English
Log in
English

Input functions in OnScale

By Kevin Chan 26 February 2020

We all live in a world of dynamic signals with frequencies stretching from kilohertz for audio signals up to the exahertz range for gamma radiation. Systems are typically designed to respond to a range of frequencies. For example, our brains are sensitive to audio signals of up to 20 kHz: signals above this range do not provide us with any useful information that we can respond to. Our eyes detect visible light in the 400–790 THz range of the electromagnetic spectrum.

When running dynamic simulations, it is very important to select the correct range of frequencies to excite the model.

How Do I Select What Input Function to Use?

The function that should be used is typically determined by the operational bandwidth of interest and the outputs that are important to the user.

For determining the characteristics of a system such as an impedance curve for a piezoelectric transducer, a short pulse is typically used. This is due to its broad range of frequency content that will excite the various resonances that may exist in the system. You can find more information on this type of simulation here!

For matching experimental waveforms with simulated waveforms, ideally we need to load the same input signal in the simulation that we have used in the experiment. In this case, you should take care when using any signals with sharp amplitude changes or high harmonic content such as square waves, as the mesh is typically limited to accurately discretizing a specific bandwidth.

What Wideband Input Functions Are Available in OnScale?

Sine Function (sine)

The sine subcommand lets you create single cycle sine wave pulses to continuous wave drives. Additional argument[GS4] [KC5] [KC6] s allow the phase shifts, offsets and time delays to generate multiple variations of sinusoids. For wideband analysis, a single sine wave cycle can be input into the system. This has the following characteristics:

Sine Wave

Note that all frequency domain plots were normalized to peak amplitude value to allow a relative comparison between input functions.

Ricker Wavelet (wvlt)

The Ricker wavelet is our main wideband input signal due to its low spectral leakage. This is important for minimizing higher harmonics that may not be meshed well.

Ricker Wavelet

Blackman-Harris Windowed Signal (blak)

Similar to the Ricker wavelet, the 2nd derivative of the Blackman-Harris window is slightly narrower band with a reduced DC component. This can sometimes be useful for simulations where a DC offset in the spectrum can cause simulation to drift, for example due to unbalanced forces.

Waveforms

Gaussian Pulse (gaus)

This is used to define a two-sided gaussian pulse. Again, it is a fairly wideband signal with low spectral leakage. The peak amplitude for this waveform shifts slightly on the lower side when using the same pulse width as the equivalent sinusoid.

Input Function

Cosine Function (astep)

Apodized step can be used to create complex step functions as well as generating a cosine response.

Time DomainFrequency Domain

* Simulations were also executed for 100 cycles at the center frequency; therefore resolving the DC component perfectly will not be possible.

I Don’t See the Signal I Want to Use. Can I Import My Own?

Yes, OnScale can read in custom waveforms using a combination of the data hist and func hist commands, which accept text and CSV files.

For additional information regarding how to set up these input functions, please search our updated Command Reference, were you’ll find a detailed explanation of all the arguments:

 

Input Function

 

Kevin Chan
Kevin Chan

Kevin is a Senior Application Engineer at OnScale. He tests and helps with the development of OnScale. His background and experience with the solvers has allowed him to work on a wide range of projects with a big focus on MEMS & RF. Kevin holds a MEng in Electronic and Electrical Engineering at the University of Strathclyde.