Noise parameters are commonly used to design low noise amplifiers (LNAs), which are generally the first stage of a radio frequency receiver. If the LNA feeds directly into a mixer, the noise parameters of the mixer can significantly raise the overall noise parameters of the system. The extraction of mixer- based system noise parameters is complicated by the frequency conversion enabled by the mixer where the noise power on the input is at a different frequency than that on the output. This paper describes the setup, noise receiver calibration and extraction of mixer noise parameters. Additionally, challenges in low-frequency noise receiver calibration are described and illustrated using measurement examples.
Recent years have seen a more widespread use of over-the-air (OTA) measurement techniques to characterize active antennas, especially at mm-wave frequencies. Because the level of integration of circuit components with the passive radiating elements increases, more OTA-type measurements need to be made without increasing the cost of test. Test engineers are challenged by issues of repeatability, mechanical positioning, and movement. It can be difficult and time consuming to correlate results across different measurement setups. In our contribution, we put together practical steps to verify OTA measurement results and show their utility with example measurement data marking typically achievable results.
In this contribution, we present a rigorous analysis based on uncertainty propagation techniques to estimate the random variation of the controlled reflection coefficient in mixed-signal load-pull test benches. A digital-to-analog converter is commonly used in these test benches to generate the baseband signal required to synthesize the high-frequency, user-defined injected wave. To study the random noise of the injected wave, which can be mapped to the noise of the controlled reflection coefficient, we employ Jacobian sensitivity functions between the baseband signal and the RF one. First, the baseband integrated rms noise of the up-converter is evaluated, and then the upconverted noise is determined via the derived transfer function. Finally, experimental results to validate the uncertainty control bound of the synthesized reflection coefficients are presented, highlighting a full coverage of the measured reflection coefficients.
An on-wafer multi-port measurement system based on a vector network analyzer (VNA) is exploited in order to perform the wideband characterization and linearization of a dual input Doherty power amplifier (PA) at millimeter-wave frequencies. By leveraging on the error-corrected wideband waves, an algorithm is proposed in order to realize modulated signal control across a 600-MHz bandwidth at the device-under-test (DUT) on-wafer plane, eventually allowing to impose a user-defined emulated splitting ratio on the dual input. The linearization performance of such an emulated PA configuration is finally evaluated.
In this article, we present a calibration approach for a microwave sensor aimed for dielectric spectroscopy of liquids. The sensor consists of a T-junction terminated with an interdigitated electrodes (IDE) structure. Therefore, the calibration technique involves de-embedding a 3-port network (i.e., the T-junction) by using only three known materials. The sensor is calibrated for 1–2 GHz frequency range using three samples of ethanol/water mixtures with 20%, 50%, and 80% ethanol content, respectively, and the permittivity extraction approach is validated using four different samples. The retrieved complex permittivity results in a 1.8% and 6.4% average relative error for the real and imaginary parts compared to coaxial probe measurements for all samples, which shows the great performance of the proposed calibration technique.
