# 2.2 SOURCE v1.5

#### SOURCE version 1.5.2

15 February 2022

Changes:

• The speed of MAT-file imports has been significantly improved.
• Improvements to time-domain signal processing to ensure proper lengths of time blocks.
• A preference was added to control how the naming and grouping of operational datasets is established upon import.
• Some bugs have been fixed related to unexpected shutdowns.

#### SOURCE version 1.5.1

4 February 2022

SOURCE version 1.5 brings new functionalities for order analysis, operational consistency, and several improvements to graphing.

Support for order cuts
Next to order spectra, SOURCE now supports single order cuts as well. Typically, single order cuts are created by processing a time measurement with a specified RPM channel, e.g. in Müller-BBM PAK or Siemens TestLab. SOURCE imports these order cuts from ATFX files. All cuts belonging to the same measurement are grouped into an “Order Cuts” dataset.

• Order cut magnitudes can be plotted over time, RPM, and derived frequency. The derived frequency is computed as $$f(t) = \mathrm{ord} * \mathrm{rpm} (t) / 60$$. These x-axes are selected by right-clicking on the X-axis. A-weighting can be applied as well, in which case the derived frequencies are used to get the right scaling per point regardless of the current X-axis.
• Order cuts can be used in analyses, just like any other operational dataset. This leads to the computation of order-based Blocked Forces and component TPA.
• Order cuts can be exported to ATFX only.

Operational Consistency
A new unique feature introduced in SOURCE v1.5 is “Operational Consistency”. Operational Consistency can reveal incorrect behavior in the operational measurements, i.e. if there are overloads or strange physical (non-linear) phenomena. It furthermore indicates whether the sensors are consistent with one another: no sensors have been moved or gotten loose. It does so by evaluating the consistency through Virtual Point Transformation of the response data. This needs geometrical data, i.e. the location and direction of the response channels, to be set correctly.

The most convenient way to use Operational Consistency in SOURCE:

1. Load in a DIRAC project and use the channels as master channels. This contains accurate geometrical information already. Also, Virtual Point groupings are automatically assigned to the master channels.
2. Map the channels of the operational datasets to the master channels as usual.
3. Right-click on the operational dataset and choose Plot > Operational Consistency. This will only be available when mapped channels are found that have a virtual point grouping.

Improved data import & automatic channel mapping

• An import dialog is now shown for DIRAC project imports. This allows to select the desired datasets from all matrix presets as defined in DIRAC. Also, one can define and review channel type mappings that are made based on the Description field of the sensors and excitations. This way, the right master channels can be generated and configured automatically.
• When importing other data formats, the user will be prompted whether or not the channels need to be “auto-mapped”. If yes, channel mappings are made based on the name of the channels.
• When importing ATFX files, SOURCE will now try to rebuild the project hierarchy to match the way the data was organized in the original application. This feature is enabled by default but can be disabled in Preferences under the category Data Import.

UFF export of operational data
UFF export is now available for any time or spectral dataset. All selected datasets are written into one *.UFF file.

Graphing improvements
The graphing functionality of SOURCE 1.5 has been substantially improved.

• There is a new “sum level plot”, which shows sum levels over time. This operates on both time and spectral data. The sum levels are calculated according to the spectral settings in the top left of the axes.
• The selection of channels and blocks/columns has been improved, along with the interaction for pinning channels.
• Added plots to existing axes will automatically be assigned a new color.
• Curve legends could already be defined dynamically using so-called “tags”, by typing for instance {Dataset.Name} or {Channel.Quantity}. The available tags are now previewed when typing “{” in the text box for legends. This feature will be further improved in the next release.

Other changes:

• Fourier transform (FFT) is now adjusted to separate the DC (0 Hz) component from a signal before any windowing is applied. This DC component is then added to the spectrum separately. This way, the DC value does not get distorted by applying a spectral window.
• In calculations of matrix-inverse and matrix multiplication where the frequency vector of the two datasets does not match, linear interpolation is now being applied. This enhances the quality of source characterizations.
• In time-domain conversion, you now have the option to apply low-pass and high-pass filtering on the spectra. This gets applied before they get converted to the time domain.