On the development and automation of a high-speed load-pull system based on Pxie modules

Husseini, Thoalfukar 2018. On the development and automation of a high-speed load-pull system based on Pxie modules. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF - Published Version Download (13MB) | Preview
	PDF - Supplemental Material Restricted to Repository staff only Download (147kB)

Abstract

Recent RF applications and research require thousands of accurate measurements to be performed within a practical time. For instance, the global model extraction of a DUT requires thousands of accurate measurements, which would take a very long time when using the traditional RF measurement systems because they are relatively slow. Moreover, the inaccessible software that is used by the traditional systems has made them a vendor-defined system, where their application cannot be extended or amended. This is contrary to the need for a flexible RF system that can be extended and modified according to user preferences. Furthermore, the traditional load-pull measurement strategies are time-consuming thanks to the iteration process and the need for the user interaction. Therefore, developing a new high-speed measurement is essential. This work demonstrates a high-speed load-pull measurement system that maintains flexibility, accuracy, speed, and high dynamic range. The system’s architecture is based on PXIe modules, where the signal Thoalfukar Husseini detection is achieved through the use of vector signal analysers (VSA) that can operate over 50 MHz frequency bandwidth. The RF signal generation employs vector signal generators (VSG) using continuous wave (CW) mode generation. The system is calibratable over a 100 dB dynamic range and the measurement speed approaches 200 measurements/ sec at 10K samples/average. Due to the accessibility of the raw measured data and the customisable written software, statistical information has been employed to monitor the quality of the measurements and the status of the system. Moreover, an automated active load-pull measurement has been implemented on this measurement system. The automated process has been achieved by exploiting the load-based Cardiff behavioural model. This model is used to predict the required injected signals a21 to emulate a load impedance at the DUT reference plane, wherein the results show the ability of the model to achieve a load-target with an error less than -35 dB. The prediction of the DUT’s response by the model combined with customised software has allowed for an automated fundamental active load-pull process that requires minimum user interaction to automatically identify the optimum load conditions for the design-relevant parameters (e.g. gain, efficiency or output power over one or multi-power levels) within a few seconds. Two methods have been used to take the impact of the test-set on the generated signal into account: descriptive function and ix Thoalfukar Husseini simple look-up table. These two approaches have been implemented and verified. The results show that each model can achieve the power target with a residual error of less than 0.1 dB. The automation process has not only covered the definition of the optimum impedances over different power levels but has also identified, in a time efficient manner, the appropriate load-pull impedance space. This ensures that the model’s coefficients, which are required for predicting the DUT’s response b21 and efficiency, are accurately extracted. This approach significantly reduces the number of required measurements, and hence reduces the measurement time when compared to the traditional approach. It takes less than 42 sec to perform 1282 load-pull measurements, that define the appropriate design space (-3dB power contours) for 16 power levels while ensuring that the a- wave based Cardiff behavioural model is simultaneously and accurately extracted. For the sake of an efficient utilization of the measurement system and further reduction in the required number of measurements required to generate a global behavioural model that is compatible for CAD-tool design, a linear interpolation approach over the extracted coefficients was employed and verified. This approach has allowed x Thoalfukar Husseini further reduction in the number of measurements because there is no need to perform the load-pull measurement over a high dense grid of input drive power levels (a11), which is essential for the global model generation.

Item Type:	Thesis (PhD)
Date Type:	Submission
Status:	Unpublished
Schools:	Engineering
Uncontrolled Keywords:	Measurement System based on PXIe Modules; Load-Pull Measurement system; Automated Load-Pull Measurement; Behavioural model Extraction; Cardiff Behavioural model; Harmonic Load-Pull.
Date of First Compliant Deposit:	30 January 2019
Last Modified:	11 Dec 2020 02:51
URI:	https://orca.cardiff.ac.uk/id/eprint/119018

Actions (repository staff only)

Edit Item