Post Go back to editing

Simulation difference LTSpice 24.0.9 and 24.1.x (for Windows)

Thread Summary

The user encountered issues with ambiguous models and simulation differences in LTspice version 24.1.10 compared to 24.0.9. The model conflict can be resolved by renaming the custom model, and the voltage offset is due to C1's right pin not being connected, which was incorrectly handled in the earlier version. Adding a 100 MΩ resistor to ground after C1 emulates the desired effect.
AI Generated Content
Category: Software
Software Version: 24.1.10

The latest LTspice versions (eg.  24.1.10) for windows seem to behave differently compared to 24.0.9 with at least the following 2 issues found:

- ambiguous models. This can be fixed with a simple rename, but this was not a problem before when defining a custom model (see eg. file below)

- simulation difference. In the circuit below input and output are not identical but have an offset of about -1.8V

A test file : cmos_source_follower.asc

Edit Notes

spelling mistake
[edited by: StevenBos at 9:38 AM (GMT -4) on 17 Oct 2025]
  • Dear  ,

    LTspice always puts the line

    .model NMOS NMOS

    in the netlist if an nmos device is present. That's the cause of your model name conflict. A netlist with duplicate model names is ill-formed. Former LTspice versions were simply wrong to accept that.

    The voltage difference you observe is due to C1. It's right pin is not connected anywhere. Thus, no current can ever flow in C1, thus the voltage across C1 is always 0. Thus, the result you get from 24.0.9 is incorrect.

    Best Regards,
    Mathias

  • Makes sense, i´ll update the name to ANMOS or MYNMOS or similar.

    We wanted to remove the DC component with C1. By adding a 100 M resistor to ground after C1 to emulate a scope we get the same effect as we had with 24.0.9. I wonder why 24.0.9 worked though!