Refine
Year of publication
Document Type
Institute
In this paper, optimization of the loaded quality factor Q<sub>L</sub> for reflection-type heterojunction bipolar transistor (HBT) oscillators is investigated. The main result is an optimum relation between the S-parameter phases at the three transistor ports. A new design strategy for this type of oscillator is proposed. The analysis is verified by comparing several Ka-band monolithic-microwave integrated-circuit oscillators in GaAs HBT technology with different resonators. The measured loaded Q<sub>L</sub> values correspond to the measured phase noise of the circuits. At an oscillation frequency of 33 GHz, an excellent phase noise of -87 dBc/Hz at 100-kHz offset frequency is achieved over the whole tuning range.
AlGaN/GaN heterostructure field effect transistors (HFETs) were irradiated with 2 MeV protons, carbon, oxygen, iron and krypton ions with fluences ranging from 1 × 109 cm−2 to 1 × 1013 cm−2. DC, pulsed I–V characteristics, loadpull and S-parameters of the AlGaN HFET devices were measured before and after irradiation. In parallel, a thick GaN reference layer was also irradiated with the same ions and was characterized by X-ray diffraction, photoluminescence, Hall measurements before and after irradiation. Small changes in the device performance were observed after irradiation with carbon and oxygen at a fluence of 5 × 1010 cm−2. Remarkable changes in device characteristics were seen at a fluence of 1 × 1012 cm−2 for carbon, oxygen, iron and krypton irradiation. Similarly, remarkable changes were also observed in the GaN layer for irradiations with fluence of 1 × 1012 cm−2. The results found on devices and on the GaN layer were compared and correlated.
AlGaN/GaN heterostructure field effect transistors (HFETs) were irradiated with protons as well as carbon, oxygen, iron and krypton ions of high (68 and 120 MeV) and low (2 MeV) energy with fluences in the range from 1x107 to 1x1013 cm-2. High energy irradiation with protons, carbon and oxygen produced no degradation in devices while krypton irradiation at the fluence of 1x1010 cm-2 resulted in a small reduction of 2% in the transconductance. Similarly, for GaN samples irradiated with protons, carbon and oxygen at high energy no changes were seen by XRD, PL and Hall effect, while changes in lattice constant and a reduction in PL intensity were observed after irradiation with high energy krypton. Low energy irradiation with carbon and oxygen at a fluence of 5x1010 cm-2 results in small change in the device performance while remarkable changes in device characteristics are seen at a fluence of 1x1012 cm-2 for carbon, oxygen, iron and krypton irradiation. Similarly changes are also observed by XRD, PL and Hall effect for the thick GaN layer irradiated at the fluence of 1x1012 cm-2. The device results and GaN layer properties are strongly correlated.
The effect of proton, carbon, oxygen and krypton irradiation on AlGaN HFET devices has been studied. Irradiation was performed at 68 and 120 MeV with fluences in the range from 1 × 10^7 to 1 × 10^13 cm2 . Before and after irradiation, dc and pulsed I – V characteristics, loadpull and S -parameters of the AlGaN HFET devices were measured. A thick GaN reference layer was characterized by x-ray diffraction, photoluminescence and Hall measurements before and after irradiation. Proton, carbon and oxygen irradiation show no degradation in devices while krypton irradiation shows a small change at a fluence of 1 × 10 10 cm2 in the device characteristics. The device results are correlated with the thick GaN results.
This paper investigates the impact of possible parameter extraction errors caused by inaccurate definition of the calibration reference impedance of in-situ multiline TRL. Two calibration sets implemented on GaAs and Si/SiGe:C wafer processes were quantitatively analyzed. Obtained results demonstrated that for most practical cases, the desired 5%-level of confidence of extracted parameters of high-reflective devices can easily be achieved without additional efforts. Thus, implementation of the in-situ TRL into a characterization workflow of high-performance microwave devices can be significantly simplified.
The uncertainty calculation for on-wafer 1-port S-Parameter measurements due to the uncertainty of coplanar on-chip calibration standards is presented. Analytical expressions for the sensitivities are used and applied for typical fabrication tolerances of monolithic integrated on-chip structures. The method is verified for OSM calibration by means of simulations with a commercial calibration software as well as by measurements, where test-structures with artificial errors were used. In each case the analytically calculated deviation is compared to a numerical approach and good agreement is found. All results are given for a MMIC process on GaAs, but can be adopted for other technologies.
Fabrication and Electrical Performance of Oscillators in GaAs/GaInP-HBT MMIC Technology up to 40 GHz
(2002)
A flip-chip packaging approach for W-band GaAs chips is presented using thin-film structures on silicon as carrier substrate. Reliability investigations indicate that, depending on bump size, the CTE mismatch is not critical and an underfiller does not provide distinctive benefits. A 77 GHz VCO GaAs-HBT MMIC is flip-chip-mounted to demonstrate validity of the packaging scheme.
A MMIC Colpitts oscillator in push-push configuration with integrated frequency divider using InGaP/GaAs HBTs is presented. The output is taken from the second harmonic port while the fundamental signal is fed to a frequency divider by two thus providing a reference signal at one quarter of the output frequency. The MMIC VCO reaches state-of-the-art phase-noise performance in X-band down to -120dBc/Hz at 1MHz offset frequency at high output power and a tuning range of 4%.
In this paper, optimization of the loaded quality factor QL for reflection-type HBT oscillator is investigated. Main result is an optimum relation between the S parameter phases at the three transistor ports. To support this finding, a 35.5 GHz MMIC VCO phase noise better than -85 dBc/Hz at 100 kHz offset frequency is demonstrated.
A new method is presented how to characterize multiport devices using a two-port vector network-analyzer (VNA). Up to now, at least one of the port terminations had to be fully known to measure the S-parameters of the device. Our new measurement method overcomes this restriction. All of the device parameters and all of the port terminations are calculated from the device measurements.
Measurement uncertainties due to non-ideal calibration standards for unknown thru calibration
(2013)
Coplanar W-band push-push VCO MMICs using GaInP/GaAs HBTs are presented. One circuit operates at 77 GHz with phase noise of -92 dBc/Hz at 1 MHz offset. To our knowledge this is the first fully monolithic W-band VCO with phase noise better than -90 dBc/Hz. A second version with two varactor diodes yields an almost threefold relative tuning bandwidth.