linNet, the Software for symbolic Analysis of linear Electronic Circuits

linNet is an application to compute the transfer function of linear, electronic circuits. The computation is done symbolically, not numerically, and the result is a formula rather than a number or a series of such. The found formula is the Laplace transform of the dependencies of the voltages and currents in the circuit on the input voltages and currents.

A linear electronic circuit is a combination of the supported basic devices as listed in the table below. The circuit is input to linNet. The representation of the circuit is a list of devices with connectivity information. The interconnections are expressed by references to nodes, where a node is a point of the circuit, which normally at least two devices are connected to. This leads to a simple formal syntax, the circuit network list or simply netlist. This list can be created and maintained with a text editor; there's no graphical interface for editing a circuit.

The computed formulas are printed to the console and to the application log file and can be used for further investigation or for publications or didactic purpose.

Resistor (R)

Conductance (Y)

Capacitor (C)

Inductivity (L)

Ideal operational amplifier (op-amp) (OP)

Constant voltage source (U)

Voltage controlled voltage source (U(U))

Current controlled voltage source (U(I))

Constant current source (I)

Voltage controlled current source (I(U))

Current controlled current source (I(I))

Current probe (wire) (PI)

Table 1: Supported linear devices

To make the application somewhat more attractive it exports the computed formulas as Octave or MATLAB script code, too. Numeric evaluation becomes a simple one-line command in Octave. The formulas are exported as LTI transfer function objects so that the complete set of analysis functions from the Octave control toolbox can be applied just like that. This reaches from simple transfer function plotting to stability analyses and system response computation on arbitrary system input.

Figure 1: Simple example of a linear electronic circuit

Please refer to figure 1 as an example of how linNet works. This is a simple RLC element with a transfer function of second order. It can be represented by the following circuit netlist:

U Uin in gnd

L L in K1

C C K1 out

R R out gnd

PLOT G U_out U_in

Given this was put into file rlc.cnl, then we can run linNet:

linNet -o rlc.cnl

and would yield the output:

User-defined result G (Bode plot):

The dependency of U_out on U_in:

U_out(s) = N_U_out_U_in(s)/D_U_out_U_in(s) * U_in(s), with

N_U_out_U_in(s) = R*C * s

D_U_out_U_in(s) = L*C * s^2

                  +R*C * s

                  +1

Going to Octave and typing G to plot the transfer function (still using default device values) gives us figure 2. More plots or plots with altered device values are a matter of single commands in Octave.

Figure 2: Octave plot of transfer function of the above circuit

You will find the details of the usage of linNet, in particular installation, netlist syntax and Octave interface and an explanation of the mathematical concept of the software in the user guide. Please go to the tab Files in the header of this window to download the user guide. The download of the ready-to-use, pre-compiled software (only Windows and Linux) with or without source files can be found there as further links.

linNet means "linear network". It founds on a symbolic solver for linear equation systems. There's no way to model any non linear effects like noise, voltage or current limits, non-linear distortions or switching operations. All of these effects play an important role in real electronic circuits and a good deal even uses these effects as their principle of operation - you won't find linNet helpful for an investigation of these kind of effects or circuits. There are many numeric circuit simulation tools, which are capable to do this, in the first place the popular open source tool SPICE with all its derivates. linNet is conceptually not a competitor of these tools, although it can behave a tiny bit alike when using Octave as numeric post-processor. linNet is not the worse SPICE, linNet is different.