0
\$\begingroup\$

Is this possible to have perfect components in LTspice? No parasitic inductance, no parasitic capacitor, no resistor, etc? Is it not possible is it possible to make a model and model it with no parastic parameters?

I would like to have a perfect diode, perfect MOSFET, IGBT etc. Should I use another software for doing something like this?

If I not able to understand some phenomenoms with perfect components I won't be able to understand it with parasatic components.

\$\endgroup\$
  • 2
    \$\begingroup\$ Sometimes simulation with all "perfect" components leads to numerical or convergence issues. But in general you can use a switch of almost 0 on resistance for an ideal MOSFET, etc. \$\endgroup\$ – John D Feb 2 at 20:39
  • 2
    \$\begingroup\$ In the default LTSPICE XVII, the generic devices "D", "NPN", "PNP" etc. seem to have many of the qualities you seek. When models for real devices are substituted, parasitic capacitance, resistance are much worse. However, what you deem "parasitic" and "ideal" are subjective. For passive components, ensure that control_panel/Hacks!/ "Always default inductors to Rser=0" box is checked. \$\endgroup\$ – glen_geek Feb 2 at 20:48
6
\$\begingroup\$

LTspice solves nonlinear differential equations numerically. It's not an AI system that can figure out intelligently whether a diode is conducting or not, and thus solve your homework questions or whatever.

In general I think that models that are not "smooth" will cause convergence problems.

So a perfect resistor is fine, but a perfect diode would have to be approximated to the desired degree of accuracy by something like a arctangent function such that the diode current vs. voltage does not have an abrupt change.

A perfect capacitor or inductor would cause a singularity if you connected it to a voltage source or a current source respectively that changed instantly.

LTSpice tries to prevent you from making this mistake by helpfully ignoring you when you set rise and fall times to zero and using default values that are rather large. Similarly, the "switches" in LTspice and PSpice are made imperfect.

| improve this answer | |
\$\endgroup\$
  • 2
    \$\begingroup\$ Well said. Besides, a simulator is already "idealistic" enough since you don't have connection impedances, transmission lines have infinite bandwidth, even the "ideal" diode has a smooth inflexion point to avoid derivative problems (at least in LTspice). If you want it to be mathematically precise, you're better off researching mathematics or physics (no sarcasm, but I was tempted to mention about a spherical horse). \$\endgroup\$ – a concerned citizen Feb 2 at 22:18
2
\$\begingroup\$

In a sense all components in LTSpice are ideal, but generally networks of them are used to simulate imperfect real components, with greater or lesser fidelity.

Sometimes the defaults give you imperfect components, but you can always override that by choice. It often helps to use perfect components in a first cut proof of principle circuit, before turning up the difficulty of making it work in the real world.

Some components in the real world tend to be so imperfect, parasitic series resistance in an inductor for instance, it makes sense always to include it. Resistors and capacitors OTOH, especially when used at low frequency, tend to be OK when ideal. We know we're dealing with an RF circuit when we have to start worrying about parasitics in Rs and Cs. In an RF circuit, we also worry about parasitics in connecting wires, the nets! These behave like transmission lines.

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.