Creating a waveform

[Robert Gallagher]

Hi all,

I have created a weighted chirped waveform for use on my models. 

The issue I am encountering is when I run a waveform with a time longer than the time window gprMax truncates the waveform. In this example my time window is 3e-9 and my waveform is 5e-9. The model runs but displays the following error message.

However, when I reduce the time on my waveform to match my time window the model will not run. 

I have also tried to replicate the provided waveform structure in an attempt to resolve this. 

The model now runs but I have an additional error message 

What I am trying to determine is if the truncating of my initial waveform going to cause issues with my results? And what is the best way forward in implementing my weighted chirped waveform?

As always, any thoughts or assistance would be much appreciated.

Kind regards,

Rob.

[Robert Gallagher]

Sorry, it is difficult to see the second error message to I have attached it here.

[Craig Warren]

Dear Robert,

There are a few important points to note when you are trying to use a custom waveform:

1. You should aim to have your waveform with the same number of samples as required for the time window and time step for you simulation. You can run a simulation where you waveform has less samples, and it will be zero padded, or more samples and it will be truncated. However, in both these cases you may introduce unwanted high frequencies into your simulation due to chopping the signal abruptly, and they may not be resolved correctly. 
2. Similarly to 1, your waveform must rise smoothly from zero. Again, any abrupt changes can introduce unwanted high frequencies into your simulation which may not be resolved correctly.
3. It is also good practice to specific a time vector in your excitation file, that matches the time step and window of your simulation. It is not strictly required, but if it is not there then gprMax may use interpolation/extrapolation as necessary.

Points 1 and 2 are facts of using the FDTD method and are related to the second error message you encountered - you must choose the spatial resolution for you simulation to correctly resolve the targets/layers etc... of interest, and also to mitigate numerical dispersion. See our docs for more info on this - http://docs.gprmax.com/en/latest/gprmodelling.html#spatial-discretization

Kind regards,

Craig