Setup Module

In the Setup module, the time response functions are read into spFRF, the signal processing parameters are defined, the time range is selected, and the channels sieve is selected. Information about the time response dataset, and the references and responses in the channels sieve, is displayed in the Data Info panel. Hovering the mouse pointer over ATI Name will show some additional information about the data file or workspace imat_fn. Almost all of the setup parameters are accessible from the Setup module GUI, with the exception of the impact and exponential windows parameters, which are entered in the Windowing dialog, and some additional averaging parameters, which are entered in the Averaging dialog. If the height of the spFRF figure is such that the Triggering or Spectra panel disappears, there is a Triggering dialog and a Spectra dialog available on the Dialog menu.

There is a context menu on the plotted time response functions to bring the selected trace to the front or send it to the back. The gray patch in the upper time response plot axes is the frame cursor. Its width is equal to the Frame Length. The spectra of the time segment of data contained in the frame cursor are plotted in lower frame spectra plot axes. The Window, Amplitude Units, and Normalization setup parameters are applied to the frame spectra. There is a context menu on the frame cursor to select the Blocksize and cursor color. In addition, the context menu on the upper axes has a Frame Cursor Here selection that will move the center of the frame cursor to the mouse pointer position. On this axes context menu and also on the Setup Display menu is a Frame Cursor selection to toggle the visibility of the frame cursor on and off.

Selecting a plotted line with the mouse pointer and holding down the middle button will highlight that line and its corresponding entry in the plot legend. Selecting a line in the plot legend with the mouse pointer and holding down the middle button will highlight that line and the corresponding line in the plot axes (this feature is not available in Matlab R2014b). The associated line in the other plot axes will also be highlighted, if applicable, and the name from the legend will be displayed in the message line at the bottom of the window.

The START button at the bottom-right corner of the GUI starts the signal processing. spFRF switches to the Results module when completed. The STOP button will stop the signal processing, a file read, or a file write in progress. The ACCEPT and REJECT buttons are used with manual Averaging Acceptance to accept or reject each frame for signal processing during frame preview.

Time Range : Start & End Time – Enter the start and end limits of the Time Range to be processed. These values can also be set using the two Time Range cursors in the upper plot axes. There is a context menu on the edit boxes from which to set the full time range or to set the time range to the upper plot axes X limits.

Sampling – Enter numeric values for the Blocksize, Spectral Lines, Frame Length and Frequency Spacing. These parameters are related by Rayleigh's Criterion and Shannon's Sampling Theorem when Averaging Method is 'welch' or 'peak hold'. The blocksize is not required to be a power of two, but the minimum blocksize is 64. There is a context menu on the Blocksize edit box from which to select from the common power of two blocksizes, and on the Spectral Lines edit box from which to select  from the customary blocksize/2.56, alias-protected number of spectral lines.

When Averaging Method is 'daniell', Blocksize and Frame Length are determined by the Time Range, Averaging Lines, and Maximum FFT Factor. Spectral Lines, Frequency Spacing, and Averaging Lines are interrelated with Blocksize.

Windowing – Select the type of Window to apply to the time response data during signal processing. The choices are 'uniform' (aka, boxcar, rectangular, or none), 'hanning broad', 'hanning narrow', 'hamming', 'flattop', 'half sine', 'exponential', or ' impact\expo' (impact and exponential). Windowing is only applicable when Averaging Method is 'welch' or 'peak hold'. Window is 'uniform' when Averaging Method is 'daniell'. When Window is 'half sine', Overlap is set to 67% if Trigger Method is 'free run'.[1]

The impact and exponential window parameters are entered in the Windowing dialog shown below, which is opened from the Dialog menu or from the context menu on the Window popup menu. The impact window is applied to the reference channels, and the exponential window is applied to all channels. If Window is 'impact/expo', Triggering Method must be 'channel'.

Impact Width is entered as percentage of the time frame length. The exponential window decay is defined by Expo Cutoff and Expo Endval. The exponential window starts at a value of 1 at time 0, and Expo Cutoff and Expo Endval define a second point on the exponential decay, as shown below.

Expo Cutoff is entered as a percentage of the time frame length (T) and Expo Endval is entered as a percentage of unity. The exponential window time constant is computed such that the window decays from a value of 1 at time 0 to the value defined by Expo Endval (y-value) at the time in the frame length defined by Expo Cutoff (x-value).

Averaging : Method – Select the Averaging Method, as either 'welch', 'daniell', or 'peak hold'. The 'welch' averaging method, which may also be known as "RMS" or "stable" averaging, is the implementation of the spectral estimation technique developed by P.D. Welch in 1967. This is the ubiquitous procedure of accumulating auto- and cross-spectral functions of overlapped, windowed time frames. The 'peak hold' averaging method is processed in the same manner, except the maximum of the auto-spectral functions at each spectral line is retained. Only auto-spectra measurements are permitted with peak hold averaging.

The 'daniell' averaging method is an implementation of the spectral estimation technique developed by P.J. Daniell in 1946. In this method, the entire time records are Fourier transformed, which leads to spectra with a very large number of spectral lines. The auto- and cross-spectral functions are computed for these big-blocksize spectra. Since the individual spectral lines are statistically independent, and if the spectra are reasonably smooth as a function of frequency, the same variance reduction as in the Welch method can be obtained by averaging bands of adjacent spectral lines. This condensation of the spectral lines results in averaged spectral functions with a more typical number of spectral lines.

For the Daniell method, the time signals must begin and end at the same value, in which case no windowing is required, or for that matter allowed, and the resulting spectral functions will have no leakage errors or distortion caused by windowing. Refer to [2] for more information on the Welch, Daniell, and other spectral estimation techniques.

As implemented in spFRF, the sampling frequency of the time response functions and the selected Time Range determine the Blocksize (NBLK) for the Daniell averaging method and the Spectral Lines (NS) specifies the number of spectral lines of the resulting, averaged spectral functions. The number of spectral lines in the big-blocksize spectra (NBIG) is half of the Blocksize and the Averaging Lines (NL) defines the number of these spectral lines that will be averaged together to create one of the NS spectral lines in the averaged spectral functions. That is, NBIG = 1.28(NS x NL), where the nominal 1.28 factor accounts for the alias-protected segment of the spectral lines. However, NS and NL may not be evenly divisible into NBIG/1.28 In this case, the time records will be zero-padded such that NBIG/1.28 is evenly divisible by NS and NL. The Blocksize, Frame Length, and Freq Spacing are adjusted accordingly. A Tukey window is used to weight the spectral lines in each of the averaged bands.

Although the Daniell method was introduce over twenty years before the Welch method, it has only recently become practical to implement for the general population. The Daniell method requires computing the discrete Fourier transform of the entire time record, which could potentially be several hundreds of thousand time points. MATLAB uses a highly sophisticated FFT algorithm that factorizes the length of the transform to optimize the processing speed. The transform is fastest when the factors are small prime numbers, but there are no restrictions on NS and NL to enforce this condition. Instead, the time records are further zero-padded such that the largest transform factor is no larger than the value specified by the Maximum FFT Factor. The Blocksize, Frame Length, and Freq Spacing are adjusted accordingly. The additional spectral lines created by this zero-padding are truncated from the spectra before condensing the bands of NL spectral lines to produce the averaged spectral functions with NS spectral lines.

Additional averaging parameters are selected or entered in the Averaging dialog shown below, which is opened from the Dialog menu or from the context menu on the Averaging Method popup menu.

Averaging : Remove DC Bias – Select Remove DC Bias as either 'on' or 'off'. If this option is 'on', the mean of the entire time record is subtracted from the time record to remove the DC bias.

Averaging : Maximum FFT Factor – Enter a numeric value for the Maximum FFT Factor for the Daniell averaging method. The time records will be zero-padded such that the largest discrete Fourier transform factor is no larger than this value. This parameter is only applicable when Averaging Method is 'daniell'.

Averaging : Tukey Window Taper – Enter a numeric value between 0 and 1 for the Tukey Window Taper. A Tukey window is used to weight the spectral lines averaged for the Daniell averaging method and this parameter sets the ratio of the taper sections to the constant section. For a taper of 0, the Tukey window becomes a uniform window and for a taper of 1, the Tukey window becomes a Hanning window. This parameter is only applicable when Averaging Method is 'daniell'.

Averaging : Acceptance – Select the frame Acceptance method as either 'all' or 'manual'. If Averaging Acceptance is 'manual', the Triggering Method must be 'channel'. Manual acceptance is intended to be used for processing impact test data, previewing each frame with the chance to accept or reject the frames included in the averaging. If Averaging Acceptance is 'all', all frames are accepted without previewing. This parameter is only applicable when Averaging Method is 'welch' or 'peak hold'. Averaging Acceptance is 'all' when Averaging Method is 'daniell'.

Averaging : Frames & Overlap – Enter a numeric value for the number of Frames or the Overlap. These parameters are related by the Frame Length and Time Range. These parameters are only applicable when Averaging Method is 'welch' or 'peak hold'.

Averaging : Lines – Enter a numeric value for the number of Averaging Lines for the Daniell averaging method. Averaging Lines, Spectral Lines, and Frequency Spacing are interrelated with Blocksize. This parameter is only applicable when Averaging Method is 'daniell'.

Exclude Time Range Cursors – Segments of the time response can be excluded from the signal processing with "exclude time range cursors." When exclude time range cursors are defined in the spFRF Setup module, any overlapped frames containing any portion of an excluded time range are skipped during the processing. Excluding time ranges is only applicable when Averaging Method is 'welch'.

An exclude time range cursor is a patch cursor in a pale reddish hue that is similar to the Frame Cursor except that its width can be changed graphically. The patch has a context menu to remove the cursor or to stretch it left or right. To stretch the cursor to the right, select Stretch Right from the context menu and then click on the cursor patch. The left edge of the cursor is anchored and the right edge moves with the mouse pointer. The Stretch Left operation anchors the right edge of the cursor and the left edge moves with the mouse pointer.

Adding and removing exclude time range cursors are available from the Display menu for the Setup module or from the context menu on the upper plot axes, and Ctrl+E is the hotkey for adding a cursor. A cursor added from the Display menu is created at the center of the plot axes and a cursor added from the axes context menu is centered at the mouse pointer. The default width of an added cursor is the smaller of half a frame length or 90% of the axes x-limits.

Triggering : Method – Select the Triggering Method, as either 'free run' or 'channel'. If Triggering Method is 'free run', Averaging Acceptance must be 'all', and if it is 'channel', overlapping is not available. Triggering Channel, Slope, and Delay Type can be selected, and Triggering Level and Delay Value can be entered when Triggering Method is 'channel'. The Triggering parameters are only applicable when Averaging Method is 'welch' or 'peak hold'. Triggering Method is 'free run' when Averaging Method is 'daniell'.

Triggering : Channel – Select the Trigger Channel from the list of reference and response channels. Trigger Channel can also be selected from the context menus on the References and Responses listboxes.

Triggering : Slope – Select the Triggering Slope, as either 'positive' or 'negative'.

Triggering : Level – Enter a numeric value for the Trigger Level as a percentage of the maximum amplitude of the trigger channel response in the frame.

Triggering : Delay Type and Delay Value – Select the Triggering Delay Type, as either 'samples', 'time 'or 'percent'. The units of the numeric value entered for the Trigger Delay Value are dependent on the Trigger Delay Type, as specified in the table below. A negative Trigger Delay Value is a pretrigger. The maximum allowed delay is 100% of the frame or blocksize.

Trigger Delay Type

Trigger Delay Value

'samples'

number of samples

'time'

seconds

'percent'

percent

 

Spectra : Normalization – Select the Normalization attribute for the spectral functions, as either 'EU2' for power spectra, 'EU2/Hz' for power spectral density (PSD), or 'EU2s/Hz' for energy spectral density (ESD).

Spectra : Amplitude Units– Select the Amplitude Units attribute for the spectral functions, as either 'rms', 'peak', or 'half-peak'.

Spectra : Clear Lower & Clear Upper – Enter numeric values in Hertz for the Clear Lower and Clear Upper frequencies. The X limits of the lower, frame spectra, plot axes and Results module plot axes are set to the Clear Lower and Clear Upper frequencies. The spectral functions are truncated to the Clear Lower and Clear Upper frequency range when written to a file or a workspace imat_fn or sent to other IMAT applications or plots. There is a context menu on the edit boxes to set the Clear Lower and Clear Upper frequencies to the full Nyquist bandwidth (0 -Nyquist) , the customary alias-protected bandwidth (0 – Span), or the lower plot axes X limits..

Measurement : Spectral Functions – Check the spectral function measurements (FRF, Coherence, Auto Spectra, etc.) that are to be computed during signal processing.

Measurement : FRF Method – Select the FRF Estimation Method, as either 'H1', 'H2', 'H3', 'Hv', or 'Hd'. 'H2' and 'H3' are only available for single reference measurements. 'Hd', the singular value decomposition FRF estimator.[4] When 'Hd' is selected the FRF basis listbox is visible. Both reference and response channels can be used to estimate a set of orthogonal basis vectors that are used to derive estimates of the reference and response spectra. The spectra estimates are then used to calculate the FRF and a generalized formulation for coherence that can also include the coherence between the references (i.e., “Reference Coherence”).

Plot Function – The only Plot Function available in the Setup module is 'Time Response'.

References – Select the reference channels to plot. Use the context menu to select the attributes listed or displayed in the legend, disable the selected channels, send the selected channels to the Responses sieve, add the selected channels to the Hd FRF basis, make one selected channel the Trigger Channel, enable all references, sieve by coordinate trace, or open the Channels dialog. When enabling all references, only channels with a data type of excitation force are enabled if the "References by excitation force data type" auto sieve on read preference is selected, otherwise all disabled channels are enabled.

FRF Basis – Select the FRF basis channels to plot. Use the context menu to select the attributes listed or displayed in the legend, remove the selected channels from the Hd FRF basis, disable the selected channels, make one selected channel the Trigger Channel, select all references for the Hd FRF basis, select all drive points (references and matching responses) for the Hd FRF basis, select all virtual channels (i.e., channels with response direction of 'VIRT') for the Hd FRF basis, select the Hd FRF basis by coordinate trace, or open the Channels dialog.

Responses – Select the response channels to plot. Use the context menu to select the attributes listed or displayed in the legend, choose a sorting option, disable the selected channels, send the selected channels to the Reference sieve, add the selected channels to the Hd FRF basis, make one selected channel the Trigger Channel, enable all channels, sieve by coordinate trace, find a channel, or open the Channels dialog. When enabling all responses, channels with a data type of excitation force are not enabled if the "References by excitation force data type" auto sieve on read preference is selected otherwise all disabled channels are enabled.

Selecting Find Channel from the context menu presents an input dialog into which a response coordinate (e.g., 1234X+ or 1234X) or node (e.g., 1234) is entered. Any channel with the specified coordinate or node will be found and plotted. The direction sense (i.e., +/−) is not required and if omitted, any response coordinate with the positive or negative sense of the specified direction will be found.

The sorting options are intended to rank the channels by some data quality metric from "worst" to "best." The Sort by options are None, Dynamic RMS, and Dynamic Crest Factor. The dynamic RMS is the RMS of the signal after subtracting the mean, that is, the RMS of the dynamic part of the signal. The dynamic crest factor is the ratio of the dynamic peak to the dynamic RMS, where the dynamic peak is the peak of the signal after subtracting the mean. Sorting by dynamic RMS sorts the channels from lowest to highest. Putting the channels with the smallest response at the top of the list should aid in finding nonfunctional channels. Sorting by dynamic crest factor sorts the channels from highest to lowest. This data quality metric is often useful to find responses that contain spurious glitches.

Slide Show – The slide show is an automated means to sequentially plot time response functions by incrementing one response channel. The slide show controls on the toolbar are meant to have similar functionality to the buttons on a DVD player remote control. The slide show can step one response forward or reverse, or play the responses forward or reverse sequentially. In the slide show play mode, the response channel is automatically incremented on a timed interval. The Slide Show Pause Interval is set in Preferences dialog.

Tools > Setup > Send Plot Channels to UIPLOT – Sends the plotted time response functions in the upper plot axes, truncated to the Time Range, to UIPLOT.

Tools > Setup > Send Plot Channels to IMAT Plot – Sends the plotted time response functions in the upper plot axes, truncated to the Time Range, to an IMAT plot figure.

Tools > Setup > Send Plot Channels to spVIEW Time Response – Sends the plotted time response functions in the upper plot axes, truncated to the Time Range, to the spVIEW Time Response module. If spVIEW is already running, the functions are sent to this instance of spVIEW, otherwise a new spVIEW is started.

Tools > Setup > Send Frame Spectra to UIPLOT – Sends the plotted frame spectra in the lower plot axes, truncated to the Clear Lower and Clear Upper frequency range, to UIPLOT.

Tools > Setup > Send Frame Spectra to IMAT Plot – Sends the plotted frame spectra in the lower plot axes, truncated to the Clear Lower and Clear Upper frequency range, to an IMAT plot figure.

Tools > Setup > Send Frame Spectra to to spVIEW Spectra – Sends the plotted frame spectra in the lower plot axes, truncated to the Clear Lower and Clear Upper frequency range, to the spVIEW Spectra module. If spVIEW is already running, the functions are sent to this instance of spVIEW, otherwise a new spVIEW is started.

Tools > Setup > Signal Processing – This tools menu has a number of signal processing operations.

> Truncate to Time Range – The time response functions are truncated to the time range between the two Time Range cursors in the upper plot axes, which are at the values specified by the Start Time and End Time. The truncated time response functions are read into spFRF, replacing the previous functions.

> Remove DC Bias – The mean of the time response functions between the two Time Range cursors is subtracted from the time response functions to remove the DC bias. The modified time response functions are read into spFRF, replacing the previous functions.

> Resample – The time response functions are resampled using the MATLAB resample function as

Y = RESAMPLE(X,P,Q,N,BTA)

See the help documentation on this function for more detailed information. An input dialog is presented to define the input arguments to the function. Follow the prompts on the dialog to enter numeric values for the P, N, and BTA input arguments. The resampled time response functions are read into spFRF, replacing the previous functions.

> Decimate – The time response functions are decimated using the IMAT decimate function as

G = DECIMATE(F,FACTOR,METHOD[,ARGS])

See the help documentation on this function for more detailed information. An input dialog is presented to define the input arguments to the function. Follow the prompts on the dialog to the FACTOR, METHOD and optional ARGS input arguments. The decimated time response functions are read into spFRF, replacing the previous functions.

> Interpolate – The time response functions are interpolated using the IMAT interp function as

G = INTERP(F,'inc',INCREMENT,'lin',METHOD[,EXTRAPTYPE])

See the help documentation on this function for more detailed information. An input dialog is presented to define the input arguments to the function. Follow the prompts on the dialog to enter the INCREMENT, METHOD, and optional EXTRAPTYPE input arguments. The interpolated time response functions are read into spFRF, replacing the previous functions.

> IIR Filter – The time response functions are filtered with an infinite impulse response (IIR) filter using the IMAT filteri function as

OUT = FILTERI(IN,TYPE,WP,NPOLES[,'silent'])

where TYPE specifies the type of IIR filter: high-pass, low-pass, band-pass, or band-stop). See the help documentation on this function for more detailed information. An input dialog is presented to define the input arguments to the function. Follow the prompts on the dialog to enter theWP, NPOLES, and optional'silent' input arguments. The filtered time response functions are read into spFRF, replacing the previous functions.

> FIR Filter – The time response functions are filtered with a finite impulse response (FIR) filter using the IMAT filterf function as

OUT = FILTERF(IN,TYPE,WP,[,WS][,'silent'] [,'nopad'])

where TYPE specifies the type of FIR filter: high-pass, low-pass, band-pass, or band-stop). See the help documentation on this function for more detailed information. An input dialog is presented to define the input arguments to the function. Follow the prompts on the dialog to enter theWP and optional WS, 'silent' and 'nopad' input arguments. The filtered time response functions are read into spFRF, replacing the previous functions.

Tools > Setup > Batch Processing – This tool is used for batch processing a number of ATI files with the same signal processing setup. There is a checkbox on the dialog to enable the 'Retain current sieve' Auto Sieve on Read preference for batch processing. Use the Browse button to select a number of ATI file to batch process. The selected files in the listbox will be batch processed. Select the Write to destination as either 'adf file' or 'workspace', and the 'append' or 'replace' option. If the 'append' option is selected, use the adjacent [] button to select the Filename or Variable. A file or variable selection dialog is presented from which to select it. If the 'replace' option is selected, the spectral function results for each ATI file are written to a separate AFU file or imat_fn, replacing it if it already exists. If the Write to destination is 'adf file', the .ati extension of the filename is replaced with .afu. If the Write to destination is 'workspace', _afu is appended to the variable name. Push the START button to batch process the selected files in the listbox. As each file is processed, it is unselected from the list. Batch processing can be stopped with the STOP button and then restarted for the remaining files in the list. 

The ATI Auxiliary Function is an optional function of the user's creation that is called just before the signal processing is started. This function could be used for filtering the time responses, for example. The AFU Auxiliary Function is an optional function of the user's creation that is called just before the spectral function results are written. There is a batch processing auxiliary function template m-file found in the spfrf/tools folder that explains the input arguments with which these functions are called from the batch processing function.  To view this m-file, type edit batch_processing_auxfunc_template in the MATLAB Command Window.

Tools > Setup > Waterfall Processing – This tool controls the spFRF signal processing procedure to compute auto-spectra waterfalls over the Time Range. The Waterfall Processing GUI shown below is used to specify the waterfall processing parameters. Enter the number of Slices (i.e., the number of time segments for which to compute spectra), the Slice Length and the Slice Overlap. Also enter the number of Frames and the Frames Overlap for signal processing within each slice. Select the Plot Waterfall checkbox to plot the waterfalls after processing. The Triggering Method is set to 'free run' and the Averaging Acceptance is set to 'all'. Push the START button to begin the waterfall processing.

The Waterfall Processing tool now takes over control of the spFRF operation to sequentially increment the Time Range cursors to each slice, perform the signal processing in that slice, and accumulate the auto-spectra for every slice. Two red, dashed-line cursors mark the overall Time Range for the waterfall processing. The resulting auto-spectra imat_fn array will be the number of channels by number of slices. When the waterfall processing is completed, or stopped with the STOP button, a variable selection dialog is presented from which to save the auto-spectra waterfall imat_fn array to the MATLAB workspace. An info structure with the same variable name as the imat_fn, but with _info appended, is also saved to the workspace.

If the Plot Waterfall checkbox was selected, the waterfalls are then plotted in an IMAT 3-D plot figure, such as the example below. The plotted channel can be selected from the red spFRF menu. If the plotted waterfall becomes something other than blue lines outlining opaque white plates, pushing the blue button on the toolbar resets the plot properties.

Tools > Setup > Frame Cursor Movie – This tool is used to step the frame cursor sequentially across the Time Range at a constant increment and rate. The frame spectra in the lower plot axes are updated at each increment of the frame cursor. Enter the Frame Position, in seconds, and the left edge of the frame cursor will move to that time position. Enter the Frame Increment, in seconds, to specify how far the frame cursor moves for each increment. Enter the Pause Interval, in seconds, to specify how long the frame cursor pauses before stepping to the next increment. The controls at the bottom of the GUI are meant to have similar functionality to the buttons on a DVD player remote control. The frame cursor movie can step one increment forward or reverse, or play the increments forward or reverse sequentially. In the play mode, the position of the frame cursor is automatically incremented on a timed interval. There is a context menu on the Frame Position edit box to refresh its value to the current left edge time position of the frame cursor. This will synchronize the Frame Cursor Movie GUI with the current position of the frame cursor if it had been moved manually after the Frame Cursor Movie GUI was opened. The frame cursor movie will stop when it reaches either edge of the Time Range.

 

 

Tools > Setup > FRF Basis Virtual Channels – This tool generates virtual channels to be used as FRF basis channels for the Hd FRF method. Virtual degrees of freedom are generalized coordinates associated with constraint shapes in a similar way as modal coordinates are associated with mode shapes. Instead of assuming the physical degrees of freedom x are equal to a linear combination of mode shapes, the assumption is that they are a linear combination of the constraint shapes, i.e., x = . The constraint shapes are determined by singular value decomposition of the response channels time data, filtered to the Clear Lower and Clearer Upper frequencies, such that UUT = TTT, where each column of T contains the time history of one response channel and the constraint shapes are the columns of the left singular vector matrix U. A plot of the normalized cumulative summation of the singular values, as shown below, indicates how much of the time data is captured by a given set of constraint shapes. Select the appropriate number of virtual channels with the full-crosshair cursor. Then the corresponding constraint shapes Ũ are used to create time response functions for the virtual channels from (t) =ŨTx(t). The response node attribute for the virtual channels are numbered sequentially and the response direction attribute is 'VIRT'. The virtual channels along with the reference channels replace the current FRF basis sieve .

Tools > Setup > List Exclude Time Ranges – This tool list the exclude time ranges in a table dialog, sorted by the range start time. Selecting one of the ranges sets the x-limits of the time response plot axes in the spFRF Setup module to show that exclude time range cursor. The selected exclude time range can also be removed from the dialog.

Tools > Setup > Read Partial ADF File – This tool has options to read part of an ATI file. A file selection dialog is presented from which to browse and select the file to read.

> Measurement Run – If more than one measurement run (i.e., data acquisition start) is stored to the same ATI file, each data acquisition start increments the measurement run attribute of the records stored to the file. The file headers are read, and then a dialog such as the one below is presented from which to select the Measurement Run to read from the ATI file.

> Channels – A dialog, such as shown the one below, is presented from which to select the records (i.e., channels) to read from the ATI file.

> Time Range – Only the portion of time response functions within the range of the abscissa values as specified by the Time Range are read from the ATI file. To use this tool, an ATI file needs to have been read into spFRF to be able to set the Time Range; otherwise, the entire file is read. The way that this tool might be used is to first read one channel from a file (that is too large to read in its entirety, for example), set the Time Range, and then read all of the channels for just the selected Time Range.

> Time Range & Measurement Run – Combines these above two tools.

> Time Range & Channels – Combines these above two tools.

Tools > Setup > Read Next Measurement Run – This tool increments the Measurement Run attribute of the current ATI file by one and reads the next Measurement Run from the same ATI file.