The applet plots pairs of real data files, analyzes these, outputting either their correlation function or the Fourier transform.
Circular autocorrelation is also carried out on each file, or cross-correlation between the data pair.
The lower display plots the input data and the upper display plots, either the correlation function, or the Fourier transform as selected using the upper radio buttons.
There are a number of data examples of DF error files available in the Data Source drop-down menu for examination of their dominant modal frequencies. The examples are based on error files extracted from the amplitude comparison DF applet.
The primary application is the analysis of DF error files to examine the contribution of the various features in the DF system which control the DF accuracy performance. However, the correlation and Fourier algorithms perform real processes and adapt to any length and so the applet can be used as a general purpose analyzer
The Mode Scale Factor and Mode Center Adjust scrollbars expand and recenter the Mode Display for easing mode analysis.
User data of any length meeting the data specification can be loaded for analysis. User data is entered via the data console (Control+V), activated with the Load User Data button, and selected from the Data Source menu User Data entry.
The Data Source drop-down menu includes example DF error runs of 6-port DF systems using the Amplitude Ratio, Quadratic Fit and Arc-tangent algorithms. In addition, there is a Sine/Cosine field comprising an odd number entries to demonstrate the more general application of the Fourier Transfom applet.
The mode number represents the frequency equivalent to the number of cycles occurring over the length of the data sample.
A loadable example file is test file. This can either be edited once opened or used as a data formatting guide to the load file specification.
Mouse movement in the azimuth - error space displays plotted data values with respect to the x-axis, azimuth position.

User Notes:
In default example, the error data cycles through six periods in 360°. This shows that the data is derived from a 6-port DF system and demonstrates the error contribution of the antenna pattern shape. Different algorithms are variously affected by pattern shape, but it is possible to use the resulting measure to influence the antenna design process.
Energy is spread through to higher modal frequencies and confirms that the antenna pattern deviates from the optimum required by the processing algorithm. The antenna pattern is actually distorted by pattern ripples responsible for the higher modes - these may arise due to the installation or reflections from adjacent antennas, for example.
To clarify the display, mean data offsets are removed using the 'Adjust Mean' scrollbar. In the default example, when the mean offset is balanced the maximum Fourier component corresponds to mode number six. Viewing the correlation function with the mean error removed also confirms six cycles. It also shows that the error is negatively correlated when delayed by half the antenna spacing.