Çift diyodlu simetrik olmayan bir GUNN osilatörünün çalışma karakteristiklerinin teorik ve deneysel incelenmesi

Abstract

Çok diyodlu osilatör devrelerinin analizine ilişkin yayınlanmış ça lışmalarda diyodlar basitleştirilmiş devre modelleri ile belirlenmekte dir. Ayrıca osilatörün pasif düzeni de çoğu kez diyodları. yüke göre simetrik konumda kılacak bir yapıda olmaktadır. Bu tezde işe diyodları yüke göre simetrik olmayan bir konumda bulunan iki diyodlu Gunn osila- törünün analizi yapılmıştır. Bu analizde diyodlar deneysel olarak be lirlenen yarı - lineer elemanlar ile modellenmiştir. Osilatörün deney sel olarak belirlenmiş devre modelinin analizinden ayarlı kısa devre komşularına göre osilasyon frekansı ve çıkış gücünün değişimleri elde edilmektedir. Osilatörün pasif düzenine ilişkin parametre değerleri çalışmada ge liştirilen bir ölçme yöntemi ile ölçülerek belirlenmektedir. Bu yöntem de transmisyon borusunun bir kesitinde bulunan süreksizliğe karşı düşen T - devre modeli elemanları ölçülmektedir. Süreksizliğin kayıpları bu devre modelinde seri kollardaki elemanlarda değerlendirilmektedir. Ge liştirilen ölçme yönteminde ayarlı kısa devre İle süreksizlik kesitinde seri ve paralel rezonans koşullarının oluşturulması deney sırasında göz lemle yapılabilmektedir. Bunun için bir ucu ayarlı kısa devre ile son landırılmış transmisyon borusunda, diğer uçtaki yansıma katsayısının a - yarlı piston konumlarına göre değişimlerinin bir devre analizöründe göz lemlenmesi yeterli olmaktadır. Osilatörün pasif düzeni içinde en önemli yeri tutan besleme filtreli düzenin parametreleri yukarıda belirtilen yöntemle ölçülmektedir. Bu du rumda diyod tutucu düzen bilinen farklı iki reaktansla sonlandırılarak, T - devre modeli parametreleri bulunmaktadır. Ölçülen bu parametreler ve -VI- bilinen sonlandırma reaktanslarından hareketle diyod tutucu düzenin üç kapılı devre modeline ilişkin parametre değerleri hesaplanmaktadır,. Tezde ölçülen yansıma katsayıları, rastlantısal hatalar, ölçü düzeni ve yöntem hataları bakımından ayrıca ele alınarak tekrar değerlendiril mektedir. Besleme filtreli diyod tutucu düzen parametreleri ölçüldükten sonra diy odlar tek sinisoidal girişli açıklayıcı fonksiyonlarla tanımlanmış yarı lineer elemanlarla modellenmektedir. Bu modelin belirlenmesinde kul lanılacak diyodlarla tek diyodlu osilatör devreleri gerçeklenerek bu dev relerde osilasy on frekansı ve çıkış gücü ölçülmektedir.: Ayrıca osilatörün parametre değerleri belirli devre modelinden, diyod uçlarına gelen toplam pasif düzen admitansı hesaplanmaktadır. Bu durumda diy od admi tansı, diyod uçlarına indirgenmiş pasif düzen admitansının zıt işaretli değerine eşit olmaktadır. Çalışmada teorik analizden önce iki diyodlu osilatörün, ayarlı kısa dev re konumlarına göre, osilasyon frekansı ve çıkış gücü değerleri ölçülmekte dir. Analiz, osilatörün iki diyodlu devre modelinden tek diyodlu devre mode line indirgenerek yapılmaktadır. Bu durumda devredeki parametre değerleri ölçme ile elde edilen eğrilerle belirlendiğinden, osilasyon frekansı ve çıkış gücünün devre modelinden hesaplanmasında grafik yöntemler kullanıl maktadır. Tezde verilen analiz ile iki diyodlu Gram osilatöründe güç ve frekans karakteristiklerinin, osilatörün modellenmesindeki ve model parametreleri nin değerlendirilmesindeki doğruluğa bağlı olarak, iyi bir yaklaşıklıkla çıkarılabileceği, gösterilmiş bulunmaktadır. Frekans vs güç eğrilerinin hesaplanabilmesi, bu karakteristiklere ilişkin frekans bandı, mod atlama sı ve osilasyon gücünün yüke en az kayıpla aktarılması sorunlarına, iki diyodlu osilatörlerde teorik olarak çözüm aranmasına olanak vermektedir.

In all the published papers about the analysis of the multiple-diode oscillator circuits, the diodes have been represented by the simplified circuit models. Furthermore the passive circuit of the oscillator has been assumed to possess certain symmetrical structure «In this thesis, the analysis of the two-diode Gunn oscillator in that the diodes were mount ed in the axial direction of the rectangular cavity is accomplished with the utilization of the quasi- linear approximators for the Gunn diodes. Using this analysis, the frequency and power characteristics of the os cillator are found as a function of the sliding short-circuit location. In the analysis, the quasi-linear approximators for the diodes are de fined by the single-sinusoid input describing function which is found from an experimental measurement at the X-band.The other aspect of the analysis that might be leading to some difficulties is the presence of the loading effect of the diodes between each other. Since various pa rameter values of the oscillator are given by the curves that determined from the measurements, the equations related to the oscillation conditions and the stability of the solutions are studies by the graphical proce dures. For this purpose, as a first task, the two-diede oscillator circuit has to be resembled to the that of the single-diode oscillator. In this situation, the application of the graphical procedures has hem drived in the literature for some oscillators in which the diode admittance had been limited by only the oscillation voltage across the diode. In this study the diode admittance are defined by some curves related to the os cillation frequency and voltage amplitude across the diode. The various parameter values of the passive circuit are determined_x000B_-VIII- with a measurement technique which is developed during the course of this study. In this technique the parameters of the S network belonging to a discontinuity in the waveguide could have been measured. In the cir cuit model, the losses of the discontinuity are represented in the series arms of the T network. To obtain it's elements values, the driving-point impedance of the waveguide whose other end has been closed with a slid ing short-circuit is measured using common measurement method, achieving the series and parallel resonance conditions at the discontinuity plane. In the measurement technique developed here, the resonance conditions were achieved with a sliding short -circuit. Short circuit positions were located during experiment. The variations of the reflection coefficient for the sliding short locations were observed using a network analyzer. The waveguide post diode mounting structure with a bias filter at the upper end of the post has occupied an important part of the oscil lator passive circuit. The effect of the bias filter on the post para meters have been causeddifficulties for the theoretical analysis. In this case, the parameters of the T network belonging to the post struc ture are measured while the post have been terminated with two differ ent reactances. From the measured values and the terminating reactances the parameters of the three-port circuit model related to the waveguide post are computed. For the effect of the bias filter on the parameters, it was observed that the bias filter had provided an inductive element in series and shint arm of the T network. The discontinuities at the experimental setup in which the reflec tion coefficient have been measured cause some errors in the measured value's. To reduce these errors, an error model of the measurement setup has been defined. The parameter values of the model have been found by_x000B_-IX- the measurements that had been made with using some calibrated loads. The errors, that have caused by the measurement method, at the measured vaiues are estimated to determine the case in which the measurement method- errors were large. In this large error case, the mean values pos sess less error. To reduce the random errors in the measured values, a large number of the measurements are made at the same experimental set- up and then the mean values are selected among these values. The total error of the measurementtechnique developed in this thesis after all the error eliminating procedures is discussed on the post. structure whose parameter values have been calculated theoretically. Hence, it is seen that the. experimental values were obtained agree to within about 10 percent with the theoretical values. In the thesis the active devices of the oscillator are modelled by the quasi-linear elements defined with using the single-sinusoid input describing functions. In the measurement technique to estimate these quasi-linear elements, the oscillation frequency and output power are measured as the diode was oscillated in the cavity. Then the passive circuit impedance in the diode plane of the cavity is calculated with reducing the total passive circuit of the oscillator to the diode ter minals. The oscillation voltage across the diode is found from the os cillation output power and the passive circuit model of the oscillator. The diode admittance that defines the quasi-linear element equals to the reduced passive circuit admittance with reverse sign. Because of the diode admittance has determined -the response of the* diode to the single-sinusoid input, there is an oscillation with single frequency in the oscillator. To clarify this, the X-band are swept by using a frequency meter that existed in the measurement setup to measv-_x000B_ure the oscillation frequency. The diode admittance measured, in the first step belongs to the pack aged diode. But, in the place of the admittance at the analysis, it is more available that the usage of the expanded diode admittance, that is to occupy the waveguide post parameters in addition to the packaged diode admittance* In the thesis »some experimental results of the two-diode Gunn os cillator are also discussed. For the experiment an oscillator is built with measured diodes and waveguide posts. In the experiment the varia tions of the oscillation frequency and output power with the short-cir cuit locations are measured. The load impedance of the two-diode oscil lator belonging to the different oscillation modes is determined in another measurement. The purpose of the analysis of the two-diode Gunn oscillator is to find the curves of the oscillation frequency and output power with the short-circuit locations. Thus, in the first task, the analytical relations due to the oscillation modes are found. Then, with the aid of these re lations, the single-diode oscillation circuit model was drived in the form that was more suitable to find the solutions by graphical proce dures. After then, the solutions of these relations are investigated by the curves that have been provided by the experimental values. Since the Gunn- diode admittance in the equivalent single-diode os cillation circuit model depends the frequency and voltage amplitude, the element admittance is defined with a surface in the complex plane. Thus it is rather difficult to deduce the solutions from the inter - sections in the admittance plane of the curves belonging to the pas sive circuit and negative resistive element. In this case, to find out_x000B_-XI- the solutions from the intersection points, the relation of the oscil lation condition is separated into the it's real and imaginary compo nents and then the curves with single parameter are plotted in a plane, From here, the solutions are investigated by the intersection points of these curves. (The solution is determined by the intersection points that have the equal values of the same parameters. The parameters of the same frequency intersection points have the values of the voltage amplitude. In the solution points, these amplitude values that were be longed to the each component of the complex oscillation condition for the same frequency have equal values. Therefore the parameter values of the solution point are the oscillation frequency and oscillation voltage amplitude of the two-diode oscillator.., When the curves of the oscillation frequency and output power due to the experimental and theoritical results are 'compared, the follow ing conclusions are found t i) The errors of the curve of the oscillation frequencies re lated to the short-circuit locations are about -1 percent. ii) The errors of the curve of the oscillation output power, related to the short-circuit locations are about -10 percent. In addition,an error about 1mm due to the short-circuit locations are found at the same time. As another conclusion of this analysis, it shown that certain as pects of the two-diode oscillator have Seen predicted with some er rors limited by the accuracy of the oscillator circuit model and it's parameter values determined by the. experiments. In this way, it is possible to draw some general conclusions about the influence of the linear part on the behavior of the total circuit._x000B_-XII- Then, synthesis of a linear network giving the total circuit optimal properties in the sence of power combining and frequency band should be achieved for the asymmetrical oscillator passive circuit.