Function Generators vs. Arbitrary Waveform Generators
Filed under Generators
This technical guide analyzes the architectural trade-offs between standard Direct Digital Synthesis (DDS) function generators and point-by-point arbitrary waveform generators (AWGs). Selecting the correct signal source requires an understanding of how these instruments store, clock, and output waveform data. Standard function generators are highly efficient for periodic, standard waveforms like sines, squares, and ramps. However, when application demands shift toward custom, non-periodic, or high-fidelity custom signals without phase noise or aliasing, point-by-point arbitrary waveform architectures become necessary. This article compares the underlying technologies, ranging from classic benchtop DDS units to ultra-fast modular systems, to assist engineers in making an informed instrument selection.
- Key Architectures
- Direct Digital Synthesis (DDS), Point-by-Point (Trueform)
- Memory Ranges
- 64 Kpoints to 2 GSa
- Typical Sample Rates
- 200 MSa/s to 12 GSa/s
- Vertical Resolution
- 12-bit to 14-bit
Direct Digital Synthesis (DDS) Technology
Direct Digital Synthesis (DDS) is the standard architecture for traditional function generators. It uses a fixed-frequency reference clock, a phase accumulator, and a lookup table to generate periodic waveforms. The phase accumulator determines the phase angle step size, allowing the generator to output precise frequencies with exceptional frequency resolution.
For example, the Keysight Technologies 33250A uses a 12-bit, 200 MSa/s DDS architecture. It can produce sine and square waves up to 80 MHz, and pulse waveforms up to 50 MHz. However, because DDS relies on a fixed clock and phase-step skipping to adjust the output frequency, it can introduce jitter or skip points when reproducing custom arbitrary waveforms. The arbitrary waveform frequency on the 33250A is limited to 25 MHz with a 64 Kpoints memory depth.
Point-by-Point and Trueform Architectures
To overcome the limitations of DDS when generating custom signals, modern instruments employ point-by-point synthesis. Instead of skipping points to change frequency, a point-by-point arbitrary waveform generator alters the sample clock rate, ensuring that every single sample stored in memory is played back sequentially without skipping or interpolation errors.
The Keysight Technologies 33622A utilizes Trueform technology to achieve this point-by-point performance. Running at a 1 GSa/s sample rate with a 14-bit vertical resolution, it eliminates the skipped points and aliasing typical of traditional DDS generators. It supports a standard memory of 4 MSa per channel (upgradeable to 64 MSa per channel). This architecture ensures that complex, custom waveforms are reproduced with high fidelity, maintaining signal integrity even for waveforms up to 120 MHz.
High-Performance Modular AWGs
For advanced RF, microwave, and high-speed digital communications testing, standard benchtop instruments may lack the required bandwidth and sample rate. High-performance modular AWGs, such as those built for AXIe systems, provide the necessary speed, vertical resolution, and memory depth.
The Keysight Technologies M8190A is a modular AWG operating in a 2-slot AXIe form factor. It delivers up to 12 GSa/s with 12-bit vertical resolution, or up to 8 GSa/s with 14-bit resolution. With an analog bandwidth of DC to 5 GHz (Direct DAC) and a memory depth of up to 2 GSa per channel, it can synthesize highly complex, wideband signals that are impossible to generate on standard DDS or basic benchtop AWG hardware.
Key Selection Parameters
When choosing between a function generator and an AWG, evaluate these critical parameters:
- Frequency and Sample Rate: Traditional function generators focus on maximum frequency (such as 80 MHz on the 33250A), while AWGs focus on sample rate (such as 1 GSa/s on the 33622A or up to 12 GSa/s on the M8190A).
- Vertical Resolution: Higher vertical resolution (e.g., 14-bit versus 12-bit) provides a lower noise floor and better amplitude accuracy.
- Memory Depth: DDS generators typically have shallow arbitrary memory (64 Kpoints on the 33250A), whereas dedicated AWGs offer millions of points (4 MSa on the 33622A) or billions of points (2 GSa on the M8190A) to capture long, non-repeating sequences.
Example instruments
Frequently asked questions
- What is the main difference between DDS and Trueform point-by-point playback?
- DDS architecture uses a fixed clock and skips waveform points to vary the output frequency, which can lead to jitter and distortion in custom waveforms. Trueform and point-by-point architectures vary the sample rate directly, playing back every single point in memory sequentially without skipping, resulting in higher fidelity.
- When is a high-speed modular AWG necessary?
- A modular system like the AXIe-based Keysight M8190A is necessary when you require multi-gigasample-per-second sampling (up to 12 GSa/s), massive memory depth (up to 2 GSa), and wide analog bandwidths (up to 5 GHz) for RF and microwave signal generation.