Düşük gerilim devreleri ve simülasyon programları ile analizi

Abstract

Günümüzde devre simulasyon programları.tasarımcılara hem zaman bakımından hem de ekonomik yönden çok yararlı olmaktadır. Bu programlarla hiç bir harcama yapmadan düşündüğünüz tasarımları yapabilir, değiştirebilirsiniz. Ancak yapılan bu analizlerin, doğruluklarının iyi olması gerekir ki tasarımların gerçekleştirilmesi safhasında gereksiz harcama yapılmamış olsun. Bunun için kullanılan simulasyon programının analizler için yeterli olduğu kabulü ile devredeki elemanları temsil edecek model parametrelerinin doğru olarak verilmesi gerekir. Bu da Üretimde kullanılan devre elemanlarının parametrelerinin iyi bir doğrulukla belirlenmesi ile mümkün olur. Bu çalışmada EXAR BlOl-NPN ve EXAR B103-PNP bipolar tranzistorl arının, SPICE, TIME-1 ve SPICE-3C simulasyon programlarında kullanılan model parametreleri belirlenmiş ve bu parametrelerden düşük gerilimle beslenen devrelerde baskın olan parametreleri alınarak birkaç düşük gerilim devresinin analizi yapılmıştır. Analiz sonuçlarıyla karşılaştırmak ve analizlerde kullanılan simulasyon programlarının nasıl yakınsadığını görmek açısından da devrelerin laboratuvar ölçümleri yapılmıştır.

Computer-aided design CCADD enables the circuit de signer to do things which are not possible with other techniques. Using the computer, he can : Observe waveforms and frequency responses of voltages and currents without loading the circuit as a probe would in an actual circuit. Predict the performance of an IC at high frequencies, without the par asi tics a breadboard introduces. Use ideal devices selectively, such as one with an infinite bandwidth or a very large gain, to isolate the effects of various de vice parameters on the circuit performance or to do futuristic "blue sky" analysis. Feed into a circuit ideal waveforms, such as extremely fast pulses or a mixture of pulses and sinusoids. Separate out dc circuitry in order to un derstand the basic part of the circuit. Open a feedback loop without disturbing the dc levels. Determine the poles and zeros of a trans fer function for even large circuits. - v -Do noise, sensitivity, worstcase and statistical analysis. The circuit designer today has programs available which allow him to do a wide variety of analysis. How ever, the different programs with their differing input formats, rules, notations and device models can be very confisung and discouraging to the inexperienced user. Fortunately, the program input formats and rules are normaly well documented, so if mistakes are made they are relatively easily detected Celt her by the computer, the program or the user!). The biggest problem, however, lies in the lack of standardization of the notation and device models being used and the measurement of the mo del parameters. The model parameters are those parameters used in the model equations to describe the device for a given set of operating conditions. Some programs are very flexible and allow some model parameters to be specifi ed indirectly. EM2, EM3, GP are non-linear models which are based on the EMI, Ebers-Moll model. EMI is the original Ebers-Moll model. It is a non-linear dc model only. EM2 is the next level of complexity. With the EMI model as its basis, it provides a first order model of the non-linear charge-storage effects and ohmic resistance. EM3 is the third level of complexity. It inclu des such second-order effects as basewidth modulation, ft and t variations with current, a better representation of the distributed col lector -base junction capacitance and an improved temperature dependence. GP is the Gummel-Poon model as implemented in the program SPICE. It differs from the Integral Charge-Control model by Gummel and Poon mainly in the input parameters required - vi -Do noise, sensitivity, worstcase and statistical analysis. The circuit designer today has programs available which allow him to do a wide variety of analysis. How ever, the different programs with their differing input formats, rules, notations and device models can be very confisung and discouraging to the inexperienced user. Fortunately, the program input formats and rules are normaly well documented, so if mistakes are made they are relatively easily detected Celt her by the computer, the program or the user!). The biggest problem, however, lies in the lack of standardization of the notation and device models being used and the measurement of the mo del parameters. The model parameters are those parameters used in the model equations to describe the device for a given set of operating conditions. Some programs are very flexible and allow some model parameters to be specifi ed indirectly. EM2, EM3, GP are non-linear models which are based on the EMI, Ebers-Moll model. EMI is the original Ebers-Moll model. It is a non-linear dc model only. EM2 is the next level of complexity. With the EMI model as its basis, it provides a first order model of the non-linear charge-storage effects and ohmic resistance. EM3 is the third level of complexity. It inclu des such second-order effects as basewidth modulation, ft and t variations with current, a better representation of the distributed col lector -base junction capacitance and an improved temperature dependence. GP is the Gummel-Poon model as implemented in the program SPICE. It differs from the Integral Charge-Control model by Gummel and Poon mainly in the input parameters required - vi -TABLE-1 Parameters that use in analysis of low-voltage circuits for SPICE Simulation Program. TABLE-2 Parameters that use in analysis of low-voltage circuits for TIME-1 Simulation Program. - viii -TABLE-3 Parameters that use in analysis of low-voltage circuits for SPICE-3C Simulation Program. To measure the basic model parameters ft, I, the ^ ' f' so' classical measurement techniques are used [11] To de - ter mine the other model parameters M, N, C. N and R. r ' » 2* EL CB' new methods which are presented [6] are used, that are not specified has a default value. Parameters All of the large signal parameters are obtained from Ln(Ic), Ln(lB) versus Vbe curve for Vcb=0 as given fig-1, from ft versus Ln CIcD curve as given fig-2 and from the variations of the h-parameters with the collector current and with the col lector -emi t ter voltage. From the ft ver- sus Ln(Ic) curve, ft and from Ln(Ic) versus Vbe curve, I are obtained directly. Using the variations of the so * ** measured h-parameters with I and V, the model parame - ters M and R can' be determined. With computer programs - ix -V N model parameters are determined. Ln(I) Pf 'BE Ln(Ic) Figure-l Ln(Ic) and Ln(lB) Figure-2 (3 versus Ln(Ic) versus Vbe curve. curve. In this thesis measurement, of model parameters which are used in low-voltage circuits is given in Section-3. In Section-2 low-voltage circuits and analysis of these with simulation programs are given and are discussed.