In RF testing, an essential attribute of every RF signal generator is the maximum output power it can supply to a device under test (DUT) while maintaining spectral purity and level accuracy. The ability to deliver a pure, accurate signal at +25 dBm or greater ensures improved measurement accuracy. Plus, it enables testing of extreme or unusual operating conditions.
These capabilities can simplify the testing of high-power amplifiers, overcome losses within automated-testequipment (ATE) systems, and address the attenuation of signals within long cable runs. Ultimately, using a signal generator with high output power can reduce the cost, size, and weight of the resulting test system or configuration.
SIMPLIFY AMPLIFIER TESTING
How can a high-output RF signal generator help when testing an amplifier? Consider a traveling-wave tube (TWT) microwave amplifier, which can produce more than 100 W or +50 dBm of output power and require input levels of +25 dBm or greater.
Unfortunately, most of today’s microwave-signal generators can’t provide leveled outputs with that much power. As a result, the only option is to connect the signal generator to an external microwave preamplifier and the support equipment necessary to monitor, level, and calibrate the signal delivered to the TWT amplifier.
It’s important to note that to prevent damage to unique or highvalue DUTs such as satellite systems and components, the highpower signal generator should include a power-clamp feature with a response time of 30 s or faster. It should also allow the user to set minimum and maximum allowable power levels.
The most failsafe approach is to implement this capability in hardware rather than software. If the instrument is inadvertently reset, the hardware implementation ensures that it will not return to a state of maximum power. Linking the power-clamp feature to external or internal power leveling provides additional protection.
So with the welfare of the DUT in mind, let’s examine a commonly used configuration for the testing of high-power amplifiers (Fig. 1). The signal generator output, which is typically +10 to +23 dBm, is fed into a preamp with 30 to 35 dB of gain. The preamp can be either a broadband device or one that matches the frequency range of the amplifier under test (AUT).
To keep the input power as flat as possible, the preamp output is fed into a leveling coupler, which provides a proportional sampling point that’s fed into the signal generator’s automatic level control (ALC). This type of dynamic mismatch correction is necessary due to the high likelihood of variation in the frequency response of the preamp. To adequately characterize the AUT gain, the coupler’s output signal must have a level accuracy of at least ±0.5 dB at the amplifier input.
The amplifier drives a water- or oil-cooled load. As with the coupler, the output of the load is proportional to the AUT output, bringing the signal into a range that’s easily measured with commercial power sensors and power meters.
This approach has three notable shortcomings. For one, the configuration is complex. For another, it requires costly high-power test accessories. Finally, its overall accuracy depends on the precision of every element within the system. Cost and complexity will increase if multiple narrowband preamps are needed to maintain proper input power across the AUT frequency range.
A signal generator with high output power enables the use of a simpler test configuration (Fig. 2). If the signal generator can provide at least +25 dBm of output power, then it’s possible to remove the preamp and external leveling coupler. This reduces system cost and improves performance.
To maintain the desired ±0.5-dB level accuracy across the AUT frequency range, the signal generator can obtain correction factors from the power meter. For greater level accuracy, the external leveling coupler can be added to this configuration (see “Optimizing Measurement Accuracy”). This enables accurate measurement of AUT input power, as well as external power leveling for the signal generator (Fig. 3). It retains some of the cost and complexity of the original configuration (Fig. 1, again). However, it eliminates the cost of either one broadband preamp or multiple narrowband preamps.
OVERCOMING ATE-SYSTEM SIGNAL LOSSES
A typical automatic test-equipment (ATE) system accumulates signal-power losses throughout a variety of system elements: cabling; switches; and the passive couplers, combiners, isolators, and so on that allow for signal sharing. The availability of greater power from the signal generator can overcome these losses and thereby ensure greater measurement accuracy. Extra power also makes it possible to insert filters and signal monitors, improving overall measurement quality.
The decision to include a high-power signal generator brings wide bandwidth and relatively low-noise amplification of stimulus signals to the ATE system. Ultimately, using such a source reduces system cost by eliminating narrowband amplifiers and the associated switching systems from the overall test setup.
When testing antennas or satellite subsystems, the signal source may reside a significant distance from the device under test. On an antenna test range, for example, the transmit antenna could be placed on a tower that’s 15 to 80 feet tall (Fig. 4). As for satellite subsystems, they may be placed into a thermal/vacuum chamber. But the test system will be outside the chamber, which may be quite large, and the access ports may be far above the floor in a high-bay building.
In such cases, the most common solution is to use long runs of coaxial cable. However, these cause considerable losses in RF power and, of course, these losses increase dramatically at higher frequencies. Although the attenuation can vary widely depending on cable quality, typical values for 100 feet of coax are 45 dB at 12 GHz and 70 dB at 20 GHz.