Soil compaction has been recognized for decades as a crop yield limiting factor. In a compacted layer (also known as a “hardpan”), the soil density is usually higher and the soil strength greater than in the surrounding soil. Hardpans increase the penetration resistance and consequently limit the development of an extensive root system. Other effects are reduced soil pore size and reduced water and air permeability. Cooperative AAES and ARS studies are using sound to help rapidly determine the depth of hard pans for more effective management of this problem.

The so-called cone penetrometer is the American Society of Agricultural Engineers (ASAE) standardized device to determine the penetration resistance of the soil and, in doing so, determine hardpan depth. A cone shaped object is pressed into the soil at a constant speed and the penetration resistance is recorded as a function of depth. The overall depth of the hardpan across the field is obtained by repeating the cone penetrometer test, usually in a locally randomized fashion. The tests are very time consuming and an untrained user can produce erroneous results by not maintaining a constant speed. Also, it is a “point-to-point” method, which limits its applicability in a sensor based precision tillage scheme.

Solid model of the tine with cone fitted on the front. The device on the back is a load cell that is used to simultaneously measure the force exerted by the soil on the cone.

Although the cone penetrometer can reveal the hardpan depth across a field, alleviation of the hardpan is still performed at a constant depth. This implies that in certain locations the hardpan may not be disrupted at all, while in others the soil is disturbed deeper than necessary, leading to significant energy waste. There is a need for a measurement system that determines the hardpan depth on-the-fly. This allows for automatic adjustment of the implement to an optimal depth where the hardpan is disrupted and no “overkill” takes place.

A simple and inexpensive approach to on-the-fly hardpan detection was developed at the Auburn University Department of Biosystems Engineering and the USDA-ARS National Soil Dynamics Laboratory. It is based on the measurement of sound, produced by a cone being drawn through the soil. The hypothesis was that the sound level would be proportional to the soil density (more particles would produce more sound) and soil strength, since it requires more energy to break up harder aggregates.

The proposed method is purely empirical; no attempts were made to analyze and explain the sound producing mechanics of a cone sliding through a medium. The objective was merely to determine whether the information obtained from a simple acoustic measurement system would be sufficient to determine hardpan depths. Experiments were conducted in a Decatur clay loam soil in the soil bins of the USDA-ARS National Soil Dynamic Laboratory in Auburn.

Three soil density levels were introduced by varying the number of passes of a compression wheel. Case 1, (“no pass”) means that no hardpan was installed. Case 2 (“single pass”) means that a single pass of the compression wheel was used. Case 3 (“double pass”) implies that two passes of the compression wheel were used. In cases 2 and 3, the hardpan was installed at 25.4-cm (10-inch) depth. Cone penetrometer readings were taken throughout the soil bins and they confirmed that the hardpan was installed at the intended depth level. Also, during experiments, the location of the implement as a function of time was measured, and combination yielded the Cone Index (CI) as a function of time. The measurement arrangement used in the experiments is shown in Figure 1.

Click to see Figure 1.

An aluminum cone was mounted on a shaft that was bolted on a tine with a sharp frontal edge. A microphone was fitted into the cone using rubber grommets to prevent direct contact sound with the tine. The wires of the microphone were fed through the shaft and upwards through a channel that was welded on the back of the tine (not shown). The data recording took place using MatLab's Data Acquisition Toolbox under control of a dedicated MatLab program.

During experiments, the tine was started at the soil surface level and was gradually entered into the soil until it reached a depth of 30.5 cm (12 inches). At the end of the run, the tine must penetrate through the hardpan and this was expected to give higher sound levels in a certain frequency window range.

To assess the overall performance of the system, the CI as a function of time was compared to the sound levels as a function of time. The sound signal was filtered using a frequency window. This window was determined from constant depth experiments, which revealed differences in sound levels in the highest frequency spectrum. To listen to the sound of the tine being drawn through the soil click here.

As an example, the data is given for a single pass test run. The sound data shows that there is a very good correlation between sound level (in the detection edge range) and Cone Index (see Figure 2.) There seems to be a small shift in time (approximately two seconds), which is most likely caused by the time difference between the movement of the cone and the starting of the data collection program.

The acoustic measurement system as developed showed great potential to determine the hardpan depth accurately and in real-time. The data acquisition system is inexpensive and the filtering process is straightforward.

It was found that the plow pan “revealed itself” in the highest information carrying range of the frequency spectrum. This is no coincidence, there is a mechanism at work that allows the higher frequency sound waves (that contain more energy) to propagate into the sensor. More in-depth research could explain why this is the case.

The method does not have a solid mathematical foundation, which limits its scientific significance. However, it has significant value in the attempt to minimize the use of fossil fuels in tillage operations.

Top            Table of Contents