SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 27 February 2006
Published in Soil Sci Soc Am J 70:537-540 (2006)
DOI: 10.2136/sssaj2005.0176N
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Logsdon, S. D.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Logsdon, S. D.
Agricola
Right arrow Articles by Logsdon, S. D.
Related Collections
Right arrow Water Content
Right arrow Time Domain Reflectometry, TDR
Right arrow Soil Methods/Instrumentation

Soil Physics Note

Experimental Limitations of Time Domain Reflectometry Hardware for Dispersive Soils

S. D. Logsdon*

National Soil Tilth Lab., 2150 Pammel Dr., Ames, IA 50011

* Corresponding author (logsdon{at}nstl.gov)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Time domain reflectometry (TDR) has been used to determine soil water content. Multiple attachments and long cables may result in unreliable data for dispersive soils, but these attachments are necessary for automated field monitoring at multiple sites and depths. The objective of this study was to experimentally determine the effect of attachments on the TDR waveform for a parallel probe with a balun. The probe with parallel waveguides and balun was successively attached to the front panel of the cable tester (Level 1), to a transient suppressor (Level 2), to the first level multiplexer (Level 3), or to a second level multiplexer (Level 4). Each attachment level significantly reduced frequency bandwidth, and Level 4 frequency bandwidth was about half that of Level 1. The greatest attenuation decrease was between Levels 1 and 2. Attenuations due to TDR hardware make it difficult to obtain useful waveforms for determining soil water content.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
TIME DOMAIN reflectometry has been used to determine soil water content (Ferré and Topp, 2002; Robinson et al., 2003). Multiple attachments and long cables often result in unreliable data, especially for fine-textured, high-charge soils. Unfortunately using TDR for automated field measurements of soil water content has relied on attachments such as multiplexers (for multiple depths and site), cables (for multiple sites), and transient suppressors (old systems). Logsdon (2000) showed that hardware limitations and waveform analysis were the major obstacles to accurate TDR with many hardware attachments.

Interest has developed in dissecting TDR signal response from each component of the system (Feng et al., 1999; Lin, 2003; Heimovaara et al., 2004), usually limited to front panel, cable, probe head, and waveguide embedded in the soil, and sometimes a layered soil. Other field attachments (transient suppressor, multiplexers at multiple levels, low loss cable vs. high loss cable at multiple levels, balun within head) have not been examined. Most have considered quasi-coaxial probes (Feng et al., 1999; Lin, 2003). Schaap et al. (2003) examined a two-wire (parallel) probe (with no balun), combining the equations for the parallel probe with coaxial equations for the cable. The objective of this study is to experimentally determine the effect of several hardware attachments on the TDR waveform.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Webster clay loam soil (Fine-loamy, mixed, superactive, mesic Typic Endoaquoll) was packed into two PVC pipes 70-mm i.d. and 360-mm long at bulk densities of 1.10 Mg m–3 and water content of 0.288 m3 m–3. A similar column was filled with distilled water. Parallel waveguides (0.3 m long) with 1:1 baluns (Spaans and Baker, 1993) were embedded in each soil sample. The probe dimensions are waveguide thickness = 3.175 mm, waveguide spacing = 30 mm. This gives a spacing/thickness ratio of 9.4, which is <10 as recommended by Knight (1992). The fraction of electrical field (J) within the 70-mm diam. PVC pipe was determined from (Knight, 1992)

Formula 1[1]
in which m is the waveguide diameter, s is spacing, and f is measurement fractional radius [= (70–30)/2 = 20 mm]. The calculated fraction for the plane parallel with the waveguides is 0.63 for these probes in 70-mm diam. PVC tubes. Similar calculations for the perpendicular plane (f = 35 mm) result in a calculated fraction of 0.82. This meant that at least 63 to 82% of the electrical field was contained within the PVC tube assuming no attenuation. Due to high attenuation by soil, this calculated value would be a minimum.

All three columns and peripheral attachments were placed into a temperature controlled chamber. Attachments included an additional waveguide for measurements in air and water, the Tektronix 1502B cable tester, cables, transient suppressor, two multiplexers (Campbell Scientific), and a short attachment to use at the end of the waveguide in air. The Tektronix was attached to an external computer. To run the system, DOS software was used that had been developed by Dr. Heimovaara (personal communication, 2004), along with batch files developed by Dr. Schaap (personal communication, 2004), to collect 16 384 points of data (0.1 m/div), starting at 0.5 m before the front panel. The open and short measurements were used to show the input signal (Schaap et al., 2003):

Formula 2[2]
in which o and s subscripts indicate open and short signals from the program developed by Heimovaara (1994) and modified by Schaap (personal communication, 2004). Electrical length was estimated from divergence of open and short reflection coefficients, {rho}, for each level. Since each level had more than one hardware addition, the separation of the hardware pieces was determined by examining the open signal for changes indicating impedance mismatch.

Level 1 attachment was the probe with associated head and 3.7-m cable (RG58A) attached directly to the front panel of the cable tester. Level 2 attachment was the probe and associated head and cable attached to the transient suppressor, which was attached to the front panel with a 0.9-m cable (RG58A). Level 3 attachment was the probe and associated components attached to the first level multiplexer, which was attached to the transient suppressor with a 1.8-m cable (RG58A), and the other components. Level 4 attachment was the probe and associated components attached to the second level multiplexer, which was attached to the first level multiplexer with a 15-m low loss cable (RG8) and components. These attachment levels are summarized in Fig. 1 .


Figure 1
View larger version (15K):
[in this window]
[in a new window]
  		Fig. 1. Layout of the four attachment levels. The RG58A cable between the cable tester and transient suppressor is 0.9 m and between the transient suppressor and multiplexer one is 1.8 m. The RG8 low loss cable between multiplexers one and two is 15 m. The probe consists of 3.7-m RG58A cable, head containing 1:1 balun, and 0.3-m waveguides.

 
Then approximately 100 g of water was added to each column, which was then equilibrated for 5 d, and the data were collected again at the three temperatures, but only for the soils. Then approximately 100 g of water was added to each column again and the measurements repeated after equilibration for 4 d. The soil water content was determined by weighing each column immediately after the measurement and at the end of the experiment. Composite soil subsamples were taken before packing the columns and at the end of the experiment to determine gravimetric water content. Some of the 100-mL additions of water inadvertently drained from the columns, and a small amount may have evaporated even though the columns were sealed at both ends.

To differentiate attenuation in the water or soil, the reflection coefficient was measured before the waveform encountered the soil (Point a), at the peak (just before start of soil, Point b), at the valley (end of soil Point c), and at long time (Point d). The differences between points a and b (Step 1), b and c (Step 2), and c and d (Step 3) were used to characterize attenuation for attachment level, temperature, and water content (Fig. 2 ). Time (or length) to beginning of water or soil (Point b), and travel time (between Points a and b) were determined. The reflection coefficient differences and the times for soil were compared across attachment level, water content, and temperature using analysis of variance (completely randomized design).


Figure 2
View larger version (19K):
[in this window]
[in a new window]
  		Fig. 2. Effect of attachments on TDR waveform, shown for one replicate, water content of 0.352 m3 m–3 and 15°C. Level 1 is the probe (and associated cable and head) attached to the cable tester front panel, Level 2 adds a cable and transient suppressor, Level 3 adds a cable and multiplexer, and Level 4 adds a low loss cable and second multiplexer. The waveform times were adjusted so that all levels appeared in the same frame. Points pre-a, b, c, and post-d indicate the areas chosen to compare reflection coefficient differences.

 
Several definitions of rise time, {tau}, are given in the literature; for this study, the definition given by Hilhorst (1998) is used, "the time needed to reach 0.66 of the amplitude of the reflected step." Frequency bandwidth was calculated as 1/(2{pi}{tau}) as given by Hilhorst (1998). These calculations are most appropriate for lossless systems, but were still used as a basis for internal comparison even for the dispersive soils in this study. These values are not meant to be absolute values for frequency bandwidth to compare with literature values because of the dispersive nature of the samples and because of the variety of definitions for rise time.

The travel time, tt, was corrected by 44 ps within the head after the balun, and converted to water content, {theta}, by the Topp et al. (1980) equation:

Formula 3[3a]

Formula 4[3b]
in which c is the speed of light (3 x 108 m s–1), L is sample length, and {varepsilon}a is the apparent permittivity.

Calculated frequency bandwidth, and calculated water contents were tested using analysis of variance across attachment level, water content, and temperature. Similar statistics were calculated for frequency bandwidth and {varepsilon}a of water, treating temperature as replicates.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
The difference between calculated electrical length and physical length (Table 1) was due to variations in the impedance and permittivity of each component (not shown). The impedance and permittivity are interrelated and a range of values have been back-fit from various data sets (Feng et al., 1999; Lin, 2003; Heimovaara et al., 2004).


View this table:
[in this window]
[in a new window]
  		Table 1. Mean electrical length and physical length for hardware system sub-components.

 
Adding a transient suppressor and associated cable significantly reduced the reflection coefficents differences for water by 1/2 to 1/3 (Table 2), with additional reductions due to multiplexers and long cables. Reece (1998) observed that a much larger resistance was associated with a transient suppressor than for other attachments. Decreasing these differences would decrease the sensitivity for waveform analysis. The square root of apparent permittivity, {varepsilon}a1/2, determined from travel time for water showed little effect of transient suppressor or multiplexers, but the long cable dramatically increased {varepsilon}a1/2. Calculated frequency bandwidth for water was not significantly affected by level of additional hardware (Table 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Attachment level main effects on water reflection coefficient differences, frequency bandwidth, and square root of apparent permittivity ({varepsilon}a1/2) across temperatures.

 
As expected, each attachment level created more distortions (Fig. 2) in the soil waveforms. Attenuation was apparent from the transient suppressor, but additional reflections were obvious from both levels of multiplexers. Typical waveforms (Fig. 2) for the parallel probe with a 1:1 balun revealed a large impedance mismatch in the head, which joined the cable (~50 ohm) and the probe (~360 ohm), and caused a large increase in reflection coefficient. The reflection coefficient fell rapidly in the moist soil environment.

The reflection coefficient differences were again significantly reduced by adding a transient suppressor (Table 3, Fig. 2); lesser additional reductions were caused by adding multiplexers and a long cable. These soil results were similar to those shown by water (Table 2). The soil water content main effect showed that the Step 2 reflection coefficient differences for 0.385 m3 m–3 was significantly larger than for 0.288 m3 m–3 (not shown). Temperature main effects showed that the Step 2 reflection coefficient differences for 25°C was significantly larger than for 5°C, and that Step 3 reflection coefficient differences for 5°C was significantly smaller than for 15 or 25°C (not shown).


View this table:
[in this window]
[in a new window]
  		Table 3. Attachment level main effects on soil reflection coefficient differences, travel time, and frequency bandwidth, across water contents and temperatures.

 
Adding attachments to a cable tester significantly reduced calculated frequency bandwidth (Table 3), and the frequency bandwidth was for Level 4 was about half that of Level 1. The water content main effect showed that the calculated frequency bandwidth for 0.288 m3 m–3 was significantly greater than for the 0.352 or 0.385 m3 m–3 (192 vs. 166 or 163 MHz). The temperature main effect showed that the calculated frequency band for 5°C (195 MHz) was significantly greater than for 15°C (171 MHz), which was significantly greater than for 25°C (156 MHz).

The water content from the Topp et al. (1980) equation overestimated actual water content (Table 4), which commonly occurs for dispersive soils. For the soils, Level 4 reduced the calculated water content, opposite of the effect of Level 4 on the {varepsilon}a for water. Within each level, there were significant stepwise water content effects, as expected (not shown). Within Levels 2 and 4, the calculated water content was significantly greater for 25°C than for 5°C, and within Level 1, the calculated water content for 5°C was significantly less than for 15 or 25°C (now shown). The trend for calculated soil water content was the opposite of that shown for the {varepsilon}a1/2 of water, both of which were calculated from the travel time. There was a decrease for soil at the fourth level, but an increase for water at the highest level. The differences were apparently due to difficulties in determining travel time.


View this table:
[in this window]
[in a new window]
  		Table 4. Mean actual and calculated water contents (Eq. [3b]) for each level of the soil samples, across temperatures.

 
Attenuations and distortions due to TDR hardware have made it difficult to obtain useful waveforms for determining the square root of apparent real permittivity (Logsdon, 2000, 2005), which is used to determine soil water content. Logsdon (2000) determined that most of the difficulties were due to hardware limitations, and inability for internal waveform routines to correctly determine travel time for rounded, attenuated waveforms (Logsdon, 2005).


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Using TDR for automated field monitoring of soil water content usually requires one or more layers of multiplexers (for multiple depths or sites), long cables for sites not close together, and one older system used a transient suppressor. Each of these hardware attachments attenuated the signal, caused impedance mismatches that result in spurious reflections, and reduced frequency bandwidth. The TDR has proved useful for laboratory and manual field determination of soil water content, but field used of TDR for automated monitoring of soil water content often results in disappointing data (Bridge et al., 1996; Logsdon, 2005). For this study, additional problems were associated with the parallel probe and balun. Although early studies suggested a balun would be appropriate for a parallel probe (Spaans and Baker, 1993), later studies showed that a balun was not necessary for a two-wire probe (Nissen et al., 2003). The 1:1 balun was superior to older balun configerations (Spaans and Baker, 1993).

Several studies (Feng et al., 1999; Lin, 2003, Heimovaara et al., 2004) have back-fit hardware and sample properties using a layered approach, but have fitted a range of combined parameters that match the data well. Instead this study presented waveform data for increasing layers of hardware. Sample permittivity properties could instead be determined from systems that do not have so many layers and can subtract the cable effect, that is, vector network analyzers (Campbell, 1990; Hook et al., 2004; Huisman et al., 2004; Logsdon, 2005), or more accurate time domain systems (Anis and Jonscher, 1993; Ishida et al., 2000).

Received for publication June 7, 2005.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Logsdon, S. D.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Logsdon, S. D.
Agricola
Right arrow Articles by Logsdon, S. D.
Related Collections
Right arrow Water Content
Right arrow Time Domain Reflectometry, TDR
Right arrow Soil Methods/Instrumentation

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome