In the development stages of a circuit for a new PCB application, there are usually several iterations to the circuit, including testing, redesigning, building new circuits, etc. These many changes can be costly, and it is not uncommon for a project to have 4–8 changes before it can be released to the market. One item that can substantially reduce the number of changes and the associated costs is the use of a good circuit simulation software.
There are many very good software programs on the market that will allow a circuit designer to predict the electrical performance of a circuit. The predicted electrical performance simulation is done on a model of the circuit and is often related to impedance and insertion loss. There are many other circuit attributes that can be simulated, but there are usually some differences between the actual circuit performance and the predicted performance of the simulated model. Sometimes, these differences are small; other times, the differences are very significant.
Before the designer inputs the data into the simulation software, they need to make sure they know the details behind how the data was generated. Since all simulation models are not the same due to the unique desired performance of the circuit, the data being input may not be appropriate for a specific model. The inaccurate information could be caused by the input from the user, who is defining the model. Sometimes, it could be an oversight from the user, or it could be inaccuracies from the information which the user assumes to be accurate. One area of potential inaccuracy for these software programs is the dielectric constant (Dk) value for a circuit material. Although, even when the Dk value is accurate, the user may use it inappropriately due to a misunderstanding of how the Dk value was obtained and what it represents.
To read this entire column, which appeared in the April 2020 issue of Design007 Magazine, click here.