Working With Nonlinear Capacitors

Micro-Cap 11 handles user-defined nonlinear capacitors in two different ways. This article will discuss the two methods and the tradeoffs between them. Method 1: Using the CAPACITANCE attribute. To see how this works, look at this sample file. The top capacitor has its capacitance defined as the expression: 3n/(1+V(C1)) The capacitance is defined as 3n/(1+V(C1)) and the charge expression is not defined. Method 2: Using the CHARGE attribute. Now consider the bottom capacitor. Its charge is defined with the expression 3n*LN(1+V(C2)) The capacitance is not defined and the charge is defined as 3n*LN(1+V(C2)). Both methods use consistent expressions in that the derivative of the charge expression is equal to the capacitance expression. However, the preferred approach is the Method 2. We shall see why later. Run transient analysis and you get these plots. Both capacitors exhibit the same voltage waveforms. Here is the plot of capacitor currents. As you can see both capacitors produce the same voltage and current waveforms. How about the capacitance and charge plots? Well, here they are: The capacitance and charge curves of C1 are both correct but the charge curve for C2 is not. That's because the charge expression for it is not supplied and must be estimated as C(V1)*V(C1), which in this case is a poor approximation. That is one reason why Method 2 is superior for nonlinear capacitors. There are other, more subtle, reasons which perhaps we'll discuss in a future article. Those other reasons have to do with Method 1 being a more robust technique and more likely to converge. So if you can, use the charge expression method.