1,721,045 research outputs found

    Millimeter-wave load-pull techniques

    Full text link
    In this talk, the challenges in the realization of on-wafer large-signal measurement systems in W-band will be summarized and a recent implementation of a W-band on-wafer load-pull system will be presented. The availability of W-band large signal set-ups opens several possibilities, ranging from technology assessment and large-signal device modeling at sub-THz frequencies, to W-band MMIC design and characterization. Direct and accurate load-pull measurements at W-band are thus crucial in the development of sub-THz integrated circuits. The main issue at these frequencies is the realization of highly reflective loads at the on-wafer reference planes. In-situ tuners can solve this problem, but they need to be integrated with the device under test and this is not always possible. Passive mechanical tuners could be a solution, but the losses between tuner and probe tip need to be compensated in some way, otherwise the reachable reflection coefficients will present severe magnitude limitations. An active load clearly overcomes this limit and in addition, with an active load, measurement directional couplers can be placed in real-time configuration and achieve higher accuracy than purely passive systems. Active loads can be implemented in open- or closed-loop configuration. The realized system is the first load-pull system to implement a 94 GHz load by means of an active loop exploiting frequency conversion techniques. This has advantages in cost and load stability. The system is based on a HP8510 vector network analyzer. Its performances in terms of reachable load, load stability, speed and residual uncertainty will be shown and compared to a 40 GHz load-pull system based on a PNA-X. Finally, the first measurements performed on high performance InP double heterojunction bipolar transistors (DHBTs) and GaN high electron mobility transistors (HEMTs) will be presented

    A Comparison of Uncertainty Evaluation Methods for On-Wafer S-Parameter Measurements

    Full text link
    An experimental analysis of on-wafer S-parameter uncertainties is presented. Recently, two different approaches, based either on differential numerical programming or on a fully analytical solution have been introduced. In order estabilish their suitability, a careful comparison is here given for on-wafer meaurements. Through this comparison, possible limitations and causes of errors are also highlighted. Finally, the uncertainty evaluation of the 16-term error model is here presented for the first tim

    On-Wafer Calibration Algorithm for Partially Leaky Multiport Vector Network Analyzers

    Full text link
    A new solution for multiport vector network analyzer calibration is presented in this paper. The error model is divided in two separate leaky halves, each of them with crosstalk terms, but without the leakage between the two sides. This error model is particularly useful for on-wafer measurements, when multisignal probes are employed and the crosstalk among probe fingers may dramatically affect the measurement accuracy.We will show that, with a simple choice of calibration standards, the new procedure takes the same time of a classical two-port line-reflect-match or thru-reflect-line calibration. The proposed algorithm is verified with measurements and simulations in both coaxial and on-wafer environments

    VALUTAZIONE E RIDUZIONE DELLE INCERTEZZE RESIDUE DI TARATURA PER BANCHI DI LOAD-PULL ALLE ONDE MILLIMETRICHE

    No full text
    In questa memoria viene presentato un metodo per la valutazione delle incertezze residue di taratura di sistemi di load-pull a microonde. Si riporta qui per la prima volta un confronto completo tra sistemi di tipo real-time e non-real-time alle onde millimetriche. Per la valutazione delle incertezze, vengono prese in considerazione e confrontate due cifre di merito. Si mostrano quindi le differenze ottenute usando le due diverse metodologie per la valutazione di incertezze, chiarendo in quali casi è opportuno usare l’una o l’altra. Per quanto riguarda i sistemi non-real-time, grazie ad un simulatore messo a punto per lo scopo, vengono forniti i valori tipici di incertezza di tali sistemi alle onde millimetriche. Infine viene proposta una nuova metodologia per ridurre le incertezze residue di taratura nei sistemi nonreal- time. Tale tecnica è basata sull’ottimizzazione delle misure di un dispositivo thru. Come verifica, vengono mostrati gli effetti di tale ottimizzazione su misure reali di un transistore, effettuate a 40 GHz
    corecore