Acquisition of long waveforms using WinTDR '98

 

By: Guy Serbin, Department of Geological and Environmental Sciences, Ben Gurion University of the Negev, P.O. Box 653, Beer Sheva 84105 ISRAEL (email: serbin@mail.bgu.ac.il)

 

Introduction

WinTDR '98 may be used to collect long TDR waveforms (beyond 251 data points) for such purposes as the determination of the frequency domain dielectric permittivity of soil-water mixtures (Heimovaara, 1994, Friel and Or, 1999). The method entails collecting several consecutive screens of data which may be pieced together later for analysis via fast Fourier transforms (FFT), etc.

This manual explains how to set up WinTDR '98 such that numerous long waveforms may be collected in as painless a manner as possible.

Sampling of long waveforms can take anywhere between 5 minutes to a half an hour or more depending on the number of probes and screens being sampled. Due to this water content values, temperature, etc. must remain static during measurement as any change in these will affect the waveform.

 

How to collect long waveforms in WinTDR '98

 

Let's just say that you are interested in collecting a waveform (or several from a number of probes via multiplexer) of 2510 data points. This is the equivalent of 10 screens of data. Each screen of data requires its own setup file. Files for single probes should have only one probe defined. WinTDR appears to measure each configured probe in the list irrespective of the fact that you may only want a single measurement. Furthermore WinTDR does not appear to use the parameters defined for the specific probe and thus may cause data corruption. Setup files need be identical except for the following:

  1. Cursor location. Each setup file should have the cursor location (for a specific probe) which is offset by 10 x Distance/Division (i.e., if your D/D is equal to 0.1 m then each cursor value should be off 1 meter from the previous setup file).
  2. Reading intervals. All setup files should take readings at least 30 seconds after the previous setup file. In general offset readings by 1 minute for every group three probes or less being measured. Measurements must not overlap one another for ANY of the setup files (see chapter 7 of the WinTDR manual).
  3. Waveform filenames. Each probe in the setup files must not only differ from other probes in the file (for multiplexed probes); they must also differ from files for the same probe in a different screen (i.e., if you have a probe #9 and to screens the waveform filenames should be 9_1.wav, 9_2.wav, 9_3.wav, etc.).

Setup files may be generated in two manners, one being slightly tedious and the other being extremely tedious. The slightly tedious manner involves setting up the initial (1st screen) TDR setup file in WinTDR and then using a text editor or word processor to make subsequent files, altering only the fields required (namely cursor location, file names and acquisition times). The extremely tedious method involves using WinTDR to generate all these files.

Initial setup will follow the following manner:

  1. Setup of probes, raw waveform file names, acquisition times and intervals (you must set this at "Fixed times").
  2. You should also go into the "Preferences/Wave File Format" menu. The "Wrap Wave" option should be unchecked and the "Wave 2nd Line" option should be checked.

 

Included below is a sample setup file.

  1. D/D values (2 for 0.1 m and 3 for 0.25 m) are teal colored
  2. Cursor locations are in millimeters and are colored red
  3. Waveform file names colored dark green
  4. Sample reading start and end times in seconds since 0 time on the computer (for the specific computer that I was using it is Jan. 1, 1970 at 12:00:00 AM) and are in blue.
  5. The time interval between readings is also in seconds (here 14400 seconds of every 4 hours) in magenta.

An explanation as to how to alter this file is given at the end of the setup file.


 

[Version Section]

File Version = 2, 0, 0

 

[Analysis Section]

Analysis Method = 1

Slope Anchor = 250

 

[Header Section]

Document Type = 2

Horizontal Units = 1

Vertical Units = 1

Auto Gen Wct File = 0

Auto Gen Wav File = 1

Waveform Save Freq = 1

No TDR Wave Sample = 0

Analysis Nbr Size = 0

Analysis Peak Tol = 0

 

[Mux Section]

Mux Count = 1

Mux 1 = 0, 1, 2, 0, 3, 8, 2

 

[Probe Section]

Probe Count = 6

Probe 1 = 3, 1, 1, 19.1, 122.6, 2, 58, 4000, 10312, 0,DEFAULT.WCT,3_01.WAV

Probe 2 = 3, 2, 1, 19, 108.2, 2, 60, 4000, 10450, 0,DEFAULT.WCT,9_01.WAV

Probe 3 = 3, 3, 1, 19.3, 115.7, 3, 64, 4000, 11005, 0,DEFAULT.WCT,7_01.WAV

Probe 4 = 3, 4, 1, 19.3, 116.4, 2, 73, 4000, 10771, 0,DEFAULT.WCT,8_01.WAV

Probe 5 = 3, 5, 1, 19.9, 108.2, 3, 34, 4000, 11601, 0,DEFAULT.WCT,11_01.WAV

Probe 6 = 3, 6, 1, 19.2, 117.5, 2, 48, 4000, 12280, 0,DEFAULT.WCT,1_01.WAV

 

[Time Section]

Cycle Type = 2

Time Count = 1

Frequency %l

Is Single 1 = 0

Start Time 1 = 917479800

End Time 1 = 918169477

Interval Time 1 = 14400

Probe 1 = 0

 

[Wave File Format Section]

Is Julian = 0

Is Two Lines = 0

Is Standard = 1

Date Delimiter = /

Time Delimiter = :

Section Delimiter = ;

Other Delimiter = ,

Save in 13 bits = 1

 


To alter this file for the next screen you would have to change the following:

  1. Cursor distance to 5000 if the D/D=2 (0.1 m) or 6500 for D/D=3 (0.25 m), i.e. add 1000 ´ 10 ´ D/D to each subsequent waveform setup file.
  2. All "*_1.WAV" files should be altered to "*_2.WAV"
  3. The start time should be offset by 3 minutes, namely 180 seconds. Add 120 to 917479800.
  4. Should you desire you may alter the interval time to whatever value, so long as it is the interval in seconds.

You may also want to look at the setup files that I have included, for work with and without multiplexers. The multiplexer files have *mux_?.tdr file names.

 

Acquiring long waveforms

 

Once the setup files are ready you will need to do the following (preferably well before the setup files are to start taking measurements:

  1. Go to the directory in which the waveforms are to be saved. If you have not defined this in the setup files then this will be in the directory in which the setup files reside. Delete, move to another directory or rename any data files that are defined in the setup files as this may either slow down the importation and joining of the waveforms or cause errors in this process.
  2. Open each file. I suggest that you minimize them also.
  3. Click on each setup file. Go to the "Scheduled readings" menu. Click on "Stop scheduled readings" and then "Start scheduled readings". You have to do this or else WinTDR will not collect waveforms. This has to be done for every setup file.
  4. After the waveforms have been acquired you may then import the data into a spreadsheet such as Quattro Pro, Excel, etc. The files are usually comma and semicolon delimited. If you import to Excel 97 then two macros are available.

 

Importation of waveforms into Excel 97 and use of the macros

 

Waveform files should be imported into Excel 97 such that the rightmost worksheet is the first screen of the waveform with each sequential screen to it's left. The last screen should be the leftmost. The active cell in each worksheet must be "A1". From here you should open up the file "WinTDR macros.xls". Go to the Tools/Macro/Macros menu or press ALT+F8. Choose 'wintdr macros.xls'!WinTDRtraceImport. You will notice that a dialog box will open up. In this box you must input the number of screens being imported. After doing that you should press the "Import!" button. Hit cancel should you not desire to import the data. This macro may take several minutes to run based upon the speed of your machine as well as the quantity and type of data (the macro identifies a few different types of WinTDR wav file formats and imports accordingly).

Do not do anything else on your computer while this macro is running. I found that this macro did not work properly (i.e., ruined data sets) when I used other applications (such as ICQ, Netscape, etc.) while the macro was running.

Also, each data file (i.e., screen) must be identical with respect to the number of readings, reading times, dates, etc. as the macro counts the number of data sets in the last worksheet/screen and then assumes that all other sheets are identical. If this is not the case you will have to alter the data files such that they are identical prior to running the macro.

After the macro is finished you should examine the data to see that it imported all waveforms correctly. You may delete any extraneous sheets from the Excel file.

A sample Excel 97 workbook, 11_2.xls, may be used to test the macro.

A second macro is provided for computation of the reflection coefficient. This macro, 'wintdr macros.xls'!ComputeRC, is extremely inefficient and should only be used on fast machines or with large data sets. The averages the values of the first 6 data points in a trace and assumes it to be the impedance/voltage of a 50 ohm coaxial line. This macro then computes the reflection coefficient r via:

where Z50 and Zd are the impedances of the 50 ohm coaxial line and data points respectively.

To run the macro you must select the cell containing the first data point of the waveform. Upon running the macro you will see that a dialog box opens. You must input in the box the number of waveforms which you desire to process, the column spacing for the waveforms, whether or not you want to insert columns for the reflection coefficient (the macro writes to the column directly to the right of the data and such that any data in those cells will be overwritten), and how many data points are in the waveforms.

However, as previously stated this macro is very inefficient. This being the case I have provided a sample equation in the dialog box with which you may copy and paste to the worksheet. It assumes that the first data point is in cell "B2", but should this differ just alter the cell values in the equation, or just type in:

=(B2-AVERAGE(B$2:B$7))/(B2+AVERAGE(B$2:B$7))

where B2 denotes the first data point of the waveform and B$2:B$7 the absolutely referenced first six data points of a waveform. After the equation has been tailored for one waveform it may them be copied for all additional ones without any further modifications.

 

Sources

  1. T. J. Heimovaara, W. Bouten, and J. M. Verstraten, “Frequency Domain Analysis of Time Domain Reflectometry Waveforms 2. A Four-Component Complex Dielectric Mixing Model for Soils,” Water Resources Research, vol. 30, pp. 201-209, 1994.
  2. R. Friel and D. Or, “Frequency Analysis of Time Domain Reflectometry with Application to Dielectric Spectroscopy of Soil Constituents,” Geophysics, vol. 64, pp. 1-12, 1999.

Click here for WinTDR '98 waveform import macro file, wintdrmacros.zip

Click here for the Visual Basic for Applications source code, should you desire to modify it for a different version of Excel or whatever.

Click here for sample WinTDR '98 setup files (single probe and multiplexed) and a sample Excel 97 data file which is ready for import


Up to the top of the page

Back to Guy Serbin's TDR page

Questions, comments, suggestions? Email me.