering. Higher clock rates extend the usable output frequency range but increase power consumption and component performance requirements.

Switching speed can approach nanosecond scale since frequency changes require only loading a new tuning word and waiting for it to propagate through the pipeline. This speed enables applications like frequency-hopping spread spectrum that require rapid frequency changes.

Advanced DDS Techniques

Modern DDS implementations incorporate numerous enhancements beyond the basic architecture:

Spectral shaping techniques reduce spurious content through dithering, randomization of truncation errors, and optimized lookup table design. These methods spread spurious energy into the noise floor rather than allowing it to concentrate at discrete frequencies.

Integrated modulation enables direct implementation of amplitude, frequency, and phase modulation by modifying accumulator values or lookup table addressing. Built-in modulation functions simplify system design and provide precise modulation control.

Multi-channel DDS chips include multiple synthesis cores sharing a common clock, enabling phase-coherent generation of multiple frequencies for applications like I/Q modulation, phased arrays, and multi-tone testing.

Higher-order synthesis extends DDS output range by generating mixing products with analog oscillators. A DDS at lower frequency modulates a high-frequency carrier, achieving frequency coverage beyond direct DAC capabilities while maintaining DDS resolution and switching speed.

AWG Architecture

An arbitrary waveform generator centers on waveform memory that stores the desired output samples. A sample clock advances through this memory, outputting successive samples to a DAC that reconstructs the analog waveform. The memory depth determines how long a waveform can be before repeating, while the sample rate determines the maximum frequency content and timing resolution.

Memory organization varies among instruments. Simple AWGs use linear memory accessed sequentially. Advanced instruments provide segmented memory, where multiple waveform segments can be stored and combined in various sequences. This architecture enables complex test patterns without requiring enormous memory capacity.

The DAC resolution, typically 8 to 16 bits, determines amplitude resolution and dynamic range. Higher resolution reduces quantization noise but typically requires slower conversion rates due to circuit complexity. The conversion rate directly limits the maximum output frequency and edge rate.

Output reconstruction filtering smooths the stair-step DAC output into a continuous waveform. Filter design involves trade-offs between frequency response flatness, out-of-band rejection, and time-domain characteristics like overshoot and ringing.

Waveform Creation Methods

AWG waveforms originate from several sources:

Mathematical definition creates waveforms from equations describing the desired signal. Standard waveforms like sine and square are trivially computed. Complex signals might involve summing multiple frequency components, applying window functions, or computing modulation envelopes. AWG software typically provides equation editors or programming interfaces for mathematical waveform creation.

Waveform capture transfers real-world signals into AWG memory. An oscilloscope captures a signal of interest, and file transfer loads these samples into the AWG for playback. This technique enables recreation of specific anomalies, interference patterns, or complex signals that would be difficult to describe mathematically.

Simulation tools generate waveforms representing modulated signals, protocol sequences, or system outputs. Software like MATLAB or specialized communications tools creates sample data that the AWG reproduces electrically. This flow enables testing receivers and systems with precisely defined signals before actual transmitters are available.

Standard file formats facilitate waveform exchange between instruments and software. Common formats include raw binary samples, CSV text files, and specialized formats like those defined for specific instrument families. Understanding format requirements ensures successful waveform transfer.

AWG Performance Parameters

Several specifications characterize AWG capability:

Sample rate determines the maximum frequency content and timing resolution. Nyquist theory requires the sample rate to exceed twice the highest frequency component. Practical implementations typically sample at three to five times the maximum output frequency for adequate reconstruction.

Memory depth limits waveform complexity and duration. A 16 million sample memory at 1 GS/s provides 16 milliseconds of waveform before repeating. Deep memory enables long data patterns, complex modulation, and fine frequency resolution for periodic signals.

Analog bandwidth of the output stage may limit frequency response independent of sample rate. A DAC might run at 2 GS/s, but if the output amplifier has 500 MHz bandwidth, higher frequency components will be attenuated.

Timing resolution determines how precisely waveform features can be positioned. Clock jitter, interpolation accuracy, and memory addressing granularity all affect timing precision. Sub-nanosecond edge placement requires careful attention to these factors.

Effective number of bits (ENOB) captures overall dynamic range including noise and distortion, typically several bits less than the DAC resolution. ENOB varies with frequency, generally degrading at higher output frequencies.

AWG Applications

Arbitrary waveform generators address diverse application needs:

• Communications testing: Generate modulated carriers with specific symbol patterns, impairments, and interference for receiver characterization
• Radar simulation: Create complex pulse trains with precise timing, chirp modulation, and Doppler shifts
• Sensor simulation: Reproduce sensor outputs for testing processing systems without physical sensors
• Power electronics: Generate switching waveforms and transients for converter testing
• Medical electronics: Simulate physiological signals for equipment verification
• Education: Demonstrate waveform concepts with precisely controlled examples

Amplitude Modulation

Amplitude modulation (AM) varies the carrier amplitude in proportion to the modulating signal. The modulation depth, expressed as a percentage, indicates how much the amplitude varies from its unmodulated level. At 100% modulation depth, the carrier amplitude varies from zero to twice its nominal value.

Generator implementation of AM typically uses analog multipliers or variable-gain amplifiers following the carrier source. The modulating signal controls the gain, impressing its waveshape onto the carrier envelope. For DDS and AWG sources, AM can be computed digitally by scaling samples according to the modulation function.

AM signal quality depends on modulator linearity, which determines how faithfully the modulating signal appears in the carrier envelope. Nonlinearity creates distortion products that spread the signal spectrum beyond the intended bandwidth. Carrier suppression measures how completely the carrier amplitude goes to zero at 100% modulation.

Frequency and Phase Modulation

Frequency modulation (FM) varies the carrier frequency in proportion to the modulating signal. The frequency deviation specifies the peak frequency excursion from the carrier frequency. Phase modulation (PM) varies the carrier phase, producing similar spectral characteristics but with different relationships between modulating frequency and deviation.

DDS sources implement FM and PM naturally by modulating the frequency tuning word or adding phase offsets to the accumulator. The digital nature of this modulation provides precise control over deviation and linearity. Analog implementations use voltage-controlled oscillators with careful attention to linearity and tuning bandwidth.

Wideband FM, where the modulation index (deviation divided by modulating frequency) exceeds unity, produces complex spectra with many significant sidebands. Carson's rule estimates the bandwidth as twice the sum of deviation and modulating frequency. Generators intended for wideband FM testing must provide sufficient carrier frequency range and deviation capability.

Digital Modulation Formats

Digital communications systems use modulation formats designed for spectral efficiency and robustness against impairments. Testing these systems requires generators capable of producing precisely defined modulated signals:

Phase-shift keying (PSK) encodes data in discrete carrier phase states. Binary PSK uses two phases 180 degrees apart. Quadrature PSK (QPSK) uses four phases at 90-degree intervals, doubling the data rate for the same bandwidth. Higher-order PSK formats pack more bits per symbol at the cost of reduced noise margin.

Quadrature amplitude modulation (QAM) varies both amplitude and phase to create a two-dimensional constellation of symbol points. 16-QAM provides 4 bits per symbol, 64-QAM provides 6 bits, and 256-QAM provides 8 bits. The dense constellations require high signal-to-noise ratios and linear amplification.

Orthogonal frequency-division multiplexing (OFDM) distributes data across many closely spaced subcarriers. Used in WiFi, LTE, and digital broadcasting, OFDM signals have high peak-to-average power ratios that stress amplifier linearity. AWGs are particularly suited to OFDM generation due to the complex waveform structure.

Vector signal generators combine I/Q modulation with arbitrary baseband generation, enabling any linear modulation format. The baseband I and Q signals define the amplitude and phase of the carrier at each instant. This architecture provides maximum flexibility for current and future modulation formats.

Impairment Generation

Real-world signals suffer impairments that testing must evaluate. Signal generators add controlled impairments to ideal signals:

• Additive noise: Gaussian noise at specified signal-to-noise ratios tests receiver sensitivity and demodulation accuracy
• Phase noise: Close-in frequency instability tests oscillator tracking and demodulation phase reference
• I/Q imbalance: Gain and phase differences between quadrature paths test image rejection and constellation accuracy
• Nonlinear distortion: Amplifier compression effects test demodulator robustness to distorted signals
• Multipath fading: Delayed and scaled copies of the signal simulate wireless propagation environments
• Frequency offset: Carrier frequency error tests receiver tuning and tracking capabilities

Linear and Logarithmic Sweeps

Linear sweeps change frequency at a constant rate, adding the same frequency increment per unit time throughout the sweep. This mode suits applications where frequency spacing is important, such as identifying closely spaced resonances or making linear frequency axis plots.

Logarithmic sweeps change frequency at a constant percentage rate, covering each decade in equal time. This mode matches human perception of frequency and displays well on logarithmic frequency scales. Audio and acoustic measurements typically use logarithmic sweeps that efficiently cover wide frequency ranges.

Step sweeps change frequency in discrete increments rather than continuously. The frequency dwells at each step for a specified time, enabling measurements that require settling or integration time. Stepped sweeps are essential when the measurement system cannot track continuous frequency changes.

Sweep Parameters and Modes

Key sweep parameters include:

Start and stop frequencies define the sweep range. Alternatively, center frequency and span specify the same information in a form convenient for narrow-band analysis around a frequency of interest.

Sweep time determines how quickly the frequency changes. Faster sweeps improve throughput but may not allow adequate settling of filters or other frequency-selective elements in the measurement path. The minimum useful sweep time depends on the bandwidth of the device under test.

Trigger modes control sweep execution. Single sweep completes one pass and stops. Continuous sweep repeats indefinitely for real-time observation of changing responses. Triggered sweep awaits an external trigger before starting, enabling synchronization with other events.

Marker output generates a pulse when the sweep crosses specified frequencies. These markers trigger oscilloscopes, counters, or other instruments to make measurements at specific frequencies within the sweep.

Swept Measurement Considerations

Accurate swept measurements require attention to several factors:

Resolution bandwidth of the measurement receiver must accommodate the sweep rate. A receiver with 1 kHz resolution bandwidth requires roughly 1 millisecond to respond to each frequency component. Sweeping faster than this rate underestimates response peaks and distorts the measured frequency response.

Tracking errors occur when the receiver cannot follow the source frequency precisely. Phase-locked tracking receivers maintain accurate frequency alignment but may have limited tracking bandwidth. The sweep rate must remain within the receiver's tracking capability.

Amplitude flatness of the swept source affects measurement accuracy. Variations in source amplitude with frequency appear as device response variations. Calibration procedures normalize out source flatness, but residual errors remain. Source specifications should indicate amplitude flatness across the sweep range.

Harmonics and spurious from the source can corrupt measurements of nonlinear devices or systems with harmonic-related responses. Low-distortion sources or harmonic filtering may be necessary for critical measurements.

Chirp Signals

Chirp signals represent special swept waveforms where frequency varies within each waveform period. Linear chirps, also called frequency ramps, find extensive use in radar, sonar, and spread-spectrum communications.

The chirp rate (frequency change per unit time) and bandwidth determine the signal characteristics. Time-bandwidth product, the duration times the bandwidth, indicates the processing gain achievable through matched filtering. Large time-bandwidth products enable radar systems to achieve high resolution while maintaining adequate signal energy.

AWGs generate chirp signals by computing the appropriate instantaneous frequency at each sample time. The phase accumulates as the integral of frequency, creating the characteristic parabolic phase function of a linear chirp. Nonlinear chirps with optimized frequency profiles reduce sidelobe levels in the matched filter output.

Burst Modes and Parameters

Generators typically offer several burst configurations:

N-cycle burst generates a specified number of waveform cycles, then stops or holds at a defined level. The waveform begins at a specified starting phase and can end at the natural cycle completion or at a defined stopping phase. This mode suits applications requiring coherent bursts with controlled phase relationships.

Gated burst generates output while an external gate signal is active. The output follows the gate timing, enabling external control of burst duration. Gate response time affects how precisely the output follows the gate signal.

Triggered burst awaits a trigger event, generates the specified burst, and awaits the next trigger. This mode enables synchronization of bursts with external events. Trigger delay shifts the burst timing relative to the trigger edge.

Burst parameters include carrier frequency, number of cycles or burst duration, burst repetition rate or wait time between bursts, and starting/ending phase. Some generators provide carrier phase coherence across bursts, maintaining a predictable phase relationship regardless of burst timing.

Envelope Shaping

Abrupt burst edges contain high-frequency components that may cause spurious responses in receivers or interference with adjacent channels. Envelope shaping applies amplitude tapering to smooth burst edges, reducing spectral splatter while maintaining most of the burst energy.

Common envelope shapes include:

• Raised cosine: Tapers amplitude following a cosine function, providing smooth edges with good spectral containment
• Gaussian: Produces minimum time-bandwidth product, optimal for some applications
• Trapezoidal: Linear rise and fall with flat top, simple to specify and analyze
• Custom: User-defined shapes for specific requirements

The rise and fall times trade off spectral containment against timing precision. Longer transitions produce cleaner spectra but reduce the usable flat portion of the burst. Application requirements determine the appropriate balance.

Burst Timing Considerations

Accurate burst timing requires understanding several latency and uncertainty sources:

Trigger latency delays burst start from the trigger edge. This delay is typically fixed and can be compensated, but jitter in the trigger response adds timing uncertainty.

Carrier phase coherence determines whether bursts start at the same carrier phase. Coherent bursts simplify phase measurements and maintain spectral characteristics. Non-coherent bursts start at random carrier phases, appropriate for amplitude-only testing.

Minimum burst spacing limits how closely bursts can follow each other. Generator hardware requires time to reset between bursts, establishing minimum off time. Applications requiring tightly spaced bursts must verify generator capability.

Inter-burst behavior specifies output level between bursts. Options include zero output (off), holding the last sample value, or transitioning to a specified DC level. The choice affects baseline and offset behavior in subsequent processing.

Segment-Based Sequencing

In segment-based sequencing, the AWG memory contains multiple waveform segments, each with its own samples and length. A sequence table defines the order and repetition of segments. Playing the sequence outputs segments according to the table, creating arbitrarily long patterns from finite memory.

Each sequence step typically specifies:

• Segment selection: Which waveform segment to play
• Repeat count: How many times to repeat this segment before proceeding
• Next step: Where to continue after completing this step
• End action: Whether to stop, hold, or return to sequence start

This structure enables efficient storage of repetitive patterns. A serial data stream with repeated idle periods stores the idle pattern once and references it multiple times in the sequence. Only the unique portions require dedicated memory.

Conditional Branching

Advanced sequencers support conditional branching based on external inputs or internal conditions. The sequence can take different paths depending on trigger states, enabling interactive test scenarios:

Dynamic pattern selection chooses among multiple response patterns based on external conditions. A device under test might receive different stimuli depending on its state, with the sequence responding to state indicators.

Wait for trigger pauses sequence execution until an external event occurs. This synchronizes AWG output with device behavior, ensuring stimulus arrives at appropriate times in the device's operating cycle.

Conditional loops repeat segments until an exit condition is met. The sequence can idle in a waiting state, outputting a specified pattern until triggered to proceed.

Applications of Sequencing

Sequence capabilities address various testing requirements:

Protocol testing generates message sequences with headers, payloads, and timing that exercise protocol handling. Different message types appear in the sequence according to the test scenario. Error injection inserts corrupted messages at specified points.

Stress testing creates worst-case patterns that maximize power consumption, crosstalk, or other stress conditions. By identifying the specific transitions that create stress and sequencing them efficiently, long stress patterns become practical.

Radar and sonar simulation generates pulse trains with varying pulse repetition intervals, jitter, and dropouts. Environmental simulations add multiple target returns with appropriate delays and Doppler shifts.

Power supply testing creates load transient sequences that exercise regulation and transient response. Complex load profiles with varying step sizes, rates, and patterns characterize supply behavior across operating conditions.

Sequence Development Tools

Creating complex sequences benefits from software tools that abstract low-level segment and step definitions:

Graphical editors display sequence structure visually, enabling intuitive construction and modification. Timeline views show when each segment plays and how loops affect timing.

Programming interfaces enable algorithmic sequence construction from mathematical descriptions or data files. Scripts generate sequences that would be tedious to create manually.

Protocol-aware tools understand specific protocol structures, automatically creating appropriate segments and sequences from high-level descriptions. The user specifies message content and timing; the tool handles waveform details.

Simulation and verification capabilities preview sequence output before hardware execution. Timing analysis identifies conflicts or impossible configurations before they cause generator errors.

Pulse Parameters

Key pulse specifications include:

Period and frequency determine the repetition rate for continuous pulse trains. Duty cycle relates pulse width to period, expressed as a percentage.

Pulse width specifies the duration of the high (or low) state. Width specifications range from picoseconds for high-speed generators to seconds or longer for general-purpose instruments.

Rise and fall times measure how quickly the output transitions between levels. Fast edges require high-bandwidth output stages and careful attention to signal integrity. Edge speeds may differ between rising and falling transitions.

Timing resolution indicates the smallest increment for specifying delay, width, and period. High-resolution generators enable precise timing adjustments for setup/hold testing and jitter insertion.

Jitter specifications characterize timing uncertainty in the generated edges. Random jitter adds Gaussian-distributed timing variations. Deterministic jitter sources include reference clock imperfections and internal switching artifacts.

Multi-Channel Pulse Generation

Many applications require multiple precisely timed pulse outputs. Multi-channel pulse generators provide several outputs with controlled timing relationships:

Independent channels have separately programmable parameters, enabling different frequencies, widths, and amplitudes on each output. Common triggering synchronizes the channels when required.

Delay between channels positions one pulse relative to another. Setup and hold time testing requires edges at precisely controlled intervals. Channel-to-channel skew specifications indicate timing match between outputs.

Complementary outputs provide both true and inverted versions of a pulse, useful for differential signaling or push-pull driving.

Pulse Pattern Generation

Beyond regular pulse trains, pattern generators produce sequences of bits according to defined patterns:

Pseudorandom bit sequences (PRBS) generate patterns with noise-like spectral characteristics while remaining deterministic and repeatable. Standard sequences like PRBS-7, PRBS-15, and PRBS-31 are widely used for bit error rate testing and stress testing of digital links.

User-defined patterns specify arbitrary bit sequences for protocol testing or specific worst-case conditions. Pattern depth limits the maximum pattern length before repetition.

Data encoding applies coding schemes like 8b/10b, NRZ, or Manchester to raw data patterns. The generator produces the encoded waveform directly, simplifying testing of encoded interfaces.

External Reference and Clock

Generators accept external references to lock their internal timebase to a master clock. This synchronization ensures frequency accuracy and enables phase-coherent operation with other instruments sharing the same reference.

Common reference frequencies include 10 MHz (standard laboratory reference), 100 MHz (for higher-frequency systems), and various frequencies specific to application domains. Reference input specifications include frequency range, level requirements, and locking bandwidth.

Some generators accept external sample clocks, running their DACs directly from the external clock rather than an internal synthesized clock. This mode enables precise phase alignment with other clocked systems at the cost of losing internal frequency control.

Trigger Inputs and Outputs

Trigger inputs control waveform initiation, gating, and sequencing:

Edge triggering responds to rising, falling, or both edges of an external signal. Trigger level and hysteresis settings determine when edges are recognized.

Level triggering activates output while the trigger exceeds (or falls below) a threshold, enabling gate-style control.

Trigger delay shifts waveform start relative to the trigger edge. Both positive delays (waveform follows trigger) and negative delays (waveform precedes trigger, using internal pretrigger generation) may be available.

Trigger outputs provide synchronization pulses for other equipment. Marker outputs indicate specific waveform positions, enabling oscilloscopes or analyzers to capture signals at known points in complex waveforms.

Multi-Generator Synchronization

Complex test systems may require multiple generators operating in synchronization. Approaches include:

Master-slave configurations where one generator provides timing references to others. The master's sample clock or trigger outputs drive corresponding inputs on slave generators.

Common reference distribution locks all generators to a shared frequency standard. Phase alignment may require additional calibration to account for cable delays and internal path differences.

Trigger distribution provides simultaneous start signals to multiple generators. Matched cable lengths and careful attention to trigger circuit delays minimize channel-to-channel timing differences.

Dedicated synchronization buses on some instrument families provide integrated multi-unit operation with automatic timing calibration and simplified configuration.

Selecting a Waveform Generator

Choosing appropriate signal generation equipment requires matching capabilities to application needs:

• Determine required frequency range, including both the highest frequency and the frequency resolution needed
• Specify amplitude range and resolution, considering both the maximum output and the smallest increment needed
• Evaluate waveform types needed: standard functions only, or arbitrary waveforms with specific memory and sample rate requirements
• Consider modulation requirements: types of modulation, internal sources, and external modulation input specifications
• Assess timing features: trigger modes, synchronization options, and burst capabilities
• Review interface options for remote control and waveform loading
• Compare spectral purity specifications if measurements are sensitive to source distortion or noise

Signal Integrity Considerations

Delivering clean signals to the device under test requires attention to the complete signal path:

Cable quality affects signal fidelity, particularly at high frequencies. Use appropriate cables with adequate bandwidth and shielding. Connectors must maintain impedance and minimize discontinuities.

Termination prevents reflections that distort waveforms. Match the cable and load impedances. When load impedance differs from the cable, consider source termination or matching networks.

Ground loops introduce interference that may appear as signal distortion or noise. Single-point grounding or ground isolation may be necessary in sensitive measurements.

EMI considerations become important when generators produce high-frequency or high-amplitude signals. Shielded enclosures and filtered power connections prevent radiation and conduction of unwanted signals.

Calibration and Verification

Maintaining generator accuracy requires periodic calibration and verification:

Frequency calibration verifies timebase accuracy against a traceable standard. Aging of internal references causes slow drift that calibration corrects.

Amplitude calibration ensures output levels are accurate across the frequency range. Temperature variations and component aging affect amplitude accuracy.

Verification procedures confirm performance meets specifications between calibrations. Quick checks of key parameters identify problems before they affect measurements.

Adjustment procedures correct deviations found during calibration. Modern generators often include internal calibration routines that adjust correction factors without hardware modification.