Conferences

Addressing Frequency Drift in Tunable Filters

Narrowband resonant circuits exhibit degradation in performance and frequency shift when component characteristics vary due to environmental effects. When the variation is predictable and repeatable, it can be addressed at the design stage by using electronic calibration methods and mechanical sealing. When the variation occurs unpredictably and especially during operation, sophisticated solutions like built-in-test or auto-correction may be needed to counter. This article presents data on such variation and effects on the performance of an X-band tunable bandstop filter as well as the effectiveness of actions to address the variation.

Design of RF/microwave Amplifiers for Maximum Power Using Harmonic Balance Simulation when the only Available Data are the Transistor's Small Signal S-parameters

This presentation describes a method of designing RF/microwave power amplifiers using harmonic balance simulation when no non-linear models are available but only S-parameters or linear models. The design method described has its basis in the Cripps’s Load-Line Approach and Pieter Abrie’s power parameters method and could be considered as an addition to them.

Fast-tuned Active High-Q X-Band Filters

Modern warships need to rapidly detect incoming RF threats, but the Electronic Support Measures (ESM) sensors can be blinded by on-board systems such as Frequency Agile Active Electronic Scanned Array (FA AESA) radars.  These high duty cycle emitters present a serious problem for ESM systems, as traditional protection methods such as simple band-stop filters or pulse blanking will significantly degrade the Probability Of Interception (POI) of the threat signals.  This paper will first present a Fast Tuned Notch Filter (FTNF) that provides protection against FA AESA systems with negligible impact on threat POI.  The paper will then present a second generation design that takes the original concept and moves it from a large, high-component-count, duroid-based design to a 2mm square piece of GaAs die with significantly improved performance.

Intrinsic Cree GaN HEMT Models Allow More Accurate Waveform Engineered PA Designs

Following a brief review of the advantages of GaN HEMT transistors compared to other technologies and an overview of Cree’s well-proven GaN HEMT large-signal models, a new 6 port, so-called “intrinsic”, transistor model is introduced. This model gives direct access to the drain current generator ports as well as automatically “de-embedding” package parasitics (for packaged devices) and drain-to-source capacitance. Such a model allows a more accurate simulation of voltage and current waveforms and dynamic load-lines at the “intrinsic” transistor. With the increasing popularity of wave-form engineered PA’s such as Class F, inverse Class F, continuous Class F etc. to improve efficiencies and instantaneous bandwidths it is important to be able to properly match the transistor(s) and terminate the harmonics such that the intrinsic transistor drain voltage and current waveforms emulate ideal conditions such as half-sinusoid and square in anti-phase. Three examples of the use of this new modeling approach will be presented. The first example investigates the waveforms and load-lines of a narrow-band, 120 watt, S-Band amplifier using the Cree CGH40120F transistor where three harmonic impedances are used both in the input and output matching. The second example shows the use of the 6 port model in a 2 GHz amplifier, employing the Cree CGH40025F transistor, with respect to output mis-match robustness. VSWR’s of 10:1 with arbitrary phase are presented to the PA and the maximum excursions in drain voltage and current are determined. In addition, since the model incorporates a self-heating feature, transistor channel temperature can also be monitored. Such analyses allow the designer to ascertain whether a PA under VSWR mis-match conditions is likely to fail owing to extreme voltage or current overdrive or over-heating. The final example involves a low-power Doherty PA (using Cree CGH40010F transistors) where the voltage and current waveforms as well as dynamic load-lines of both the main and peaker amplifiers are investigated as a function of RF input drive level. The example clearly indicates the power of the new 6 port model in allowing visualization of waveforms, particularly as the peaker amplifier “turns on” and the effect of the active load pull of the peaker on the waveforms at the main amplifier.

Measurement of Fast Settling Microwave Synthesizers

Microwave synthesisers employed in electronic-warfare applications are required to possess performance attributes that place extreme demands on the test and measurement methods utilized. Specifically, the task of measuring frequency switching speed over a large bandwidth with sufficient fidelity is non-trivial. This paper presents a method of real-time analysis of frequency settling time using automated test routines. Issues related to achieving the required measurement fidelity are discussed, along with a summary of the theoretical performance bounds for frequency estimation.

Non-Invasive Testing of RF Packaged Devices

As the world of Integrated Circuitry soars higher in frequency, and complexity, but forever smaller in size, how do we the test these devices? The answer may lie with an anisotropically conductive material that is referred to as PariPoser Interconnects. PariPoser material makes electrical connections uniformly between opposing contact areas using conductive columns that are regularly distributed within a sheet of advanced polymer. Pariposer material has been demonstrated as being a very low loss interconnection at frequencies well above 60 GHz, and has resulted in a range of interconnection products for use in both test and production applications.  

On the use of the Auxiliary Generator Technique to Improve Non-linear Microwave Circuit Analysis and Design

We present a simulation technique which is very easily implemented in AWRDE Microwave Office and can help Harmonic Balance to converge to a new set of stationary solutions, which may be present in a real circuit but do not show up during normal simulations due to intrinsic limitations of the frequency domain methods such as the Harmonic Balance algorithm. We discuss non convergence problems with Harmonic Balance and likely causes for differences between simulated and measured results. We give some basic insights into the different approaches of the time domain and frequency domain analysis tools.

Wideband Class B Amplifiers for Split Frequency Envelope Modulated RF Power Amplifiers

Envelope modulation is an efficient means of amplifying high dynamic range signals like LTE. Overall efficiency is the product of the RF amplifier’s and the envelope modulator’s efficiencies. In the split frequency modulator architecture, the linear amplifier has the largest effect on overall performance. This paper examines a number of linear amplifier architectures. For high efficiency, a class G complementary current mode amplifier achieves 72% modulator efficiency at 17W output power with a 3MHz bandwidth LTE signal. For bandwidth a class B totem pole amplifier achieves 62% efficiency at 9W output power with a 20MHz bandwidth LTE signal.