Troubleshooting errors

If you encounter an error when running a simulation, always try checking the "print" file. This is a file with extension .flxprt that is saved in the same directory as your model. It records all operations performed by the solvers and can be very helpful when it comes to identifying problems.

Below are some errors you might encounter along with tips on how to solve them.

Bad floating point numbers found in ...

The model is going unstable, most likely due to meshing or absorbing boundary conditions. Try:

  • Increasing the number of elements in your model so that the structure is accurately represented. Preview your model to ensure there are no bad elements that may be causing issues.
  • Switching absorbing boundary conditions (absr) for impedance-matched boundaries (impd).
Grid spacing not monotonic

The model's keypoints are not monotonic. They must either be entirely non-increasing or entirely non-decreasing.

Note: For information about keypoints, see What are Keypoints?

Try looking in the .flxprt file to find the xyz node coordinates and the element spacing.

Invalid input

This occurs when you have invalid code syntax. Often this is because you haven't declared a primary command before using an associated subcommand or you have failed to close a loop.

The .flxprt file stops "printing" at the point where the error occurred. As such, look here to find the source of the problem.

Note: To check the syntax of a command, right-click the command in Analyst and select Show Help for ...

No primary command found

This occurs if you haven't declared a primary command before using an associated subcommand. Alternatively, you may have failed to close a loop.

The .flxprt file stops "printing" at the point where the error occurred. As such, look here to find the source of the problem.

Check that all primary commands have been declared and that the code syntax is correct.

Matrix too large for single precision – try iterative solver

This means that the electric window is too big for the single-precision version of the code. There are a few ways to fix this:

  • Change to the double-precision version by going to  > Settings > Solver Settings and clicking Double Precision.
  • Try a different solver via the piez slvr command. dcgd is usually the most efficient.
  • Optimize your electric window using piez wndo auto piez.
No electrodes defined

Check the following:

  • Preview your model and ensure that your model is correct, especially around the electrode area. It is possible that your mesh may not be fine enough to model the electrodes, or perhaps you changed the geometry or mesh size after you defined the electrode.
  • If you are working in Designer and using primitive shapes, check that their precedence is correct and that one is not overwriting another.
  • Ensure that you are using piez node/nod2 correctly.

The following image shows what can happen when you change the mesh size after defining an electrode:

no_nodes_defined_electrode.png

You can visualize the electrodes using the following code, after the piez command:

grph
     type standard
     line on
     plot piez
     end

Make sure the electrodes are where you want them to be!

Tip: You can use plot matr piez to plot both the model and the electrodes.

No suitable compressor found on system for fourcc cvid

This codec problem has one of two causes: either you minimized the embedded graphics window during the model run or the system doesn't have the standard video codec.

You can set a codec using the following command:

grph
     set imag avi menu

This will show a pop-up window when the first imag subcommand is encountered, listing all the video-encoding options available on the system.

Some pressure loaded faces contained massless nodes

This normally occurs when you try to apply a pressure load to a surface that is void. This is not allowed in OnScale.

The easiest way to prevent this is to use plod sdf2 to apply the pressure as the interface between two materials. You can create dummy materials to help with this.

Surface in contact with void material
This comes from the extrapolation command (extr) and is usually due to the extrapolation boundary being next to a void material. Check your extr nodecommand.