Direct Digital Synthesizers

Frequency synthesizers are one of the principal building blocks of precise time and frequency system. Direct digital synthesizers DDS have become increasingly important as the need for small, low power, high resolution, wide frequency range, fast settling, high spectral purity synthesizers has grown. This paper reviews the various types of DDSs and the principles of their design. DDS architectures fall into 6 major categories pulse output, sine output, triangle output, phase interpolation, jitter injection, and fractional divider or pulse snatching DDSs. The details, design principles, and performance characteristics of these categories are reviewed and the pros and cons of one category verses another are presented. included is a discussion of the relationships between the design parameters of the various types of DDSs and synthesizer performance parameters such as spectral purity, phase jitter, frequency range, frequency resolution, and settling time. Of prime importance in the design of these DDSs are the requirements on spurious sideband and phase jitter levels. A theory for predicting these spur and jitter levels and of relating these levels to DDS design parameters is presented. It is shown that DDS spurs can be understood as coming from harmonic distortion in the DDS output which is aliased down to lower frequencies due to the discrete stepped nature of DDS operation. The relationship of the sizes and frequencies of these aliasing spurs to DDS parameters is discussed showing that spur considerations place fundamental limitations on the permissible frequency range of the DDS. A quantitative theory explaining how jitter injection reduces spur levels is also presented.