Microwave frequency synthesizers provide the stable signals needed in a wide range of modern electronic systems, including communications systems, civilian and miliary radars, and electronic-warfare (EW) systems. They are available in many configuarations, from the tiny integrated circuits (ICs) embedded into cellular telephones to the rugged, rack-mount enclosures used in naval shipboard radar systems. Microwave frequency synthesizers can be compared in terms of many different performance parameters, although one of the clearest differentiators is switching speed.

The switching speed of a microwave frequency synthesizer is a function of circuit topology and synthesizer technology. Microwave and RF synthesizers generate signals based on two types of architecture: indirect and direct synthesis. Indirect synthesizers, which provide the slower frequency switching speed of the two architectures, rely on a variable-frequency tuned oscillator, such as voltage- controlled oscillator (VCO) or YIG-tuned oscillator, that is stabilized in phase to a more precise, lower-frequency reference oscillator. The tuning speed of an indirect frequency synthesizer method is in the range of tens of microseconds at best and usually in the range of milliseconds.

One of the fastest commercial indrect frequency synthesizers is the IBS Series from Elcom Technologies (Fig. 1). The IBS Series synthesizers, which are available from 0. to 6.0 GHz or 0.1 to 20 GHz, achieve switching speeds of 10 to 100 µs with low phase noise and superb spectral purity. The switching speed is a function of the distance between two frequencies, with full-band switching possible in only 100 µs.

In contrast, a direct frequency synthesizer essentially selects an output by means of high-speed switching among different generated frequencies. The frequencies are produced by multiplying and dividing a low-noise reference source and selecting among signals processed through banks of filters. Unlike a VCO or YIG oscillator which must be tuned to a new frequency by means of a change in voltage or current, respectively, the frequencies in an indirect synthesizer are always "on," and can be selected subject to the switching-speed limitations of solidstate PIN-diode or FET switches, which settle in a matter of nanoseconds.

In recent years, direct-digital synthesis-(DDS) technology has improved to the point where these VHF/UHF sources are often incorporated within indirect or direct frequency synthesizers to provide fine frequency steps with fast-switching speed. Combined with traditional directsynthesis techniques, a DDS can form part of a fast-switching frequency synthesizer solution capable of settling to a new frequency in under a microsecond.

The UFS Series (Fig. 2) of broadband direct frequency synthesizers from Elcom Technologies, for example, combines DDS technology for small frequency steps with a direct analog synthesizer architecture to achieve frequency switching speed of 200 ms with 1-Hz resolution over a standard frequency range of 300 MHz to 18 GHz (Fig. 3). Since the UFS Series synthesizers have been designed as a collection of modules that can be added and subtracted as more or less frequency resolution and range is needed, they can be supplied in a traditional rack-mount enclosure as well as a compact VXI format.

It is important to apply a consistent definition of switching speed when comparing frequency synthesizers from different manufacturers. Switching speed can be applied to both output amplitude and frequency and is the delay time required by the synthesizer to change between two frequencies or two power levels or both. The delay is a combination of the time required for the synthesizer's dedicated processor to react to a command, the switch/ blanking time, the dwell time, and the settling time. The switch/blanking time is the delay from when a parallel BCD word is sent to the synthesizer and the new frequency is stable within 0.1 rad of the final output phase or within a given tolerance (±x Hz, depending on the frequency step size) of a new frequency. Blanking refers to the capability of turning off the synthesizer's RF output power during the transition from one frequency to the next, in order to avoid harmonic or spurious signals appearing at the output port during switching.

Frequency-switching characteristics can be further defined in terms of whether or not outputs are phase coherent or phase continuous (Fig. 3). With phasecoherent switching, a synthesizer can shift from one frequency to the next and then back to the original frequency, resuming with the phase that it would have had at that initial frequency had it been running at that frequency continuously. This capability is only possible in a direct analog synthesizer when all output frequencies are generated simultaneously and the synthesizer switches among these possible output frequencies. Phase-coherent switching capability is critical to certain applications, such as coherent pulse Doppler radar systems that use coherent pulse detection for predetection integration. At lower frequencies, phasecoherent synthesizers are also used in NMR/MRI spectrometry systems for medical and material-analysis applications.

Changes in frequency are phase continuous when they do not cause discontinuities in the phase (or amplitude) of the output signal. The first phase value after a frequency change is an increment of the last phase value before the change. A synthesizer capable of phase-continuous switching exhibits almost no transient behavior or noise when switching from one frequency to the next (Fig. 4). This is important in some radar and EW systems when it is necessary to generate an analog-like synthesized sweep, such as the generation of linear frequency modulation (FM) or minimum-shiftkeying (MSK) modulation. Such output signals exhibit phase transitions that are smooth with very little noise.